Molecular Psychiatry (2010) 15, 364–371 & 2010 Nature Publishing Group All rights reserved 1359-4184/10 $32.00 www.nature.com/mp ORIGINAL ARTICLE Differential effects on T-cell function following exposure to serum from schizophrenia smokers M Herberth1, DN Krzyszton1, D Koethe2, MR Craddock1, E Bulger1, E Schwarz1, P Guest1, FM Leweke2 and S Bahn1 1Institute of Biotechnology, University of Cambridge, Cambridge, UK and 2Department of Psychiatry and Psychotherapy, University of Cologne, Cologne, Germany

Cigarette smoking is more prevalent in subjects with schizophrenia compared to those with other psychiatric disorders or the general population and could therefore affect molecular pathways that impact the pathophysiology of this disorder. As smoking is also known to suppress immune responses, we investigated the effects of ‘smoking-conditioned’ serum obtained from schizophrenia and control subjects on healthy T cell in vitro. We found that T-cell proliferation was significantly increased following exposure to serum from smoking schizophrenia patients whereas no effect was observed when using serum from smoking control subjects or non-smoking patients and controls. We eliminated the possibility that these effects were due to quantitative differences in cigarette consumption as serum levels of the stable nicotine metabolite cotinine were similar in schizophrenic and control smokers. Molecular characterization showed that serum from patient smokers increased expression of T-cell activation markers CD69high, CD25high, co-stimulatory molecules CD26 þ , CD27 þ and CD28 þ , and decreased T-cell receptor complex components TCRa/b and CD3. Moreover, analysis of supernatants collected after T-cell exposure to serum from smoking patients showed a time-dependent decline in interleukin (IL)-2 levels, suggesting that the proliferation effect is promoted by enhanced IL-2 processing. These results suggest that cigarette smoking has selective effects on serum components that, in turn, lead to altered immune function in schizophrenia patients relative to healthy subjects. Further studies aimed at characterizing these components could result in a better understanding of the onset and aetiology of schizophrenia and potentially lead to novel therapeutic strategies. Molecular Psychiatry (2010) 15, 364–371; doi:10.1038/mp.2008.120; published online 11 November 2008 Keywords: schizophrenia; T-cell signaling; smoking; cytokine; cotinine; activation marker

Introduction Schizophrenia patients have a reduced life expec- tancy of approximately 20% compared to the general Compared to the general population and other population with coronary heart disease accounting for psychiatric disorders, cigarette consumption is high- 10 1,2 about two-thirds of the mortality rate. This disease est among schizophrenia patients. Their depen- susceptibility may be exacerbated through the sup- dence on nicotine is approximately three times that of pressive effect of cigarette smoking on the immune the general population with a prevalence reaching 11,12 3 system. almost 90%. Investigations of smoking behaviour Nicotine, the major active chemical in cigarettes, among schizophrenia patients have shown that they has been shown to impair antigen-driven T-cell smoke more, exhibit stronger addictive behaviours response and inhibit proliferation in rodents as well and encounter more difficulties with smoking cessa- 13–16 4,5 as humans. T lymphocytes have integral function tion compared to healthy cigarette consumers. in cell-mediated immunity and are involved in host Patients may use cigarettes as a form of self-medica- 6 defence against infectious agents. Peripheral changes tion, as a consequence of the stimulatory effect of in the immune system of schizophrenia patients, nicotine on the mesolimbic dopamine system that such as decreased cellular immune response17,18 and improves positive and negative symptoms as well as 19 7,8 altered cytokineexpression, have been reported. cognitive functions such as memory and attention. However, these immunological findings are incon- However, cigarette smoking is a well-known risk 9 sistent and controversial due to different methodo- factor of cardiovascular and respiratory disease. logical approaches, the heterogeneity of sample sourcing and acquisition, as well as lack of controls Correspondence: Dr S Bahn, Institute of Biotechnology, Univer- for confounding factors such as smoking.20 sity of Cambridge, Tennis