Gene expression signatures in neutrophils exposed to glucocorticoids a new paradigm to help explain

Gene expression signatures in neutrophils exposed to glucocorticoids:A new paradigm to help explain ‘‘neutrophil dysfunction’’in parturient dairy cowsJeanne L.Burton a ,b ,*,Sally A.Madsen a ,Ling-Chu Chang a ,Patty S.D.Weber a ,Kelly R.Buckham a ,Renate van Dorp b ,Mary-Clare Hickey c ,Bernadette Earley caImmunogenetics Laboratory,Department of Animal Science,Michigan State University,1205E Anthony Hall,East Lansing,MI 48824,USAbCenter for Animal Functional Genomics,Department of Animal Science,East Lansing,MI 48824,USA cAnimal Health and Welfare Department,Teagasc,Grange Research Center,Co Meath,IrelandAbstractNeutrophils are the first line of immunity against most pathogens that infect cattle.These normally short-lived white blood cells develop from myeloid-lineage cells in bone marrow.Upon maturation,bone marrow neutrophils are released into the circulation where they marginate on inflamed blood vessel endothelial cells and migrate through them into the area of infection.Once migrated,neutrophils do not reenter the circulation,but rather,perform their bactericidal functions and die by apoptosis in the tissue.The cytokine and hormonal milieu of the blood and extracellular tissue fluid can influence neutrophil development and immunity-related activities,but the molecular basis of these phenotypic changes and physiological benefits or drawbacks of them are poorly understood.In the current paper,we review new gene expression information that resulted from two of our functional genomics studies designed to evaluate effects of glucocorticoid hormones on bovine neutrophils.This work provides one model to describe complex changes that occur in neutrophils as the cells respond to glucocorticoids,which might act to alter the cells’functional priorities and tip the delicate balance between health and disease during stress,including at parturition.A bovine immunobiology microarray and real time RT-PCR were used to study blood neutrophils collected during the natural surge of endogenous glucocorticoid (cortisol)in parturient dairy cows and bone marrow neutrophils collected from glucocorticoid (dexamethasone)-treated dairy steers.The gene expression signatures we observed led us to perform additional phenotyping of the neutrophils and correlation analyses,which together painted a picture suggesting that glucocorticoids have key roles in modulating neutrophil development,life span,and tissue defense functions during parturition and hormone therapy.Based on these observations,we postulate that glucocorticoids orchestrate adaptive changes in the entire neutrophil system that support increased cell numbers and longevity in blood and heightened remodeling activity in tissues,while at the same time decreasing some important antimicrobial defense activities of the cells.Thus,our functional genomics studies have enabled us to elucidate multiple consequences of neutrophil exposure to glucocorticoids,highlighting a probable role for this interaction in the/locate/vetimmVeterinary Immunology and Immunopathology 105(2005)197–219*Corresponding author.Tel.:+15173539702/9815(Laboratory);fax:+15173531699.E-mail address:burtonj@ (J.L.Burton).0165-2427/$–see front matter #2005Elsevier B.V .All rights reserved.doi:10.1016/j.vetimm.2005.02.012induction of parturition and partly explaining why some parturient dairy cows may experience heightened incidence and severity of in flammatory diseases like mastitis.#2005Elsevier B.V .All rights reserved.Keywords:cDNA microarray;Neutrophils;Bone marrow cells;Cortisol;Dexamethasone1.BackgroundElevated blood levels of glucocorticoid hormone during stress and exogenous administration of the steroid have been associated with increased suscept-ibility to infectious diseases in cattle that affect production,including mastitis in parturient dairy cows and shipping fever in transportation stressed feedlot animals (Roth and Kaeberle,1982;Burton and Erskine,2003).Our groups are interested in the effects of glucocorticoids on bovine neutrophils because of the critical role these leukocytes play in innate immune defense against such diseases.We have studied bovine neutrophils following administration of the potent glucocorticoid,dexamethasone (Burton et al.,1995;Burton and Kehrli,1995;Weber et al.,2001,2004;Chang et al.,2004)and during parturition when there is a surge in blood