Advance Access publication on June 22, 2005 Extreme regression

JPART18:543–571 Collaborative Governance in Theoryand PracticeChris AnsellAlison GashUniversity of California,BerkeleyABSTRACTOver the past few decades,a new form of governance has emerged to replace adversarial and managerial modes of policy making and implementation.Collaborative governance,as it has come to be known,brings public and private stakeholders together in collective forums with public agencies to engage in consensus-oriented decision making.In this article, we conduct a meta-analytical study of the existing literature on collaborative governance with the goal of elaborating a contingency model of collaborative governance.After review-ing137cases of collaborative governance across a range of policy sectors,we identify critical variables that will influence whether or not this mode of governance will produce successful collaboration.These variables include the prior history of conflict or cooperation, the incentives for stakeholders to participate,power and resources imbalances,leadership, and institutional design.We also identify a series of factors that are crucial within the collaborative process itself.These factors include face-to-face dialogue,trust building,and the development of commitment and shared understanding.We found that a virtuous cycle of collaboration tends to develop when collaborative forums focus on‘‘small wins’’that deepen trust,commitment,and shared understanding.The article concludes with a discus-sion of the implications of our contingency model for practitioners and for future research on collaborative governance.Over the last two decades,a new strategy of governing called‘‘collaborative governance’’has developed.This mode of governance brings multiple stakeholders together in common forums with public agencies to engage in consensus-oriented decision making.In this article,we conduct a meta-analytical study of the existing literature on collaborative governance with the goal of elaborating a general model of collaborative governance. The ultimate goal is to develop a contingency approach to collaboration that can highlight conditions under which collaborative governance will be more or less effective as an Early versions of this article were presented at the Conference on Democratic Network Governance,Copenhagen,the Interdisciplinary Committee on Organizations at the University of California,Irvine,and the Berkeley graduate seminar Perspectives on Governance.We thank the participants of these events for their useful suggestions and Martha Feldman,in particular,for her encouragement.We also thank two anonymous reviewers for their thoughtful and useful comments.Address correspondence to the author at cansell@ or aligash@.doi:10.1093/jopart/mum032Advance Access publication on November13,2007ªThe Author2007.Published by Oxford University Press on behalf of the Journal of Public Administration Researchand Theory,Inc.All rights reserved.For permissions,please e-mail:journals.permissions@approach to policy making and public management.1In conducting this meta-analytic study,we adopted a strategy we call ‘‘successive approximation’’:we used a sample of the literature to develop a common language for analyzing collaborative governance and then successively ‘‘tested’’this language against additional cases,refining and elaborating our model of collaborative governance as we evaluated additional cases.Although collaborative governance may now have a fashionable management cache´,the untidy character of the literature on collaboration reflects the way it has bubbled up from many local experiments,often in reaction to previous governance failures.Collabo-rative governance has emerged as a response to the failures of downstream implementation and to the high cost and politicization of regulation.It has developed as an alternative to the adversarialism of interest group pluralism and to the accountability failures of manageri-alism (especially as the authority of experts is challenged).More positively,one might argue that trends toward collaboration also arise from the growth of knowledge and in-stitutional capacity.As knowledge becomes increasingly specialized and distributed and as institutional infrastructures become more complex and interdependent,the demand for collaboration increases.The common metric for all these factors may be,as Gray (1989)has pointed out,the increasing ‘‘turbulence’’faced by policy makers and managers.Although Susskind and Cruikshank (1987),Gray (1989),and Fung and Wright (2001,2003)have suggested more general theoretical accounts of collaborative governance,much of the literature is focused on the species rather than the genus .The bulk of the collaborative governance literature is composed of single-case case studies focused on sector-specific governance issues like site-based management of schools,community po-licing,watershed councils,regulatory negotiation,collaborative planning,community health partnerships,and natural resource comanagement (the species).2Moreover,a num-ber of the most influential theoretical accounts of this phenomenon are focused on specific types of collaborative governance.Healey (1996,2003)and Innes and Booher (1999a,1999b),for example,provide foundational accounts of collaborative planning,as Freeman (1997)does for regulation and administrative law and Wondolleck and Yaffee (2000)do for natural resources management.Our goal is to build on the findings of this rich literature,but also to derive theoretical and empirical claims about the genus of collaborative governance—about the common mode of governing.DEFINING COLLABORATIVE GOVERNANCEWe define collaborative governance as follows:A governing arrangement where one or more public agencies directly engage non-state stakeholders in a collective decision-making process that is formal,consensus-oriented,and deliberative and that aims to make or implement public policy or manage publicprograms or assets.This definition stresses six important criteria:(1)the forum is initiated by public agencies or institutions,(2)participants in the forum include nonstate actors,(3)participants engage directly in decision making and are not merely ‘‘consulted’’by public agencies,(4)the 1Thomas (1995)develops a contingency perspective on public participation,though it aims more broadly and is developed from the perspective of public managers.2A smaller group of studies evaluates specific types of collaborative governance at a more aggregated level (for example,see Beierle [2000],Langbein [2002],and Leach,Pelkey,and Sabatier [2002]).Journal of Public Administration Research and Theory544Ansell and Gash Collaborative Governance in Theory and Practice545 forum is formally organized and meets collectively,(5)the forum aims to make decisionsby consensus(even if consensus is not achieved in practice),and(6)the focus of collab-oration is on public policy or public management.This is a more restrictive definition thanis sometimes found in the literature.However,the wide-ranging use of the term has,as Imperial notes,been a barrier to theory building(Imperial2005,286).Since our goal is to compare apples with apples(to the extent possible),we have defined the term restrictivelyso as to increase the comparability of our cases.One critical component of the term collaborative governance is‘‘governance.’’Much research has been devoted to establishing a workable definition of governance that is bounded and falsifiable,yet comprehensive.For instance,Lynn,Heinrich,and Hill (2001,7)construe governance broadly as‘‘regimes of laws,rules,judicial decisions,and administrative practices that constrain,prescribe,and enable the provision of publicly supported goods and services.’’This definition provides room for traditional governmental structures as well as emerging forms of public/private decision-making bodies.Stoker,onthe other hand,argues:As a baseline definition it can be taken that governance refers to the rules and forms that guidecollective decision-making.That the focus is on decision-making in the collective impliesthat governance is not about one individual making a decision but rather about groups ofindividuals or organisations or systems of organisations making decisions(2004,3).He also suggests that among the various interpretations of the term,there is‘‘baseline agreement that governance refers to the development of governing styles in which bound-aries between and within public and private sectors have become blurred’’(Stoker1998, 17).We opt for a combined approach to conceptualize governance.We agree with Lynn, Heinrich,and Hill that governance applies to laws and rules that pertain to the provision ofpublic goods.However,we adopt Stoker’s claim that governance is also about collective decision making—and specifically about collective decision making that includes bothpublic and private actors.Collaborative governance is therefore a type of governance inwhich public and private actors work collectively in distinctive ways,using particular processes,to establish laws and rules for the provision of public goods.Although there are many forms of collaboration involving strictly nonstate actors,ourdefinition stipulates a specific role for public agencies.By using the term‘‘public agency,’’our intention is to include public institutions such as bureaucracies,courts,legislatures,andother governmental bodies at the local,state,or federal level.But the typical public in-stitution among our cases is,in fact,an executive branch agency,and therefore,the term‘‘public agency’’is apt.Such public agencies may initiate collaborative forums either tofulfill their own purposes or to comply with a mandate,including court orders,legislation,or rules governing the allocation of federal funds.For example,the Workforce InvestmentAct of1998stipulates that all states and localities receiving federal workforce develop-ment funds must convene a workforce investment board that comprised public and privateactors in order to develop and oversee policies at the state and local level concerning job training,under-and unemployment.According to our definition,these workforce invest-ments boards are mandated to engage in collaborative governance.Although public agencies are typically the initiators or instigators of collaborative governance,our definition requires participation by nonstate stakeholders.Some scholars describe interagency coordination as collaborative governance.Although there is nothing inherently wrong with using the term in this way,much of the literature on collaborativegovernance uses this term to signal a different kind of relationship between public agencies and nonstate stakeholders.Smith (1998,61),for example,argues that collaboratives in-volve ‘‘representation by key interest groups.’’Connick and Innes (2003,180)define collaborative governance as including ‘‘representatives of all relevant interests.’’Reilly (1998,115)describes collaborative efforts as a type of problem solving that involves the ‘‘shared pursuit of government agencies and concerned citizens.’’We use the term ‘‘stakeholder’’to refer both to the participation of citizens as indi-viduals and to the participation of organized groups.For convenience,we will also here-after use the term ‘‘stakeholder’’to refer to both public agencies and nonstate stakeholders,though we believe that public agencies have a distinctive leadership role in collaborative governance.Our definition of collaborative governance also sets standards for the type of participation of nonstate stakeholders.We believe that collaborative governance is never merely consultative.3Collaboration implies two-way communication and influence be-tween agencies and stakeholders and also opportunities for stakeholders to talk with each other.Agencies and stakeholders must meet together in a deliberative and multilateral process.In other words,as described above,the process must be collective .Consultative techniques,such as stakeholder surveys or focus groups,although possibly very useful management tools,are not collaborative in the sense implied here because they do not permit two-way flows of communication or multilateral deliberation.Collaboration also implies that nonstate stakeholders will have real responsibility for policy outcomes.Therefore,we impose the condition that stakeholders must be directly engaged in decision making.This criterion is implicit in much of the collaborative gov-ernance literature.Freeman (1997,22),for example,argues that stakeholders participate ‘‘in all stages of the decisionmaking process.’’The watershed partnerships studied by Leach,Pelkey,and Sabatier (2002,648)make policy and implementation decisions on a range of ongoing water management issues regarding streams,rivers,and watersheds.Ultimate authority may lie with the public agency (as with regulatory negotiation),but stakeholders must directly participate in the decision-making process.Thus,advisory committees may be a form of collaborative governance if their advice is closely linked to decision-making outcomes.In practice (and by design),however,advisory committees are often far removed from actual decision making.We impose the criteria of formal collaboration to distinguish collaborative gover-nance from more casual and conventional forms of agency-interest group interaction.For example,the term collaborative governance might be thought to describe the informal relationships that agencies and interest groups have always cultivated.Surely,interest groups and public agencies have always engaged in two-way flows of influence.The difference between our definition of collaborative governance and conventional interest group influence is that the former implies an explicit and public strategy of organizing this influence.Walter and Petr (2000,495),for example,describe collaborative governance as a formal activity that ‘‘involves joint activities,joint structures and shared resources,’’and Padilla and Daigle (1998,74)prescribe the development of a ‘‘structured arrangement.’’This formal arrangement implies organization and structure.Decisions in collaborative forums are consensus oriented (Connick and Innes 2003;Seidenfeld 2000).Although public agencies may have the ultimate authority to make 3See Beierle and Long (1999)for an example of collaboration as consultation.Journal of Public Administration Research and Theory546Ansell and Gash Collaborative Governance in Theory and Practice547 a decision,the goal of collaboration is typically to achieve some degree of consensus among stakeholders.We use the term consensus oriented because collaborative forumsoften do not succeed in reaching consensus.However,the premise of meeting together ina deliberative,multilateral,and formal forum is to strive toward consensus or,at least,tostrive to discover areas of agreement.Finally,collaborative governance focuses on public policies and issues.The focuson public issues distinguishes collaborative governance from other forms of consensus decision making,such as alternative dispute resolution or transformative mediation. Although agencies may pursue dispute resolution or mediation to reduce social or politicalconflict,these techniques are often used to deal with strictly private conflicts.Moreover,public dispute resolution or mediation may be designed merely to resolve private disputes.While acknowledging the ambiguity of the boundary between public and private,we restrict the use of the term‘‘collaborative governance’’to the governance of public affairs.Our definition of collaborative governance is meant to distinguish collaborative gov-ernance from two alternative patterns of policy making:adversarialism and managerialism (Busenberg1999;Futrell2003;Williams and Matheny1995).By contrast with decisionsmade adversarially,collaborative governance is not a‘‘winner-take-all’’form of interest intermediation.In collaborative governance,stakeholders will often have an adversarial relationship to one another,but the goal is to transform adversarial relationships into more cooperative ones.In adversarial politics,groups may engage in positive-sum bargainingand develop cooperative alliances.However,this cooperation is ad hoc,and adversarial politics does not explicitly seek to transform conflict into cooperation.In managerialism,public agencies make decisions unilaterally or through closed de-cision processes,typically relying on agency experts to make decisions(Futrell2003; Williams and Matheny1995).Although managerial agencies may take account of stake-holder perspectives in their decision making and may even go so far as to consult directlywith stakeholders,collaborative governance requires that stakeholders be directly includedin the decision-making process.A number of synonyms for collaborative governance may cause confusion.For ex-ample,‘‘corporatism’’is certainly a form of collaborative governance as we define it. Classic definitions of corporatism(like Schmitter’s)emphasize tripartite bargaining be-tween peak associations of labor and capital and the state.Typically,these peak associa-tions have a representational monopoly in their sector(they are‘‘encompassing’’).If westart with this narrower definition of corporatism,collaborative governance is the broader term.Collaborative governance often implies the inclusion of a broader range of stake-holders than corporatism,and the stakeholders often lack a representational monopoly overtheir sector.The term‘‘associational governance’’is sometimes used to refer to the more generic mode of governing with associations,but collaborative governance may not even include formal associations.The Porte Alegre project,for example,is a form of collabo-rative governance that includes individual citizens in budgetary decision making(Fung and Wright2001).Sometimes the term‘‘policy network’’is used to describe more pluralistic forms ofstate-society cooperation.A policy network may include both public agencies and stake-holder groups.Moreover,policy networks typically imply cooperative modes of deliber-ation or decision making among actors within the network.Thus,the terms policy networkand collaborative governance can refer to similar phenomena.However,collaborative governance refers to an explicit and formal strategy of incorporating stakeholders intomultilateral and consensus-oriented decision-making processes.By contrast,the co-operation inherent in policy networks may be informal and remain largely implicit (e.g.,unacknowledged,unstated,nondesigned).Moreover,it may operate through infor-mal patterns of brokerage and shuttle diplomacy rather than through formal multilateral processes.Collaborative governance and public-private partnership can also sometimes refer to the same phenomenon.Public-private partnerships typically require collaboration to func-tion,but their goal is often to achieve coordination rather than to achieve decision-making consensus per se.A public-private partnership may simply represent an agreement between public and private actors to deliver certain services or perform certain tasks.Collective decision making is therefore secondary to the definition of public-private partnership.By contrast,the institutionalization of a collective decision-making process is central to the definition of collaborative governance.Finally,a range of terms are often used interchangeably with collaborative gover-nance.Such terms include participatory management,interactive policy making,stake-holder governance,and collaborative management.We prefer the term governance to management because it is broader and encompasses various aspects of the governing process,including planning,policy making,and management.The term collaborative is also more indicative of the deliberative and consensus-oriented approach that we contrast with adversarialism or managerialism than terms like participatory or interactive.A MODEL OF COLLABORATIVE GOVERNANCEArmed with a working definition of collaborative governance,we collected a wide range of case studies from the literature.We did this in the typical fashion:we systematically reviewed journals across a wide range of disciplines,including specialist journals in public health,education,social welfare,international relations,etc.We also conducted key word electronic searches using a wide variety of search terms,including those described above and many more (e.g.,‘‘comanagement,’’‘‘public participation,’’‘‘alternative dispute res-olution’’).Of course,we also followed up on the literature cited in the cases we discovered.Ultimately,our model is built on an analysis of 137cases.Although international in scope,our search was restricted to literature in English,and thus,American cases are overrepre-sented.Even a cursory examination of our cases also suggests that natural resource man-agement cases are overrepresented.This is not due to any sampling bias on our part but rather reflects the importance of collaborative strategies for managing contentious local resource disputes.Most of the studies we reviewed were case studies of an attempt to implement collaborative governance in a particular sector.As you might imagine,the universe of cases we collected was quite diverse and the cases differed in quality,methodology,and intent.Although our definition was restrictive so as to facilitate comparison of apples with apples,representing this diversity was also one of our goals.We perceived experiments with collaborative governance bubbling up in many different policy sectors,with little sense that they were engaged in a similar governance strategy.Surely,we felt,these diverse experiments could learn from each other.Yet this diversity proved a challenge.Our original intention to treat these cases as a large-N data set subject to quasi-experimental statistical evaluation was not successful.Since it is useful for both scholarsJournal of Public Administration Research and Theory548Ansell and Gash Collaborative Governance in Theory and Practice549 and practitioners to understand how we arrived at our conclusions,we briefly report on the problems we encountered in conducting our meta-analysis.Early attempts at systematic coding were frustrating,and we soon developed an understanding of our dilemma.Although scholars studying collaborative governancehad already made some important theoretical statements,the language used to describewhat was happening was far from standardized.We found ourselves groping tofinda common language of description and evaluation even as we were trying to‘‘code’’studies.Add to this challenge a severe problem of‘‘missing data’’—a reflection of thehighly varied motivations of the researchers—and we concluded that a quasi-experimental approach was ill advised.Ultimately,we moved toward a meta-analytic strategy that wecall successive approximation.We selected a subset of our cases and used them to developa common‘‘model’’of collaborative governance.4We then randomly selected additional subsets of case studies.The second subset was used to‘‘test’’the model developed in thefirst round and then to further‘‘refine’’the model.A third sample of cases was used to testthe second-round model,and so on.The appendix provides a list of the studies evaluated ineach of four successive rounds of evaluation.Successive approximation has the advantage of both refining the conceptual modelwhile providing some of the evaluative‘‘discipline’’of a quasi-experimental study.How-ever,we are under no illusion that this process yielded‘‘the one’’model of collaborative governance.There was a large element of art involved in both specifying and evaluatingour model.As we proceeded,we were overwhelmed by the complexity of the collaborative process.Variables and causal relationships proliferated beyond what we felt would ulti-mately be useful for policy makers and practitioners.Therefore,our model representsa conscious attempt to simplify as much as possible the representation of key variablesand their relationships.This goal of simplification led us to stress common and frequentfindings across cases.This approach strengthens the generality of ourfindings but dis-counts less universal or frequently mentionedfindings from the literature.Toward the endof our analysis,we were ourselves in disagreement about how to represent key relations.We used thefinal round of case analysis to settle these differences.One other important clarification needs to be made before we introduce ourfindings.Our survey of the cases quickly disabused us of the notion that we could use our analysis to answer the question:‘‘Is collaborative governance more effective than adversarial or managerial governance?’’Very few of the studies we reviewed actually evaluated gover-nance outcomes.This is not to say that the comparison between collaborative,adversarial,and managerial governance is not relevant to these studies.Experiments with collaborative governance were typically driven by earlier failures with adversarial or managerial approaches.But systematic comparisons were rarely explicitly made.What most studiesdid try to do was understand the conditions under which stakeholders acted collaboratively.Did they engage in good faith negotiation?Did they pursue mutual gains?Did they achieve consensus?Were they satisfied with the process?In other words,most studies in the collaborative governance literature evaluate‘‘process outcomes’’rather than policy or management outcomes.Figure1provides a visual representation of our centralfindings.The model has fourbroad variables—starting conditions,institutional design,leadership,and collaborative4To avoid recreating the wheel,ourfirst subset was not randomly selected but included many of the most prominent theoretical statements about collaborative governance.process.Each of these broad variables can be disaggregated into more fine-grained vari-ables.Collaborative process variables are treated as the core of our model,with starting conditions,institutional design,and leadership variables represented as either critical contributions to or context for the collaborative process.Starting conditions set the basic level of trust,conflict,and social capital that become resources or liabilities during col-laboration.Institutional design sets the basic ground rules under which collaboration takes place.And,leadership provides essential mediation and facilitation for the collaborative process.The collaborative process itself is highly iterative and nonlinear,and thus,we represent it (with considerable simplification)as a cycle.The remainder of the article describes each of these variables in more detail and draws out their implications for a contingency model of collaborative governance.STARTING CONDITIONSThe literature is clear that conditions present at the outset of collaboration can either facilitate or discourage cooperation among stakeholders and between agencies and stake-holders.Imagine two very different starting points.In one,the stakeholders have a history of bitter division over some emotionally charged local issue and have come to regard each other as unscrupulous enemies.In the other,the stakeholders have a shared vision for what they would like to achieve through collaboration and a history of past cooperation and mutual respect.In both cases,collaboration may be difficult,but the first case must over-come problems of distrust,disrespect,and outright antagonism.We narrowed the critical starting conditions down to three broad variables:imbalances between the resources orFigure 1A Model of Collaborative GovernanceJournal of Public Administration Research and Theory550。