Court Road, Cambridge, CB2 1QT UK. E-mail: [email protected] To investigate whether smoking has a distinct effect Received 29 July 2008; revised 18 September 2008; accepted 9 on the molecular pathways associated with the October 2008; published online 11 November 2008 immune response in schizophrenia, we examined Healthy T-cell response to in vitro stimulation M Herberth et al 365 the functional response of normal T cells in vitro to Syndrome Scale (PANSS) rating scale (PANSS total: serum obtained from first-onset anti-psychotic naive 100.7 (±24.0), positive symptoms: 25.4 (±4.6), patients and healthy controls (HCs). The main objec- negative symptoms: 22.7 (±8.8), general symptoms: tive was to determine whether cell-extrinsic serum 52.6 (±11.6)). Serum samples were produced by factors from schizophrenia smokers evoke distinct allowing blood to coagulate for 2 h at room disease-associated differential T-cell responses. temperature, centrifuged at 4000 g, 5 min and the supernatant was collected. Samples were stored at À80 1C until use. Materials and methods Sample collection T-cell isolation and stimulation Peripheral blood was collected from consenting Subjects for T-cell isolation. CD3 þ pan T cells were participants into 9 ml EDTA S Monovette tubes isolated from peripheral blood from healthy (Sarstedt). PBMCs (peripheral blood mononuclear volunteers recruited at the University of Cambridge cells) were isolated by centrifugation over Ficoll- showing no comorbidities such as diabetes, heart Paque Plus (GE Healthcare) and washed in Dulbecco’s disease, thyroid disease, autoimmune disease or any phosphate-buffered saline (DPBS; Invitrogen, Paisley, recent infections. Volunteers with a psychotic family Renfrewshire, UK). T cells were obtained by negative history or drug abuse were also excluded. All subjects selection using the MACS human pan T-cell isolation gave written consent. kit II in association with LS MACS separation columns (Miltenyi Biotech, UK) according to the Subjects for serum collection. The study was manufacturer’s specifications. T-cell purity (routinely approved by the local ethics committee and above 95%) was assessed by flow cytometry (CyAn; subsequently performed at the University Hospital DakoCytomation, UK) measuring expression of the of Cologne. All patients and healthy volunteers gave T-cell marker CD3-e. CD3 þ T cells from healthy their written informed consent. We consecutively donors were cultured in triplicate in RPMI 1640 recruited 24 subjects, comprising 12 HCs with no medium (Sigma, Poole, Dorset, UK) supplemented family history of schizophrenia or detectable medical, with 1% glutamine, penicillin, streptomycin (referred psychiatric or neurological history and 12 patients to as incomplete medium) in the absence or presence suffering from first-episode paranoid psychosis of 1 mgmlÀ1 anti-CD3 (clone OKT3). Serum (10%) (Table 1). All patients were drug naive for anti- from a panel of schizophrenia patients and matched psychotic medication. HCs were matched for age, controls was added to each well before incubation at gender, smoking and ethnicity. None of the HCs 37 1C, 5% CO2 for 48 h. Cells were pulsed with developed schizophrenic symptoms during the 0.037 MBq 3H-thymidine (Amersham Biosciences, follow-up period of at least 1 year. HCs were Little Chalfont, Bucks, UK) allowing incorporation carefully screened for medical disorders (such as into DNA for a further 24 h incubation and DNA diabetes, heart disease, thyroid disease, autoimmune harvested onto 96-well filter plates (PerkinElmer, Seer disease or any recent infections) and current or Green, Bucks, UK). T-cell proliferation was deter- previous psychiatric illnesses using the severe mined by measuring the incorporation of 3H-thymi- combined immunodeficiency for Diagnostic and dine into progeny DNA using a liquid scintillation Statistical Manual of Mental Disorders, fourth counter (Packard, Isotech, Chesterfield, Derbe, UK). edition; no previous or current diagnosis of illicit The intra- and inter-assay variability was < 15% drug abuse or dependence and no cannabis use coefficient of variation (CV). within the past 6 months before the study were allowed. Psychopathology of the patients was T-cell surface marker analysis assessed on the day of the blood withdrawal and To corroborate the difference in proliferation between measured by using the Positive and Negative the patient and control smokers we treated T cells