concentrations of the endogenous glucocorticoid,cortisol (Preisler et al.,2000;Weber et al.,2001;Madsen et al.,2002,2004;Chang et al.,2004).Pronounced neutrophilia occurs in both cases,showing that glucocorticoids profoundly alter neutrophil homeostasis.The vast majority of circulating cells in the early phase of glucocorticoid-induced neutrophilia are segmented neutrophils that have de-marginated from blood vessels due to down-regulated expression of surface adhesion molecules (Burton et al.,1995;Weber et al.,2004)and,possibly,up regulation of anti-adhesion molecules (Goulding et al.,1998;La et al.,2001).However,signi ficant release of mature (segmented and band)neutrophils from bone marrow is clearly detectable in cattle !24h following glucocorticoid challenge and may contri-bute to additional waves of neutrophilia later in the steroid response (Burton and Kehrli,1995;Weber et al.,2004).Thus,when studying in vivo effects of glucocorticoids on bovine neutrophils,it is relevant to do so in both blood and bone marrow pools of the cells.In cattle,circulating neutrophils are sensitive to glucocorticoids because they possess high expressionof glucocorticoid receptors (GR).In most mammalian cells,including bovine neutrophils (Chang et al.,2004),GR exists in two main isoforms,GR a and GR b (Hollenberg et al.,1985).However,only GR a has glucocorticoid binding capacity (Bamberger et al.,1996;Kino et al.,2001).We have shown that bovine blood neutrophils have $8-fold higher GR a mRNA abundance compared with the cells ’more immature counterparts developing in bone marrow (Fig.1).Thus,bovine blood neutrophils are predicted to beJ.L.Burton et al./Veterinary Immunology and Immunopathology 105(2005)197–219198Fig.1.Bovine blood neutrophils express $8-fold higher abundance of mRNA for the hormone-binding alpha isoform of the glucocor-ticoid receptor (GR a )than immature neutrophil-lineage cells devel-oping in bone marrow.Data were generated by quantitative slot blot analysis with b -actin as the control gene (Madsen et al.,2002),GR a and b -actin cDNA probes described in Weber et al.(2001),and RNA samples assayed in duplicate.The RNA samples were obtained from bone marrow neutrophil-lineage cells and blood neutrophils of healthy steers (n =12)using Percoll density gradients (Weber et al.,2004).Bone marrow F1cells contained $70%early-immature myeloblasts and promyelocytes,F2cells $55%late-immature myelocytes and metamyelocytes,and F3cells $70%band and segmented neutrophils.The P -value indicates the level of signi fi-cance for the main effect of granulocyte maturation fraction,obtained by general linear models analysis of the data,and a and b above individual bars indicate signi ficant (P 0.05)differences in mean (ÆS.E.M.)GR a mRNA abundance by maturation fraction.highly sensitive to changes in circulating glucocorti-coid concentrations.In cells not exposed to stress or therapeutic levels of glucocorticoids,most GR a are located in the cytoplasm as complexes with accessory proteins that maintain the receptors in a high af finity hormone-binding state (Bamberger et al.,1996).When glucocorticoid concentrations rise inside the cell due to elevated extracellular concentrations,hormone binding to GR a activates the receptor causing it to dissociate from its protein complex.This allows GR a to translocate into the cell ’s nucleus (Eicher and Burton,2004).In the nucleus,hormone-bound GR a employs multiple mechanisms to change expression of hormone sensitive genes,including binding to DNA and other transcription factors to in fluence rates of gene transcription and to mRNAs already existing in the cytoplasm to affect their stability (Bamberger et al.,1996;Newton,2000;Beato and Klug,2000;Almawi and Melemedjian,2002).In previous studies,we demonstrated that blood neutrophils of dairy cows express functional GR a proteins,staining brightly with fluorescently labeled hormone when collected during mid and late lactation and exiting the cells ’cytosol in response to the cortisol surge at parturition (Preisler et al.,2000;Fig.2a).We also showed that glucocorticoid-induced loss of GR a from the cytosol is time-and dose-dependent when cultured neutrophils are treated with the steroid in vitro (Chang et al.,2004and our unpublished data).In both scenarios of endogenous and exogenous gluco-corticoid challenge,GR a loss from the neutrophils ’cytosol is associated with acute and pronounced changes in expression of two genes that