Number of Lymph Nodes Retrieved is an Important Determinant of Survival of Patients with Stage II and Stage III Colorectal CancerKenjiro Kotake 1,2,*,Satoshi Honjo 2,Kenichi Sugihara 2,Yojiro Hashiguchi 2,Tomoyuki Kato 2,Susumu Kodaira 2,Tetsuichiro Muto 2and Yasuo Koyama 21Department of Surgery,Tochigi Cancer Center and 2Registry Committee,Japanese Society for Cancer of the Colon and Rectum,Tochigi-ken,Japan*For reprints and all correspondence:Kenjiro Kotake,Department of Surgery,Tochigi Cancer Center,4-9-13Yohnan,Utsunomiya,Tochigi-ken 320-0834,Japan.E-mail:kkotake@tcc.pref.tochigi.lg.jpReceived June 30,2011;accepted October 21,2011Objective:The number of lymph nodes retrieved is recognized to be a prognostic factor of Stage II colorectal cancer.However,the prognostic significance of the number of lymph nodes retrieved in Stage III colorectal cancer remains controversial.Methods:The relationship between the number of lymph nodes retrieved and clinical and pathological factors,and significance of the number of lymph nodes retrieved for prognosis of Stage II and III colorectal cancer were investigated.A total of 16865patients with T3/T4colo-rectal cancer who had R0resection were analysed.Results:The arithmetic mean of the number of lymph nodes retrieved of all cases was 20.0.The number of lymph nodes retrieved were varied according to several clinical and patho-logical variables with significant difference,and the greater difference was observed in scope of nodal dissection.Survival of Stages II and III was significantly associated with the number of lymph nodes retrieved.Five-year overall survival of the patients with 9of the number of lymph nodes retrieved and those with .27differed by 6.4%for Stage II colon cancer,8.8%for Stage III colon cancer,12.5%for Stage II rectal cancer and 10.6%for Stage III rectal cancer.With one increase in the number of lymph nodes retrieved,the mortality risk was decreased by 2.1%for Stage II and by 0.8%for Stage III,respectively.The cut-off point of the number of lymph nodes retrieved was not obtained.Conclusions:The number of lymph nodes retrieved was shown to be an important prognos-tic variable not only in Stage II but also in Stage III colorectal cancer,and it was most promin-ently determined by the scope of nodal dissection.A cut-off value for the number of lymph nodes retrieved was not found,and it is necessary to carry out appropriate nodal dissection and examine as many lymph nodes as possible.Key words:colorectal cancer –number of lymph nodes retrieved –lymph node dissectionINTRODUCTIONLymph node metastasis is the most important determinant of survival in localized colorectal cancer (CRC),and it has been used as an indicator for postoperative adjuvant chemo-therapy.In recent years,it has been recognized that in add-ition to lymph node metastasis,the number of lymph nodes retrieved (NLNR)is closely related to prognosis of Stage II CRC (1–4).Some clinical guidelines recommendpostoperative adjuvant chemotherapy in Stage II CRC with a small NLNR to prevent recurrence (5,6).However,the precise mechanism how the NLNR works on prognosis has not been clarified,because the NLNR is thought to be de-pendent on the scope of lymph node dissection,quality of pathologic examination and individual differences in the number of lymph node present.Significance of the NLNR in Stage III CRC remains controversial (1,7–10).Clinical#The Author 2011.Published by Oxford University Press.All rights reserved.For Permissions,please email:journals.permissions@Jpn J Clin Oncol 2012;42(1)29–35doi:10.1093/jjco/hyr164Advance Access Publication 18November2011at United Arab Emirates University on May 9, 2014/Downloaded fromevidence on the NLNR should be formulated,also allowing for the fact that laparoscopic surgery has been rapidly spreading in Japan since the late 1990s (11).The aim of this study is not only to investigate the rela-tionship between the NLNR and clinical and pathological factors,but also to clarify significance of the NLNR on prog-nosis of Stage II and III CRC,using the database for the Japanese Society for Cancer of the Colon and Rectum (JSCCR),which contains information of a large number of patients with CRC recorded in accordance with the clinical and pathological classification of the JSCCR (12).PATIENTS AND METHODSThe JSCCR has a registration system which has started in 1980.The member institutions which are located all over Japan voluntarily have been registering the clinical and pathological information of patients with CRC who were treated in each institute.The database currently contains in-formation about 140000CRC patients treated between 1974and 2002(13).A total of 16865patients who had T3or T4colorectal adenocarcinoma and underwent R0resection accompanied by D2or D3lymph node dissection between 1985and 1994were extracted from the database.Because laparoscopic surgery for CRC was introduced in 1992(11),nearly all the cases in the present study were treated with open surgery.During this period,chemotherapeutic regimens of neither 5-fluorouracil and leucovorin nor oxaliplatin were available in Japan.In colon cancer,D2dissection means removal of peri-colic nodes and intermediate nodes,and D3dissection means removal of main lymph nodes at the root of the re-gional artery in addition to D2dissection.In rectal cancer,D2dissection is removal of both peri-rectal nodes locating in the mesorectum and lymph nodes along the middle rectal artery between the proper rectal fascia and the pelvic nerve plexus,and D3dissection includes removal of the main nodes around the inferior mesenteric artery and lateral nodes dissection which means removal of internal iliac and obtur-ator lymph nodes in addition to D2dissection.The classifi-cation of JSCCR clearly states the scope of dissection (12).Exclusion criteria were as follows:unspecified age,syn-chronous multiple cancers,no pathologic examination of lymph nodes,unknown followed-up status and receiving perioperative radiotherapy.To eliminate possible data entry errors at the time of registration,patients with 100or more lymph nodes retrieved were also excluded.The correlation between the histologically proven NLNR and the following variables were investigated:year of treat-ment,age,gender,tumour site,histological grade,depth of tumour invasion,degree of lymph node metastasis,tumour size,scope of lymph node dissection,postoperative adjuvant chemotherapy and the number of patients registered per insti-tution.The depth of tumour invasion and degree of lymph node metastasis described by the JSCCR classification (12),which is not translated into the UICC-TNM classification system,is given in Table 1.S TATISTICAL M ETHODSFirst,the NLNR was investigated in relation to age,gender,tumour site,depth of tumour invasion,stage and other clinic-al and pathological characteristics using multiple linear re-gression with mutual adjustment of these factors to calculate adjusted means due to each level of them (14).Secondly,an association of the NLNR in Stage III with the number of metastasized lymph nodes was investigated by the Kruskal–Wallis test for the categorized NLNR and by the use of Spearman’s correlation coefficient for the NLNR treated as continuous variable.Thirdly,survival analysis was done according to the quartiles of the NLNR for patients with Stages II and III by the Kaplan–Meier methods,and statis-tical significance for the difference was tested by log-rank test with length of the follow-up period being truncated at 5years.Fourthly,the Cox proportional hazard model was employed to examine an effect of the NLNR on survival adjusted for other clinical variables in multivariate analyses.Table 1.Classification of depth of tumour invasion and lymph node metastasis according to the rules by the JSCCR Depth of tumour invasion a sm Invasion beyond the mucosa but within the submucosa mpInvasion beyond the submucosa but within the proper muscle layerFor further invasion of cancers arising from the intestine covered with the serosa ss Invasion beyond the proper muscle layer but not exposed on the serosal surfacese Invasion exposed on the serosal surface siInvasion to other organs or structuresFor further invasion of cancers arising from the intestine not covered with the serosa a1Invasion beyond the proper muscle layer but penetrating not deeplya2Invasion beyond the proper muscle layer and penetrating deeply ai Invasion to other organs or structures Lymph node metastasisn0No lymph node metastasisn1Metastasis to lymph nodes that are located along the marginal artery and adjacent to the colonn2Metastasis to lymph nodes that are located along the course of the major vessels supplying the arean3Metastasis to lymph nodes that are located around the superior or inferior mesenteric vessels,and/or metastasis to lateral lymph nodes for some cases with rectal cancern4Metastasis to lymph node more distant than that for n3JSCCR,Japanese Society for Cancer of the Colon and Rectum.asm,mp,a group consisting of ss,a1and a2,another group consisting of se,ai and si are corresponding to the UICC pT1,pT2,pT3and pT4,respectively.30Number of lymph nodes retrievedat United Arab Emirates University on May 9, 2014/Downloaded fromDummy variables in the models were assigned to levels of relevant variables other than a reference category of each variable.Interactions between the NLNR and degree of lymph node metastasis (presence or absence),scope of lymph node dissection and tumour site (colon vs.rectum)on the risk for death due to CRC were examined by comparing the full model with the model having also the interaction terms;the likelihood ratio statistics obtained from 22times difference in log-likelihood between the two models was tested as x 2statistics.Because the interaction between the NLNR and the lymph node metastasis was statistically sig-nificant as shown in the latter,the multivariate analysis by the Cox model was performed for the Stage II and III cancers separately.Finally,martingale residuals (15)were computed based on the model including the year of treat-ment,age,gender,depth of tumour invasion,histological grade,tumour site,tumour size,lymph node dissection,post-operative chemotherapy (done vs.not done)and the number of registrations from institution.The residuals were plotted against the NLNR for Stages II and III,separately.All statis-tical analyses were performed using STATA Statistical Software (STATA Corporation,College Station,TX,USA),and P ,0.05was regarded as denoting statistical signifi-cance otherwise specified.RESULTSFigure 1shows the distribution of the NLNR for 16865patients.A total of 336857lymph nodes were retrieved,with a median of 17.0,an arithmetic mean of 20.0with a standard deviation of 14.0and a mode of 10.Because the NLNR was distributed with the right skewness,a log-transformed value of the NLNR was used in most statistical analyses and the geometric mean was calculated instead of the arithmetic mean in the following section.As shown in Table 2,the NLNR was varied according to categories of selected variables including age,gender,degree of tumour invasion,histological grade,tumour site,tumour size,scope of lymph node dissection,degree of lymph node metastasis and number of registration per insti-tute.The greater difference in the NLNR was observed in scope of lymph node dissection.The NLNR was divided into four subgroups (quartiles): 9,10–16,17–26and 27.The number of metastatic nodes for each subgroup is shown in Table 3.The number of positive lymph nodes was increased with the increasing NLNR.The same was observed in each site.Even when handling the both number as continuous variables,the two were weakly correlated to each other (Spearman’s correlation coefficient ¼0.0813,P ,0.01).Overall survival of Stages II and III was significantly asso-ciated with the NLNR.The same results were observed even if it was investigated according to tumour sites (Table 4).When comparison was made between patients with 9of the NLNR and those with 27,the 5-year overall survival differed by 6.4%for Stage II colon cancer,8.8%for Stage III colon cancer,12.5%for Stage II rectal cancer and 10.6%for Stage III rectal cancer.The 5-year disease-specific survival differed similarly between those with 9of the NLNR and with 27(P ,0.001,data not shown).Scope of the lymph node dissection and tumour site did not influence the effect of the NLNR on the risk for death (P for interaction ¼0.91and 0.58,respectively,when the NLNR was categorized;and 0.19and 0.84,respectively,when the NLNR was treated as a continuous variable),but lymph node metastasis influenced the effect of the NLNR (P ,0.01when the NLNR was categorized or treated as con-tinuous).Therefore,the Cox proportional hazard model was employed for the Stage II and III CRC separately examining CRC mortality risk due to the increasing NLNR adjusted for age,sex,scope of the dissection and other factors (Table 5).Stages II and III with 27of the NLNR had the decreased risk by 54and 25%when compared with those with 9,re-spectively.With one increase in the NLNR,CRC mortality risk was decreased by 2.1%for Stage II and by 0.8%for Stage III,respectively.For both Stages II and III,the cut-off point was not obtained because the plots of the martingale residuals were fairly linear (data not shown).DISCUSSIONIn the present study,it was shown that the NLNR was a crit-ical prognosticator for not only Stage II but also Stage III CRC.The arithmetic average of the NLNR in our study was comparable to the reports based on data from single institu-tion (16–18).On the other hand,the NLNR was around 10in the studies that examined large-scale database,such as the NCDB and SEER (2,10,19,20).Theoretically,the NLNR may be influenced by the extent of surgery or scope of lymph node dissection,technique of pathologic examination and individual difference intheFigure 1.Number of lymph nodes retrieved per patient in 16865colorectal cancer patients using the JSCCR database.JSCCR,Japanese Society for Cancer of the Colon and Rectum.Jpn J Clin Oncol 2012;42(1)31at United Arab Emirates University on May 9, 2014/Downloaded fromTable 2.Adjusted means of the number of lymph nodes retrieved (NLNR)according to clinical and pathological variables Variables Number of cases Adjusted means of NLNR 95%confidence interval P value1985148315.52Reference 1986167815.3614.5916.160.6891987165115.0714.3215.870.2631988167914.7013.9615.480.0391989185815.0514.3215.820.2311990194815.3214.5816.100.6151991164115.3314.5516.140.6361992156316.3015.4717.170.0671993155416.4015.5617.280.0401994181015.7915.0016.610.510Trend0.001Age (years)16–4054220.5219.2421.88,0.00141–54348717.2416.7117.78,0.00155–64549915.54Reference 65–74480514.7214.3115.15,0.00175þ253213.5913.1214.08,0.001Trend,0.001Sex Male 971615.13Reference Female 714915.9215.5716.28,0.001Depth of invasion ss/a1989215.14Reference s/a2606716.0015.6316.39,0.001si/ai 90615.3514.5916.150.600Trend0.001HistologyWell differentiated 841516.00Reference Moderately differentiated 714314.9914.6515.34,0.001Others 130714.6914.0615.35,0.001SiteRight colon 481717.60Reference Left colon 528113.5713.1813.97,0.001Rectum 676715.6015.1716.04,0.001Diameter (mm)10–1915910.859.6912.14,0.00120–2982710.9610.3911.58,0.00130–39221512.6212.1513.11,0.00140–49332414.18Reference 50–59332115.8515.3216.40,0.00160–69249117.2216.6017.87,0.00170–79150118.1017.3318.90,0.00180–8997119.0018.0520.00,0.001Continued32Number of lymph nodes retrievedat United Arab Emirates University on May 9, 2014/Downloaded fromnumber of nodes present.The present study clearly showed that the scope of lymph node dissection affected the NLNR. More lymph nodes are to be harvested for D3than for D2 dissection if the pathological examination technique is con-stant.The NLNR was dependent on factors other than scope of lymph node dissection in the present investigation,and thefindings on several factors were comparable with those from previous studies including gender,age,degree of lymph node metastasis,grade of differentiation,tumour site, depth of tumour invasion and size(2,3,9,10,17,19,21).Associations of those factors with the NLNR may be explained as follows:(i)the reported association between the NLNR and lymph node metastasis may be possibly related to the larger size of metastatic nodes than that of non-metastatic ones(19,22),(ii)proliferation of lymph tissue around the ileocecal region and the longer removed bowel may be associated with the larger NLNR for the right than for the left colon and(iii)lateral dissection for rectal cancer should be associated with the increased NLNR.Although we have no direct evidence showing the contribution ofTable2.ContinuedVariables Number of cases Adjusted means of NLNR95%confidence interval P value90–9954219.5518.3220.87,0.001 100–19972919.3018.2020.48,0.001Trend,0.001 Scope of lymph node dissectionD2624412.31ReferenceD3*******.6617.2518.08,0.001 Grade of lymph node metastasisn0914315.00Referencen1461515.9615.5516.38,0.001 n2260116.4315.9016.97,0.001 n350614.6013.6715.600.429Trend,0.001 Reported chemotherapyNo587015.61ReferenceYes1099515.3815.0115.750.219 Number of registrations per institution1–1957512.2611.5013.07,0.001 20–49139714.9314.3115.590.518 50–99337414.7814.3315.250.806 100–199603014.72Reference200þ548917.2716.8117.75,0.001Trend,0.001Table3.Number of lymph node retrieved and that of metastatic lymph nodesNLNR Total Number of metastatic lymph nodes(mean)Right colon Left colon Rectum1–9 2.1 2.2 1.9 2.110–16 2.9 3.2 2.7 2.717–26 3.4 3.9 2.9 3.127þ 4.1 4.8 3.4 3.6 Overall 3.2 3.6 2.65 3.03Table4.Five-year overall survival rates of the patients with Stage II and III colorectal cancer according to the number of lymph nodes retrievedNLNR Stage II Stage IIIColon Rectum Colon Rectum1–981.071.664.253.910–1684.379.468.155.817–2684.682.968.854.827þ87.484.173.064.5Log-rank0.0006,0.00010.0003,0.0001Jpn J Clin Oncol2012;42(1)33at United Arab Emirates University on May 9, 2014/Downloaded frompathological technique to the NLNR (2),it is difficult to find other speculation other than differentiated precision of patho-logical examination for the variation in the NLNR by insti-tute (23).It remains unknown why age and gender were related to the NLNR as shown not only in the present but also in other studies (2,3,10).Studies have reported that the NLNR is influenced also by inflammation (21),diverticulum (21)and obesity (19,23,24).In this manner,the NLNR is determined most prominently by the extent of surgery,but is influenced to some degree by other factors.In the present study,patients with the larger NLNR had a significantly better prognosis,and the increased NLNR was associated with decreased risk of death.Furthermore,there was a positive correlation between the NLNR and the number of lymph node metastases (16).When more lymph nodes are examined,the opportunity for discovering a meta-static lymph node is believed to increase and the likelihood of under-staging may decrease.However,the finding that patients with the larger NLNR may have the better chance for survival,regardless of the number or metastasized nodes,should be more important (10).This cannot be explained based solely on stage migration.Because the present study is retrospective,a residual confounding effect from other related clinical factors should remain after statistical adjust-ment for those.We found no clear statistical evidence for interaction between the scope of lymph node dissection and the NLNR on the risk for death while the NLNR was signifi-cantly larger for D3than for D2.This finding may explain not only the valid staging but also a possible resection of occult metastasis to lymph nodes is potentially related to im-provement in prognosis.The effects of the NLNR on prognosis,however,cannot be explained based solely on the extent of surgery and pathological examination.The reason for this is that few surgeons would alter the extent of surgery and few patholo-gists would modify examination techniques based on gender,age and histological grade.Thus,a third factor,such as individual differences in lymph nodes as defined by host –tumour relationships,may be considered to be involved.Leibl et al.(21)hypothesized that ‘lymph node neo-formation’established by micro-vessels and peri-capillary lymphocyte accumulation may contribute to indi-vidual variation in the NLNR.Further studies are required to explain the association of individual difference in lymph node and prognosis.For proper assessment of resection and staging,the NLNR is a critical clinical issue to solve.With UICC and AJCC TNM classification,based on the recommendation by the Working Party Report to the World Congress of Gastroenterology in 1999(25,26),at least 12lymph nodes should be examined to ensure N0.However,this recommen-dation has few pieces of evidence.Swanson et al.(2)com-mented that 13or more of lymph nodes should be examined to be judged as N0,Wong (17)and Tepper (1)reported 14,Goldstein did 17(16)and Le Voyer et al.(10)did 20.In these studies,no coherent logic leading to the proper NLNR has been established.In the present study,a cut-off value for the NLNR was not determined.This finding was comparable to a large-scale SEER study (4),as well as a study conducted by individual institution (16).In addition to the NLNR,recent studies demonstrate that a ratio of metastatic lymph node to NLNR (i.e.lymph node ratio)has prognostic significance for estimating overall survival for the patients with colon cancer (27–29),although the cut-off value of the ratio remains undetermined.Therefore,at the present,it is required to dissect and examine as many regional lymph nodes as possible for reliable staging.Patients with the small NLNR following proper dissection and pathological examination should be considered at high risk of recurrence.The mainstay of treatment against CRC is surgery.Local control with proper surgery and appropriate adjuvant therapy based on reliable pathological staging are essential.The NLNR may be a potential indicator for postoperative adju-vant therapy.Table 5.Hazard ratio of patients with Stage II and III colorectal cancer according to NLNR analysed by Cox’s proportional model NLNRStage II Stage III Number of casesHazard ratio a 95%CI P valueNumber of cases Hazard ratio 95%CI P value1–92294 1.00Reference 1758 1.00Reference 10–1623330.630.530.76,0.0119490.910.80 1.030.1417–2622900.590.490.72,0.0119690.920.81 1.050.2327þ22260.460.370.57,0.0120460.750.650.86,0.01Trend ,0.01Trend ,0.01Increase by 1node0.9790.9720.985,0.010.9920.9890.996,0.01CI,confidence interval.aHazard ratio of patients with Stage II and III CRC according to NLNR analysed by Cox’s proportional model.34Number of lymph nodes retrievedat United Arab Emirates University on May 9, 2014/Downloaded fromFinally,we should mention the limitations of the study. Since the present study was retrospectively done and infor-mation on cases without follow-up was not used for the ana-lyses,the effect of the NLNR on the survival in the present study is possibly over-or underestimated. CONCLUSIONSIn the present analysis of a database containing information on patients with20lymph nodes retrieved on average,the NLNR was shown to be an important prognostic variable in Stage II and III CRC,especially in patients with Stage III, as well as the number of metastatic nodes.The NLNR was determined most prominently by the scope of lymph node dissection and by other patient’s profile and characteristics of tumour.A cut-off value for the NLNR was not found,and it is necessary to carry out proper dissection and examine as many lymph nodes as possible(4,9).Basic and clinical studies investigating the relation between the varying NLNR across individuals and oncological outcome should be reinforced.FundingJapanese Society for Cancer of the Colon and Rectum. Grant-in-Aid from the Ministry of Health,Labour and Welfare of Japan.Conflict of interest statementNone declared.References1.Tepper JE,O’Connell MJ,Niedzwiecki D,Hollis D,Compton C,Benson AB,3rd,et al.Impact of number of nodes retrieved on outcome in patients with rectal cancer.J Clin Oncol2001;19:157–63.2.Swanson RS,Compton CC,Stewart AK,Bland KI.The prognosis ofT3N0colon cancer is dependent on the number of lymph nodes examined.Ann Surg Oncol2003;10:65–71.3.Sarli L,Bader G,Iusco D,Salvemini C,Mauro DD,Mazzeo A,et al.Number of lymph nodes examined and prognosis of TNM stage II colorectal cancer.Eur J Cancer2005;41:272–9.4.Cserni G,Vinh-Hung V,Burzykowski T.Is there a minimum number oflymph nodes that should be histologically assessed for a reliable nodal staging of T3N0M0colorectal carcinomas?J Surg Oncol2002;81: 63–9.5.NCCN Clinical Practice Guidelines in Oncology.Colon Cancer.V.3.2010./professionals/physician_gls/PDF/colon.pdf6.National Cancer Institute.Colon Cancer(PDQ)./cancertopics/pdq/treatment/colon/healthprofessional7.Prandi M,Lionetto R,Bini A,Francioni G,Accarpio G,Anfossi A,et al.Prognostic evaluation of stage B colon cancer patients is improved by an adequate lymphadenectomy:results of a secondary analysis of a large scale adjuvant trial.Ann Surg2002;235:458–63.8.Caplin S,Cerottini JP,Bosman FT,Constanda MT,Givel JC.Forpatients with Dukes’B(TNM Stage II)colorectal carcinoma, examination of six or fewer lymph nodes is related to poor prognosis.Cancer1998;83:666–72.9.Johnson PM,Porter GA,Ricciardi R,Baxter NN.Increasing negativelymph node count is independently associated with improved long-termsurvival in stage IIIB and IIIC colon cancer.J Clin Oncol2006;24:3570–5.10.Le Voyer TE,Sigurdson ER,Hanlon AL,Mayer RJ,Macdonald JS,Catalano PJ,et al.Colon cancer survival is associated with increasingnumber of lymph nodes analyzed:a secondary survey of intergroup trialINT-0089.J Clin Oncol2003;21:2912–9.11.Inomata M,Yasuda K,Shiraishi N,Kitano S.Clinical evidences oflaparoscopic versus open surgery for colorectal cancer.Jpn J ClinOncol2009;39:471–7.12.Japanese Society for Cancer of the Colon and Rectum.JapaneseClassification of Colorectal Carcinoma.2nd English edition.Tokyo:Kanehara&Co.,Ltd2009.13.Kotake K,Honjo S,Sugihara K,Kato T,Kodaira S,Takahashi T,et al.Changes in colorectal cancer during a20-year period:an extendedreport from the multi-institutional registry of large bowel cancer,Japan.Dis Colon Rectum2003;46:S32–43.14.Armitage P,Berry G.Statistical Methods in Medical Research.3rd edn.Oxford:Blackwell Science1994.15.Therneau TM,Grambsch PM,Fleming RE.Martingale based residualsfor survival models.Biometrika1990;77:147–60.16.Goldstein NS.Lymph node recoveries from2427pT3colorectalresection specimens spanning45years:recommendations for aminimum number of recovered lymph nodes based on predictiveprobabilities.Am J Surg Pathol2002;26:179–89.17.Wong JH,Johnson DS,Hemmings D,Hsu A,Imai T,Tominaga GT.Assessing the quality of colorectal cancer staging:documenting theprocess in improving the staging of node-negative colorectal cancer.Arch Surg2005;140:881–6;discussion886–7.18.Hashiguchi Y,Hase K,Ueno H,Mochizuki H,Kajiwara Y,Ichikura T,et al.Prognostic significance of the number of lymph nodes examinedin colon cancer surgery:clinical application beyond simplemeasurement.Ann Surg2010;251:872–81.19.Baxter NN,Virnig DJ,Rothenberger DA,Morris AM,Jessurun J,Virnig BA.Lymph node evaluation in colorectal cancer patients:apopulation-based study.J Natl Cancer Inst2005;97:219–25.20.Joseph NE,Sigurdson ER,Hanlon AL,Wang H,Mayer RJ,MacDonald JS,et al.Accuracy of determining nodal negativity incolorectal cancer on the basis of the number of nodes retrieved onresection.Ann Surg Oncol2003;10:213–8.21.Leibl S,Tsybrovskyy O,Denk H.How many lymph nodes arenecessary to stage early and advanced adenocarcinoma of the sigmoidcolon and upper rectum?Virchows Arch2003;443:133–8.22.Mo¨nig SP,Baldus SE,Zirbes TK,Schro¨der W,Lindemann DG,Dienes HP,et al.Lymph node size and metastatic infiltration in coloncancer.Ann Surg Oncol1999;6:579–81.ler EA,Woosley J,Martin CF,Sandler RS.Hospital-to-hospitalvariation in lymph node detection after colorectal resection.Cancer2004;101:1065–71.24.Go¨ro¨g D,Nagy P,Pe´ter A,Perner F.Influence of obesity on lymphnode recovery from rectal resection specimens.Pathol Oncol Res2003;9:180–3.pton CC,Fielding LP,Burgart LJ,Conley B,Cooper HS,Hamilton SR,et al.Prognostic factors in colorectal cancer.College ofAmerican Pathologists Consensus Statement1999.Arch Pathol LabMed2000;124:979–94.26.Hammond ME,Fitzgibbons PL,Compton CC,Grignon DJ,Page DL,Fielding LP,et al.College of American Pathologists ConferenceXXXV:solid tumor prognostic factors-which,how and so what?Summary document and recommendations for implementation.CancerCommittee and Conference Participants.Arch Pathol Lab Med2000;124:958–65.27.Mammen JM,James LE,Molloy M,Williams A,Wray CJ,Sussman JJ.The relationship of lymph node dissection and colon cancer survival inthe Veterans Affairs Central Cancer Registry.Am J Surg2007;194:349–54.28.Wang J,Hassett JM,Dayton MT,Kulaylat MN.Lymph node ratio:rolein the staging of node-positive colon cancer.Ann Surg Oncol2008;15:1600–8.29.Chen SL,Steele SR,Eberhardt J,Zhu K,Bilchik A,Stojadinovic A.Lymph node ratio as a quality and prognostic indicator in stage IIIcolon cancer.Ann Surg2011;253:82–7.Jpn J Clin Oncol2012;42(1)35at United Arab Emirates University on May 9, 2014/Downloaded from。