Table 1 Demographic characteristics of serum and T-cell samples

Demographic parameter Study 1 Study 2

T cells Serum samples T cells Serum samples

Patient Control Patient Control

Number (n)18661166 Age (year, mean±s.d.) 27.9±8.1 28.3±2.9 28.2±3.3 28.4±6.8 33.2±4.8 32.2±5.0 Sex (male/female) 9/9 3/3 3/3 4/7 3/3 3/3 Race (Caucasian/Asian) 17/1 6/0 6/0 10/1 6/0 6/0 Smoking (yes/no) 3/15 4/2 4/2 3/8 4/2 3/3

Molecular Psychiatry Healthy T-cell response to in vitro stimulation M Herberth et al 366 from 12 healthy donors with four representative Cotinine analysis serum samples from these two categories. Owning to Cotinine serum levels of 12 patient and 12 control logistical reasons the number of serum samples was samples were quantified by direct ELISA (Calbiotech, limited by the amount of standard T cells that Spring Valley, CA, USA) following the manufacturer’s could be obtained from a single donor. Cells in the protocol. The inter- and intra-assay variation was presence of serum were cultured in incomplete < 15% CV for standard and < 20% CV for the analysed medium as described above for 72 h. T cells were samples. In brief, 10 ml of serum and cotinine standard labelled in staining buffer (DPBS supplemented with was added in duplicate to each well and incubated 10% fetal calf serum) at 4 1C in the dark with the with 100 ml cotinine enzyme conjugate for 1 h at room indicated monoclonal antibodies conjugated to fluor- temperature. Wells were washed with distilled water escein isothiocyanate, phycoerythrin, PE-Cy5 or allo- before addition of 100 ml TMB substrate solution. This phycocyanin-Alexa Fluor 750. The antibodies used in was incubated for 30 min at room temperature before this study were produced against CD3, CD8 reaction was stopped with stop solution. Plates were (Dako Cytomation), CD26, CD27, CD28 (Immunotech, read at dual wavelengths of 450 and 650 nm on a plate Marseille, France), CD4, TCRa/b (BD Bioscience, reader (Bio-Rad). Oxford, Oxon, UK), CD25 (eBioscience, San Diego, CA, USA) and CD69 (Serotec, Kiddlington, Oxon, Statistical analysis UK). Antibody concentrations were titrated before Data were analysed in GraphPad Prism (La Jolla, CA, use. Cells were washed twice in staining buffer and USA). Statistical analyses for the cell-surface marker resuspended at a final concentration of 5 Â 106 cells expression and cytokine ELISA studies were per- per ml. Samples were analysed using the CyAn ADP formed using Wilcoxon matched pair’s test. P-values Flow Cytometer.2 (Dako Cytomation) equipped with of less than 0.05 were considered significant. Using summit version 4.1 software. The flow cytometer was the same method, statistical significance was calcu- calibrated with SpectrAlign Alignment beads (Dako lated for proliferation study 1 and 2 independently. Cytomation) to ensure accurate and uniform para- Pearson’s correlation coefficients were determined to meters. The inter-assay variability was < 15% CV. investigate the effect of age, gender, smoking and Lymphocytes were gated on CD3 þ T cells and 10 000 cotinine concentrations on proliferative responses. events were recorded. Data were processed and Using the same method, the correlation between analysed with FlowJo software 7.2.2 for Windows cotinine concentrations and the quantity of cigarettes (Tree Star, Ashland, OR, USA). The expression of cell- smoked per day was assessed. Statistical significance surface markers was calculated as median fluores- of the cotinine ELISA study was determined by non- cence intensity (FI) to determine cell-surface antigen parametric Mann–Whitney U-test. P-values of less density and frequency (%) of antigen presentation for than 0.05 were considered significant. CD4 þ and CD8 þ T-cell-distinct positive and nega- tive subpopulations. For single populations (% = 100), Results the median FI was taken into account. Effect of patient serum on T-cell proliferation The proliferative effects of patient and control serum on normal T cells incubated in the presence or Cytokine analysis absence of anti-CD3 were determined by 3H-thymi- Supernatants from T cells (n = 15) stimulated with dine incorporation. Because the number of serum anti-CD3 in the presence of patient (n = 3) or control- samples per experiment was limited by the amount of smoker serum (n = 3) were collected after 0, 24 and T cells obtained from a single donor, proliferation 48 h of culture, centrifuged and stored at À80 1C measurements were divided over two separate studies before use. The cytokines interleukin (IL)-2, IL-10 as indicated in Table 1. Both studies were performed and INF-g were quantified by commercially avai- under identical conditions; however, different serum lable enzyme-linked immunosorbent assays (ELISA; and T-cell sample cohorts were investigated in each eBioscience) according to the manufacture’s specifi- study, thus the second validates findings obtained in cations. All ELISAs were specific for the relative the first study. We found a significant correlation cytokine and inter- and intra-assay variation was less between smoking and proliferation (P < 0.03) that was than 10% CV in each case. In brief, T-cell super- independent of the smoking status of healthy T-cell natants were diluted 1:2 in dilution buffer and added donors (P = 0.3) and of age or gender. Serum from together with 100 ml of each standard diluent to the schizophrenia smokers induced a 1.2-fold (study 1) plate. This was incubated at 4 1C overnight to achieve and 1.4-fold (study 2) higher response compared to maximal sensitivity. Plates were washed 5 Â in assay serum from HC smokers (Figure 1a) and a similar 1.2- diluent and subsequently incubated with detection fold (study 1) and 1.5-fold (study 2) increase com- antibody, avidin-horseradish peroxidase and sub- pared to serum from schizophrenic non-smokers strate solution according to the manufacturer’s proto- (Figure 1c). No difference in proliferation was found col. The reaction was terminated with 50 ml stop between non-smokers regardless of health status solution and results were quantified at 450 nm using a (Figure 1b) nor between HC smokers and non-smokers plate reader (Bio-Rad, Birmingham, Warks, UK). (Figure 1d), suggesting that the altered proliferative

Molecular Psychiatry Healthy T-cell response to in vitro stimulation M Herberth et al 367 Study #1 Study #2 Smokers Non-smokers Patients Controls 200000 abcd p*** p** p** p** 150000

100000

50000

FC=1.2 FC=1.4 FC=1.0 FC=1.2 FC=1.0 FC=1.0 FC=0.9 FC=1.5 0 [H]-thymidine incorporation [counts/min]