regulate neutrophil behavior,including margination (CD62L;Weber et al.,2001,2004)and programmed cell death (Fas;Chang et al.,2004).Thus,GR a activation by glucocorticoids in circulating neutrophils causes distinctive phenotypic changes in the cells via altered expression of hormone-sensitive genes.However,these changes are relatively short-lived ($48h),possibly due to the fact that GR a activation rapidly decreases abundance of its own mRNA (Fig.2b)byJ.L.Burton et al./Veterinary Immunology and Immunopathology 105(2005)197–219199Fig.2.The surge in blood cortisol during bovine parturition (a)is correlated (r =À0.71;P =0.06)with rapid down regulation of glucocorticoid receptor-alpha (GR a )mRNA abundance in blood neutrophils (b),possibly due to the ability of glucocorticoid to reduce the rate of transcription of its own receptor gene (c).Blood for serum cortisol assays and neutrophil GR a mRNA abundance assessment (a and b)was collected from three primiparous Holstein cows on days À7,0,0.25,and 1relative to parturition (on day 0).Serum cortisol was measured in duplicate by EIA (Assay Designs,Inc.,Ann Arbor,MI)and GR a mRNA by quantitative real time RT-PCR (Madsen et al.,2004).GR a mRNA abundance on days 0,0.25,and 1were relative to expression on day À7(b)and calculated using the 2ÀDD Ct method (Livak and Schmittgen,2001).Nuclear run on assay was used to generate data in panel (c).Brie fly,Percoll-isolated blood neutrophils of four healthy Holstein steers (3–4months of age)were left untreated (Control)or treated with 10À7M of dexamethasone (Dex)for 4h prior to isolation of the cells ’nuclei,which were snap frozen before being used to transcribe new radi-olabeled mRNA in vitro.GR a and b -actin (control)cDNA probes (Weber et al.,2001)were spotted in excess on nylon membrane strips and the blots probed with the newly synthesized,radiolabeled mRNA.Densitometry was used to quantify abundance of tran-scribed GR a and b -actin mRNA,and the GR a transcription ratewas expressed as the mRNA abundance ratio of GR a :b -actin.P -values associated with each graph show the level of signi ficance for the main effect of parturition (a and b)or treatment (c),and a –c above individual bars indicate signi ficant (P 0.05)differences in means (ÆS.E.M.)shown in each plot.repressing the transcription rate of its gene(Fig.2c). Thus,choices of blood sampling times relative to glucocorticoid challenges in vivo and in vitro are critical in experiments designed to study neutrophil gene expression changes in response to the hormone.Results of our studies cited above led us to ask several questions about possible additional impacts of glucocorticoids on the bovine neutrophil system.In particular,we wanted to know what other genes might be affected in blood neutrophils during the cortisol surge at parturition.This is important because such knowledge may elucidate why blood neutrophils have depressed bactericidal activities that correspond with increased susceptibility to opportunistic infectious diseases,such as coliform mastitis,during parturition (reviewed in Kehrli and Harp,2001;Burton and Erskine,2003;Burvenich et al.,2003).Also,though neutrophil-lineage cells developing in bone marrow express low levels of GR a(Fig.1),we sought to understand if glucocorticoid challenge could never-theless impact gene expression in this only renewable source of blood neutrophils.To begin to address these questions we collected blood neutrophils from dairy cows around the parturient cortisol surge,and bone marrow cells from dairy steers treated with a high dose of dexamethasone.Total RNA isolated from these cells was used to obtain preliminary gene expression profiles by cDNA microarray and quantitative real time RT-PCR(Q-RT-PCR)analyses.2.Gene expression profiling of blood neutrophils during the surge in cortisol at parturition In ourfirst cDNA microarray experiment,blood from four primiparous Holstein cows was collected close to the surge in parturient cortisol.These animals were healthy and did not experience retained placenta, metritis,or mastitis at or following parturition. Neutrophils were enriched from the blood samples to94–98%purity using Percoll density gradients(as in Weber et al.,2001)and their gene expression patterns profiled using our group’s third generation BOTL(for bovine total leukocyte)cDNA microarrays(clones searchable at ).As we reported in Madsen et al.