Original ArticleSox11modulates neocortical development by regulating the proliferation and neuronal differentiation of cortical intermediate precursorsYongzhe Li,Jianjiao Wang,Yongri Zheng,Yan Zhao,Mian Guo,Yang Li,Qiuli Bao,Yu Zhang,Lizhuang Yang*, and Qingsong Li*Department of Neurosurgery,The2nd Affiliated Hospital,Harbin Medical University,Harbin150086,China*Correspondence address.Tel:þ86-451-86605040;E-mail:liqingsong1973@(Q.L.)/lizhuangyang@(L.Y.)Neural precursor cells play important roles in the neocor-tical development,but the mechanisms of neural progeni-tor proliferation,neuronal differentiation,and migration, as well as patterning are still unclear.Sox11,one of SoxC family members,has been reported to be essential for em-bryonic and adult neurogenesis.But there is no report about the roles of Sox11in corticogenesis.In order to in-vestigate Sox11function during cortical development,loss of function experiment was performed in this study. Knockdown of Sox11by Sox11siRNA constructs resulted in a diminished neuronal differentiation,but enhanced proliferation of intermediate progenitors.Accompanied with the high expression of Sox11in the postmitotic neurons,but low expression of Sox11in the dividing neural progenitors,all the observations indicate that Sox11induces neuronal differentiation during the neocor-tical development.Keywords neocortical development;neural stem cell; proliferation;neuronal differentiation;Sox11Received:March16,2012Accepted:April22,2012 IntroductionDuring the development of mammalian neocortex,neural precursor cells(NPCs),including radial glial cells and intermediate precursor(IP)cells,have been proven to play essential roles in neurogenic processes,such as prolifer-ation,neuronal differentiation,neuronal migration,lamin-ation,and patterning of neocortex[1–5].In the ventricular zone(VZ)of the cortical telencephalon where the cerebral cortex forms,radial glial cells divide symmetrically to maintain the neural progenitor population,or divide asym-metrically to give birth to one neural progenitor and one immature neuron;during the cortical neurogenesis,radial glial cells give rise to IP cells which migrate into the subventricular zone(SVZ)and the intermediate zone(IZ) where IP cells either directly differentiate into neurons and migrate radially into cortical plate(CP),or divide once or twice before differentiating into immature neurons[5–8]. Although many transcription factors and signaling path-ways have been proven to play important roles during these processes,there are still many uncovered mechanisms related to neocortical development.SRY-box(Sox)proteins are a family of transcription factors,which includes20family members defined as SoxA,B,C,D,E,F,G,H subpopulations[9,10].Among all Sox family members,SoxB and SoxC family members have been reported to be essential for NPC proliferation and differentiation[11–13].Sox11,together with Sox4and Sox12,belongs to SoxC subfamily,and is highly expressed in some subtypes of precursors and post-mitotic neurons [14,15].By in ovo electroporation in chicken neural tube, overexpression of Sox4or Sox11led to an induction of neuronal markers,whereas knockdown of Sox11expres-sion repressed the endogenous expression of these neuron-specific markers[14].In adult murine hippocampus,Sox11 has been demonstrated to be expressed in the doublecortin (DCX)-expressing precursor cells and immature neurons from adult neurogenic niche,and is required for neuronal differentiation in adult hippocampal neurogenesis[15,16]. Interestingly,Sox4and Sox11are also required for repro-gramming of astroglia into neurons[16].Furthermore, Sox4and Sox11have been identified as survival factors during spinal cord development[17].Although Sox11has been proven as an important tran-scription factor during the neural development and neurogen-esis,there is no report about its function during neocortical formation and development.In this study,in situ hybridiza-tion analysis of embryonic mouse brains showed that Sox11 was highly expressed in cortical post-mitotic neurons during embryonic development.Knockdown of Sox11expression by electroporating Sox11siRNA construct during cortical neurogenesis resulted in an impaired neuronal differentiationActa Biochim Biophys Sin2012,44:660–668|ªThe Author2012.Published by ABBS Editorial Office in association with Oxford University Press on behalf of the Institute of Biochemistry and Cell Biology,Shanghai Institutes for Biological Sciences,Chinese Academy of Sciences.DOI:10.1093/abbs/gms045.Advance Access Publication11June2012Acta Biochim Biophys Sin(2012)|Volume44|Issue8|Page660 by guest on June 25, 2014 / Downloaded fromand enhanced proliferation of IP cells,suggesting its critical roles in the cortical development.Materials and MethodsAnimalFifteen CD1wild-type mice(Laboratory Animal Center, Harbin Veterinary Research Institute)were used in this study.For embryo staging,the midday of vaginal plug for-mation was regarded as embryonic day0.5(E0.5).All animal procedures used in this study were conducted under the Animal Protocol of Harbin Medical University.In situ hybridizationMouse Sox11mRNA fragment was amplified according to Sox11mRNA open reading frame(ORF)using the follow-ing primer pair:50-GCTGGAAGATGCTGAAGGAC-30 and50-GCTGCTTGGTGATGTTCTTG-30(amplification product size:582bp).Digoxygenin(DIG)-labeled anti-sense mRNA probe was produced by in vitro transcription. In situ hybridization on cryosections prepared by Leica cryostat(Leica,Wetzlar,Germany)was performed as described[18].Briefly,the sections were hybridized with hybridization buffer(1ÂSSC,50%formamide,0.1mg/ml salmon sperm DNA solution,1ÂDenhart,5mM EDTA, pH7.5)at658C overnight and washed with wash buffer (1ÂSSC,50%formamide,0.1%Tween-20).After block-ing for2h with blocking buffer(1ÂMABT,2%blocking solution,20%heat-inactivated sheep serum),sections were labeled with anti-DIG antibody(1:1500;Roche Diagnostics,Mannheim,Germany)at48C overnight and washed with1ÂMABT and staining buffer(0.1M NaCl, 5mM MgCl2,0.1M Tris-HCl,pH9.5),stained with BM purple(Roche)at room temperature until ideal intensity. The images of in situ hybridization were collected using a Leica digital camera under a dissection scope(Leica).Sox11siRNA construct cloningThe siRNA oligos against Sox11were designed according to mouse Sox11ORF and the protocol of pSilencer TM 1.0-U6siRNA Expression Vector(Ambion,Austin,USA). After annealing,siRNA oligos were inserted into the pSilencer1.0vector.Sox11siRNA constructs were mixed equally as cocktail for further in utero electroporation.In utero electroporationIn utero electroporation was performed as described[19]. Briefly,electroporation was conducted on E13.5embryo cortex,and the tissues were harvested24h later at E14.5or 4days later at E17.5.Plasmid DNAs were prepared using EndoFree Plasmid Maxi Kit(Qiagen,Valencia,USA) according to the manufacturer’s instructions,and diluted to 2.5m g/m l with loading dye.DNA solution was injected into the lateral ventricle of the cerebral cortex,and electro-porated for five50-ms pulses at35V using an ECM830 electrosquareporator(BTX,Hawthorn,USA).The follow-ing constructs were used:pSilencer and pSilencer–mSox11–siRNA cocktail.Western blot analysisProtein samples were harvested from the brain tissues elec-troporated with Sox11siRNA and empty pSilencer vector(only the green region)4days after electroporation atE17.5,and lysed with RIPA lysis buffer with protease in-hibitor mixture at48C for1h.The protein samples wereboiled in sodium dodecyl sulfate(SDS)loading buffer for5min before loading onto10%SDS-polyacrylamide elec-trophoresis gel at20m g/lane.After running the gel,pro-teins were transferred onto nitrocellulose membrane.For immunoblotting,membrane was blocked with5%non-fatmilk powder in PBST[phosphate-buffered saline(PBS)with0.05%Tween-20,pH7.4]and incubated at48C over-night with primary anti-rabbit antibodies against Sox11(1:1000;Abcam,Cambridge,USA)and Actin(1:500;Sigma-Aldrich,St Louis,USA).After washing with PBST, membrane was incubated with specific horseradish peroxidase-conjugated secondary antibody for1h at room temperature and followed by extended washes with PBST. Immunoblot reactions were visualized using chemilumines-cent substrate on the BioMax light films(Kodak, Rochester,USA).BrdU incorporationTo access proliferation of NPCs in developing cortex,onedose of BrdU(50m g/g body weight)was administrated by intraperitoneal(i.p.)injection to pregnant mice1h before sacrifice.Tissue preparation and immunohistochemistryMouse brains were in utero electroporated at E13.5and collected1or4days after electroporation at E14.5orE17.5.After fixed in4%paraformaldehyde in PBS at48C overnight,brain tissues were dehydrated by incubating in30%sucrose in PBS,and embedded in Tissue-Tek (Sakura,Torrance,USA)and stored at2808C until use.Brains were coronally sectioned(10m m/section for immu-nohistochemistry;16m m/section for in situ hybridization)by Leica cryostat(Leica).For immunohistochemistry,sections were incubated in heated(95–1008C)antigen recovery solution(1mM EDTA,5mM Tris,pH8.0)for15220min for antigen re-covery,and cooled down for20230min on ice.Before applying antibodies,sections were blocked in10%normalgoat serum in PBS with0.1%Tween-20for1h.Sectionswere incubated with primary antibodies at48C overnightand visualized using goat anti-rabbit IgG-Alexa-Fluor-488,Sox11regulates neural precursor proliferation and neuronal differentiationActa Biochim Biophys Sin(2012)|Volume44|Issue8|Page661by guest on June 25, 2014/Downloaded fromgoat anti-chicken IgG-Alexa-Fluor-488and/or goat anti-mouse IgG-Alexa-Fluor-546(1:200;Molecular Probes, Grand Island,USA)for2h at room temperature. 40,6-diamidino-2-phenylindole was used to fluorescently stain cell nuclei.Images were captured using a Leica digital camera under a fluorescent microscope(Leica). Following primary antibodies were used:anti-chicken anti-body anti-GFP(1:1000;Abcam);anti-rabbit antibodies anti-GFP(1:1000;Rockland,Gilbertsville,USA), anti-Ki67(1:500;Abcam),anti-Tbr2(1:500,Abcam), anti-Tbr1(1:500;Abcam),anti-Ctip2(1:1000;Abcam); anti-mouse antibodies anti-5-bromo-20-deoxyuridine (BrdU)(1:50;DSHB,Iowa City,USA),NeuN(1:300; Millipore,Billerica,USA),and Cux1(1:300;Santa Cruz Biotechnology,Santa Cruz,USA).Statistical analysisAt least three brains and at least three sections from each brain were collected for antibody labeling,cell counting, and statistical analysis.Positive cells were counted in the same-width electroporated cortical region.The percentage of double-positive cells for cell type markers and green fluorescent protein(GFP)in total GFP-positive cells were presented in this study.Statistical analysis was performed by unpaired Student’s t-test,and data were presented as mean+standard deviation(SD).ResultsSox11is highly expressed in post-mitotic neuronsIn situ hybridization analysis was performed to investigate the expression pattern of Sox11during neocortical devel-opment.Results showed that at E12.5,Sox11was highly expressed in the early-born neurons in the cortex and gan-glionic eminence,while weakly expressed in the proliferat-ing regions VZ and SVZ[Fig.1(A,A0)].At E15.5,Sox11 is highly expressed in two layers among cortex—early-born neurons,which localized near the IZ,and late born neurons,which localized near to the top of the CP.Among the VZ and SVZ,Sox11was weakly expressed in these neurogenic regions[Fig.1(B,B0)].Sox11siRNA construct cocktail efficiently knockdowns Sox11expression in the developing cortexBased on Sox11mRNA coding region,two Sox11siRNA oligos were designed according to Ambion pSilencer TM 1.0-U6siRNA Expression Vector protocol[Fig.2(A)]and cloned into pSilencer1.0vector.To test knockdown effect, two Sox11siRNA constructs were mixed as cocktail,and electroporated into E13.5murine cortex.The empty pSilencer vector was electroporated as control.Four days after electroporation,the electroporated brain tissues(green regions under a fluorescence microscope)were harvested and proteins were extracted for western blot analysis [Fig.2(B)].Western blot results showed that Sox11siRNA cocktail largely inhibited the Sox11expression in the em-bryonic cortex after electroporation[Fig.2(C)]. Knockdown of Sox11enhances the proliferation of intermediate progenitorsThe effect of Sox11-knockdown on neural progenitor pro-liferation is tested by electroporating Sox11siRNA cocktail into E13.5mouse embryos,and brains were collected24h after electroporation.One hour before sacrifice of mice, one dose of BrdU was i.p.injected into mice to label the dividing progenitors[Fig.3(A)].Antibody against GFP was applied to label the electroporated cells;antibodies against BrdU(to labels the S phase of cell cycle)and Ki67 (to labels the all cell cycle except G0)were applied to immunostain proliferating cells;antibody recognizing Tbr2 was applied to mark IP cells.Results showed that compared with the brain tissues electroporated with empty vector,those electroporated with Sox11siRNA have significant higher percentage of both BrdUþ/GFPþcells(38.0%+4.0%versus30.6%+1.8%; n¼23)[Fig.3(B)]and Ki67þ/GFPþcells(44.3%+3.2% versus32.6%+2.8%;n¼23)[Fig.3(C)]in total GFPþcells,and the percentage of Tbr2þ/GFPþdouble-positive cells in total GFPþcells in Sox11-knockdown groupwas Figure1Expression pattern of Sox11during the mouse cortical development In situ hybridization analysis of Sox11expression was performed in embryonic day12.5(E12.5)murine cortex(A and A0), E15.5murine cortex(B and B0).Scale bar:500m m in A and B;100m m in A0and B0.Sox11regulates neural precursor proliferation and neuronal differentiationActa Biochim Biophys Sin(2012)|Volume44|Issue8|Page662 by guest on June 25, 2014 / Downloaded fromalso higher than empty vector group (45.8%+2.9%versus 35.2%+2.9%;n ¼23)[Fig.3(D)],suggesting that knockdown of Sox11results in an increased proliferation of IP.Since knockdown of Sox11caused enhanced neural progenitor proliferation,it will be very interesting to inves-tigate whether knockdown of Sox11could also affect neur-onal differentiation.Knockdown of Sox11impairs neuronal differentiation by keeping neural precursors in cell cycleBy electroporating Sox11siRNA construct cocktail into E13.5mouse brains,the effect of loss of function of Sox11on neuronal differentiation was analyzed 4days after elec-troporation [Fig.4(A)].Antibody against GFP was applied to label electroporated cells.Antibodies against different neuronal markers NeuN (to label total mature neurons),Tbr1(to label early-born neurons in layer VI),Ctip2(to label neurons in layer V),and Cux1(to label late-born pyr-amidal neurons of the upper layers II–IV)were applied to label mature neurons at different CP layers.The percen-tages of Ki67þ/GFP þ,and Tbr2þ/GFP þdouble-labeled cells in the total electroporated GFP þcells were also ana-lyzed to determine what percentage of electroporated cells still stays in progenitor status 4days after electroporation.Results showed that 4days after electroporation,murine brains electroporated with Sox11siRNA construct cocktail have remarkable less NeuN þ/GFP þ(17.7%+3.2%versus 28.3%+5.4%;n ¼12)[Fig.4(B)],Ctip2þ/GFP þ(11.3%+1.1%versus 14.3%+2.5%;n ¼12)[Fig.4(C)],Tbr1þ/GFP þ(16.9%+1.5%versus 22.8%+2.8%;n ¼12)[Fig.4(D)],and Cux1þ/GFP þ(44.3%+1.9%versus 48.1%+2.9%;n ¼12)[Fig.4(E)]double-positive cells in total GFP þcells compared with the control brains electropo-rated with empty vector.At the meantime,the percentages of Ki67þ/GFP þ(29.3%+3.4%versus 24.0%+2.4%;n ¼15)[Fig.5(A,B)]and Tbr2þ/GFP þ(26.8%+1.9%versus 22.5%+1.8%;n ¼14)[Fig.5(C,D)]double-positive cells in total GFP þcells from the Sox11siRNA-electroporating group were significantly higher than those in the control group.Interestingly,both in SVZ (25.2%+2.0%versus 22.0%+1.4%;n ¼14)[Fig.5(E)]and IZ (1.7%+1.7%versus 0.4%+1.0%;n ¼14)[Fig.5(E)],the percentages of Tbr2þ/GFP þcells in total GFP þcells from Sox11-knockdown group were higher than those in the control group.All these observations suggest that knockdown of Sox11results in an impaired neuronal differentiation,but increased neural progenitor pool.DiscussionSox proteins are a family of transcription factors with high-mobility group-type DNA-binding domain,and have been subdivided into eight subfamilies A–H [10,20].Among all Sox proteins,SoxB1proteins,including Sox1,Sox2,and Sox3,have been shown to participate in maintenance of undifferentiation status of embryonic and adult neural stem cells (NSCs)[21].SoxE proteins,including Sox8,Sox9,and Sox10,are expressed and play their roles inneuralFigure 2Sox11siRNA constructs knockdown the expression of Sox1124h after in utero electroporation in E13.5murine brains (A)Sox11siRNA oligo sequences.(B)The schematic description of in utero electroporation.(C)Western blot analysis of Sox11expression in brain tissues electroporated with Sox11siRNA constructs in comparison with the brain tissues electroporated with empty vector pSilencer.Antibody against actin was used to demonstrate equal protein loading.Sox11regulates neural precursor proliferation and neuronal differentiationActa Biochim Biophys Sin (2012)|Volume 44|Issue 8|Page 663by guest on June 25, 2014/Downloaded fromcrest cells [22,23].SoxC proteins,including Sox4,Sox11,and Sox12,are expressed in both central nervous system (CNS)and peripheral nervous system,and play roles in the establishment of neuronal properties [14,24–26].High-throughput RNA sequencing analysis of transcrip-tome profiling of embryonic and neonatal murine cortex revealed that SoxC proteins,Sox4and Sox11,were highly expressed in the embryonic stage,a neurogenic period,rather than postnatal stage [27].Immunohistochemical ana-lysis of SoxC proteins revealed that Sox11as well as Sox4were co-expressed in the neurogenic region in the embry-onic and adult murine brain,such as the SVZ of the lateralventricle,the subgranular zone of the hippocampal dentate gyrus,and the rostral migratory stream [15,16].Further in-vestigation showed that Sox11was expressed in DCX-expressing precursor cells and immature neurons in the adult neurogenic niches [15,16].In this study,to inves-tigate functions of Sox11in the neocortical development,the expression pattern of Sox11was analyzed during the embryonic neocortical development using in situ hybridiza-tion technique.At E12.5,the starting time point of cortical neurogenesis,Sox11was highly expressed in the upper layer of cerebral cortex,which is formed by the early-born immature neurons and some NPCs.At E15.5,the peakofFigure 3Knockdown of Sox11results in an enhanced proliferation of intermediate procursors (A)The schematic description of in utero electroporation for proliferation assay.E13.5mouse embryonic brains were electroporated and collected 24h later at E14.5.One hour before sacrifice,one dose of BrdU was i.p.injected to label the proliferating cells.(B–D)Photomicrographs and quantification of percentage of BrdU þ/GFP þ(yellow)double-positive cells (B),Ki67þ/GFP þ(yellow)double-positive cells (C),and Tbr2þ/GFP þ(yellow)double-positive cells (D)in total GFP þ(green)cells.Statistical significance was assessed by Student’s t -test.***P ,0.001.Scale bar:50m m.Sox11regulates neural precursor proliferation and neuronal differentiationActa Biochim Biophys Sin (2012)|Volume 44|Issue 8|Page 664by guest on June 25, 2014/Downloaded fromcortical neurogenesis,Sox11was highly expressed in the top of the CP,formed by late-born neurons,lower layer of CP,formed by early-born neurons,as well as IP,formed by immature neurons and IP cells.All these observations indicated that Sox11might be involved in corticalneurogensis.Figure 4Knockdown of Sox11results in an impaired neuronal differentiation (A)The schematic description of in utero electroporation for neuronal differentiation assay.E13.5mouse embryonic brains were electroporated and collected 4days later at E17.5.(B–E)Photomicrographs and quantification of percentage of NeuN þ/GFP þ(yellow)double-positive cells (B),Tbr1þ/GFP þ(yellow)double-positive cells (C),Ctip2þ/GFP þ(yellow)double-positive cells (D),and Cux1þ/GFP þ(yellow)double-positive cells (E)in total GFP þ(green)cells.Statistical significance was assessed by Student’s t -test.***P ,0.001.Scale bar:100m m.Sox11regulates neural precursor proliferation and neuronal differentiationActa Biochim Biophys Sin (2012)|Volume 44|Issue 8|Page 665by guest on June 25, 2014/Downloaded fromPrevious studies showed that as a transcription factor induced by proneural protein basic helix-loop-helix as NPCs develop into immature neurons,Sox11plays important role in embryonic and adult neurogenesis [14].Overexpressing Sox11in the VZ of chicken spinal cord resulted in a strong ectopic expression of neuronal markers Tuj1and MAP2,while knocking down the expression of Sox11showed op-posite effects,leading to a suppression of endogenous ex-pression of multiple neuronal markers [14].All these indicated the essential roles of Sox11in the neurogenesis.In another study,overexpressing Sox11in in vitro cultured adult NSCs by retrovirus resulted in an increased generation of DCX þimmature neurons,and promoted neurogenesis of adult NSCs [15].Furthermore,using transgenic mouse model as well as retrovirus system,Lie’s group hasdemonstrated the roles of SoxC,including Sox4and Sox11,in the neurogenesis in adult hippocampus [16].Deletion of Sox4/Sox11inhibited the neurogenesis of adult hippocampal NSCs both i n vitro and in vivo ,whereas ectopic expression of Sox4and Sox11promoted neurogenesis of cultured adult hippocampal precursors in vitro [16].Lie’s group observed that under in vitro culture condition,deletion of Sox4and Sox11led to a decreased proliferation of differentiating adult NSCs,while in vivo in adult hippocampus,many Sox4/Sox11-deficient cells also expressed stem/precursor marker Sox2,but not proliferation marker Ki67,IP marker Tbr2,nor astroglial marker glial fibrillary acidic protein (GFAP)[16].They suggested that Sox4/Sox11-deleted cells remained a non-proliferative,GFAP 2,but Sox2-expressing status,but cell fate of these cells remainedunclear.Figure 5Knockdown of Sox11keeps neural progenitors stay in proliferation status during neurogenesis.(A,B)Photomicrographs and quantification of percentage of Ki67þ/GFP þ(yellow)double-positive cells in total GFP þ(green)cells.(C,D)Photomicrographs and quantification of percentage of Tbr2þ/GFP þ(yellow)double-positive cells in total GFP þ(green)cells.(E)Quantification of percentage of Tbr2þ/GFP þ(yellow)double-positive cells in total GFP þ(green)cells at the IZ and the SVZ.(F)Summary of Sox11functions during the embryonic cortical development.Radial glial cell was shown as green;IP cell was shown as yellow;neuron was shown as red.Statistical significance was assessed by Student’s t -test.*P ,0.05;***P ,0.001.Scale bar:100m m.Sox11regulates neural precursor proliferation and neuronal differentiationActa Biochim Biophys Sin (2012)|Volume 44|Issue 8|Page 666by guest on June 25, 2014/Downloaded fromAlthough the function of SoxC proteins,including Sox11,in neurogenesis has been demonstrated in different species and different neural regions,there is no report about roles of Sox11in neocortical development.In this study,loss of function of Sox11has been performed using siRNA and in utero electroporation technique to investigate roles of Sox11in neurogenesis during corticogenesis. Similar to the previous reports,knockdown of Sox11 caused a diminished neuronal differentiation:4days after knockdown of Sox11,the percentages of different neuronal markers,including NeuN,Tbr1,Ctip2,and Cux1,positive cells in total electroporated cells were significantly decreased compared with the control group.Interestingly, more Ki67þand Tbr2þcells were observed in Sox11-knockdown cells4days after electroporation. Analysis of the percentage of Tbr2þ/GFPþcells from the SVZ and the IZ separately showed that not only in SVZ, but also in the IZ,Sox11-knockdown cells had more Tbr2þco-expressed cells than empty vector-electroporated cells.Proliferation analysis was performed1day after elec-troporation.Ki67þ,Tbr2þ,as well as S-phase marker BrdUþcells were counted,which showed knockdown of Sox11led to an enhanced proliferation of IP cells in Sox11-ablated cells in the developing cortex.All these observations indicated that during neocortical formation, suppressing the expression of Sox11could keep Sox11-abalated cells stay in the proliferating and NPC status[Fig.5(F)].Our observation of the function of Sox11on neurogenesis is consistent with the results of pre-vious studies from different groups.But regarding the effects of Sox11on proliferation and cell fate,our results were not completely identical with theirs,which might be due to different cell origins,different developing stages, and detection methods.It has been shown that SoxC pro-teins,Sox4,Sox11,and Sox12,are expressed complemen-tary to stem/precursor markers SoxB1proteins,Sox1, Sox2,and Sox3:all three SoxB1proteins are expressed in most neural progenitors in developing and adult CNS,and maintain them in a precursor state.SoxC proteins are mostly expressed in post-mitotic differentiating neurons, and induce expression of different neuronal markers Tuj1 and MAP2[14,28].Further investigation using chromatin immunoprecipitation technique revealed that SoxC proteins share a high number of target genes with SoxB1proteins, suggesting that SoxC proteins and SoxB1proteins modu-late cell fate of NPCs via a common sets of target genes involved in CNS development,including Sox proteins, Notch signaling molecules,Wnt protein,transforming growth factor-b family members,fibroblast growth factors and Pax transcription factors[28].Further real-time PCR analysis revealed that among these shared target genes, ectopic expression of SoxB1protein upregulated the ex-pression of NPC genes,while forced expression of SoxC resulted in an increased expression of neuronal genes,and competitive binding activities of SoxB1and SoxC balancethe transition between NPCs and post-mitotic neurons[28]. Together with the findings from Lie’s group—many Sox4/Sox11-deficient cells also expressed stem/precursor markerSox2—knockdown of Sox11maintain NPCs in undifferenti-ated state.Further investigation of spatiotemporal expressionand functions of Sox11at different neural regions and differ-ent developing periods is expected to allow insights intoSox11-regulating neural development and neural disorders. FundingThis work was supported by a grant from the DoctoralFund of the Second Affiliated Hospital of Harbin Medical University(1352008-17).References1Gage FH.Mammalian neural stem cells.Science2000,287:1433–1438.2Go¨tz M and Barde YA.Radial glial cells defined and major intermediatesbetween embryonic stem cells and CNS neurons.Neuron2005,46:369–372.3Merkle FT and Alvarez-Buylla A.Neural stem cells in mammalian devel-opment.Curr Opin Cell Biol2006,18:704–709.4Temple S.The development of neural stem cells.Nature2001,414:112–117.5Go¨tz M and Huttner WB.The cell biology of neurogenesis.Nat Rev MolCell Biol2005,6:777–788.6Gupta A,Tsai LH and Wynshaw-Boris A.Life is a journey:a geneticlook at neocortical development.Nat Rev Genet2002,3:342–355.7Nadarajah B and Parnavelas JG.Modes of neuronal migration in the devel-oping cerebral cortex.Nat Rev Neurosci2002,3:423–432.8Kriegstein AR and Noctor SC.Patterns of neuronal migration in the em-bryonic cortex.Trends Neurosci2004,27:392–399.9Guth SI and Wegner M.Having it both ways:Sox protein function betweenconservation and innovation.Cell Mol Life Sci2008,65:3000–3018.10Bowles J,Schepers G and Koopman P.Phylogeny of the SOX family of developmental transcription factors based on sequence and structural indi-cators.Dev Biol2000,227:239–255.11Wegner M.SOX after SOX:SOXession regulates neurogenesis.Genes Dev2011,25:2423–2428.12Bylund M,Andersson E,Novitch BG and Muhr J.Vertebrate neurogenesis is counteracted by Sox1-3activity.Nat Neurosci2003,6:1162–1168.13Sandberg M,Ka¨llstro¨m M and Muhr J.Sox21promotes the progression of vertebrate neurogenesis.Nat Neurosci2005,8:995–1001.14Bergsland M,Werme M,Malewicz M,Perlmann T and Muhr J.The estab-lishment of neuronal properties is controlled by Sox4and Sox11.GenesDev2006,20:3475–3486.15Haslinger A,Schwarz TJ,Covic M and Chichung Lie D.Expression of Sox11in adult neurogenic niches suggests a stage-specific role in adultneurogenesis.Eur J Neurosci2009,29:2103–2114.16Mu L,Berti L,Masserdotti G,Covic M,Michaelidis TM,Doberauer K and Merz K,et al.SoxC transcription factors are required for neuronal dif-ferentiation in adult hippocampal neurogenesis.J Neurosci2012,32:3067–3080.17Thein DC,Thalhammer JM,Hartwig AC,Crenshaw EB,III,Lefebvre V, Wegner M and Sock E.The closely related transcription factors Sox4andSox11regulates neural precursor proliferation and neuronal differentiationActa Biochim Biophys Sin(2012)|Volume44|Issue8|Page667by guest on June 25, 2014/Downloaded from。