3 P C P C P C P C NS S NS S NS S NS S # of sera 4 4 4 3 2 2 2 3244 2 2 433 Figure 1 The effect of patient and control serum on T-cell proliferation. Plotted are the pair-wise comparisons of T-cell proliferative responses following exposure to (a) serum from healthy control (C) or schizophrenia (P) smokers (S), (b) serum from healthy control and patient non-smokers (NS), (c) serum from non-smoking and smoking patients and (d) serum from non-smoking and smoking controls. Both studies were performed under identical conditions at different time points using independent serum and T-cell sample cohorts (study 1, n = 18; study 2, n = 11). # of sera, the number of serum samples averaged per T-cell donor for each indicated group; FC, fold change; **P < 0.005, ***P < 0.0005. response was not the result of smoking alone. In groups were found for CD4 þ and CD8 þ single addition, anti-psychotic medications or their metabo- populations or CD4 þ /CD8 þ ratios (data not shown). lites could not have been responsible for the effects as the serum samples used for both studies were obtained from drug-naive patients. Patient serum effects on T-cell receptor complex molecules Measurement of serum cotinine levels The T-cell proliferative response to antigen recogni- We measured cotinine levels in serum samples to tion is tightly regulated by CD3 and by degradation determine if the differential effect in proliferation was and downregulation of the T-cell receptor (TCR) associated with the rate of smoking. Cotinine, the complex TCRa/b.23 After stimulation, the median FI major nicotine metabolite, was chosen as the most of the TCRa/b receptor was decreased 1.4-fold reliable indicator of nicotine levels as it resides in the following patient serum exposure from smokers in blood stream between 15 and 17 h compared with CD4 þ cells (P < 0.005) and 1.3-fold in CD8 þ cells only 2 and 4 h for nicotine.21,22 Serum cotinine (P < 0.02), compared with the effects seen after concentrations correlated strongly with the self- exposure to control serum from smokers (Table 2). reported number of cigarettes smoked per day in In addition, the CD3 antigen was decreased 1.2-fold patients (R2 = 0.59) as well as HCs (R2 = 0.69). How- following treatment with patient serum in CD4 þ ever, no significant differences in cotinine levels were cells compared to controls (P < 0.007). No difference found between healthy and patient smokers and was observed in CD8 þ cells. cotinine concentrations did not correlate with the proliferative responses in either the patient (R2 = 0.00) or control (R2 = 0.02) group (data not shown). Patient serum effects on co-stimulatory molecules Effects on cell-surface marker expression T-cell activation was also evaluated by measuring the For further molecular characterization of the T-cell expression of co-stimulatory molecules: CD26, a activation response, we measured expression levels of dipeptidyl peptidase IV cell-surface protease; CD27, various T-cell surface markers using four-colour a member of the tumour necrosis factor (TNF) staining. This was determined as the median FI and receptor family and CD28, a member of the B7 co- frequency (%) for cases in which distinct positive (for stimulatory family.24–26 All receptors showed signi- example, CD69high) and negative (for example, ficant upregulation in response to patient serum CD69low) populations were present. In cases where exposure from smokers in stimulated CD4 þ and population frequencies were 100% (for example, CD8 þ cells compared to the effects seen with CD26 þ ), only the median FI was taken into control serum from smokers (CD4 þ CD26 þ , 1.2-fold; consideration. All antigens were examined for CD8 þ CD26 þ , 1.3-fold; CD4 þ CD27 þ , 1.2-fold; unstimulated and stimulated CD4 þ and CD8 þ CD8 þ CD27 þ , 1.1-fold; CD4 þ CD28 þ , 1.3-fold; subpopulations. No differences between the two CD8 þ CD28 þ , 1.7-fold) (Table 2).

Molecular Psychiatry Healthy T-cell response to in vitro stimulation M Herberth et al 368 Table 2 Changes in expression of T-cell surface markers after T-cell activation

Marker (a) Frequencies (%)

CD4 þ CD8 þ

Ctrl* Pt* Fold change P-valuea Ctrl* Pt* Fold change P-valuea

CD69high 63.0±4.8 80.1±1.5 1.3 m 0.0005 55.0±4.1 66.9±3.1 1.2 m 0.0010 CD69low 36.5±4.9 19.8±1.5 0.5 k 0.0005 37.6±3.8 24.4±1.9 0.6 k 0.0015 CD25high 31.4±3.2 40.9±4.8 1.3 m 0.0244 30.3±5.1 42.7±6.0 1.4 m 0.0068 CD25low 64.6±4.7 53.6±5.3 0.8 k 0.0322 69.0±5.2 56.4±6.0 0.8 k 0.0068

(b) Median fluorescence intensities CD3 182.4±16.0 167.1±15.7 0.9 k 0.0015 168.8±12.1 168.8±14.9 1.0 2 0.9697 TCR 12.4±2.1 8.9±1.4 0.7 k 0.0049 10.2±1.5 7.8±1.5 0.8 k 0.0122 CD69high 68.8±6.4 86.6±4.6 1.3 m 0.0024 52.3±2.3 56.0±2.5 1.1 m 0.0120 CD69low 3.9±0.2 3.1±0.2 0.8 k 0.0049 4.8±0.4 3.8±0.4 0.8 k 0.0244 CD25high 292.3±62.0 326.0±67.8 1.1 2 0.2061 238.8±50.7 231.3±58.4 1.0 2 0.7646 CD25low 4.0±0.6 4.8±0.5 1.2 2 0.1475 10.0±1.1 14.87±2.0 1.5 m 0.0068 CD26 12.8±0.5 16.0±1.5 1.3 m 0.0425 12.3±1.0 18.1±2.0 1.5 m 0.0034 CD27 28.4±1.5 33.6±2.2 1.2 m 0.0068 27.0±2.5 33.7±4.2 1.2 m 0.0161 CD28 16.7±3.2 23.3±4.7 1.4 m 0.0024 12.3±2.3 16.5±3.3 1.3 m 0.0010

*Mean±s.e.m. m upregulated, k downregulated or 2 no change in T cells (n = 12) following exposure to serum from smoking patients (n =2) compared to smoking controls (n = 2). aCalculated by Wilcoxon matched pairs test.

a 150 b 100 p*** p** p** p* 80

high 100

high 60

40

50 [%] CD69

median FI CD69 20

0 0 CP C P CPCP CD4 CD8 CD4 CD8 Figure 2 Expression of CD69high on CD4 þ and CD8 þ cells. Plotted are the pair-wise comparisons of the expression of CD69high on stimulated T cells treated with patient (P) or control (C) serum from smokers. Expression measured as (a) the median fluorescent intensities (FI) and (b) frequencies (%) of CD69high on CD4 þ and CD8 þ cells. *P < 0.05, **P < 0.005, ***P < 0.0005.

