Signal Inhibition by the Dual-Specific Phosphatase 4 Impairs T Cell-Dependent B-Cell Responses With

Signal inhibition by the dual-speciﬁc phosphatase 4 PNAS PLUS impairs T cell-dependent B-cell responses with age

Mingcan Yua,b,GuangjinLia,b, Won-Woo Leec,MingYuand,DapengCuie,CorneliaM.Weyanda,b, and Jörg J. Goronzya,b,1

aDepartment of Medicine, Stanford University School of Medicine, Stanford, CA 94305; bDepartment of Medicine, Palo Alto Department of Veterans Affairs Health Care System, Palo Alto, CA 94304; cDepartment of Microbiology and Immunology, Seoul National University College of Medicine, Seoul 110-799, South Korea; dSchool of Industrial and Systems Engineering, Georgia Institute of Technology, Atlanta, GA 30332; and eLowance Center for Human Immunology, Emory University, Atlanta, GA 30322

Edited by Rino Rappuoli, Novartis Vaccines, Siena, Italy, and approved February 22, 2012 (received for review June 17, 2011)

T cell-dependent B-cell responses decline with age, suggesting Defects in T-cell responses have been mostly attributed to the defective CD4 T-cell function. CD4 memory T cells from individuals naïve T-cell compartment that contracts in size and diversity be- older than 65 y displayed increased and sustained transcription of cause of declining thymic production with age (10–12). CD8 the dual-specific phosphatase 4 (DUSP4) that shortened expression memory cells show ample evidence of immune aging with a loss of of CD40-ligand (CD40L) and inducible T-cell costimulator (ICOS) central memory cells and changes in gene expression, such as the (both P < 0.001) and decreased production of IL-4, IL-17A, and IL-21 loss of CD28 and the gain in expression of negative regulatory – (all P < 0.001) after in vitro activation. In vivo after influenza vacci- molecules (13 15). In contrast, defects in CD4 memory T-cell fi + − nation, activated CD4 T cells from elderly individuals had increased responses have escaped a de nition. CD4 CD28 T cells are only DUSP4 transcription (P = 0.002), which inversely correlated with the infrequently seen with age. If they are present, they are usually associated with an inflammatory disease (16, 17). CD4 memory T- expression of CD40L (r = 0.65, P = 0.002), ICOS (r = 0.57, P = 0.008), cell subset distribution is stable with age, and most elderly indi- and IL-4 (r = 0.66, P = 0.001). In CD4 KO mice reconstituted with viduals have a large fraction of CD4 central memory T cells and DUSP4 OT-II T cells, DUSP4 had a negative effect on the expansion fi P lack the expansion of oligoclonal CD4 effector T cells that is of antigen-speci c B cells ( = 0.003) and the production of ova- characteristic for CD8 T cells (18). In murine systems, CD4 fi P speci c antibodies ( = 0.03) after immunization. Silencing of DUSP4 memory cells generated early in life have a better functional P < P in memory CD4 T cells improved CD40L ( 0.001), IL-4 ( = 0.007), profile than those cells generated late in life (19); however, this and IL-21 (P = 0.04) expression significantly more in the elderly than phenomenon has not been characterized at the molecular level. young adults. Consequently, the ability of CD4 memory T cells to Telomere shortening has been postulated to limit memory T-cell support B-cell differentiation that was impaired in the elderly (P = responses and may reach a critical level in humans (20). 0.004) was restored. Our data suggest that increased DUSP4 expres- Efforts to improve vaccine efficacy are currently mostly fo- sion in activated T cells in the elderly in part accounts for defective cused on improving vaccine formulation. Adjuvanted vaccines adaptive immune responses. (for example, the oil in water emulsion MF59) hold promise (21). High-dose vaccines have been used with some success in