Pictures of Appetizing Foods Activate Gustatory Cortices for Taste and RewardW.Kyle Simmons 1,Alex Martin 2and Lawrence W.Barsalou 11Department of Psychology,Emory University,Atlanta,GA 30322,USA and 2Cognitive Neuropsychology Section,Laboratory of Brain and Cognition,National Institute of Mental Health,Bethesda,MD,USAIncreasing research indicates that concepts are represented as distributed circuits of property information across the brain’s modality-specific areas.The current study examines the distributed representation of an important but under-explored category,foods.Participants viewed pictures of appetizing foods (along with pictures of locations for comparison)during event-related pared to location pictures,food pictures activated the right insula/operculum and the left orbitofrontal cortex,both gustatory processing areas.Food pictures also activated regions of visual cortex that represent object shape.Together these areas contribute to a distributed neural circuit that represents food knowledge.Not only does this circuit become active during the tasting of actual foods,it also becomes active while viewing food pictures.Via the process of pattern completion,food pictures activate gustatory regions of the circuit to produce conceptual inferences about taste.Consistent with theories that ground knowledge in the modalities,these inferences arise as reenactments of modality-specific processing.Keywords:concepts,fMRI,insula/operculum,knowledge,orbitofrontal cortex IntroductionHow are concepts for everyday objects represented in the brain?Based on accumulating lesion and neuroimaging evi-dence,an object concept is represented as a distributed circuit of property representations across the brain’s modality-specific areas (Martin,2001;Martin and Chao,2001;Thompson-Schill,2003).On encountering a physical object,relevant modalities represent it during perception and action.As the object is processed,association areas partially capture property informa-tion on these modalities,so that this information can later be reactivated during conceptual processing,when the object is absent (Damasio and Damasio,1994;Simmons and Barsalou,2003).Although these conceptual reenactments share import-ant commonalties with mental imagery,there are also important differences.Mental imagery typically results from deliberate attempts to construct conscious vivid images in working memory.In contrast,the perceptual reenactments that underlie conceptual processing often appear to lie outside awareness,resulting instead from relatively automatic and implicit pro-cesses.Of primary interest,these reenactments occur in responses to words and other symbols,and play central roles in the representation of conceptual knowledge (Barsalou,1999,2003a,b;Barsalou et al.,2003a,b).The category of tools illustrates the distribution of property representations across modality-specific brain areas.When people use a hammer,a distributed set of brain areas becomes active to represent the hammer’s properties,including its visual form (ventral occipitotemporal cortex),the physical actionsused to manipulate it (ventral premotor cortex and intraparietal sulcus),and the visual motion that results (middle temporal gyrus)(Beauchamp et al.,2002;Chao et al.,1999;Chao and Martin,2000;Damasio et al.,2001;Grafton et al.,1997;Handy et al.,2003;Johnson-Frey,2004;Martin et al.,1995;Perani et al.,1995).As just described,the brain’s association areas capture this distributed set of modality-specific states for later concep-tual use.On subsequent occasions,when no hammers are present,reenactments of these states represent hammers con-ceptually (e.g.during language comprehension and thought).In the experiment reported here,we explored the distributed property account for the category of foods.Foods constitute a central category for humans,not only in perception and action,but in higher cognition (Ross and Murphy,1999).Previous research on food concepts has addressed the visual properties of fruits and vegetables,relative to the visual properties of other object categories (McRae and Cree,2002).Here,we focus instead on the tastes of high-caloric,high-fat processed foods,such as cheeseburgers and cookies (see Fig.1).We focus on taste properties because the tastes of foods are at least as important as their visual appearances.We focus on processed foods because they are central to the modern diet and because they are associated with strong gustatory and appetitive responses that underlie how people select and consume them.If a distributed circuit of property information represents food knowledge,then viewing a food picture should not only activate brain areas that represent visual properties of the pictured food,but should also activate brain areas that represent how the food is likely to taste and how rewarding it would be to eat.Once one part of the distributed circuit becomes active by viewing the picture,the remainder should become active via the conceptual inference process of pattern completion across the circuit.Given the central role that such inferences play in normal food selection and consumption,it is essential to understand their bases in the brain.Furthermore,given the extensiveness of eating disorders,obesity and other food-related problems,it is important to understand how people generate taste and reward inferences to the broad array of food representations available in modern culture.We presented pictures of food and non-food entities (loca-tion pictures)to subjects undergoing event-related fMRI and predicted that a distributed circuit of brain areas would become active to represent the visual and gustatory properties of the pictured foods.Regarding the visual properties of foods,a large literature demonstrates that ventral temporal regions underlie the representation of objects’visual form properties (Ishai et al.,1999,2000).Thus,we expected regions of the inferior temporal and fusiform gyri to respond to the distinctive visual properties of the pictured foods.Analogously,location pictures shouldÓThe Author 2005.Published by Oxford University Press.All rights reserved.For permissions,please e-mail:journals.permissions@Cerebral Cortex October 2005;15:1602--1608doi:10.1093/cercor/bhi038Advance Access publication February 9,2005at Indian Institute of Science Education and Research, Kolkata (IISER-K) on February 28, 2011 Downloaded fromactivate parahippocampal gyrus,given that this region responds to the visual-spatial properties characteristic of buildings and landmarks(Aguirre et al.,1998;Epstein and Kanwisher,1998; Epstein et al.,1999).Most importantly,the current study attempted to demonstrate that pictures of visual objects,in this case foods,can produce taste inferences.If the distributed account of concept represen-tation is correct,then multiple modality-specific regions should become active when people represent foods conceptually.Not only should visual areas become active to represent a food’s unique visual properties,gustatory areas should become active to represent how the food tastes.Once people access knowledge for a pictured food,an inference is produced about how it tastes. Even though people are not actually tasting the food,their gustatory system becomes active to represent this inference. Specifically,we predicted that simply viewing pictures of appetizing foods(relative to locations)should activate two brain regions that commonly respond to actual taste stimuli in psychophysical neuroimaging studies(Francis et al.,1999;de Araujo et al.,2003a,b;O’Doherty et al.,2001b).Thefirst area, a region in the insula/operculum,is known to represent how foods actually taste(Rolls et al.,1988;Rolls and Scott,2003; Scott et al.,1986).The second area,a region in orbitofrontal cortex(OFC),is known to represent the reward values of tastes (Gottfried et al.,2003;Rolls et al.,1989).Here we demonstrate that simply viewing pictures of processed foods activates both brain regions in much the same way that taste stimulants do in psychophysical studies.Materials and MethodsSubjectsNine right-handed,native-English-speaking volunteers from the Emory University community participated in the scanning study(six female and three male;age range,18--45years).All participants completed a health questionnaire prior to scanning and none of the participants indicated a history of neurological problems.In accordance with protocols prescribed by Emory University’s Institutional Review Board,all partici-pants read and signed an informed consent document describing the procedures and possible risks.Sixteen native-English-speaking volunteers from the Emory commu-nity participated in the stimulus selection study(ten female and six male;age19--46years).None of these volunteers participated in the later brain imaging experiment.As with the imaging participants,all participants read and signed an informed consent document describing the procedures and possible risks in accordance with protocols prescribed by Emory University’s Institutional Review Board. Experimental DesignBefore beginning the brain imaging phase of the study,32types of foods and35types of locations were selected as candidate materials.The foods(e.g.cheeseburger,spaghetti,cookie,etc.)in the list were chosen because they are all encountered frequently in American society.In addition,only processed foods that are relatively high in fat and calories were used.No fruits or vegetables were included.The locations(e.g. house,mall,school,etc.)in the list were chosen because they are all types of places that participants in the study might visit frequently. The foods and locations were equated for familiarity by having volunteers(none of whom participated in the brain imaging experi-ment)provide familiarity ratings for the35types of locations,and32 types of foods.Ratings were made on a1--7scale,with1indicating that a type of food or location was completely unfamiliar and7indicating that it was extremely familiar.Based on these ratings,15food and15 location types were selected such that no reliable familiarity differences existed between the two groups of stimuli.Between six and ten pictures for each type of food and location were then collected.A group of16participants viewed all259pictures and rated each for how typical it was of its respective food or location type.Ratings were made on a1--7scale,with1indicating that a picture was not at all typical of its food or location type and7indicating that it was very typical.For each type of food or location,the three most typical pictures were selected for use in the imaging study,thus yielding a total of90picture stimuli(45foods,45locations)equated for typicality.All of the food and location pictures depicted non-unique entities that would not be individually recognizable to the participants.Finally,23location pictures and22food pictures were randomly selected to create phase-scrambled images that were presented during scanning asfiller items(see Fig.1).During scanning,participants viewed food,location and scrambled pictures.For each picture,participants used a response pad to provide yes/no judgments as to whether it was the same or different as the preceding picture.The pictures were presented in the center of the screen for2s each.Interspersed among picture presentations were variable(‘jittered’)interstimulus intervals(mean=5.7s,range=2--20s) that were included to optimize estimation of the event-related fMRI response.During these interstimulus intervals,participants saw afix-ation cross presented in the center of the screen.Participants were instructed that when they saw thefixation cross they should continue attending to the screen and prepare for the next picture presentation. Prior to beginning data collection,participants performed an abbre-viated practice run to insure that they understood the task instructions. Functional data were collected in three scanning runs.The trial lists for the three runs were counterbalanced across participants.During each run,participants saw16food and16location pictures.Fifteen picture presentations from each category were novel pictures,while one picture was repeated to maintain the participants’attention to the picture repetition detection task.In other words,one location picture and one food picture was repeated in each scanning run.Across the three scanning runs,each subject saw three food picture repetitions and three location picture repetitions.Subjects were told in advance that repeated stimuli would occur in each run.Knowing this and given that the repeated stimuli occurred infrequently,this task requires subjects to pay close attention to each picture presentation to insure that they did not miss a repetition trial.The data from the repetition trials in each run were not analyzed given that they were only included to ensure that participants remained attentive to the task.Subjects were highly accurate at repetition detection(Mean correct=98.8%,SD= 0.94).Each5min8s run consisted of4min48s of the repetition detection task,followed by an additional20s restperiod.Figure1.Examples of location,food,and scrambled image stimuli.Cerebral Cortex October2005,V15N101603at Indian Institute of Science Education and Research, Kolkata (IISER-K) on February 28, Downloaded fromImage Acquisition and AnalysisPictures were back-projected onto a screen located at the head of the scanner and were viewed through a mirror mounted on the head coil. Stimulus presentation and response collection was controlled using Presentation software(v.0.70,).In each of the three imaging runs,154gradient echo recalled MR volumes depicting BOLD contrast were collected with a3T Siemens Trio scanner.Each volume consisted of34contiguous,2mm thick slices in the axial plane(T E=30ms,T R=2000ms,flip angle=90°,FOV= 192mm2,64364matrix).Voxel size at acquisition was33332mm, but was33333mm after spatial normalization.Prior to statistical analyses,image preprocessing was conducted in SPM99(Wellcome Department of Neurology,UK,http://www.fil.ion. ).To reduce motion-related signal changes between volumes, each participant’s scans were realigned and resliced using sinc in-terpolation.Volumes were then normalized to a template EPI scan and finally smoothed in the axial plane using a6mm isotropic Gaussian kernel.Subsequent statistical analyses were also conducted using SPM99. First,individual subjects’data were analyzed using multiple regression. For each subject,event-related changes in neural activity were modeled using afinite impulse response model corresponding to picture stimuli presentation and convolved to the standard SPM hemodynamic re-sponse function.Interstimulusfixation periods having variable durations served as the signal baseline.Global effects were removed by pro-portional scaling and the data were low-passfiltered.Condition effects at the subject level were then assessed with orthogonal contrasts comparing neural activity for food and location pictures.These contrast images,one for each participant,were then analyzed in a second-level random effects analysis of the foods--locations and locations--food contrasts using one sample t-tests.A statistical significance threshold of P<0.005(uncorrected for multiple comparisons)and a spatial extent threshold of at least seven contiguous voxels(corresponding to P<0.05 uncorrected)was used in the random effects analyses.There are at least two reasons why the use of uncorrected P-values in the present study is warranted.First,the activations reported here were identified using random effects analyses which take into account both within-and between-subjects variance.Not only does this allow the results to be generalized to the population from which subjects were drawn,but it also makes the analyses inherently robust statistically. Secondly,based on much previous research reported in the literature (see Introduction and Discussion),we started with a priori hypotheses that the insula/operculum and OFC would be active in the food--location contrasts.Additionally,given that both food and location pictures depicted common objects,both conditions should activate regions in the ventral temporal cortex known to represent objects’visual form properties.More specifically,however,we predicted that the fusiform/ parahippocampal gyrus would be active in the locations--foods contrasts. To be reported here as significant,any other areas of activity would need to be active at the P<0.05level with correction for multiple comparisons.No other areas reached this level of statistical significance. ResultsViewing food pictures for two s in a simple picture-matching task activated gustatory cortex.Specifically,food pictures, relative to location pictures,activated a region of the right insula/operculum,an area that psychophysical research has shown represents the tastes of foods(extent threshold,P= 0.004;see Table1and Fig.2).Importantly,this region was not only significantly more active for food pictures than for location pictures,but it was also reliably activated relative to thefixation baseline(one-tailed,P=0.033).In addition,food pictures,relative to location pictures, activated two regions in the left OFC that psychophysical research has shown represents the reward values of tastes. One of these regions was located in the lateral portion of the OFC(extent threshold,P=0.05;Fig.3);the other,located more superiorly,stretched into the anterior aspect of the cingulate cortex(extent threshold,P=0.01).While the lateral OFC region was reliably activated relative to thefixation baseline(P< 0.001),the more superior OFC/anterior cingulate region was not(one-tailed,P=0.155).Viewing food pictures,relative to location pictures,also produced robust activity in ventral occipitotemporal cortex, bilaterally.Two of these areas were located in the right hemisphere;one extending from the inferior occipital gyrus forward into the inferior temporal gyrus(extent threshold,P= 0.02)and the other located more anteriorly in the inferior temporal gyrus(extent threshold,P=0.035).Additional activity was observed in the left hemisphere,stretching from inferior occipital gyrus into the fusiform and inferior temporal gyri (extent threshold,P=0.001).In addition to producing significantly more activity than location pictures,food pictures reliably activated each of the ventral temporal areas above the signal baseline(P<0.0001).In constrast,and consistent with previous reports(Aguirre et al.,1998;Epstein and Kanwisher,1998;Epstein et al.,1999), location pictures,relative to food pictures,produced bilateral activity extending from the medial portion of the fusiform gyrus into parahippocampal gyrus(see Table1).Activity in these regions was not only greater for locations than foods,but was also reliably activated relative to the signal baseline(P<0.001 for both hemispheres).DiscussionThesefindings support the hypothesis that a distributed circuit of brain regions represents conceptual knowledge about foods.As Figure4a,b illustrates,viewing food pictures activated two brain regions that lie in close proximity to gustatory regions active during psychophysical studies of taste perception (Francis et al.,1999;de Araujo et al.,2003a,b;O’Doherty et al.,2001b).As Figure4a illustrates,food pictures activated the insula/operculum very near regions that become active when people actually taste glucose,sucrose,salt,or umami.As Figure4b similarly illustrates,food pictures also activate OFC very near regions that become active when people experience taste stimuli directly.The close proximity of the regions active for food pictures to well-established gustatory areas suggests that food pictures automatically activate gustatory areas to produce conceptual inferences about taste properties.The two taste areas observed here are associated with dif-ferent functions in the gustatory system.The insula/operculum receives projections from the ventroposterior medial nucleus of Table1Regions showing differential responses to food and location picturesContrast Side/location MNI coordinates Peak T Px y zFoods[locations R insula36ÿ69 5.92\0.001 L OFCÿ2133ÿ18 6.60\0.001L OFC/anterior cingulateÿ1845ÿ6 5.08\0.001aR inferior temporal gyrus48ÿ45ÿ12 5.05\0.001R inferior temporal gyrus48ÿ66ÿ9 5.99\0.001L fusiformÿ48ÿ60ÿ18 4.690.001 Locations[foods L fusiformÿ21ÿ39ÿ1214.50\0.001 R fusiform27ÿ42ÿ159.61\0.001 L,left;R,right.a While this region was significantly active for food pictures relative to location pictures,it was not reliably active relative to thefixation baseline.1604Food Pictures Activate Gustatory Cortex Simmons et al.at Indian Institute of Science Education and Research, Kolkata (IISER-K) on February 28, Downloaded fromthe thalamus (Rolls and Scott,2003),the main subcortical processing area for gustatory input,and has been associated with taste per se .The OFC,in contrast,receives projections from the insula/operculum (Rolls and Scott,2003)and has been associated with the reward values of specific tastes.Specifically,electrophysiological studies in monkeys show that the firing rates of neurons in insula/operculum are not modulated by hunger and satiety,suggesting that they represent taste in-dependent of reward (Rolls et al.,1988).Conversely,the firing rates of neurons in OFC are modulated by hunger and satiety,suggesting that they represent the current reward value of tastes (Rolls et al.,1988).Thus,when a monkey is hungry,the firing rate of OFC neurons is high,given that the reward value of food is high.Similarly in humans,greater activation occurs in gustatory OFC before participants are satiated than after (Gottfried et al.,2003).Taste reward areas are located in a different OFC region than the reward areas for other stimuli (Elliot et al.,2000;O’Doherty et al.,2001a;Rolls,2000).For example,the caudal OFC responds to olfactory rewards (de Araujo et al.,2003b;Zaldand Pardo,2000;O ngu r et al.,2003),whereas the inferiormedial OFC responds to abstract rewards (e.g.money)(O’Doh-erty et al.,2001a).Interestingly,the inferior medial OFC hasa markedly different cytoarchitectonic structure than the morelateral aspect of the OFC where taste activations occur (O ngu ret al.,2003).Thus,the OFC areas active in the present study appear to represent the reward value of tastes,rather than reward in general.As Rolls (2000,p.285)notes,‘it is important to realize that it is not just some general ‘‘reward’’that is represented in the oritofrontal cortex,but instead a very detailed and information-rich representation of which particu-lar reward or punisher is present’.Laterality of the Taste ActivationsFood pictures activated the right insula/operculum,and the left OFC.Our a priori prediction was that food pictures would activate both regions bilaterally.Examination of the psycho-physical taste literature,however,clarifies the laterality of our results.First,consider the insula/operculum.Although many psychophysical taste studies observe bilateral activity in this area,the response is typically stronger and more spatially extensive on the right (Small et al.,1999).This may explain why we only found right insula/operculum activation for food pictures.Indeed,lowering the cluster size threshold in our random effects analysis (but not the P -value threshold)revealed significant activity in a region of the left frontaloperculumFigure 2.Viewing food pictures elicits activity in insula/operculum.A high-resolution anatomical scan showing activity in right insula/operculum associated with viewing pictures of food items.The bar graph displays the average percent signal change in the right insula/operculum cluster for all nine subjects during a period between 4and 14s post-stimulus.The y -axis indicates percent signal change relative to signal baseline,with error bars representing ±1SEM of the subjects.The data shown in the bar graph were obtained in the random effects contrast of foods [locations with P \0.005.Cerebral Cortex October 2005,V 15N 101605at Indian Institute of Science Education and Research, Kolkata (IISER-K) on February 28, 2011 Downloaded from(–48,21,12)that is commonly activate in psychophysical taste studies (Small et al.,1999).Although this cluster of activity was smaller in magnitude and size relative to the activation seen on the right,it suggests that our findings are consistent with the general trend in the psychophysical taste literature for greater insula/operculum activation in the right hemisphere than in the left.With respect to the OFC,we found significant activations only on the left.It is noteworthy that studies in the psycho-physical taste literature are inconsistent with regard to later-ality,with bilateral activity reported only in approximately half of the studies.Again,lowering the cluster size threshold (but not the P -value threshold)on the random effects analysis revealed significant activity in the right OFC (15,45,–3)in nearly the identical location as seen on the left (–18,45,–6).Perhaps the best explanation,however,for why we observe activity in the left OFC comes from a recent finding by Kringelbach et al.(2003).These researchers identified an area in the left OFC where activity was correlated with subjects’ratings of taste pleasantness.Interestingly,the area theyidentified is approximately one centimeter from the activity we observed in the lateral OFC.Given that we only showed pictures of highly appetizing foods,it makes sense that we would observe activity very near the left OFC region that tracks taste pleasantness.ConclusionThe findings reported here indicate that the gustatory system produces taste responses to pictures of foods,not just to actual foods.Other studies have reported similar results.A previous neuroimaging study on pictures of foods found activation in areas near those observed here (insula and OFC),but using a blocked design with fixed-effects analyses (Killgore et al.,2003).Indeed,still other research has found that even words for tastes activate taste areas (Simmons,W.K.,Pecher,D.,Hamann,S.B.,Zeelenberg,R.and Barsalou,L.W.,under review;see Fig.4b ).In general,pictures and words appear to activate property inferences for food tastes and rewards,thus grounding conceptual knowledge in modality-specific brainareas.Figure 3.Viewing pictures of foods elicits activity in left OFC.A high-resolution anatomical scan showing activity in left OFC associated with viewing pictures of food items.The bar graph on the left displays the average percent signal change in the left OFC for all nine subjects during a period between 4and 14s post-stimulus.The bar graph on the right displays the average percentage signal change in the left OFC/anterior cingulate cluster for all nine subjects during a period between 4and 14s post-stimulus.The y -axis indicates percent signal change relative to signal baseline,with error bars representing ±1SEM of the subjects.The data shown in the bar graphs were obtained in the random effects contrast of foods [locations with P \0.005.1606Food Pictures Activate Gustatory CortexSimmons et al.at Indian Institute of Science Education and Research, Kolkata (IISER-K) on February 28, 2011 Downloaded fromIn the experiment reported here,taste inferences arose even when subjects performed fast superficial processing of food stimuli.Subjects were required to only assess whether the current picture exactly matched the previous picture,each presented for only 2s.No categorization or other form of conceptual processing was required.Furthermore,the large majority of trials required the subject to note that the current picture differed from the previous picture,a judgment that could have potentially interfered with making conceptual inferences.In general,the fact that taste inferences were produced under this particular set of task conditions attests to their strength and ubiquity.Consistent with previous findings,the experiment here indicated that conceptual representations are distributed across the brain areas that underlie their processing in perception and action.Because different categories are associated with differ-ent distributions of multimodal properties (McRae and Cree,2002),different categories rely on different configurations of brain areas for conceptual representation.As reviewed earlier,much work has shown that thinking about tools activates brain areas that process visual form,visual motion,and object manipulation.Analogously,we have shown here that thinking about food activates brain areas that process taste,taste reward and food shape.Thus our findings support the view that thebrain areas representing knowledge for a particular category are those typically used to process its physical instances.Besides having implications for theories of distributed con-ceptual representation,these findings have implications for various societal issues related to food,such as eating disorders,obesity and advertising.Taste inferences in the gustatory system,as observed here,arise in response to a wide variety of food stimuli in the environment and in the media.In eating disorders and obesity,the perception of foods and food pictures,as well as thoughts of food,may be associated with dysfunctional inferences about taste and reward.Conversely,behavioral,cognitive and pharmacological interventions may,in part,restore the gustatory activity underlying inferences about taste and reward to more normal forms.NotesThis work was supported by NIMH grant 1F31MH070152-01to K.S.and National Science Foundation grants SBR-9905024and BCS-0212134and Emory University research funds to L.W.B..We are grateful to Melissa Armstrong and Christine Wilson for their assistance in stimulus preparation.Correspondence should be addressed to Lawrence W.Barsalou,Department of Psychology,Emory University,532North Kilgo Circle,Atlanta,GA 30322,USA.Email:barsalou@emory.ed.Figure 4.(a )Locations of peak right hemisphere insula/operculum activations reported in taste perception studies.(b )Locations of peak left OFC activations across various tasks.The squares in the insula/operculum at Z =20and Z =ÿ9represent peak activations observed when participants taste sucrose,whereas the square in the lateral OFC at Z =ÿ10is the peak activation in the area observed to respond to the combination of gustatory and olfactory stimuli,and thus is a likely candidate for being the center of flavor representation (de Araujo et al.,2003).The square in the insula/operculum at Z =13indicates an area of common activation when participants tasted either glucose or salt (O’Doherty et al.,2001).The squares in the insula/operculum at Z =10and in the OFC at Z =ÿ6indicate the peak activations observed when participants taste umami (de Araujo et al.,2003).The squares in the insula/operculum at Z =5and in the OFC at Z =ÿ18represent peak activations when participants tasted glucose (Francis et al.,1999).Diamonds in the inferior medial OFC represent peak activations observed when participants receive abstract rewards (O’Doherty et al.,2001).The circle in the OFC at Z =ÿ10represents peak activation observed when participants verify the taste properties of concepts using strictly linguistic stimuli (Simmons,Pecher,Hamann,Zeelenberg,and Barsalou,under review).Finally,the circles in the insula/operculum at Z =9and in the OFC at Z =ÿ18and Z =ÿ6indicate the activation peaks observed in the present study when participants viewed food pictures.When necessary,coordinates reported in other studies were converted from Talairach to MNI space.Cerebral Cortex October 2005,V 15N 101607at Indian Institute of Science Education and Research, Kolkata (IISER-K) on February 28, 2011 Downloaded from。