(2004),expression changes (P<0.05)for302genes were detected over the time the cows transitioned from the dry period(dayÀ7relative to parturition)through parturition(day0)and into thefirst day of lactation(days0.25and1). Pronounced changes occurred for most affected genes at parturition and6h postpartum(day0.25),when blood cortisol concentrations and neutrophilia were highest(Fig.2a and Madsen et al.,2004).2.1.Apoptosis gene expression profiles during the cortisol surge at parturitionThe largest ontological cluster of affected known genes(42genes)in our microarray study encoded apoptosis regulatory proteins.Quantitative real time RT-PCR was used to substantiate expression changes for some of the best known of these apoptosis genes (Fig.3).Percoll enriched blood neutrophils from three additional primiparous parturient cows were used for RNA isolations and cDNA synthesis in this work,as per our described methods(Madsen et al.,2004;Weber et al.,2004).Beta-actin was selected as the control gene for Q-RT-PCR because its amplification efficiency was the same as that for all test genes and abundance of its mRNA in bovine blood neutrophils does not change through the peripartum period(Weber et al.,2001; Madsen et al.,2002).Using this approach,we verified that expression of genes encoding4death inducing signaling complex(DISC)proteins(i.e.,FADD,Daxx, FLASH,and RIP),which are associated with Fas activation by its ligand(FasL;Fig.3),were profoundly down regulated during parturition(Madsen et al., 2004).In addition,parturition down-regulated gene expression for the pro-apoptotic Bcl-2family member, Bak,while simultaneously up regulating expression of genes encoding the anti-apoptotic Bcl-2homologue, A1,and the potent pro-survival chemokine,IL-8 (Fig.3;Madsen et al.,2004).These preliminary gene expression data supported a hypothesis that parturition reprograms neutrophils for extended survival in blood. In addition,IL-8is known to enhance cell survival by signaling the down regulation of Bak gene expression (Grutkoski et al.,2002).Therefore,parturition may also induce an autocrine short-loop in neutrophils whereby induction of IL-8suppresses Bak expression to support continued survival of the cells.In turn, survival induction in neutrophils could partly explain the neutrophilia that occurs in conjunction with the cortisol surge at parturition(Fig.2a;Preisler et al., 2000;Weber et al.,2001).J.L.Burton et al./Veterinary Immunology and Immunopathology105(2005)197–219 200Since publishing those original data we have Q-RT-PCR profiled expression patterns for an additional six apoptosis regulatory genes,including pro-apoptotic Fas,FasL,and DAP5,and anti-apoptotic TRAF6, TANK,and BAFF(Fig.3).Thefirstfive genes listed were significantly down regulated between parturition and thefirst day of lactation,while BAFF was up regulated at parturition(Fig.4).BAFF is a relatively newly identified member of the growing family of TNF ligands and is best known for its critical role in development of long-lived mature peripheral B cells (Chin et al.,1999;Mak and Yeh,2002;Mackay and Ambrose,2003).Its pronounced up regulation in blood neutrophils at parturition argues in favor of cell survival.Down regulation of the TRAF-6and TANK genes is not necessarily congruent with a survival gene expression signature during parturition because protein products of these genes act to desensitize cells to pro-apoptotic signals from activated TNF-a receptors(Fig.3;Bradley and Pober,2001;Wu and Arron,2003).However,TRAF-6and TANK act downstream of other key death inducing molecules, including Fas and FasL,which are highly potent in their ability to induce apoptosis in bovine neutrophils and were dramatically down regulated in the cells during parturition(Fig.4).Presence of DAP5proteins in cells ensures that continued translation of death effector proteins occurs even as cells undergo apoptosis (Fig.3;Henis-Korenblit et al.,2000,2002).Thus,the down regulation of DAP5gene expression that we observed during parturition(Fig.4)would be expected to have a positive impact on neutrophil survival.So,J.L.Burton et al./Veterinary Immunology and Immunopathology105(2005)197–219201Fig.3.Blood neutrophils are normally programmed to undergo rapid apoptosis through a combination of death receptor signaling initiated at the plasma membrane and spontaneous death initiated through mitochondrial signaling.For example,ligation of membrane bound Fas by its ligand (FasL)recruits a variety of adaptor proteins[the