Patient serum effects on T-cell activation were significantly augmented in stimulated T cells T-cell activation through antigen stimulation evokes after exposure to patient serum from smokers com- upregulation of distinct activation markers. We pared to that from control smokers (Figure 2). analysed the expression of the early and late activa- Consistent with this, the expression pattern of the tion markers CD25 and CD69 on CD4 þ and CD8 þ CD69low population showed decreased FI and fre- cells (Table 2). Expression of CD69high and CD25high is quency on CD4 þ (FI, P < 0.0005; %, P<0.005) and characteristic of activated cells whereas CD69low and CD8 þ (FI, P < 0.03; %, P < 0.002) cells after treatment CD25low expression is usually associated with quies- with patient serum from smokers. In addition, the cent cells. The frequency and median FI of frequency of CD4 þ and CD8 þ cells expressing CD4 þ CD69high (%, P < 0.003; FI, P < 0.0005) and CD25high (IL-2 receptor a-chain) was significantly CD8 þ CD69high (%, P < 0.01; FI, P < 0.003) phenotypes enhanced by treatment with patient serum from

Molecular Psychiatry Healthy T-cell response to in vitro stimulation M Herberth et al 369 between smoking status and level of schizotypy, where higher levels of schizotypy significantly pre- dicted the risk of being a smoker after controlling for gender and group status.31 Heavy smoking in patients has been regarded as an attempt of self-medication and there is evidence that nicotinic agonists improve clinical ratings of negative symptoms that are gen- erally resistant to treatment with dopamine antagonist anti-psychotic drugs.32 This implies that cigarette compounds may activate compensatory or beneficial mechanisms ameliorating the pathophysiology of the disease. In the current study, we have shown that T cells from healthy donors undergo increased proliferation following exposure to serum from drug-naive, first- onset schizophrenia smokers. This effect was not Figure 3 Interleukin (IL)-2 cytokine concentration de- observed after exposure to serum from control creased over time. No significant differences in IL-2 smokers or from schizophrenia non-smokers. Inter- concentration were found in unstimulated and stimulated estingly, it is documented that chronic exposure to cells after 24 h. However, after 48 h, IL-2 concentration was cigarette compounds such as nicotine suppresses reduced in supernatants (P=0.006) obtained from T cells immune functions including T-cell proliferation.33,34 treated with patient serum from smokers compared to 15 control serum from smokers.*P < 0.05. A study by Frazer-Abel et al. showed that direct addition of nicotine to T cells in vitro decreases their proliferation by prevention of cell-cycle entry in

G0–G1 transition. However, the present results imply smokers (CD4 þ CD25high, P<0.03; CD8 þ CD25high, that smoking exerts differential effects in patients P < 0.007) and the CD4 þ CD25low and CD8 þ CD25low leading to altered T-cell proliferation and that these phenotypes showed a concomitant significant effects are disease associated rather than related to reduction (CD4 þ CD25low, P<0.04; CD8 þ CD25low, smoking as no proliferative difference was observed P < 0.007) (Table 2). between healthy smokers and non-smokers. In addi- tion, preliminary crossover experiments from our Measurement of IL-2 concentration in T-cell laboratory show that T cells from schizophrenia supernatants patients show decreased proliferation only after T cells respond to TCR stimulation by the production exposure to serum from schizophrenia smokers of cytokines. To investigate whether serum from (unpublished data). Substances produced from toba- schizophrenia smokers induced a proinflammatory cco smoke may therefore interact differently with key or anti-inflammatory skewed T-cell response, we regulatory factors circulating in patient serum but analysed their ability to produce interferon-g (IFN-g) not in serum from normal subjects. The presence of and IL-10 cytokines. We also measured concentra- these factors, rather than components of cigarette tions of IL-2, a cytokine produced at an early stage by smoke, could have affected the observed difference T cells in response to antigen stimulation. Once in proliferation even though nicotine and cotinine secreted, IL-2 binds to CD25 to promote T-cell are known to trigger cholinergic anti-inflammatory proliferation.27 There were no differences in IL-10 pathways and proliferation in non-excitable cells and IFN-g concentrations in T-cell supernatants (data through activation of the a7 subtype of the nAChR not shown). Also, no difference in IL-2 concentration (a7nAChR).35–37 One previous study showed was found in unstimulated cells or cells that had been decreased levels of a7nAChR mRNA in peripheral stimulated for 24 h. However, IL-2 levels were blood lymphocytes from schizophrenia patients and decreased 1.3-fold (P < 0.006) following exposure to the authors ruled out the possibility that this resulted serum from schizophrenia smokers compared to from nicotine consumption as there was no difference control smokers after 48 h treatment (Figure 3). between HC smokers and non-smokers.38 However, they did not investigate whether there was any effect a Discussion of smoking on 7nAChR expression within the schizophrenia patient group, a possibility that should The association between smoking and schizophrenia be considered in the light of the results of the present has been highlighted on previous occasions.28,29 Most study. reports have commented on the socio-economic and T-cell activation is a complex process involving a behavioural issues of cigarette consumption in panoply of tightly regulated, dynamic events result- chronically ill and anti-psychotic-treated patients. ing in the expression of cell-surface marker molecules However, increased tobacco consumption is apparent that in turn control T-cell function and prolifera- in the early and prodromal stages of schizophrenia.30 tion.39,40 We demonstrated that increased proliferative Furthermore, a positive relationship was found response of exogenous T cells exposed to sera from