Journal of Experimental Botany,Vol.63,No.8,pp.3229–3241,2012 doi:10.1093/jxb/ers047Advance Access publication29February,2012This paper is available online free of all access charges(see /open_access.html for further details)RESEARCH PAPERSporophytic ovule tissues modulate the initiation and progression of apomixis in HieraciumMatthew R.Tucker1,*,Takashi Okada1,Susan D.Johnson1,Fumio Takaiwa2and Anna M.G.Koltunow1,†1CSIRO Plant Industry,Waite Campus,Hartley Grove,Urrbrae SA5064,Australia2Transgenic Crop Research and Development Centre,National Institute of Agrobiological Sciences,Tsukuba,Ibaraki,Japan*Present address:ARC Centre of Excellence in Plant Cell Walls,University of Adelaide,Waite Campus,Urrbrae,SA5064,Australia.y To whom correspondence should be addressed.E-mail:anna.koltunow@csiro.auReceived29July2011;Revised30January2012;Accepted31January2012AbstractApomixis in Hieracium subgenus Pilosella initiates in ovules when sporophytic cells termed aposporous initial(AI) cells enlarge near sexual cells undergoing meiosis.AI cells displace the sexual structures and divide by mitosis to form unreduced embryo sac(s)without meiosis(apomeiosis)that initiate fertilization-independent embryo and endosperm development.In some Hieracium subgenus Pilosella species,these events are controlled by the dominant LOSS OF APOMEIOSIS(LOA)and LOSS OF PARTHENOGENESIS(LOP)loci.In H.praealtum and H.piloselloides,which both contain the same core LOA locus,the timing and frequency of AI cell formation is altered in derived mutants exhibiting abnormal funiculus growth and in transgenic plants expressing rolB which alters cellular sensitivity to auxin.The impact on apomictic and sexual reproduction was examined here when a chimeric RNAse gene was targeted to the funiculus and basal portions of the ovule,and also when polar auxin transport was inhibited during ovule development following N-1-naphthylphthalamic acid(NPA)application.Both treatments led to ovule deformity in the funiculus and distal parts of the ovule and LOA-dependent alterations in the timing,position,and frequency of AI cell formation.In the case of NPA treatment,this correlated with increased expression of DR5:GFP in the ovule,which marks the accumulation of the plant hormone auxin.Our results show that sporophytic information potentiated by funiculus growth and polar auxin transport influences ovule development,the initiation of apomixis,and the progression of embryo sac development in Hieracium.Signals associated with ovule pattern formation and auxin distribution or perception may influence the capacity of sporophytic ovule cells to respond to LOA.Key words:Apomixis,auxin,gametophyte,ovule,seed.IntroductionSeed development inflowering plants requires the initiation offloral development and the formation of anthers and ovules.Within these organs,sporophytic reproductive pre-cursor cells undergo meiotic reduction and then mitosis to give rise to the male pollen grain containing two sperm cells and a multi-cellular female gametophyte(embryo sac). During double fertilization in the ovule,one sperm nucleus fuses with the egg cell to give rise to the embryo and the other with the central cell to produce the endosperm,thus forming the two major constituents of the mature seed.The ovule is possibly the most vital structure involved in seed development.It is physically connected to the maternal ovary tissues of the plant that supply hormones and nutrients via the funiculus,and incorporates specific tissues called the integuments that form the seed coat and protect the seed during its development(reviewed in Gasser et al., 1998;Yang et al.,2010).A third region in the ovule,the nucellus,supports growth of the embryo sac,which houses the egg cell and central cell and eventually the embryo and endosperm(Chaudhury et al.,2001).Different regionsª2012The Author(s).This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License(/licenses/by-nc/3.0),which permits unrestricted non-commercial use,distribution,and reproduction in any medium,provided the original work is properly cited. at Institute of Botany, CAS on April 26, 2014 / Downloaded fromof the ovule are critical for normal gamete formation and seed development.For example,mutations in the Arabidopsis SPOROCYTELESS/NOZZLE gene induce cell-fate changes in the nucellus and prevent differentiation of the megaspore mother cell,which is the progenitor cell of the embryo sac (Yang et al.,1999;Balasubramanian and Schneitz,2000).The epidermal layer of the Arabidopsis nucellus also plays a specific role via endogenous small RNA pathways to regulate the fate and number of sub-epidermal cells entering the gametophytic pathway (Olmedo-Monfil et al.,2010).Furthermore,alterations in the level or distribution of phytohormones such as auxin and ethylene (De Martinis and Mariani,1999;Pagnussat et al.,2009),and mutations in transcription factors that compromise integument growth such as BEL1(Reiser et al.,1995),WUSCHEL (Gross-Hardt et al.,2002),and INNER NO OUTER (Villanueva et al.,1999)also lead to defects in embryo sac development.Collectively,these findings suggest that development of the embryo sac is sensitive to multiple positional inputs from the surrounding sporophytic ovule tissues (Gasser et al.,1998;Bencivenga et al.,2011).Alterations in positional information within the ovule are also likely to influence the initiation and efficiency of asexual reproductive processes such as apomixis (Koltunow and Grossniklaus,2003;Curtis and Grossniklaus,2007;Tucker and Koltunow,2009).In apomictic Hieracium subgenus Pilosella species that undergo the apospory type of apomixis,sexual reproduction initiates first with the differentiation of a megaspore mother cell that undergoes meiosis (Fig.1A )to produce a tetrad of megaspores (Figs 1A ,2A,B ;Koltunow et al.,1998b ).However,these meiotic products are displaced by adjoining sporophytic cell(s)called aposporous initial (AI)cells that bypass meiosis (apomeiosis),initiate mitosis (Figs 1A ,2C )and produce an embryo sac containing cells resembling the sexual egg and central cell.Embryo and endosperm development subsequently initiate in the absence of fertil-ization (Fig.2D ;Koltunow et al.,1998b ,2000).Genetic studies in aneuploid H.praealtum (R35)and analyses of apomixis mutants suggest that the two compo-nents of apomixis in Hieracium ,apomeiosis and autono-mous seed development,are controlled by two independent dominant loci,LOSS OF APOMEIOSIS (LOA )and LOSS OF PARTHENOGENESIS (LOP ),respectively (Catanach et al.,2006).Furthermore,a recent study showed that the core LOA locus in R35is highly conserved in tetraploid H.caespitosum (C36),tetraploid H.piloselloides (D36),and diploid H.piloselloides (D18;Okada et al.,2011).Analysis of these and other Hieracium species showed that the details of the mechanism of apospory,such as the timing and frequency of AI cell initiation or the type of embryo sac growth,can vary between plants and/or sub-species (Koltunow et al.,2000,2011).In the case of D18,a higher frequency of early AI differentiation and ectopic embryo formation in integument tissues appeared to correlate with abnormal ovule twisting and defective funiculus develop-ment compared with its parental line,D36(Koltunow et al.,2000).Changes in the frequency and location of AI cells were also observed in deformed ovules from R35gamma irradiation mutants carrying deletions in unknown regions outside the LOA locus (Koltunow et al.,2011)and in D36plants expressing the rolB oncogene from Agrobacterium rhizogenes ,which increases cellular sensitivity to the phyto-hormone auxin (Koltunow et al.,2001).Auxin is known to play diverse roles in plant growth and morphogenesis including gynoecium morphogenesis,female gametophyte development,and embryo patterning (Nemhauser et al.,Fig.1.Schematic representation of ovule development in wild-type,GluB-1:RNase ,and NPA-treated apomictic Hieracium plants.(A–C)Comparison of early stages of ovule development,including megaspore mother cell initiation,megaspore selection,apospo-rous initial cell differentiation,and early embryo sac development in (A)untransformed apomictic Hieracium ,(B)transgenicGluB1:RNase apomictic Hieracium ,and (C)NPA-treated apomictic Hieracium .Yellow shading indicates sexual structures,red shading highlights apomictic structures,grey shading highlights regions of integument liquefaction,and green shading indicates theapproximate localization of DR5:GFP expression.Ovule stage numbers are indicated at the bottom of the figure.Abbreviations:i,integument;mmc,megaspore mother cell;f,funiculus;fm,functional megaspore;ai,aposporous initial cell;dm,degenerating megaspores;et,endothelium;il,integument liquefaction;aes,aposporous embryo sac;en,endosperm;em,embryo;de,degenerating endothelium.3230|Tucker et al.at Institute of Botany, CAS on April 26, 2014/Downloaded from2000;Weijers et al.,2006;Pagnussat et al.,2009).In Hieracium 35S:rolB plants it was unclear if the observed effects on altered apomictic initiation resulted from rolB influencing cellular responses to auxin,or from the alter-ations in cell signalling caused by dramatic changes in ovule morphology that were induced because of constitutive rolB expression (Koltunow et al.,2001).In this study,the influence of ovule tissues on apomixis in Hieracium was investigated by targeting the ablation of funicular cells and altering ovule auxin levels by applying an inhibitor of polar auxin transport.The treatments were assessed by cytologically examining the effects on ovule development and apomixis.The results suggest that information from sporophytic tissues,involving pathways dependent upon funiculus development and the phytohor-mone auxin,acts in parallel or upstream of the dominant LOA locus to influence the timing and progression of the apomictic process inHieracium.Fig.2.GluB-1:RNase leads to a de-regulation of apomixis in Hieracium .(A–D)Sections of ovules from untransformed D36Hieracium stained with Toluidine Blue.(A)Young ovule initiating sexual reproduction.(B)Ovule containing an aposporous initial cell displacing the sexual megaspore tetrad.The chalazal and micropylar orientation of the ovule is indicated.(C)Ovule showing an aposporous embryo sac undergoing mitosis.(D)Seed containing an embryo and endosperm.(E)GluB-1:GUS in a stage 3capitulum and a stage 4floret from Hieracium D36.The blue staining indicates expression and the dashed rings indicate the position of the ovaries.(F–J)Ovule sections from D36GluB-1:RNase line #7stained with Toluidine Blue.(F)Young ovule containing an archesporial cell and precocious aposporous initial-like cells.(G–G’)Ovule showing megaspores and multiple expanding AI-like cells in different planes via serial sections.The red dashed line indicates the approximate position of the aposporous initial cells in subsequent sections relative to the sexual structures.(H)Ovule containing aborted megaspore mother cells and aposporous initials at the micropylar end of the ovule and an enlargedoutgrowth at the chalazal end.(I)Ovule showing a chalazal embryo forming in the apparent absence of endosperm.(J)Ovule showing an aborted megaspore mother cell and chalazal aposporous initial cell.(K)Section of a mature seed containing a single embryo from D36and a constricted seed from D36GluB-1:RNase line #7.The whole mature seeds are also shown.Dashed red lines in (H)–(J)indicate regions of integumentary liquefaction.Yellow false colouring indicates cells acting in the sexual pathway,pink false colouring indicates cells acting in the apomictic pathway.Stages of ovule development are indicated in each panel.Abbreviations:aes,aposporous embryo sac;as,archesporial cell;ai,aposporous initial cell;em,embryo;en,endosperm;et,endothelium;f,funiculus;fl,floret;fm,functional megaspore;GB1:GUS,GluB-1:GUS;GB1:R,GluB-1:RNase ;i,integument;ov,ovary;mmc,megaspore mother cell;mp,micropyle;ms,megaspores;n,nucellar lobe;r,receptacle;st,stage.Bar ¼50l m in (D)and (E),300l m in (K),and 20l m in all other panels.Sporophytic ovule tissues modulate apomixis |3231at Institute of Botany, CAS on April 26, 2014/Downloaded fromMaterials and methodsPlant materialPlants used in this study were maintained by vegetative propaga-tion and grown in greenhouses under the same conditions as described previously (Koltunow et al.,1998b ).Plant materials were chosen based on the fact they have been described previously and extensively characterized via cytological methods (Koltunow et al.,1998b ,2000,2011;Okada et al.,2011),and because they contained the same core LOA locus (Okada et al.,2011).Self-incompatible,facultatively apomictic Hieracium piloselloides (tetraploid D36and diploid D18)and H.praealtum (aneuploid R35)produce >95%of seed via apospory coupled with autonomous seed formation,while H.pilosella (tetraploid P36;4x ¼2n ¼36)is a self-incompatible,obligate sexual plant that requires cross-pollination to produce seed (Koltunow et al.,1998b ,2000,2011;Okada et al.,2011).D18(2x ¼2n ¼18)was derived from D36(4x ¼2n ¼36)by autonomous embryogenesis of a rare meiotically reduced egg (Koltunow et al.,2011).The isolation and characterization of R35(3x +8¼2n ¼35)gamma deletion mutants was described previously (Catanach et al.,2006).The stages of floral development for Hieracium were used as a reference for staged tissue collections (Koltunow et al.,1998b ).Transgenes and transformationTo determine the suitability of the rice glutellin seed storage protein B1promoter (GluB-1)for funicular tissue ablation,a construct containing 1.3kb of the GluB-1promoter inserted upstream of the beta-glucuronidase (GUS )gene was transformed into Hieracium using the plant transformation vector pBI121.The GluB-1:RNase gene was made by inserting the 1.3kb GluB-1promoter fragment into a previously described CG1:RNase construct (Koltunow et al.,1998a )in place of the CG1promoter.The DR5:GFP construct has previously been used as a marker for auxin accumulation (Ulmasov et al.,1997;Dorcey et al.,2009)and was a gift from Dolf Weijers.The AtARF8:GUS construct was described previously (Goetz et al.,2006)and contains 1.8kb of the 5#upstream promoter sequence from the Arabidopsis ARF8gene fused to GUS .Constructs were electroporated into Agrobacterium strains LBA4404or AGL1for plant transformation by a floral disc method (Bicknell and Borst,1994).PCR and Southern blot analyses were used to confirm the presence,integrity,and number of inserted genes in the kanamycin-resistant plants.The self-incompatible nature,high hemizygosity,and high apomictic seed set (>95%)of apomictic Hieracium plants makes it almost impossible to generate true-breeding,homozygous trans-genic apomictic lines.Similarly,crosses between apomictic and sexual Hieracium give rise to highly diverse progeny that are not directly comparable for assessing the effect of transgenes (Koltunow et al.,2000).Therefore,for experimental consistency,primary transformed lines containing a single intact gene insertion from each species were selected for analysis (Koltunow et al.,2000).The total number of primary lines examined for each gene was 5for GluB-1:GUS in D36,8for GluB-1:RNase in D36,8for GluB-1:RNase in P36,3for DR5:GFP in R35,3for DR5:GFP in P36,3for DR5:GFP in loalop mutant 115,and 3for AtARF8:GUS in D36.Analysis of GUS activity was carried out as described previously by Tucker et al.(2003).Fluorescence analysisSamples were examined for GFP expression on a Zeiss M1Imager using filter set 42.Ovaries were dissected from staged florets in approximately 40l l 10%glycerol on microscopy slides,and gently squashed with a glass cover slip to aid visualization of the ovule.A standard set of exposure times (500,1000,2000ms)were used to capture each image via Axiovision software.N -1-naphthylphthalamic acid (NPA)applicationNPA (Alltech)was dissolved in DMSO and diluted to 0.01mM,0.1mM,1mM,10mM,and 100mM in 70%ethanol.In order to examine the effect of a pulse of NPA at different developmental stages,whole capitula containing 40–50florets were dipped either before MMC formation (pre-stage 2)or before AI formation (late stage 3),left for 24h and then rinsed in water.Ovules were collected at various times from the following day until the time of embryo sac formation.At least 10capitula were dipped at each stage taking care not to use more than three capitula per plant and capitulum samples were pooled.The experiment was repeated twice.A total of 65ovules were sectioned for analysis and at least 50ovules per stage were examined for DR5:GFP.CytologyD36GluB-1:RNase lines #7and #34and P36GluB-1:RNase lines #3-1and #7-3were the most extensively examined in this study.Ovaries from all of the transgenic lines were fixed in FAA (50%ethanol,10%formalin,5%acetic acid)and cleared using methyl salicylate to examine a greater number of samples,as per Koltunow et al.(2011).Ovaries for serial sectioning were fixed in glutaraldehyde and embedded in Spurr’s resin as described pre-viously by Koltunow et al.(1998b ).Approximately five ovules were sectioned per stage where possible.Sections and whole-mount ovules were examined using a Zeiss Axioplan microscope and digital images were collected using a Spot II camera.Figures were compiled using Adobe Photoshop and Illustrator software.ResultsDisruption of ovule funicular growth in apomicticHieracium leads to a de-regulation of apomixis during early ovule stagesTo impair communication between the ovule and other maternal tissues of the plant,a genetic cell ablation strategy was adopted that has previously been used to examine the relationships between cells and tissues during plant organ development (Mariani et al.,1990;Beals and Goldberg,1997;Koltunow et al.,2011).These studies employed chimeric gene constructs consisting of characterized pro-moters to direct the activity of a linked ribonuclease gene in a spatially and temporally defined manner.Studies in Citrus,tobacco,and rice detailed the GluB-1:GUS expression pattern during ovule and seed develop-ment (Takaiwa et al.,1996;Koltunow et al.,1998a ;Wu et al.,1998).In the Hieracium ovule,GluB-1:GUS was detected in the funiculus and the subtending ovary and receptacle cells at stages 2and 3when the ovules had differentiated a megaspore mother cell (Fig.2E ).In composite Hieracium flowers,the receptacle is the only point of contact between the individual florets and the floral capitulum.GluB-1:GUS persisted in stage 4capitula when aposporous initial (AI)cells had differentiated in the ovules,but subsequently became confined to the base of the floret and the associated receptacle cells.Expression declined until the late stages of seed development when GUS expression was detected in the developing endosperm.The spatial and temporal dynamics of the GluB-1promoter in Hieracium suggested that it would be a useful tool to disrupt funiculus development and to impair communication between the3232|Tucker et al.at Institute of Botany, CAS on April 26, 2014/Downloaded fromearly ovule and surrounding tissues during the initiation of apomixis.To maintain consistency between studies,apomictic H.piloselloides D36was selected for transformation with the GluB-1:RNase construct.D36typically initiates apo-mixis in a uniform manner,is sensitive to the activity of the rolB oncogene that induces changes in auxin sensitivity,and its derivative diploid plant D18shows funicular defects (see Supplementary Fig.S1A–C at JXB online)and de-regulated apomixis (Koltunow et al.,1998b ,2000,2001).Sexual P36was chosen as a non-apomictic control.In total,eight transgenic apomictic H.piloselloides D36plants and eight transgenic sexual H.pilosella P36plants were generated and all exhibited decreased seed set relative to the control lines.Two of the D36plants and one P36plant aborted capitulum and flower development at the very early stages and failed to initiate ovule or seed development (data not shown).In the remaining lines,flowers were viable and mature seeds were produced that had a distinctive terminal phenotype in that they were pointed and constricted at their site of attachment to the capitulum,reflecting the zone of expression of the ablation gene (Fig.2K ).Two D36plants that produced seeds,#7and #34,were selected for detailed histological analysis as they represented plants expressing the ablation gene at different levels from seed germination studies (Table 1).A further two P36plants (#3-1and #7-3),were also examined to provide a sexual control.The remaining lines were characterized in less detail but showed similar morphological and cytological phenotypes,suggest-ing that the phenotypes were due to the effects of the GluB-1:RNase transgene (data not shown).In ovules from untransformed apomictic D36plants,a single megaspore mother cell was observed at stage 2/3(Fig.2A )and an enlarged AI cell(s)was not apparent until stage 4when meiosis was complete (Fig.2B ).One or two AI cells,distinguished by their large size,an enlarged central nucleus flanked by large vacuoles,and similarity to the functional megaspore,were usually detected in D36ovules as per previous reports (Koltunow et al.,1998b ).In most ovules,one AI cell initiated embryo sac development and expanded to fill the space within the endothelium normally occupied by the sexual embryo sac by stage 5/6(Fig.2C ).During subsequent stages,embryo and endosperm develop-ment initiated in the embryo sac (Fig.2D ),and continued until the embryo was fully mature and filled the elongated seed (Fig.2K ).GluB-1:RNase line #7in apomictic D36Hieracium was the most strongly affected line since it produced only 16%viable seeds compared with 39%in line #34and 60%in untransformed D36(Table 1).Also,seeds from line #7showed the strongest constriction and deformation at the point of connection to the receptacle (Fig.2K ).To characterize reproductive defects in lines #7and #34,up to 10ovules were serially sectioned per stage between stages 2and 12,from ovule initiation to seed maturity.Distinct abnormalities were first observed during the early stages of ovule development.In contrast to untransformed D36,developing florets from GluB-1:RNase lines #7and #34showed small zones of necrosis and displaced zones of xylem formation at the base of the ovule funiculus (see Supplementary Fig.S2A,B,E–G at JXB online).Different degrees of reproductive abnormalities were observed in the two lines.In line #7,sexual processes aborted in most ovules during meiosis.Despite this,AI differentiation and expansion were prominent as early as archesporial cell initiation (Fig.2F ).Up to eight AI cells differentiated in ovules during the subsequent stages,and serial sections showed that these were spatially separated from the developing sexual cells in 50%(n ¼10)of cases (Fig.2G–G’).Approximately 25%(n ¼20)of sectioned ovules aborted at this stage.In line #34,;13%(n ¼15)of the sectioned ovules had aborted at the early stages.In viable ovules,AI cells appeared at a similar stage compared to line #7,but with lower frequency and reduced displacement.One to three AIs were evident at the early stages in close proximity to the MMC (Fig.S2C–C ’)and only in ;10%(n ¼11)of cases were they spatially separated from sexual structures.The variations between lines #7and #34suggest that different doses of GluB-1:RNase function are contributing to the differences in ovule development and seed viability.In contrast to the transgenic D36lines,only minimal effects were observed in the sexual P36GluB-1:RNase plants during the early stages,where funiculus growth appeared toTable 1.Seed formation and germination in GluB-1:RNase and untransformed H.piloselloides linesPlantFlorets/capitulum aNo.of Seed headsNo.of seedsFull seeds b set (%)Germination rate (%)4d30d D1832611206567023652562D363366103438434686065D36GB1:R c #7336137021037212621665D36GB1:R #30336351667045665263D36GB1:R #343663103537928663963D36GB1:R #373363103268137634563a 6Values show standard deviation (SD)from replicate experiments.b Full seeds indicates seeds that are large and black in colour compared to brown or white seeds that have aborted.cGluB-1:RNase .Sporophytic ovule tissues modulate apomixis |3233at Institute of Botany, CAS on April 26, 2014/Downloaded frombe abnormally elongated but no ectopic ‘AI-like’cells could be detected in the ovule (data not shown).These results indicate that the integrity of funicular and receptacle tissues beneath the ovule influences the timing,frequency,and location of AI formation in apomictic D36.GluB-1:RNase induces polar nuclei defects,extra cell divisions and ectopic reproductive zones in ovules from apomictic HieraciumIn the majority of ovules from the low expressing line #34,one of the early AI cells displaced the sexual structures and initiated mitosis at the micropylar end of the ovule.Apart from the early timing of AI formation,this was similar to D36controls.However,during later developmental stages when the funicular region became more constricted,ovule growth occasionally appeared stunted and embryo sacs were observed containing small embryos without endo-sperm (see Supplementary Fig.S2D at JXB online).In 27%(n ¼15)of sectioned ovules from line #34,embryo sacs contained polar nuclei that had failed to divide.By contrast,quiescent polar nuclei were only observed in 1%of untransformed D36samples at stage 8(n ¼68).These data suggest that the reduced seed viability in line #34is probably due to a combination of early ovule defects as well as failed endosperm development leading to embryo abortion.In ovules from D36GluB-1:RNase line #7that pro-gressed past the early stages,different classes of phenotypes were observed.In the minority of ovules,sporophytic cell divisions and growth were indistinguishable from untrans-formed D36controls,but defects in polar nuclei division were observed similar to those in line #34.In others,the ovule integument recurved and formed a micropyle,and multiple aposporous structures continued development in the distal (chalazal)region of the ovule (see Supplementary Fig.S2G at JXB online).This contrasted with D36and GluB-1:RNase line #34where only one AI cell typically continued development closer to the micropyle (see Supple-mentary Fig.S2E,F at JXB online).In the majority (60%,n ¼31)of ovules from GluB-1:RNase line #7,the funiculus became highly constricted and the endothelium and micropyle did not form (Fig.2H–J ;see Supplementary Fig.S2H–H’at JXB online).In these cases,somewhat surprisingly,ovule cells continued to divide resulting in the formation of chalazal outgrowths (Fig.2H–J ).Serial sections also showed the presence of multiple nucellar lobes containing both sexual and AI cells (see Supplemen-tary Fig.S2H–H’at JXB online).Typically,the megaspores and the adjacent initials closer to the micropylar end of the ovule had aborted (Fig.2H–J ).However,cells in the chalazal region of the ovule continued dividing to form an enlarged dome where cell degeneration (liquefaction)oc-curred,indicating that the integumentary identity of the tissue had been retained (Fig.2H–J ;see Supplementary Fig.S2H–H’at JXB online).Apomictic processes occurred in such domes (Fig.2I,J ),spatially removed from the arrested sexual and AI cells by many cell layers.In some ofthese ovules,embryos formed in isolation surrounded by liquefying tissue (Fig.2I ;see Supplementary Fig.S2G at JXB online).In other ovules,AI cells (Fig.2J )or inverted embryo sac structures containing egg-like cells (see Supple-mentary Fig.2I at JXB online)and undivided polar nuclei were observed.Often a diverse array of structures including AI cells,cellular-like endosperm cells,embryos,and embryo sacs could be observed in the same ovule through serial sections.Such ovules resembled those described for H.piloselloides D18(Koltunow et al.,2000).The frequency of GluB-1:RNase line #7ovules showing apomictic develop-ment in the integumentary domes and constriction at the micropylar end of the developing seed correlated with the mature fertile seed development shown in Fig.2K .It is therefore likely that some of the viable seeds contained embryos derived from the chalazal zones.This is somewhat reminiscent of the multiple adventitious chalazal embryos that form from embryo initial cells in seeds from Citrus (Kobayashi et al.,1979;Koltunow et al.,1995).Taken together,these observations suggest that the integrity of funicular and receptacle tissues in Hieracium is important for ovule development and apomixis,possibly because they transmit or synthesize regulatory information.Perturbations to these tissues in apomictic GluB-1:RNase D36induced developmental changes in unconnected regions of the ovule,leading to abnormal integument growth,precocious differentiation of multiple AI cells and disorga-nized embryo sac development (Fig.1B ).Similar defects in funiculus growth may contribute to a de-regulation of apomixis in H.piloselloides D18(see Supplementary Fig.S1at JXB online;Koltunow et al.,2000)and H.praealtum R35deletion mutants (see Supplementary Fig.S1at JXB online;Koltunow et al.,2011),which produce extra and/or displaced AI cells as well as abnormal embryo sacs.These findings suggest that the capacity of sporophytic cells to respond to LOA and subsequently develop into embryo sacs is modulated by the surrounding tissues of the ovule not directly involved in the apomictic pathway.Auxin maxima and auxin-responsive genes are detected in specific regions of the Hieracium ovule,including the funiculusPrevious studies in apomictic Hieracium D36plants express-ing the rolB oncogene suggested that alterations to auxin sensitivity in sporophytic tissues might contribute spatial and temporal control over the apomictic process (Koltunow et al.,2001).In Arabidopsis ovules,auxin contributes to integument growth (Schruff et al.,2006),embryo sac polarity (Pagnussat et al.,2009),and vascular development (Ohashi-Ito et al.,2010).Several approaches were used to determine if auxin signalling might be a sporophytic factor influencing ovule development and apomixis in Hieracium .Analysis of AtARF8:GUS in ovules from transgenic Hiera-cium D36plants showed that GUS activity was detected predominantly in the funiculus at stage 4/5,and at later stages in the whole ovule as well as connective tissues of the receptacle (Fig.3A,B ).In Arabidopsis ,ARF8is involved in3234|Tucker et al.at Institute of Botany, CAS on April 26, 2014/Downloaded from。