immunosenescence | signaling | aging VZV vaccination to prevent zoster flares and postherpetic neu- IMMUNOLOGY ralgias, and they have also been used in exploratory studies fl ith increasing age, the ability of the immune system to of in uenza vaccinations (22, 23). However, these approaches protect against new antigenic challenges or control chronic alone have limitations. A two-pronged approach, also targeting W the responding T-cell population, is likely necessary. In the infections erodes (1, 2). Incidence and severity of viral infections current study, we hypothesized that signaling defects in memory increase, and the response to prophylactic vaccination declines. fl CD4 T-cell responses in the elderly can be targeted to improve More than 90% of all in uenza-related deaths in the United vaccine responses. We found an increased induction of the dual- States occur in the elderly (3). Vaccine-induced protection for specific phosphatase 4 (DUSP4) in CD4 memory T cells from 65- fl in uenza infection is between 20% and 50% in the elderly de- to 85-y-old individuals that prevented differentiation into effec- ∼ pendent on age and study compared with 90% in young adults tive T-helper cells for B cells. In vitro as well as in vivo studies (4). In a review of 31 vaccine antibody response studies, the odds documented that the expression of DUSP4 in T cells is an im- ratios to seroconvert or develop seroprotective antibody titers portant regulator of T cell-dependent B-cell responses and that in elderly vs. young adults ranged from 0.24 to 0.59 (5). Epide- silencing of DUSP4 expression can at least partially restore the miological studies did not find an impact of the increasing immune defects in the elderly. compliance with annual flu vaccination on seasonal mortality between 1968 and 2001, further questioning their efficacy (6). Results The mechanisms accounting for this defective vaccine response Age-Related Differences in Activation-Induced Gene Expression of have not been identified but likely involve B as well as T cells. Memory CD4 T Cells. Vβ2+ CD4 memory T cells from four 20- to Given that the vast majority of adults have had previous expo- 35-y-old and four 70- to 75-y-old individuals were stimulated with sure to influenza and that antigenic drifts and shifts of the in- toxic shock syndrome toxin (TSST) presented by myeloid den- fluenza virus involve B- rather than T-cell epitopes, the T-cell defect seems to lie in CD4 memory rather than naïve T-cell function (7). A decline in the frequency and function of virus- Author contributions: M. Yu, C.M.W., and J.J.G. designed research; M. Yu, G.L., W.-W.L., specific memory CD4 T cells is also responsible for the in- and D.C. performed research; M. Yu, G.L., W.-W.L., M. Yuan, C.M.W., and J.J.G. analyzed creasing incidence of herpes zoster with age caused by the data; and M. Yu and J.J.G. wrote the paper. Varicella zoster virus (VZV) (8). VZV is an α-herpes virus that The authors declare no conflict of interest. causes chicken pox in children and establishes latency in senso- This article is a PNAS Direct Submission. rineural ganglions. On reactivation of VZV from latency, virus is Data deposition: The data reported in this paper have been deposited in the Gene Ex- transported along neuronal axons to the skin, causing herpes pression Omnibus (GEO) microarray database (accession no. GSE36476). zoster. Immune surveillance is critical for maintaining latency. 1To whom correspondence should be addressed. E-mail: [email protected]. The incidence of zoster reactivation correlates with age, ranging See Author Summary on page 5561 (volume 109, number 15). from 2 in 1,000 patient years in middle-aged adults to 10 after This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10. the age of 65 y and 15 in individuals older than 75 y (9). 1073/pnas.1109797109/-/DCSupplemental.