Molecular Psychiatry Healthy T-cell response to in vitro stimulation M Herberth et al 370 smoking schizophrenia patients was also associated long-term disability, social deprivation or drug treat- with increased expression of the activation markers ment. CD69high and CD25high and co-stimulatory molecules Considering the complexity of the serum proteome, CD26 þ , CD27 þ and CD28 þ , along with down- and the fact that tobacco smoke contains more than regulation of the TCR complexes CD3 þ and TCRa/b. 4000 chemicals besides nicotine, there is considerable These findings suggest that, in vivo, T cells from scope for future studies to characterize the molecular schizophrenia patients are activated in response to pathways involved in T-cell function. The molecular unidentified circulating factors associated with cigar- sequela linked to the presence of confounding factors ette smoking. Long-term activation could then lead to such as smoking is still poorly understood. Our desensitization of the proliferation responses as preliminary findings point towards the existence of indicated by a reduced proliferative rate of schizo- smoking-related molecular pathways underlying the phrenia patient-derived T cells, implying that patient disorder, which could open new avenues for biomar- T cells have become refractory to stimulation and ker discovery strategies. The results also emphasize respond significantly less to in vitro stimula- the importance of considering confounding factors tion.17,18,41 T-cell activation in the patient’s body, as such as substance abuse to avoid difficulties in shown here through elevated levels of CD25high and assessing disease-relevant alterations. The combined CD69high, could furthermore help to recruit T cells to application of functional cell studies and multi-tiered cross the blood–brain barrier to foster communication profiling experiments should have great potential to between the central nervous system and the immune elucidate the physiological systems involved in the system. Previous studies have shown that T cells have onset and aetiology of schizophrenia and could lead beneficial effects on a damaged central nervous to the discovery of biomarker candidates with system.42,43 Smoking may therefore be an attempt to therapeutic potential. Much of the recent work enhance the deficient immune response in schizo- undertaken within our laboratory has focused on phrenia patients to maintain brain homoeostasis. proteomic profiling of cerebro spinal fluid and serum We also found an increased expression of the IL-2 from schizophrenia patients at various stages of receptor a-chain (CD25) and a time-dependent de- disease. We have found a number of significantly cline in IL-2 concentration in T-cell supernatants after altered proteins and peptides that may be involved in treatment with serum from schizophrenic smokers. or are linked to the current findings.44,45 Once secreted, IL-2 works as an autocrine factor by Taken together, these results show that T cells from binding to CD25, promoting proliferation and survival healthy donors respond differently to serum from of recently activated T cells such as CD25 þ regula- patient smokers compared to control smokers. Despite tory T cells. Interestingly, IL-2 concentrations were similar cotinine concentrations in serum, this phy- comparable between the patient and control smoking siological difference is reflected in disturbed T-cell groups demonstrating that cells were equally able to function including upregulation of activation markers produce IL-2 after 24 h of stimulation. This suggests and increased IL-2 processing. This suggests that sera that the observed decline in IL-2 levels in the patient from schizophrenia patients contain one or more smoker group after 48 h is most likely due to factors that, in combination with specific products of enhanced processing of IL-2 potentially by increased tobacco smoking, can lead to an altered immune levels of CD25 receptors. This result further supports response. Identification and characterization of these the possibility that proliferation of T cells following components in the context of these findings could exposure to patient serum from smokers is driven by lead to the discovery of novel biomarkers for early enhanced IL-2 signalling. diagnosis and new therapeutic strategies. The most intriguing finding of the current study was the observed increase in T-cell proliferation after exposure to serum from schizophrenic smokers. We have demonstrated that this effect was not due to Acknowledgments differences in nicotine metabolism as cotinine levels were not significantly different in the sera of schizo- We acknowledge all volunteer patient and control phrenic and control smokers. We also excluded the blood donors. We thank Nigel Miller and Prof Anne possibility that the effect was due to alterations in the Cooke, Department of Pathology, University of Cam- inflammatory response by showing that the concen- bridge for their intellectual input and helpful advice. trations of IL-10 and IFN-g in serum from schizo- We also thank Prof Chris Lowe, director of Institute of phrenic and HC smokers were comparable. One Biotechnology, University of Cambridge and all advantage of the present study comes from the use members of the Bahn Laboratory for helpful discus- of samples from drug-naive, first-onset patients, sions. We are grateful to Dr Fuller Torry (Stanley thereby ruling out the potentially confounding factor Medical Research Institute, SMRI) for constructive of drug treatment and, most importantly, allowing for help with the paper. This research was supported by investigation of disease-related pathologies at an early SMRI and the National Alliance for Research into stage of illness. In this case a given abnormality might Schizophrenia and Depression (NARSAD). SB is a represent a primary disease-associated pathology NARSAD Essel Investigator. MH is supported by the rather than one that has arisen as a consequence of Cambridge European Trust.