Vol.22no.122006,pages1437–1439doi:10.1093/bioinformatics/btl116 BIOINFORMATICS ORIGINAL PAPERGenome analysisPseudoPipe:an automated pseudogene identification pipeline Zhaolei Zhang1,Nicholas Carriero2,Deyou Zheng3,John Karro4,Paul M.Harrison5and Mark Gerstein3,Ã1Banting and Best Department of Medical Research,Donnelly CCBR,University of Toronto,160College Street, Toronto,ON M5S3E1,Canada,2Department of Computer Science,Yale University,New Haven,CT06520,USA, 3Department of Molecular Biophysics and Biochemistry,New Haven,CT06520,USA,4Department of Biology, 506Wartik,Pennsylvania State University,University Park,PA16802,USA and5Department of Biology,McGill University,Stewart Biology Building,1205Dr Penfield Avenue,Montreal,QC,H3A1B1,CanadaReceived on December28,2005;revised on March1,2006;accepted on March22,2006Advance Access publication March30,2006Associate Editor:Keith A CrandallABSTRACTMotivation:Mammalian genomes contain many‘genomic fossils’i.e.pseudogenes.These are disabled copies of functional genes that have been retained in the genome by gene duplication or retrotrans-position events.Pseudogenes are important resources in understanding the evolutionary history of genes and genomes.Results:We have developed a homology-based computational pipe-line(‘PseudoPipe’)that can search a mammalian genome and identify pseudogene sequences in a comprehensive and consistent manner. The key steps in the pipeline involve using BLAST to rapidly cross-reference potential‘‘parent’’proteins against the intergenic regions of the genome and then processing the resulting‘‘raw hits’’–i.e.eliminat-ing redundant ones,clustering together neighbors,and associating and aligning clusters with a unique parent.Finally,pseudogenes are clas-sified based on a combination of criteria including homology,intron-exon structure,and existence of stop codons and frameshifts. Availability:The PseudoPipe program is implemented in Python and can be downloaded at /Contact:Mark.Gerstein@ or zhaolei.zhang@utoronto.caINTRODUCTIONPseudogenes are those sequences in the genome that bear similarity to specific protein coding genes,but nevertheless are unable to produce functional proteins due to existence of frameshifts,prema-ture stop codons or other deleterious mutations(Mighell et al., 2000).It is reported that human genome has8000–12000pseudo-genes and mouse genome has5000(Zhang et al.,2004,2003).Most of these pseudogene sequences were the results of LINE1mediated retrotransposition or genome duplication.Pseudogenes are impor-tant,as they are in essence genomic fossils that can be used to infer the ancestral sequence and evolutionary history of genes that are present today.They can often contribute to cross hybridization in high-throughput genomics experiments.Some pseudogenes reportedly have regulatory roles(Hirotsune et al.,2003).In this paper,we describe our tested pseudogene identification algorithm(Zhang et al.,2004,2003).The definition of pseudogene is somewhat ambiguous as it is more difficult to confirm non-functionality than confirm functionality.Our method is designed and best used to detect those pseudogenes that are unable to be translated into proteins.The algorithm has been implemented into a standalone software package,‘PseudoPipe’.METHODS AND ALGORITHMProgram outlineThe general dataflow in PseudoPipe has been described previously(Zhang and Gerstein,2004;Zhang et al.,2003).The inputs to the program are the genomic sequence after repeat-masking,the comprehensive and non-redundant set of protein sequences in the genome,and the chromosomal coordinates of the functional genes.The last piece of information is needed to efficiently distinguish pseudogene candidates from functional genes.The output of PseudoPipe is the complete annotation of the pseudogenes in the genome in question:their chromosomal location,nucleotide sequences, name and sequence of the parent gene,and alignment of the pseudogene with the functional gene.Homology searchThefirst step in the annotation pipeline is to identify all the regions in the genome that share sequence similarity with any known protein,using BLAST(E-value1·10À4)(Altschul et al.,1997).We then partition the BLAST hits by chromosome and strand direction.Significant overlaps with functional gene annotations are then removed.In PseudoPipe,‘signifi-cant overlap’is defined as either complete overlap or overlap of!30bps between a hit and a functional gene(Fig.1A).Eliminating redundanciesThe next step is to eliminate redundant and overlapping BLAST hits,in places where a given chromosomal segment has multiple hits.We divide these overlapping hits into two categories depending on whether they match to the same or different query proteins(Fig.1B).Thefirst scenario arises because the BLAST program is prone to break continuous long sequenceÃTo whom correspondence should be addressed.ÓThe Author2006.Published by Oxford University Press.All rights reserved.For Permissions,please email:journals.permissions@ The online version of this article has been published under an open access ers are entitled to use,reproduce,disseminate,or display the open access version of this article for non-commercial purposes provided that:the original authorship is properly and fully attributed;the Journal and Oxford University Press are attributed as the original place of publication with the correct citation details given;if an article is subsequently reproduced or disseminated not in its entirety but only in part or as a derivative work this must be clearly indicated.For commercial re-use,please contact journals.permissions@homologies into short overlapping fragments.The second scenario occurs because the BLAST query sequences include homologous proteins or protein domains.For example,in Figure 1B,both query protein A and B have BLAST hits in the same genomic region.At this step,we do not try todetermine which query protein (A or B)is the ‘real parent’of the pseudogene candidate.Instead,we first separate the BLAST hits that match to distinct queries,and then partition them into disjoint sets that do not have any overlap at all (Fig.1C).The partitioning is transitive,i.e.if hit X overlaps Y and Y overlaps Z,then X,Y and Z are all placed into the same set even if X and Z do not overlap.Within each set,we sort the hits by their BLAST E -values and remove any hits that either completely or partially overlaps with another hit but with much worse BLAST E -values.Merging the hitsHere,we merge the overlapping and successive BLAST hits into a continu-ous pseudogene structure (Fig.1C).We first analyze the overlapping BLAST hits inside a single disjoint set,and merge them into a single ‘super hit’or ‘pseudo-exon’.We then select those neighboring disjoint sets of hits that match the same query protein.Based on the distance between the hits on the chromosome (G c in Fig.1C)and the distance on the query protein (G q ),we determine whether these merged hits belong to the same pseudogene structure.These gaps G c can arise from (1)low complexity or very decayed regions of the pseudogene that are discarded by BLAST,(2)short DNA sequences inserted into the pseudogene,(3)ancestral intron sequence in the duplicated pseudogenes and (4)repetitive elements.These four scenarios can be distinguished by comparing the length of G c and G q ,and by calculating the repeat content of the gaps between the neighboring hits.Determining paternity of the pseudogenesIn this step,we resolve the paternity ambiguity of the pseudogenes,i.e.determine among the paralogous query proteins which one most likely gave rise to the pseudogene.To do this,we consider (1)the sequence identity between the pseudogene and the query proteins at either DNA or translated amino acid sequence levels,(2)the best BLAST E -value associated with any of the original hits (before merging),(3)the length of the protein sub-sequence that yields the pseudogene.In effect,we are assuming that after millions of years of evolution under neutral selection,the pseudogene remains more similar to the modern form of the original parent gene,than to paralogous genes.Refining alignmentBLAST is a tool that is intended for fast,heuristic homology search instead of accurate sequence alignment.It does not accommodate frame shifts,thus it may break a potential pseudogene into smaller fragments.Therefore,in PseudoPipe,each remaining hit is re-processed using a more accurate dynamic programming alignment program,specifically tfasty from the FASTA suite (Pearson,1997).A pseudo-exon is extended in both directions with a small buffer region (30nt),which is then aligned to the query protein to achieve an optimal alignment,to calculate accurate sequence similarity and to annotate positions of disablements (frame shifts and stop codons),as well as insertions and deletions.Classification of pseudogenesThe PseudoPipe program applies a set of sequence identity and completeness cutoffs to report a final set of good-quality pseudogene sequences.The parameters and cutoffs in the PseudoPipe can be easily altered to produce more or fewer pseudogenes.Previously (Zhang et al .,2002,2003),we used the following cutoffs:(1)amino acid sequence identity >40%,(2)BLAST E -value lower than 1E-10and (3)the pseudogene covers 70%of the parent gene.These high-confidence pseudogenes are then classified as (1)retro-transposed pseudogenes,(2)duplicated pseudogenes and (3)pseudogeneic fragments.Retrotransposed pseudogenes lack introns,have small flanking direct repeats and a 30polyadenine tail.PseudoPipe distinguishes retro-transposed pseudogenes from duplicated pseudogenes by a combination of these features,with the emphasis on the evidence of ancient introns.Pseudogenic fragments are protein/chromosome homologies that have high-sequence similarity,but are too decayed to be reliably assessed as processed or duplicated.exons of known geneschromosomeBLAST hits of query protein ABLAST hits ofquery protein B BLAST hits of query protein Cchromosome BLAST hits ofquery protein Aquery protein(A)(B)(C)Fig.1.(A)Schematic illustrations of overlap between BLAST hits and func-tional genes.Green rectangles represent exons of annotated genes and red rectangles represent BLAST hits.We expect some of the BLAST hits to identify the proteins’parent genes or their homologs.Given the locations of these genes’exons,we eliminate any BLAST hits that overlap as pictured here.(B)Separating BLAST hits into disjoint sets based on query and position on the chromosome.Three disjoint sets are shown,which match to distinct query protein A,B and C.The first two disjoint sets map to the same region on the chromosome,they were both kept at this step.Overlapping BLAST hits within each disjoint set are filtered and removed.(C)Merging neighboring hits.BLAST hits within each disjoint set are first merged into ‘superhits’.The distance between the neighboring superhits G c is compared with the distance on the query protein,G q ;the neighboring superhits are merged together if they are determined to be part of the same ancestral pseudogene structure and the distance G c is not too great.Z.Zhang et al.1438Implementation and run timeThe program was originally written in PERL but we have re-implemented it in Python.Except for the step of whole-genome BLAST search,the annotation pipeline can be run on an entire genome in a few hours,on a reasonably robust Linux workstation(3.0GHz,1GB RAM).Multiple concurrent independent pipeline runs could be started on multiple computers,e.g.several chromo-somes can be grouped together and processed on a single computer. ACKNOWLEDGEMENTSZ.Z.acknowledges start-up fund from University of Toronto Faculty of Medicine.M.G.acknowledges funding support from the NIH/NHGRI (IU01HG003156-01).Funding to pay the Open Access publication charges was provided by the NIH(grant no.NIH/NHGRI HG02357). Conflict of Interest:none declared.REFERENCESAltschul,S.F.et al.(1997)Gapped BLAST and PSI-BLAST:a new generation of protein database search programs.Nucleic Acids Res.,25,3389–3402. Hirotsune,S.et al.(2003)An expressed pseudogene regulates the messenger-RNA stability of its homologous coding gene.Nature,423,91–96.Mighell,A.J.et al.(2000)Vertebrate pseudogenes,.FEBS Lett.,468,109–114. Pearson,W.R.(1997)Comparison of DNA sequences with protein sequences.Genomics,46,24–36.Zhang,Z.and Gerstein,M.(2004)Large-scale analysis of pseudogenes in the human genome.Curr.Opin.Genet.Dev.,14,328–335.Zhang,Z.et al.(2002)Identification and analysis of over2000ribosomal protein pseudogenes in the human genome.Genome Res.,12,1466–1482.Zhang,Z.et al.(2003)Millions of years of evolution preserved:a comprehensive catalog of the processed pseudogenes in the human genome.Genome Res.,13, 2541–2558.Zhang,Z.et al.(2004)Comparative analysis of processed pseudogenes in the mouse and human genomes.Trends Genet.,20,62–67.Pseudogene annotation in vertebrate genomes1439。