www.pnas.org/cgi/doi/10.1073/pnas.1109797109 PNAS | Published online March 20, 2012 | E879–E888 Downloaded by guest on September 29, 2021 dritic cells (mDCs) derived from young adults. Gene expression significantly increased in CD4 memory T-cell responses of 65- to was examined at 16, 40, and 72 h after stimulation using Affy- 85-y-old compared with 20- to 35-y-old healthy individuals (P < metrix arrays. Probes were identified that were not different 0.001 and P = 0.03, respectively). Reporter gene assays using before stimulation but were different at 40 or 72 h after stimu- DUSP4 promoter constructs confirmed increased transcriptional lation with a probability of >0.9; 311 probes at 40 h and 390 activity (Fig. 1D). In these experiments, CD4 memory T cells were probes at 72 h fulfilled this criterion, of which 63 probes showed stimulated in anti-CD3/anti-CD28 Ab-coated plates. After 36 h, a similar pattern at both time points. An additional 14 and 10 activated T cells were transfected with DUSP4 promoter reporter probes, respectively, that reached a probability of >0.9 only at gene constructs, and reporter gene activity was assessed 12 h after one time point were also different with a probability of >0.8 at transfection. Reporter gene activity in CD4 memory T cells from the other time point, suggesting that 87 probes representing 82 elderly individuals was significantly higher (P < 0.001). This dif- genes were differentially expressed at both time points. Of these ference was also maintained when cells were restimulated by adding probes, only 24 probes were already found to be different at 16 h, ionomycin and phorbol myristate acetate (PMA) during the last 4 h suggesting that the majority of these genes are not early activa- of culture (P = 0.003). The work by Berasi et al. (25) identified the tion genes. Expression of 60 genes was increased in elderly CD4 transcription factor early growth response protein (EGR)1 as a key memory T cells, and 22 genes decreased (Table S1). regulator of DUSP4 transcription. Indeed, EGR1 silencing com- Functional annotation clustering of these differentially expressed pletely abrogated DUSP4 expression after T cell receptor (TCR) genes identified overexpression of several metallothioneins and stimulation (Fig. 1E). EGR1 transcript levels were up-regulated in the zinc transporter ZnT1 in the elderly consistent with our recent T-cell responses from elderly individuals, suggesting that the higher results in naïve CD4 T cells (24). A second cluster included the DUSP4 expression is caused by an increased responsiveness to chemokine CXCL13 and the chemokine receptors CCR4 and transcribe EGR1 after TCR stimulation (Fig. 1F). CXCR6, suggesting that activation-induced homing patterns Western blot data paralleled the transcriptional results. DUSP4 change with age. In contrast, conventional activation markers or protein expression peaked 48 h after CD3/CD28 stimulation and cytokines, with the exception of IL-9 and IL-26, did not reach then started to decline in young individuals (Fig. 2A). A similar significance. To identify pathways that may be targeted to im- kinetics of DUSP4 protein expression was seen when CD4 prove vaccine responses in the elderly, we examined the panel of memory T cells were stimulated in the more physiological system differentially expressed genes for the presence of signaling mol- with mDC and superantigen. Compared with CD3 stimulation, ecules. DUSP4 was represented with two different probes with an signal intensity of DUSP4 was reduced, reflecting lesser cell acti- overexpression of the phosphatase at 40 and 72 h. Subsequent vation; results shown are normalized to β-tubulin. In this system, PCR studies showed that DUSP4 expression in resting naïve or DUSP4 expression was more sustained and only started to decline memory CD4 T cells was minute, and transcription was induced on day 4 (Fig. 2B). DUSP4 protein expression in elderly CD4 T within the first 40 h in both cell populations. Of interest, naïve cells was increased over the entire observation period. Fig. 2C CD4 T cells displayed a higher and more sustained induction than summarizes DUSP4 protein expression at 48 h after stimulation memory CD4 cells (Fig. 1A). The kinetics in naïve CD4 T cells with immobilized anti-CD3/anti-CD28 antibodies in 24 20- to were not dependent on age; in contrast, transcription of DUSP4 in 35-y-old and 24 65- to 85-y-old healthy individuals (P = 0.04). CD4 memory T-cell responses was reduced and shortened in p-Nitrophenyl phosphate (pNPP) phosphatase assays of immu- young adults compared with the elderly (Fig. 1A). DUSP4 tran- noprecipitates showed that the overexpressed DUSP4 in elderly script numbers 48 h after anti-CD3/anti-CD28 stimulation (Fig. CD4 T cells was functionally active. Lysates from activated CD4 1B) or 72 h after stimulation with mDC and TSST (Fig. 1C) were memory T cells of five 20- to 35-y-old and five 65- to 85-y-old

− – Fig. 1. Influence of age on gene expression in CD4 T-cell memory responses. (A) CD4 CD45RO naïve (Left) and CD45RA memory T cells (Right) were stimulated on anti-CD3/anti-CD28 Ab-coated plates. DUSP4 transcripts quantified by qPCR at indicated time points are shown as mean ± SEM of three 20- to 35-y-old (○) and three 65- to 85-y-old (●) adults. (B) DUSP4 expression in CD4 memory T cells from 11 20- to 35-y-old (open bars) and 13 65- to 85-y-old (closed bars) individuals stimulated with anti-CD3/anti-CD28 Ab for 48 h is shown. (C)Vβ2+ CD4 memory T cells from 10 20- to 35-y-old (open bars) and 10 65- to 85-y- old (closed bars) adults were stimulated with TSST-1 and DC. DUSP4 transcript numbers after 72 h are shown as mean ± SEM. (D) Activated CD4 T cells were transfected with reporter gene constructs of the DUSP4 promoter. Luciferase activity was assessed 12 h after transfection in the absence or presence of the additional 4 h of restimulation with ionomycin and PMA. Results from five 20- to 35-y-old (open bars) and five 65- to 85-y-old (closed bars) individuals are shown as mean ± SEM. (E) CD4 memory T cells were transfected with EGR1-specific siRNA, stimulated, and assessed for EGR1 and DUSP4 expression at 48 h after stimulation. One Western blot representative of four experiments is shown. (F) Samples described in B were examined for the transcription of EGR1.