Molecular Psychiatry Healthy T-cell response to in vitro stimulation M Herberth et al 371 References 24 Camerini D, Walz G, Loenen WA, Borst J, Seed B. The T cell activation antigen CD27 is a member of the nerve growth factor/tumor necrosis 1 de Leon J, Diaz FJ, Rogers T, Browne D, Dinsmore L. Initiation of factor receptor gene family. J Immunol 1991; 147: 3165–3169. daily smoking and nicotine dependence in schizophrenia and 25 Huhn J, Ehrlich S, Fleischer B, von Bonin A. Molecular analysis of mood disorders. Schizophr Res 2002; 56: 47–54. CD26-mediated signal transduction in T cells. Immunol Lett 2000; 2 Diwan A, Castine M, Pomerleau CS, Meador-Woodruff JH, Dalack 72: 127–132. GW. Differential prevalence of cigarette smoking in patients with 26 Greenwald RJ, Freeman GJ, Sharpe AH. The B7 family revisited. schizophrenic vs mood disorders. Schizophr Res 1998; 33: 113–118. Annu Rev Immunol 2005; 23: 515–548. 3 Kalman D, Morissette SB, George TP. Co-morbidity of smoking in 27 Hemar A, Subtil A, Lieb M, Morelon E, Hellio R, Dautry-Varsat A. patients with psychiatric and substance use disorders. Am J Endocytosis of interleukin 2 receptors in human T lymphocytes: Addict 2005; 14: 106–123. distinct intracellular localization and fate of the receptor alpha, 4 Dalack GW, Healy DJ, Meador-Woodruff JH. Nicotine dependence beta, and gamma chains. J Cell Biol 1995; 129: 55–64. in schizophrenia: clinical phenomena and laboratory findings. Am 28 Goff DC, Henderson DC, Amico E. Cigarette smoking in schizo- J Psychiatry 1998; 155: 1490–1501. phrenia: relationship to psychopathology and medication side 5 Ziedonis DM, George TP. Schizophrenia and nicotine use: report of effects. Am J Psychiatry 1992; 149: 1189–1194. a pilot smoking cessation program and review of neurobiological 29 Kelly C, McCreadie RG. Smoking habits, current symptoms, and and clinical issues. Schizophr Bull 1997; 23: 247–254. premorbid characteristics of schizophrenia patients in Nithsdale, 6 Kumari V, Postma P. Nicotine use in schizophrenia: The self Scotland. Am J Psychiatry 1999; 156: 1751–1757. medication hypotheses. Neurosci Biobehav Rev 2005; 29: 1021. 30 Beratis S, Katrivanou A, Gourzis P. Factors affecting smoking in 7 Jacobsen LK, D’Souza DC, Mencl WE, Pugh KR, Skudlarski P, schizophrenia. Compr Psychiatry 2001; 42: 393–402. Krystal JH. Nicotine effects on brain function and functional 31 Esterberg ML, Jones EM, Compton MT, Walker EF. Nicotine connectivity in schizophrenia. Biol Psychiatry 2004; 55: 850–858. consumption and schizotypy in first-degree relatives of indivi- 8 Myers CS, Robles O, Kakoyannis AN, Sherr JD, Avila MT, Blaxton duals with schizophrenia and non-psychiatric controls. Schizophr TA et al. Nicotine improves delayed recognition in schizophrenia Res 2007; 97: 6–13. patients. Psychopharmacology (Berl) 2004; 174: 334–340. 32 Freedman R, Olincy A, Buchanan RW, Harris JG, Gold JM, Johnson 9 Hennekens CH, Hennekens AR, Hollar D, Casey DE. Schizophre- L et al. Initial phase 2 trial of a nicotinic agonist in schizophrenia. nia and increased risks of cardiovascular disease. Am Heart J 2005; Am J Psychiatry 2008; 165: 1040–1047. 150: 1115. 33 McCue JM, Link KL, Eaton SS, Freed BM. Exposure to cigarette tar 10 Brown S, Barraclough B, Inskip H. Causes of the excess mortality inhibits ribonucleotide reductase and blocks lymphocyte prolif- of schizophrenia. Br J Psychiatry 2000; 177: 212–217. eration. J Immunol 2000; 165: 6771–6775. 11 Holt PG, Keast D. Environmentally induced changes in immuno- 34 Glader P, Moller S, Lilja J, Wieslander E, Lofdahl C-G, von logical function: acute and chronic effects of inhalation of tobacco Wachenfeldt K. Cigarette smoke extract modulates respiratory smoke and other atmospheric contaminants in man and experi- defence mechanisms through effects on T-cells and airway mental animals. Bacteriol Rev 1977; 41: 205–216. epithelial cells. Respir Med 2006; 100:818. 12 Johnson JD, Houchens DP, Kluwe WM, Craig DK, Fisher GL. 35 Rehani K, Scott DA, Renaud D, Hamza H, Williams LR, Wang H Effects of mainstream and environmental tobacco smoke on the et al. Cotinine-induced convergence of the cholinergic and PI3 immune system in animals and humans: a review. Crit Rev Toxicol kinase-dependent anti-inflammatory pathways in innate immune 1990; 20: 369–395. cells. Biochim Biophys Acta 2008; 1783: 375–382. 