浅析益母草的研究专业：中药资源与开发（2）班姓名：魏亚东学号：201014022062摘要：益母草为唇形科植物益母草的全草。

一年或二年生草本，夏季开花。

生于山野荒地、田埂、草地等。

全国大部分地区均有分布。

本文重点对益母草的化学成分、药理作用以及其在临床上的应用进行了简单的综述。

关键词：益母草、药理作用、临床应用。

正文：益母草为唇形科植物益母草Leonurur japonicus Houtt.的地上部分。

高60-120cm。

茎直立，四棱形，微被毛，一年生植物基生叶具长柄，叶柄略圆形，5-9浅裂，基部心形。

茎中部叶有短柄，掌状3浅裂。

花序上的叶呈羽状深裂或浅裂成3片，裂片全缘或具少数锯齿。

上面绿色，被粗伏毛，下面淡绿色，被柔毛及腺点。

轮伞花序腋生，具花8-15朵，花萼钟形，先端5齿裂，具刺尖，花冠唇形，淡红色或则紫色。

雄蕊4,2强。

雌蕊1，柱头2裂。

小坚果三棱形。

益母草味辛、微苦、性微寒、归心、肝、膀胱经，具有活血调经、利水消肿、清热解毒之功效。

全国各地均有野生或栽培。

1：性状鉴别鲜益母草，幼苗期无茎，茎生叶圆心形，边缘5-9浅裂，每裂片有2-3钝齿。

花前期茎呈方柱形，上部多分枝，四面凹下城纵沟，长30-60cm，直径0.2-0.5cm；表面青绿色；折断面中部有髓。

叶交互对生，有柄；叶片青绿色，质鲜嫩，揉之有汁；下部茎生叶掌状3裂，上部叶羽状深裂或浅裂成3片，裂片全缘或具少数锯齿。

气微，味微苦。

干益母草，茎表面灰绿色或黄绿色；体轻，质韧，断面中空有髓。

叶片灰绿色，多皱纹。

破碎，易脱落。

轮伞花序腋生，小花淡紫色，花萼筒状，花冠二唇形。

以质嫩、叶多、色灰绿者为佳；质老、枯黄、无叶者不可供药用。

2:化学成分全草含益母草碱（leonurine，约0.05％，开花初期仅含微量，中期逐渐增高）、水苏碱（stachydrine）、芸香碱、延胡索酸、亚麻酸、p-亚油酸、月桂酸、苯甲酸、油酸等。

在全草中还可以分离提取得到5,7,3`,4`,5`,—五甲氧基黄酮（5,7,3`,4`,5`,—pentamethoxyflavone）、汉黄芩素（wogonin）、洋芹素-7-O-葡萄糖苷（apigenin-7-O-glucopyranoside）、大豆素（daidzein）、槲皮素（quercetin)[1]。

Nephrol Dial Transplant (2009)1321 13doi:10.1093/ndt/gfp500Recurrence of nephrotic syndrome after transplantation in a mixed population of children and adults:course of glomerular lesions and value of the Columbia classification of histological variants of focal and segmental glomerulosclerosis (FSGS)Guillaume Canaud 1,2,Daniel Dion 3,Julien Zuber 1,2,Marie-Claire Gubler 4,Rebecca Sberro 1,2,Eric Thervet 1,2,Renaud Snanoudj 1,Marina Charbit 5,R ´e mi Salomon 2,5,Frank Martinez 1,Christophe Legendre 1,2,Laure-Helene Noel 3and Patrick Niaudet 2,51Department of Kidney Transplantation,Necker Hospital,149rue de S`e vres,75015,Paris,2Universit´e Paris Descartes,Rue del’Ecole de M´e decine,3Department of Pathology,4Institut National de la Sant´e et de la Recherche M´e dicale U574and 5Department of Pediatrics Nephrology,Necker Hospital,149rue de S`e vres,75015,Paris,France Correspondence and offprint requests to :Guillaume Canaud;E-mail:guillaume.canaud@nck.aphp.frAbstractIntroduction.Recurrence of nephrotic-range proteinuria in patients with idiopathic nephrotic syndrome (INS)and focal and segmental glomerulosclerosis (FSGS)on native kidneys is associated with poor graft survival.Identifica-tion of risk factors for recurrence is therefore an important issue.In 2004,Columbia University introduced a histo-logical classification of FSGS that identifies five mutually exclusive variants.In non-transplant patients,the Columbia classification appears to predict the outcome and response to treatment better than clinical characteristics alone.How-ever,the predictive value of this classification to assess the risk of recurrence after transplantation has not been addressed.Methods.We retrospectively studied 77patients with INS and FSGS on native kidneys who underwent renal trans-plantation.Of these,42recipients experienced recurrence of nephrotic range proteinuria.Results.At time of recurrence,minimal-change disease (MCD)was the main histological feature.On serial biop-sies,the incidence of MCD decreased over time,while the incidence of FSGS variants increased.The variant type ob-served in the native kidneys was not predictive of either recurrence or type of FSGS seen on the allograft.Patients with complete and sustained remission did not developed FSGS.Conclusion.In conclusion,the Columbia classification is of no help in predicting recurrence after renal transplan-tation or histological lesions in the case of recurrence of proteinuria.Keywords:FSGS;histopathology;kidney transplantation;recurrenceIntroductionSteroid-resistant idiopathic nephrotic syndrome (INS)in native kidneys is frequently associated with progression to end-stage renal disease (ESRD)[1–3].In most patients,ini-tial renal biopsy shows focal and segmental glomeruloscle-rosis (FSGS).However,in early stages,FSGS and minimal-change disease (MCD)are indistinguishable at the clinical level [4–6].Serial renal biopsies indicate that some patients with minimal changes on initial renal biopsy develop FSGS during the course of the disease.This was observed in 60%of cases in a series of 48patients with steroid-resistant min-imal change [7].After kidney transplantation,recurrence of nephrotic-range proteinuria is observed in 30%of cases and is associated with poor graft survival [8–14].Identification of risk factors for recurrence is an important issue.In 2004,Columbia University introduced a histologi-cal classification of FSGS that identifies five mutually exclusive variants:perihilar (PH),cellular (CELL),tip le-sion (TIP),collapsing (COL)and not otherwise specified (NOS)[15].Since then,several studies have correlated re-nal prognosis with histological features in native kidneys.In non-transplant patients,the Columbia variant seems to pre-dict outcome and response to treatment better than clinical characteristics alone.TIP seems to be the most steroid-sensitive variant,with COL being the least favourable vari-ant [16–20].The impact of this classification to predict the FSGS variant in the transplant has been recently studied [21].In one report,Ijpelaar et al.concluded that,in most cases,the FSGS variant observed in the transplant kidney was the same as in the native kidneys,suggesting that the histological variant of the native kidney could predict the variant observed in patients with recurrence.However,the predictive value of this classification to assess the risk ofC The Author 2009.Published by Oxford University Press [on behalf of ERA-EDTA].All rights reserved.For Permissions,please e-mail:journals.permissions@– Advance Access publication 22September 200928 by guest on October 14, 2013/Downloaded fromFig.1.FSGS variants in native kidneys(NOS,n=38;PH,n=6;TIP,n=6;CELL,n=17;COL,n=10).recurrence after transplantation and graft outcome has never been addressed.We investigated in a large series whether the Columbia classification of FSGS variants provides a relevant indicator for the risk of proteinuria recurrence af-ter renal transplantation and the type of variant in the case of recurrence.Material and methodsKidney transplant recipients with INS and FSGS on native kidneys and pro-gression to ESRD were eligible for inclusion.Patients with familial history of proteinuria or secondary forms of FSGS were excluded,i.e.ischaemia-related podocyte injury,sickle cell disease,renal hypoplasia,thrombotic microangiopathy,reflux nephropathy,human immunodeficiency virus and genetic disorders when a test was available.Patients with non-adequate biopsies were also excluded(i.e.the biopsy sample containing fewer than five glomeruli).Our study population included children(n=56)and adults(n=21) (age>16years).There were40males and37females.The77patients included in the study received a sequential quadruple immunosuppression including induction therapy(thymoglobulin,n=47;basiliximab,n=30), corticosteroids,a calcineurin inhibitor(cyclosporine,n=65;tacrolimus, n=12),azathioprine(n=53)or mycophenolate mofetil(n=24).Seventy-one patients had received a first kidney transplant;4patients had received a second transplant and2a third one.Kidney allografts were obtained from a deceased donor in69cases and from a living-related donor in8cases.No patient received pre-emptive treatment for recurrence(i.e.cyclosporine A IV or pre-operative plasmapheresis).Recurrence was defined by massive proteinuria in the nephrotic range (>3g/day for adults or>50mg/kg/day for children),with normal glomeruli and without evidence of acute or chronic rejection,glomerular deposits or allograft glomerulopathy on initial kidney biopsy.Cameron’s classification was applied to define time of recurrence:immediate(<48 h),early(<3months)and late recurrence(>3months)[10].Transplant biopsies were performed either when indicated by protein-uria and/or acute renal failure,or routinely as screening biopsies at3and 12months post-transplant.Native and transplant kidney biopsies were pro-cessed following the international recommendations for light microscopy and immunofluorescence.The median number of glomeruli per biopsy was 13.7±4.6(range6–24).Histology slides were stained with haematoxylin and eosin,periodic acid–Schiff,Masson trichrome and methenamine–silver.For each biopsy,FSGS variant was evaluated according to the Columbia classification[15].The five FSGS variant definitions are sum-marized in Figure1.All native and transplant biopsies were reviewed by three senior pathologists.As electron microscopy is not of current routine practice in France,we were able to describe foot process fusion in only one case.Nevertheless,patients with proteinuria in the nephrotic range and no features of FSGS variants on kidney biopsy were classified as having MCD.ResultsFrom January1984to December2007,77patients with INS and biopsy-proven FSGS on native kidneys received a renal transplant.Patients were classified in two categories according to whether proteinuria recurred in the nephrotic range(R group,n=42)or not(NR group,n=35).The demographic characteristics of the patients are summarized in Tables1and2.The mean age at onset of nephrotic syndrome was10.4±10years and9.2±6.9years for patients in the NR group and in the R group,respectively (P=NS).The delay between the onset of disease and ESRD was4.8±2.3years in the NR group and3.9±3.3years in the R group(P=NS).There was no difference in the immunosuppressive regimen between the two groups.Four patients in each group received a kidney from a living donor. In the NR group,all patients were genotyped for NPHS1and NPHS2,and none of them had mutation.In the R group, 30/42patients were genotyped for NPHS1and NPHS2. T wo patients had a heterozygous variant,polymorphism (pA242V)[22].On native kidney biopsies,MCD was initially observed in10(children exclusively)out of77cases,and repeat re-nal biopsies showed the development of FSGS in all these patients.All variants were present on native kidneys.The1322G.Canaud et al.Table1.Characteristics of patients in the NR groupAge at onset Delay betweenof proteinuria FSGS variant on Number of onset and Patients(years)native kidney glomeruli ESRD(years) 10.8CELL107211COL14 1.633MCD then NOS16245TIP11852NOS235610MCD then CELL1627 3.3NOS12882COL9 2.593NOS2241011CELL135112COL147120.3MCD then NOS8 1.8132NOS256143CELL177150.5NOS117165CELL192171NOS294186TIP254192NOS142204CELL166211NOS119226NOS1092318MCD then PH1342412NOS2062524MCD then NOS2212622PH1882731TIP97.22814CELL9 2.42917CELL743029COL11 3.13119NOS12 3.73216NOS1543334CELL10 5.53428NOS953518NOS9 4.5Mean±SD10.4±1014±5.6 4.8±2.3FSGS variant observed on native kidney biopsies was NOS in38cases(49.3%),CELL in17cases(22.2%),COL in 10cases(12.9%),PH in6cases(7.8%)and TIP in6cases (7.8%).There were no differences in FSGS variant repre-sentation between R and NR groups(Figure2).After kidney transplantation,recurrence of proteinuria occurred immediately in32/42cases(children n=28, adults n=4),early in9/42cases(children n=4,adults n=5)and late in1/42case(adult n=1).The treatment for recurrence consisted of intravenous(IV)cyclosporine A(CsA)+plasmapheresis(PP)in10cases,CsA IV+PP+ a high dose of steroids in3cases,CsA IV alone in10 cases,oral CsA+PP in3cases,oral CsA alone in8cases, oral CsA+PP+rituximab in1case,cyclophosphamide+ PP in1case,CsA IV+PP+rituximab in4cases,CsA IV+enalapril in1case and immunoadsorption in1case (Table2).The distribution of glomerular lesions in transplanted kidneys with recurrence of INS(group R,n=42)is sum-marized in Table2and Figure2.MCD was the main histo-logical feature observed in32out of33biopsies performed early after recurrence,D6to D60days after transplantation. Only one patient had already developed and FSGS lesion (Table2:patient38,PH variant on Day15).At Month3,an Fig.2.Distribution of FSGS variants on a native kidney of patients with or without recurrence of proteinuria in the nephrotic range.FSGS lesion was observed in11out of39cases.At Month 12,FSGS lesion was observed in14out of37cases.Sev-enteen of the42patients went into complete and persistent remission,and none of them developed FSGS(2patients1323Variants of FSGS and risk of recurrenceT a b l e 2.C o u r s e o f F S G S v a r i a n t s o n k i d n e y b i o p s i e s p e r f o r m e d i n t r a n s p l a n t r e c i p i e n t s w i t h r e c u r r e n c e o f p r o t e i n u r i a i n t h e n e p h r o t i c r a n g eT r a n s p l a n t D e l a y b i o p s y a t T r a n s p l a n t T r a n s p l a n t A g e a t b e t w e e n t i m e o f b i o p s y o n b i o p s y o n o n s e t o f F S G S v a r i a n t o n o n s e t a n d r e c u r r e n c e M o n t h 3M o n t h 12p r o t e i n u r i a n a t i v e k i d n e y (n u m b e r E S R D T i m e o f R e s p o n s e t o E v o l u t i o n a f t e r (n u m b e r o f (n u m b e r o f (n u m b e r o f P a t i e n t s (y e a r s )o f g l o m e r u l i )(y e a r s )r e c u r r e n c e T r e a t m e n tt r e a t m e n t t r a n s p l a n t a t i o n g l o m e r u l i )g l o m e r u l i )g l o m e r u l i )O t h e r b i o p s y18N O S (n =12)2D 0C s A I V +P PP a r t i a l T r a n s p l a n t e c t o m y M 12M C D M 1(n =22)M C D (n =12)M C D (n =14)22N O S (n =12)4D 0C s A I V +P PN o r e m i s s i o n T r a n s p l a n t e c t o m y M 1M C D M 1(n =16)X X 36M C D (n =14)t h e n C E L L (n =8)7D 0C s A I VP a r t i a l E S R D 9y e a r sM C D M 1(n =15)M C D (n =9)M C D (n =21)C O L M 24,N O S M 9641.5N O S (n =9)1.5D 0C s A o r a lN o r e m i s s i o n T r a n s p l a n t e c t o m y M 3M C D M 1(n =18)M C D (n =13)X510N O S (n =18)2D 0I m m u n o a d s o r p t i o nC o m p l e t e f o l l o w e d b y r e l a p s e T r a n s p l a n t e c t o m y M 48M C D D 15(n =7)M C D (n =15)M C D (n =14)T r a n s p l a n t e c t o m y M 48C O L62C O L (n =17)3D 0C s A I V +P PC o m p l e t e a n d p e r s i s t e n t M CD D 10(n =12)M C D (n =7)M C D (n =18)M C D M 6270.5C E L L (n =11)2D 3C s A I V C o m p l e t e f o l l o w e d b y r e l a p s e a t M 36T r a n s p l a n t e c t o m y M 60M C D M 1(n =14)C E L L (n =8)C E L L (n =9)C E L L M 36,T r a n s p l a n t e c -t o m y M 60N O S82M C D (n =20)t h e n P H (n =6)7D 0C s A I V +P PC o m p l e t e a n d p e r s i s t e n t M CD (n =12)M C D (n =21)91.3N O S (n =8)5D 0C s A I VC o m p l e t e a n d p e r s i s t e n t M CD (n =24)M C D (n =10)M C D M 108101.5N O S (n =10)8D 0C s A I V +P P +R i t u x i m a b C o m p l e t e a n d p e r s i s t e n t M C D D 14(n =12)M C D (n =12)M C D (n =18)M C D M 36112T I P (n =12)12D 0C s A I VC o m p l e t e a n d p e r s i s t e n t M CD D 15(n =19)M C D (n =15)M C D (n =9)128M C D (n =14)t h e n N O S (n =7)3M 6C s A o r a l P a r t i a lD i e d a t M 144M C D M 6(n =12)M C D (n =9)M C D M 36133N O S (n =8)17D 0C s A I V +P P C o m p l e t e a n d p e r s i s t e n t M C D M 1(n =22)M C D (n =18)M C D (n =8)146C E L L (n =20)4D 0C s A o r a l P a r t i a lT r a n s p l a n t e c t o m y M 36M C D (n =11)M C D (n =14)T r a n s p l a n t e c t o m y M 36N O S 152.1N O S (n =15)3.2D 1C s A I V +e n a l a p r i l C o m p l e t e a n d p e r s i s t e n t M C D (n =7)M C D (n =11)M C D M 96169C O L (n =7)4D 20C s A I V +P P N o r e m i s s i o nT r a n s p l a n t e c t o m y M 60M C D D 10(n =24)M C D (n =11)N O S (n =6)T r a n s p l a n t e c t o m y M 60N O S 178N O S (n =9)7D 0C s A o r a lN o r e m i s s i o nT r a n s p l a n t e c t o m y M 9M C D M 1(n =21)N O S (n =9)X T r a n s p l a n t e c t o m y M 9N O S 187N O S (n =11)4D 1C s A I V +P PC o m p l e t e f o l l o w e d b y r e l a p s e a t M 48M CD (n =18)M C D (n =11)M C D M 48,M C D M 961910M C D (n =11)t h e n N O S (n =12)0.4D 1C s A I VC o m p l e t e f o l l o w e d b y r e l a p s e a t M 9E S RD M 96M C D D 21(n =20)M C D (n =9)C O L (n =14)C O L M 9203N O S (n =14)9D 8C s A I V +P P +r i t u x i m a bC o m p l e t e a n d p e r s i s t e n tM C D M 1(n =12)M C D (n =15)M C D (n =9)1324G.Canaud et al.215N O S (n =16)5D 0C s A I V C o m p l e t e a n d p e r s i s t e n t M C D (n =20)M C D (n =12)M C D 5y e a r s220.5N O S (n =10)2D 3C s A I VC o m p l e t e a n d p e r s i s t e n t M CD D 22(n =11)M C D (n =7)M C D (n =14)235C O L (n =9)8D 1C s A I V C o m p l e t e a n d p e r s i s t e n t M C D D 10(n =9)M C D (n =24)M C D (n =11)245CE L L (n =10)5D 0C s A I V C o m p l e t e f o l l o w e d b y r e l a p s e o n M 24M C D M 1(n =13)CE L L (n =12)C E L L (n =9)C E L L M 242512C O L (n =11)1D 0C s A +P PC o m p l e t e a n d p e r s i s t e n t M CD (n =14)M C D (n =10)265N O S (n =18)2D 0C s A I V +P P N o r e m i s s i o nM C D M 1(n =19)C O L (n =6)C O L (n =9)C O L M 242727C E L L (n =20)5D 0C s A I V +P P C o m p l e t e a n d p e r s i s t e n t M C D D 8(n =11)M C D (n =9)M C D (n =19)2821N O S (n =16)2D 3C s A I V +P P C o m p l e t e a n d p e r s i s t e n t M C D D 6(n =23)M C D (n =14)M C D (n =15)2919C E L L (n =12)5D 0C s A o r a l N o r e m i s s i o nM C D D 10(n =18)M C D (n =16)N O S (n =18)3014C O L (n =11)3D 0C s A I V +P P +h i g h d o s e s t e r o i d s C o m p l e t e a n d p e r s i s t e n t M C D D 8(n =14)M C D (n =20)M C D (n =11)3119C E L L (n =9)1.5D 4C s A o r a l +P P P a r t i a l r e m i s s i o nM C D D 17(n =18)T I P (n =13)P H (n =12)3212P H (n =7)2D 0C s A o r a l N o r e m i s s i o n T r a n s p l a n t e c t o m yM 6M C D D 12(n =16)C O L (n =17)XT r a n s p l a n t e c t o m y M 6C O L3320T I P (n =9)2D 2C s A o r a l P a r t i a l r e m i s s i o n M C D D 10(n =11)M C D (n =16)N O S (n =14)3412C O L (n =11)1.6D 0C s A I V +P P +h i g h d o s e s t e r o i d s C o m p l e t e a n d p e r s i s t e n t M C D (n =20)M C D (n =22)3516C E L L (n =13)0.3D 4C s A I V +P P +r i t u x i m a b N o r e m i s s i o nT r a n s p l a n t e c t o m yM 36M C D D 60(n =10)M C D (n =19)N O S (n =12)T r a n s p l a n t e c t o m y M 36C O L368M C D (n =11)t h e n N O S (n =16)1D 0C s A o r a l +P P +r i t u x i m a b N o r e m i s s i o nP H (n =14)P H (n =16)3717P H (n =9)3D 0C s A o r a lN o r e m i s s i o nC E L L (n =11)C E L L (n =20)3812N O S (n =21)1D 1C s A o r a l +P P P a r t i a lP H D 15(n =17)P H (n =14)P H (n =9)3918T I P (n =15)3D 8C s A I V +P P +h i g h d o s e s t e r o i d s C o m p l e t e a n d p e r s i s t e n t M C D D 10(n =19)M C D (n =17)M C D (n =14)4016P H (n =14)1.5D 1C s A I VN o r e m i s s i o nT r a n s p l a n t e c t o m y M 6M C D D 11(n =21)C O L (n =10)XT r a n s p l a n t e c t o m y M 6C O L4122N O S (n =17)4D 10C s A o r a l +P PN o r e m i s s i o nM C D D 12(n =18)C O L (n =11)C O L (n =9)428N O S (n =16)1.2D 0C y c l o p h o s p h a m i d e +P PN o r e m i s s i o n T r a n s p l a n t e c t o m y M 24M C D D 10(n =16)M C D (n =20)C E L L (n =12)T r a n s p l a n t e c t o m y M 24C E L LM e a n ±S D 9.2±6.93.9±3.319.5±11.6D ,d a y ;M ,m o n t h ;C s A ,c y c l o s p o r i n e ;I V ,i n t r a v e n o u s ;P P ,p l a s m a p h e r e s i s ,CE L L ,c e l l u l a r ;C O L ,c o l l a p s i n g ;N O S ,m o t o t h e r w i s e s p e c i f i e d ;T I P ,t i p l e s i o n ,P H ,p e r i h i l a r ;M C D ,m i n i m a l c h a n g e d i s e a s e ;X ,r e t u r n i n d i a l y s i s .1325Variants of FSGS and risk of recurrenceTable3.Course of FSGS variants on kidney biopsies performed in transplant recipients with recurrence of proteinuria on successive kidney transplant Second allograft Third allograftBiopsy at time Biopsy at timeof recurrence Month3Biopsy Month12of recurrence Month3Month12 Native First(number of(number biopsy(number(number of(number biopsy(number Patients kidney allograft glomeruli)of glomeruli)of glomeruli)glomeruli)of glomeruli)of glomeruli) 3CELL COL MCD(n=12)NOS(n=10)NOS(n=7)MCD(n=9)MCD(n=18)MCD(n=16) 7CELL NOS MCD(n=23)MCD(n=14)X19NOS COL MCD(n=16)MCD(n=15)NOS(n=8)32PH COL MCD(n=15)MCD(n=12)TIP(n=12)36NOS PH MCD(n=9)MCD(n=20)MCD(n=14)42NOS CELL MCD(n=7)MCD(n=18)MCD(n=20)MCD(n=17)MCD(n=12)MCD(n=9) CELL,cellular;COL,collapsing;NOS,not otherwise specified;TIP,tip lesion;PH,perihilar;MCD,minimal change disease;X,return in dialysis.are on long-term permanent therapy with PP).Conversely, patients who never achieved complete and sustained re-mission developed FSGS lesions.Histological findings on native kidneys were not predictive of either recurrence or category of FSGS variant on3-month or12-month biop-sies,and only3patients(Table2:patients7,16and24) had the same variants on native kidney biopsies and on 12-month transplant biopsies.In four cases(Table2:pa-tients3,7,31and35),the FSGS variant changed during the post-transplant course.Six patients experienced a recurrence on a second allo-graft and two on a third one.Histological findings were different from native kidney on the first,second or even the third allograft(Table3).In the recurrence group,6/42patients lost their graft during the first year and12/42patients during the first5 years.The five-year graft survival in this group was71.5%. In the group exempt of recurrence,the5-year graft survival was85.5%Only one patient had an electron microscopy study(Ta-ble2:patient6)at the time of nephrotic-range proteinuria recurrence with the observation of foot process effacement. DiscussionAn early transplant biopsy in patients with recurrence of nephrotic-range proteinuria without signs of rejection shows normal glomeruli by light microscopy and no de-posit by immunofluorescence.Electron microscopy,when performed,shows widespread foot process effacement.Le-sions of FSGS develop only after several weeks as a con-sequence of persistent proteinuria.Therefore,the term of FSGS recurrence is confusing,and in this paper the re-currence was considered when massive proteinuria occurs soon after transplantation even if renal biopsies did not show histological lesion of FSGS.In the absence of glomerular immune deposits and in the absence of histological sign of rejection,the diagnosis of recurrence is almost certain. Moreover,those patients with recurrence of massive pro-teinuria who respond to therapy with complete and sus-tained remission of proteinuria do not develop FSGS as it was the case in17/42patients in our series.In native kidney,we found a distribution of FSGS vari-ants different from the adult American population,but sim-ilar to the only paediatric report available[23–25].NOS and CELL were the most frequent variants observed,with a greater representation of NOS,supporting the idea that it is the commonest variant even in a paediatric popula-tion.Interestingly,we observed TIP(7.8%)and PH(7.8%) variants in children.The fact that all FSGS variants were present in the native kidney,even the TIP lesions that are often steroid sensitive and have a more favourable outcome [15],demonstrates that all FSGS variants can lead to ESRD. Interestingly,perihilar variant,usually observed in the sec-ondary form[26],was present in native kidneys of four patients who experienced recurrence.Furthermore,after transplantation,two patients suffering from recurrence had a perihilar variant.At the time of proteinuria recurrence,MCD was the most frequently histological pattern.Only one patient(patient38) had already developed FSGS(PH variant)on Day15,on the transplant kidney.One might hypothesize that this le-sion was present in the donor kidney.But,this donor had no past history of proteinuria,and pre-transplant biopsy did not show any glomerular abnormality.Furthermore, the following course of this recipient was concordant with recurrence.Thus,we interpreted this as the result of a prob-ably uncontrolled and more aggressive disease leading to the rapid constitution of FSGS.Interestingly,early FSGS was also observed in two patients in a recent study(COL on Day7and NOS on Day10)[21].At Month3and subsequently,the incidence of MCD decreased progressively while FSGS was increasingly ob-served in those patients with persistent proteinuria.Con-versely,at1year,there was a strong association between sustained remission and the absence of FSGS.So,on se-rial biopsies,the incidence of MCD decreased over time and the incidence of FSGS variants increased.As we previously observed,patients who achieved complete remission under intensive and prolonged treatment of recurrence did not develop FSGS lesion[27].These serial biopsies indicated a correlation between the clinical status and histological findings.During this period(1984–2007),many treatments were used to treat INS recurrence with good efficiency(5-year graft survival is71.5%in our study population).Concor-dant studies had demonstrated the relative efficiency of the association of intravenous cyclosporine with plasma ex-changes to obtain and sustain complete remission[27,28].1326G.Canaud et al.Fig.3.Summary of the different patterns of FSGS variant observed on a native kidney of patients with recurrence of proteinuria in the nephrotic range,and variant observed on a transplant kidney.This series confirmed that rituximab did not dramatically change the course of recurrence,and did not consistently reach remission[29,30].The aims of this study were to determine if the Columbia classification of FSGS(1)could predict the risk of re-currence of INS,and(2)the type of FSGS variant after transplantation.Here,we showed that all FSGS variants, even the PH variant,were present in the native kidney of patients with idiopathic NS progressing to ESRD,and that no difference in their distribution was observed between recurrent and non-recurrent patients.In other words,re-currence of proteinuria/NS may be observed whatever the type of FSGS variant and conversely patients with an FSGS variant(COLL,CELL)were usually associated with a se-vere clinical course.On the other hand,all types of FSGS variants were observed in the transplant kidney and we did not find any correlation between the variant present in the native kidney and those occurring in transplanted kidney (Figure3).It should be stressed that only3/21had the same FSGS variant in native and transplanted kidneys.In-terestingly,all FSGS variants were observed in transplanted kidneys with recurrence.The discrepancy between variant native and transplant kidneys may have several explanations.Several factors may play a role in the development of FSGS in nephrotic pa-tients.First of all,the probable immunological course of the disorders may lead to the production of the pathogenic circulating factor.This factor is the same before and after transplantation,and its consequences could be modified by the immunosuppressive regimen.The long-term use of im-munosuppressive drugs,especially calcineurin inhibitors, promotes obliterative vascular lesions and ischaemic ter-ritories that can lead to the development of FSGS.The form associated with ischaemia is usually the collapsing variant[31,32].Secondly,the genetic background of the transplanted kidney,different from the recipient,may be also in question.Supporting this hypothesis,recent data have highlighted that multiple MYH9SNPs and haplotypes were recessively associated with FSGS[33].In addition,the haemodynamic condition associated with a solitary kidney could favour the development of FSGS,or more specifi-cally,the PH variant.T wo other variables render it difficult to classify the variant:firstly biopsy sampling,and sec-ondly the delay between onset of proteinuria and biopsy. In this series,biopsy specimens were serially studied and a mean of13.7glomeruli per biopsy was examined allowing a precise evaluation of FSGS variants.Indeed,late biopsies after the beginning of the disease are more difficult to in-terpret because all FSGS variants found at early stages of the disease might progress towards NOS variant[15].Following transplantation,the rate of recurrence of pro-teinuria varies between30%and50%and is associated with a poor graft outcome[8–14].Several risk factors for recurrence have been identified including older age at on-set on disease,rapid progression to ESRD and recurrence on a previous graft but none of them can clearly distin-guish between patients who will and those who will not be affected by recurrence[34].On native kidneys,the recent Columbia University classification is interesting since it combines epidemiological data on native kidneys with in-formation on the prognosis and response to treatment.The two main variants are NOS and CELL.The COL variant is associated with the worse prognosis i.e.poor or no response to treatment.In contrast,patients with the TIP variant of-ten respond to corticosteroids and rarely progress to ESRD. NOS,PH and CELL variants are associated with an inter-mediate prognosis[23,25,35].The aim of our study was to evaluate the potential value of the Columbia classification of FSGS to predict risk and the type of recurrence,ques-tions that had never been addressed.Ijpelaar et al.recently identified in21cases three distinct patterns of recurrence of FSGS in renal transplant:firstly,recurrence of the same variant of FSGS;secondly,recurrence of the same variant of FSGS after an intermediate state of MCD and a third pattern of recurrence with a different FSGS variant in the allograft[21].The authors found a high rate of concordance of FSGS variant between native and transplanted kidneys. The discrepancy with our results may be explained by sev-eral differences.On the native kidney,diagnosis of FSGS and the type of variant in our series were all performed on renal biopsies with a mean of13.7±4.6(range6–24) glomeruli,whereas it was performed on the nephrectomy specimen in6/19native kidney in the Ijpelaar study.In-deed,the predominant pattern of NOS variant observed in their study is certainly explained by the course of disease progression as it was suggested in the Columbia Classifica-tion[15].In the same way,after kidney transplantation,in the Ijpelaar study,transplant biopsies were performed late, at a mean of783.4days(range10–2555),suggesting that most biopsies at the time of recurrence,which occurred classically early after transplantation,were not available.In our series,renal biopsies were performed within the first3 months of recurrence,an explanation for the diverse type of FSGS lesion observed.In conclusion,our series is the first to report serial biopsy data in kidney transplant recipients with recurrent nephrotic syndrome.Our results suggest that the Columbia classifi-cation of FSGS is not helpful in predicting neither recur-rence nor the histological variant of FSGS when recurrence develops.Conflict of interest statement.None declared.References1.Barisoni L,Schnaper HW,Kopp JB.A proposed taxonomy for thepodocytopathies:a reassessment of the primary nephrotic diseases.Clin J Am Soc Nephrol2007;2:529–5422.Habib R,Levy M,Gubler MC.Clinicopathologic correlations in thenephrotic syndrome.Paediatrician1979;8:325–3481327Variants of FSGS and risk of recurrence。