death initiating signaling complex(DISC)molecules FADD,FLASH,RIP,and Daxx]to Fas’cytoplasmic tails,which then recruit and sequentially activate caspases8and3to effect cell death.At the mitochondrial membrane,reactive oxygen species(ROS)generated during oxidative metabolism(not shown)and pro-apoptotic Bcl-2family members Bak and Bax induce release of cytochrome c(cyt c)and Smac/DIABLO(not shown),which activate caspase9to directly and indirectly activate caspase3and cell death.This spontaneous form of apoptosis is normally uninhibited in circulating neutrophils because electron transport chain activity is high and expression of key anti-apoptotic molecules like A1(Bcl-2homologue)and IL-8(chemokine)is low or absent.Furthermore,expression of DAP-5ensures that the cells continue translating death effector proteins even as they undergo apoptotic cell death.In both pathways,cell death is mediated by caspase-induced cleavage of proteins that are critical for DNA repair,cytoskeletal stability,and plasma membrane integrity.Thus,apoptotic neutrophils are distinguished from viable cells by their reduced size,flipping of inner leaflet phosphatidylserine(PS)residues to the outer leaflet of the plasma membrane,and irreversible fragmentation of genomic DNA(hypodiploidy).Factors that reduce mitochondrial ROS production or induce expression of genes encoding the potent anti-apoptotic proteins A1,IL-8,BAFF,TRAF-6,and TANK temporarily prevent these cellular events and extend neutrophil life span by quenching mitochondrial membrane destabilizing activities of ROS,Bak,and Bax and blocking signaling pathways derived from death receptors such as Fas and TNF-a(TNF-R).while the down regulation we observed in TRAF6and TANK gene expression may lead to neutrophil apoptosis in some circumstances,it is possible that parturient physiology overrides such effects by simultaneously down regulating gene expression for pro-apoptotic Fas,FasL,DAP5,FADD,Daxx,FLASH,RIP,and Bak and up regulating gene expression for anti-apoptotic A1,BAFF,and IL-8.2.2.Apoptosis gene expression pro files correlate with blood cortisol pro files during parturition Because blood neutrophils express abundant GR a mRNA (Fig.1)and protein (Chang et al.,2004),we postulated that cortisol may have been one blood factor responsible for inducing apoptosis gene expression changes in circulating neutrophils of the parturient cows that,by in large,appeared to favor cell survival.To test this possibility,we performed correlation analyses on these data sets using the PROC CORR procedure of SAS.Results in Table 1show that blood cortisol pro files were signi ficantly (P 0.05)associated with gene expression pro files for A1,BAFF,Bak,DAP5,Fas,FLASH,IL-8,and TRAF6,and tended (P =0.06)to be associated with expression pro files for FasL and Daxx.However,other unidenti fied factors also must have participated in the regulationJ.L.Burton et al./Veterinary Immunology and Immunopathology 105(2005)197–219202Fig.4.Expression changes in six neutrophil genes (Fas,FasL,TRAF6,TANK,DAP5,and BAFF)whose protein products regulate apoptosis.Data were generated using quantitative real time RT-PCR (Madsen et al.,2004),with all samples run in duplicate and b -actin used as the control gene (Weber et al.,2001;Madsen et al.,2002).Test RNA for these assays was obtained from Percoll-isolated blood neutrophils of three primiparous Holstein cows on days À7,0,0.25,and 1relative to parturition (on day 0).Gene expression changes for days 0,0.25,and 1were relative to expression on day À7and calculated using the 2ÀDD Ct method (Livak and Schmittgen,2001).P -values above each graph are the level of signi ficance for the main effect of day relative to parturition obtained through repeated measures analysis of the data,and a –c above individual bars indicate signi ficant (P 0.05)differences in mean (ÆS.E.M.)gene expression by day relative to parturition.The PCR primers and conditions used to generate these data are listed under the Links icon on our Center for Animal Functional Genomics web site (/).of specific components of the apoptosis program in bovine neutrophils at parturition because correla-tions between cortisol and gene expression profiles for FADD,RIP,and TANK were not detected (Table1).It is unlikely that estradiol was one of these factors because we(unpublished data)and others(Winters et al.,2003)could not detect estrogen receptor proteins in bovine blood neutro-phils.Also,correlations