13 Geng Y, Savage SM, Razani-Boroujerdi S, Sopori ML. Effects of 36 Wang H, Yu M, Ochani M, Amella CA, Tanovic M, Susarla S et al. nicotine on the immune response. II. Chronic nicotine treatment Nicotinic acetylcholine receptor [alpha]7 subunit is an essential induces T cell anergy. J Immunol 1996; 156: 2384–2390. regulator of inflammation. Nature 2003; 421:384. 14 Geng Y, Savage SM, Johnson LJ, Seagrave J, Sopori ML. Effects of 37 Pavlov VA, Wang H, Czura CJ, Friedman SG, Tracey KJ. The nicotine on the immune response. I. Chronic exposure to nicotine cholinergic anti-inflammatory pathway: a missing link in neu- impairs antigen receptor-mediated signal transduction in lympho- roimmunomodulation. Mol Med 2003; 9: 125–134. cytes. Toxicol Appl Pharmacol 1995; 135: 268–278. 38 Perl O, Ilani T, Strous RD, Lapidus R, Fuchs S. The α7 15 Frazer-Abel AA, Baksh S, Fosmire SP, Willis D, Pierce AM, nicotinic acetylcholine receptor in schizophrenia: decreased Meylemans H et al. Nicotine activates nuclear factor of activated T mRNA levels in peripheral blood lymphocytes. FASEB J 2003 cells c2 (NFATc2) and prevents cell cycle entry in T cells. 03-0104fje. J Pharmacol Exp Ther 2004; 311: 758–769. 39 James FP, Dixon JLL, Favero JJ. Activation of human T lympho- 16 Kalra R, Singh SP, Kracko D, Matta SG, Sharp BM, Sopori ML. cytes by crosslinking of anti-CD3 monoclonal antibodies. J Leukoc Chronic self-administration of nicotine in rats impairs T cell Biol 1989; 46: 214–220. responsiveness. J Pharmacol Exp Ther 2002; 302: 935–939. 40 Nel AE. T-cell activation through the antigen receptor. Part 1: 17 Riedel M, Spellmann I, Schwarz MJ, Strassnig M, Sikorski C, signaling components, signaling pathways, and signal integration Moller HJ et al. Decreased T cellular immune response in at the T-cell antigen receptor synapse. J Allergy Clin Immunol schizophrenia patients. J Psychiatr Res 2007; 41: 3–7. 2002; 109: 758–770. 18 Craddock RM, Lockstone HE, Rider DA, Wayland MT, Harris LJ, 41 Matloubi H, Vodjgani M, Nasehi AA, Niknam MH, Kazemnejad A, McKenna PJ et al. Altered T-cell function in schizophrenia: a Salehi E et al. Decreased T cell response to mitogen and increased cellular model to investigate molecular disease mechanisms. PLoS anti-cytoplasmic antibody in drug-free schizophrenia patients. ONE 2007; 2: e692. Iran J Immunol 2007; 4: 32–37. 19 Potvin S, Stip E, Sepehry AA, Gendron A, Bah R, Kouassi E. 42 Kipnis J, Cohen H, Cardon M, Ziv Y, Schwartz M. T cell deficiency Inflammatory cytokine alterations in schizophrenia: a systematic leads to cognitive dysfunction: implications for therapeutic quantitative review. Biol Psychiatry 2008; 63: 801–808. vaccination for schizophrenia and other psychiatric conditions. 20 de Leon J, Diaz FJ. A meta-analysis of worldwide studies Proc Natl Acad Sci USA 2004; 101: 8180–8185. demonstrates an association between schizophrenia and tobacco 43 Moalem G, Leibowitz-Amit R, Yoles E, Mor F, Cohen IR, Schwartz M. smoking behaviors. Schizophr Res 2005; 76: 135–157. Autoimmune T cells protect neurons from secondary degeneration 21 Haley NJ, Sepkovic DW, Hoffmann D. Elimination of cotinine from after central nervous system axotomy. Nat Med 1999; 5: 49. body fluids: disposition in smokers and nonsmokers. Am J Public 44 Huang JTJ, Leweke FM, Oxley D, Wang L, Harris N, Koethe D et al. Health 1989; 79: 1046–1048. Disease biomarkers in cerebrospinal fluid of patients with first- 22 Benowitz NL, Jacob III P. Nicotine and cotinine elimination onset psychosis. PLoS Med 2006; 3: e428. pharmacokinetics in smokers and nonsmokers. Clin Pharmacol 45 Huang JTJ, Wang L, Prabakaran S, Wengenroth M, Lockstone HE, Ther 1993; 53: 316–323. Koethe D et al. Independent protein-profiling studies show a 23 Valitutti S, Muller S, Salio M, Lanzavecchia A. Degradation of decrease in apolipoprotein A1 levels in schizophrenia CSF, brain T cell receptor (TCR)-CD3-zeta complexes after antigenic stimula- and peripheral tissues. Mol Psychiatry 2007; e-pub ahead of print tion. J Exp Med 1997; 185: 1859–1864. 16 October 2007

Molecular Psychiatry