Vol.22no.152006,pages1928–1929doi:10.1093/bioinformatics/btl268 BIOINFORMATICS APPLICATIONS NOTEGenetics and population analysisSNPStats:a web tool for the analysis of association studies Xavier Sole´1,Elisabet Guino´1,Joan Valls1,2,Raquel Iniesta1and Vı´ctor Moreno1,2,Ã1Catalan Institute of Oncology,IDIBELL,Epidemiology and Cancer Registry,L’Hospitalet,Barcelona,Spain and2Autonomous University of Barcelona,Laboratory of Biostatistics and Epidemiology,Bellaterra,Barcelona,Spain Received on March6,2006;revised on May16,2006;accepted on May18,2006Advance Access publication May23,2006Associate Editor:Charlie HodgmanABSTRACTSummary:A web-based application has been designed from a genetic epidemiology point of view to analyze association studies.Main cap-abilities include descriptive analysis,test for Hardy–Weinberg equili-brium and linkage disequilibrium.Analysis of association is based on linear or logistic regression according to the response variable(quanti-tative or binary disease status,respectively).Analysis of single SNPs: multiple inheritance models(co-dominant,dominant,recessive,over-dominant and log-additive),and analysis of interactions(gene–gene or gene–environment).Analysis of multiple SNPs:haplotype frequency estimation,analysis of association of haplotypes with the response, including analysis of interactions.Availability:/SNPstats.Source code for local installation is available under GNU license.Contact:v.moreno@Supplementary Information:Figures with a sample run are available on Bioinformatics online.A detailed online tutorial is available within the application.The analysis of association between genetic polymorphisms and diseases allows identifying susceptibility genes(Cordell and Clayton,2005).The proper analysis of these studies can be per-formed with general purpose statistical packages,but the researcher usually needs the assistance of additional software to perform spe-cific analysis,like haplotype estimation,and results from different packages are difficult to integrate.We present a free web-based tool to help researchers in the analysis of association studies based on SNPs or biallelic markers. Both the selection of analysis and the output have been designed from a genetic epidemiology perspective.This application can also be used for learning purposes.We have written(in Spanish)an analysis guide with detailed explanations(Iniesta et al.,2005).A similar extensive help in English can also be found on the website. The software is used following three steps,with the possibility of performing multiple analyses in one session.The steps are as follows.(1)Data entry.Raw data in tabular form can be pasted in a window or uploaded from a textfile.Variables can be named and the user can choose thefield delimiter and the missing value code(Supplementary Figure1).SNPs should be coded as genotypes with each allele separated by a slash(e.g.‘T/T’,‘T/C’,‘C/C’).(2)Data processing.A list with the variables read by the appli-cation is presented with an initial suggestion about the type: quantitative,categorical or SNP,which can be modified(Supple-mentary Figure2).The user is prompted to select those needed for the analysis and to specify which one is the response,which may be binary(disease status)or quantitative.For categorical variables, including SNPs,the user can reorder the categories.Thefirst one will be treated as reference category in the analysis.The application assumes that the main interest is the analysis of the SNPs in relation to the response.Other variables selected with type quantitative or categorical will be added to the regression models for analysis as covariates and treated as potential confounders.(3)Analyses customization.The third step requests the selection of the desired statistical analyses that will be described later in this article(Supplementary Figure3).Regarding the statistical analysis,the association with disease is modeled depending on the response variable.If binary,the applica-tion assumes an unmatched case–control design and unconditional logistic regression models are used.If the response is quantitative, then a unique population is assumed and linear regression models are used to assess the proportion of variation in the response explained by the SNPs.The association for each SNP is analyzed in turn and adjusted for the selected covariates.If more than one SNP are selected,then the application assumes that haplotype analysis is appropriate. Haplotype frequencies are estimated using the implementation of the EM algorithm coded into the haplo.stats package(Sinnwell and Schaid,2005,/mayo/research/ biostat/schaid.cfm).Association between haplotypes and disease appropriately accounts for the uncertainty in the estimation of hap-lotypes for individuals with multiple heterozygous when phase is unknown or when missing values are present(Schaid et al.,2002). Individuals with missing values in the response,in all SNPs or in any covariate are excluded from analysis.The software main page can be found online at http://bioinfo. /SNPstats.The application uses PHP server pro-gramming language to build the input forms,upload data,call the statistical analysis procedures and process the output.The sta-tistical analyses are performed in a batch call to the R package (R Development Core Team,2005,). The contributed packages genetics(Warnes and Leisch,2005) and haplo.stats(Sinnwell and Schaid,2005,http://mayoresearch. /mayo/research/biostat/schaid.cfm)are called to perform some of the analysis.Anonymous use is guaranteed and data areÃTo whom correspondence should be addressed.Ó2006The Author(s)This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License(/licenses/ by-nc/2.0/uk/)which permits unrestricted non-commercial use,distribution,and reproduction in any medium,provided the original work is properly cited.treated as confidential.Source code for local installation(Linux and Windows)is also available under GNU license.SNPStats returns a complete set of results for the analysis,cover-ing from the descriptive statistics to the haplotype analysis.The descriptive statistics returned are the absolute frequencies and pro-portions for categorical variables,and mean,standard deviation and a list of percentiles for the quantitative ones.Always the total valid sample size and the count of missing values are displayed (Supplementary Figure4).Each SNP is described as allele and genotype frequencies.An exact test for Hardy–Weinberg equilibrium is performed(Supple-mentary Figure5).When the response variable is binary,these statistics can be displayed by each response group.The user usually will be interested in checking Hardy–Weinberg equilibrium in the control population.The analysis of association for each SNP can be performed bothfor quantitative or binary response variables.For binary responses, the logistic regression analysis is summarized with genotype fre-quencies,proportions,odds ratios(OR)and95%confidence inter-vals(CI)(Fig.1).For quantitative responses,linear regression is summarized by means,standard errors,mean differences respect to a reference category and95%CI of the differences.SNPStats can also perform analyses of interactions.For simpli-city,models with only one pair of variables interacting can be selected at a time.Three summary tables are shown(Supplementary Figure7).Thefirst one is the cross-classification that uses a com-mon reference category for both interacting variables.ORs or mean differences are estimated,together with95%CI,for all other com-binations.Next tables use the margins as reference category and estimate ORs or mean differences of one variable nested within the other one.A global test for interaction is performed,as well as a test for the interaction in the linear trend of the nested variable.This assumes that the nested variable is ordinal and tests for different trend among categories.This test might be more sensitive than the global one due to the reduction in degrees of freedom.When more than one SNP is included in the analysis,SNPStats offers the possibility of performing linkage disequilibrium(LD)and haplotype analysis.For LD,matrices with selected statistics(D,D0, Pearson’s r and associated P-values)are shown.(Supplementary Figure8).In the analysis of haplotypes,descriptive statistics show the estimated relative frequency for each haplotype(Supplementary Figure9).Cumulative frequencies are also shown to help in the selection of the threshold cut point to group rare haplotypes for further analysis.The association analysis of haplotypes is similar to that of genotypes in that either logistic regression results are shown as OR and95%CI or linear regression results with differences in means and95%CI.The most frequent haplotype is automatically selected as the reference category and rare haplotypes are pooled together in a group.The analysis of haplotypes assumes a log-additive model by default,but dominant and recessive models are available as alternative choices.When haplotypes are selected for interaction tables similar to the genotype interaction ones are shown,replacing the genotypes by haplotypes(Supplementary Figure10).This analysis of interactions and presentation of the results is unique to the available alternatives explored and is an important contribution to the analysis of genetic epidemiology studies,often focused on testing for gene–environment interactions(Lake et al.,2003).As a limitation,we are aware that the selection of the available analysis has been done for the most frequent profile but might not be adequate in some instances.We plan to implement in future versions more response types:survival data for studies of prognosis, multinomial data for categorical responses with more than two categories and paired designs(matched case–control or nested case–control).ACKNOWLEDGEMENTSFunding support from the Spanish Instituto de Salud Carlos III (networks of centres RCESP C03/09and RTICCC C03/10). Funding to pay the Open Access publication charges for this article was provided by Instituto de Salud Carlos III(FIS03/0114). Conflict of Interest:none declared.REFERENCESCordell,H.J.and Clayton,D.G.(2005)Genetic association ncet,366, 1121–1131.Iniesta,R.et al.(2005)Ana´lisis estadı´stico de polimorfismos gene´ticos en estudios epidemiolo´gicos.Gac.Sanit.,19,333–341.Lake,S.et al.(2003)Estimation and tests of haplotype–environment interaction when linkage phase is ambiguous.Human Heredity,55,56–65.R Development Core Team(2005)R:a language and environment for statistical computing.R Foundation for Statistical Computing,Vienna,Austria.ISBN 3-900051-07-0.Schaid,D.J.et al.(2002)Score tests for association between traits and haplotypes when linkage phase is ambiguous.Am.J.Hum.Genet.,70,425–434.Sinnwell,J.P.and Schaid,D.J.(2005),haplo.stats:statistical analysis of haplotypes with traits and covariates when linkage phase is ambiguous.R package version1.2.2. Warnes,G.and Leisch,F.(2005)Genetics:Population Genetics.R package version1.2.0.Fig. 1.Sample output of association models for binary response.Five inheritance models are fitted,which correspond to different parameterizations or groupings of the genotypes.Akaike’s Information Criterion(AIC)and Bayesian Information Criterion(BIC)are calculated to help the user in the selection of the best model for a specific SNP.SNPStats:analysis of association studies web tool1929。