between the cows’serum estradiol profiles and neutrophil apoptosis gene expression profiles were not significant(P>0.10; not shown).While bovine neutrophils also do not express progesterone receptor mRNA or protein(our unpublished data;Winters et al.,2003),it is possible thatfluctuations in serum progesterone around parturition contributed to the apoptosis gene expression profiles we observed because progester-one can act as a GR a antagonist(Preisler et al., 2000).Thus,it is possible that the parturient drop in blood progesterone as cortisol levels soar may have facilitated glucocorticoid activity on the expression of some apoptosis genes.In support of this possibility,we observed significant(P<0.05) correlations between serum progesterone profiles and neutrophil gene expression profiles for Fas (r=0.703),RIP(r=0.796),BAFF(r=À0.793),IL-8(r=0.724),and TANK(r=0.785).2.3.Cortisol in parturient serum is partly responsible for inducing survival in blood neutrophils To substantiate that the relationships between serum cortisol and apoptosis gene expression profiles we observed translate into changed apoptosis status of the cells,we attempted to phenotype apoptosis of neutrophils in fresh whole blood samples collected from a group of primiparous cows prepartum and at parturition.However,effects of parturition on apop-tosis status were not clear because extremely high animal variation existed in the staining of neutrophils with Annexin-V-FITC,a reagent that binds with phosphatidylserine residues that haveflipped from the inner to the outer leaflet of the plasma membrane during early stages of apoptosis(Fig.3;Vermes et al., 1995).For example,the mean percentage of Annexin-V-FITC+neutrophils in blood from16cows collected $10days prepartum was48.9%(Æ16.20;S.E.M.), with a range of19.1–83.2%.This wide animal variation in Annexin-V-FITC staining in fresh cells is likely related to normal variation in proportions of variously aged neutrophil populations in whole blood (Shidham and Swami,2000).In turn,this is a function of variation in the rates of bone marrow release,cell exit from blood into tissues,and clearance of circulating apoptotic neutrophils by the body’s phagocytic network(Homburg and Roos,1996;Savill, 1997).Thus,it was difficult to conclude whether or not neutrophils in fresh blood samples were less apoptotic overall at parturition versus prepartum.When we instead enriched neutrophils from prepartum blood samples on Percoll density gradients, 4%of the freshly isolated cells stained with Annexin-V-FITC.It thus appears that our neutrophil isolation protocol is relatively selective in its enrichment of blood neutrophils that have no overt signs of this early stage of apoptosis.In fact,the recovery of neutrophils in this procedure averages$75%when neutrophilia is not present(and higher when it is).The majority of neutrophil loss(76%)occurs during removal of the buffy coat and top one-third of the red cell pack prior to addition of the remaining red cell pack to the Percoll gradient for neutrophil isolation.A more minor loss (24%)results from neutrophils that remain on or in the Percoll layer because the cells are not dense enough to penetrate through it.The dense neutrophils that do penetrate the Percoll gradient to form the cell pellet atJ.L.Burton et al./Veterinary Immunology and Immunopathology105(2005)197–219203 Table1Pearson product moment correlations(r-values)between serumcortisol profiles and expression profiles for thirteen apoptosis reg-ulatory genes in Percoll-isolated blood neutrophils from parturientprimiparous Holstein cows(n=3)Gene name Effect on apoptosis r-value P-valueA1Anti-apoptotic0.900.002BAFF Anti-apoptotic0.750.04Bak Pro-apoptoticÀ0.820.01Daxx Pro-apoptoticÀ0.720.06DAP5Pro-apoptoticÀ0.730.05FADD Pro-apoptoticÀ0.330.32Fas Pro-apoptoticÀ0.750.04FasL Pro-apoptoticÀ0.710.06FLASH Pro-apoptoticÀ0.800.02IL-8Anti-apoptotic0.730.01RIP Pro-apoptoticÀ0.550.24TRAF-6Anti-apoptoticÀ0.750.04TANK Anti-apoptoticÀ0.650.11Gene expression changes werefirst detected by cDNA microarrayanalysis(Madsen et al.,2004)and further characterized usingquantitative real time RT-PCR(Fig.4).the bottom of the tube may thus represent the youngest population of neutrophils in blood,presumably those that were recently released from the maturation pool in bone marrow.We suspect that use of Percoll isolated cells standardized the starting population of neutrophils in our apoptosis phenotyping assay because,when allowed to age in culture over48h,these cells had significantly extended viability when collected from cows at parturition versus prepartum(Madsen et al., 2003a).This observation did support those from our microarray and Q-RT-PCR experiments,which sug-gested that the cells gain a predominantly pro-survivalgene expression signature around parturition.Addi-tionally,we have shown that sera collected from the cows at parturition(day0)supported viability in significantly more Percoll-isolated steer neutrophils over48h in culture than sera collected prepartum(day À7)or postpartum(days0.25and1;Madsen et al., 2004).This suggested,as our correlation analysis did, that parturient blood does indeed contain factors that promote extended neutrophil life span.However,given the differences in Annexin-V-FITC staining character-istics between neutrophils in fresh whole blood samples versus Percoll-isolated cells,it would seem important for investigators to carefully consider which neutrophil isolation method to use for such experiments and to characterize the apoptosis status of neutrophils employed in various assays.We selected the Percoll density gradient method to isolate neutrophils because it was the quickest way to obtain high numbers of very pure neutrophils with enough quantity of high quality RNA needed for our for our microarray and Q-RT-PCR work.We felt that this was critical to our ability to generate gene expression data from which we could draw conclusions about effects of parturition on gene expression changes in neutrophils.However,we acknowledge that our conclusions about the apoptosis gene expression signatures and phenotypes observed around parturition may apply only to younger popula-tions of the cells and not to all circulating neutrophils. With this caveat in mind,we continued with several additional experiments using Percoll-isolated neutro-phils to study potential relationships between gluco-corticoids in serum and neutrophil apoptosis phenotypes.Given our in vitro apoptosis results and the correlations we detected between blood cortisol profiles and neutrophil apoptosis gene expression around parturition(Table1),we sought to determine if the cortisol component of parturient blood is one factor capable of inducing survival in cultured bovine neutrophils.We pooled the parturient(day0)serum samples,split the pool into two aliquots,and used activated charcoal to remove cortisol(Szafarczyk et al.,1995;McCarty and Schwartz,1999)from one aliquot(Table2).Charcoal treatment also removed progesterone and estradiol from this aliquot(Table2) and,perhaps,other unidentified factors.We added these sera at20%of the volume of culture medium to Percoll-isolated steer neutrophils as the cells aged in vitro for12h or24h prior to apoptosis phenotyping. Representative data from one steer in Fig.5show that compared to the steroid extracted serum(panel b), intact parturient serum(panel a)significantly increased the proportion of viable neutrophils in culture at12h(top panels of Fig.5)and reduced DNA fragmentation in the cells at24h(bottom panels of Fig.5).To clarify which extracted steroid may have been responsible for this effect,neutrophils from additional steers were pretreated for10min with steroid receptor antagonists prior to addition of intact parturient serum for12or24h,followed by apoptosis phenotyping.Results in Fig.6a show that RU486(a GR a/progesterone receptor antagonist),but not tamoxifen(an estrogen receptor antagonist),inhibited the ability of the parturient serum to support neutrophil survival,and did so to the same extent as the steroid extracted serum.Thus,cortisol and(or) progesterone were likely factors responsible for neutrophil survival induction by parturient serum.In afinal experiment,we showed that cortisol was the predominant steroid responsible for neutrophil survi-val-induction by parturient serum because only cortisol(not estradiol or progesterone)added backJ.L.Burton et al./Veterinary Immunology and Immunopathology105(2005)197–219204Table2Concentrations(pg/ml)of cortisol,estradiol,and progesterone inpooled parturient serum and charcoal-treated parturient serum withsteroids extractedHormone Parturientserum Steroid extractedparturient serumFetal bovineserumCortisol29,000a150c985b Estradiol1,500a75c869b Progesterone900a110c370b Concentrations in commercial fetal bovine serum are included for comparison.P<0.0001for within hormone differences(a–c)in serum sources.。