E880 | www.pnas.org/cgi/doi/10.1073/pnas.1109797109 Yu et al. Downloaded by guest on September 29, 2021 PNAS PLUS

Fig. 3. DUSP4 dampens CD4 memory T-cell activation. (A) CD4 T cells from Fig. 2. Activation-induced DUSP4 expression in CD4 memory T cells healthy adults were transfected with control or DUSP4-expressing pIRES2- increases with age. (A) Kinetics of DUSP4 expression in CD4 T cells after anti- AcGFP1 vector. Transfection efﬁciencies are shown as histograms of GFP CD3/anti-CD28 stimulation were determined by Western blotting. (B) CD4 expression. The solid line represents DUSP4-pIRES2-AcGFP1–transfected cells,

memory T cells were stimulated with mDCs pulsed with the superantigens IMMUNOLOGY and the dotted line represents untransfected cells. (B) CD4 T cells from young TSST and staphylococcal enterotoxin B. DUSP4 protein expression was adults’ PBMCs were stimulated on immobilized anti-CD3/anti-CD28 anti- quantified at indicated time points after stimulation. One representative bodies for 36 h and then transfected; 12 h after transfection, DUSP4-and blot with cells from 26- and 70-y-old individuals (Upper; lane C is a control control-transfected cells were assayed for the expression of activation lysate from Jurkat cells) and mean ± SEM of band intensities normalized to markers. Results are expressed as mean ± SEM. Mean fluorescence intensity β-tubulin from four young and four old individuals at indicated time points (MFI) of 9–11 experiments. (C) Transfected cells were restimulated with PMA after stimulation are shown. (C) DUSP4 expression at 48 h after CD3/CD28 and ionomycin for 4 h, and cytoplasmic cytokine production was assessed. stimulation was compared by Western blots. A representative experiment Results are expressed as mean ± SEM of a minimum of 10 experiments with cells from 35- and 66-y-old individuals is shown (Upper). Band densities depending on the marker analyzed. normalized to β-actin of blots from 24 20- to 35-y-old (open bars) and 24 65- to 85-y-old (closed bars) individuals are shown as mean ± SEM (Lower). (D) CD4 memory T cells were stimulated with immobilized anti-CD3/anti-CD28 efficiencies between 25% and 55%. To examine the consequences for 48 h. Lysates were precipitated with anti-DUSP4 antibodies, and phosphatase activities in the precipitates were determined. Results expressed as of increased DUSP4 expression during T-cell differentiation and arbitrary units are shown as mean ± SEM from experiments with cells from mimic the findings in CD4 memory T cells from elderly individu- five young (open bars) and five older (closed bars) adults. als, CD4 memory T cells from young adults were activated on plates coated with anti-CD3/anti-CD28 antibodies for 36 h and then transfected with a DUSP4-containing or control vector healthy individuals were precipitated with anti-DUSP4 antibodies. (Fig. 3A). Cells were then assayed for the expression of activation Phosphatase activities were significantly higher in the precipitates B P D markers 48 h after the initial activation (Fig. 3 ). Expression of from the elderly CD4 memory T cells ( = 0.006) (Fig. 2 ). These CD25 was not affected by increased DUSP4. In contrast, DUSP4- data convincingly show that activation-induced transcription of overexpressing cells showed a faster decline in the cell surface DUSP4 increases with age and results in increased and more density of CD69 (P < 0.001), CD40L (P < 0.001), and ICOS (P < sustained DUSP4 protein expression in elderly CD4 memory 0.001). When cells were restimulated after 48 h with ionomycin T-cell responses. and PMA and assayed for the production of cytokines by flow Functional Consequences of DUSP4 Overexpression. To examine the cytometry, IL-2 expression was infrequent, consistent with acti- C functional consequences of increased DUSP4 expression, CD4 T vated CD4 T cells being effector cells (Fig. 3 ). DUSP4 over- cells were transfected with DUSP4. Experiments in Fig. S1 show expression neither increased the frequency (by impairing effector that transfected DUSP4 had the predicted substrate specificity cell differentiation) nor decreased IL-2 production (by interfering (26). In T cells transfected with the DUSP4-containing vector and with T-cell activation). In contrast, IL-4 (P < 0.001), IL-17A (P < then activated by CD3 and CD28 cross-linking, ERK and JNK 0.001), and IL-21 (P < 0.001) production were all suppressed by phosphorylation 10 and 30 min after cross-linking was blunted, the overexpression of DUSP4 (Fig. 3C). These data suggest that whereas phosphorylation of p38 was not affected. Results are DUSP4 impairs CD4 T-cell differentiation, with preferential in- shown on gated GFP-positive CD4 memory cells with transfection hibition of some, but not all, effector functions.

Yu et al. PNAS | Published online March 20, 2012 | E881 Downloaded by guest on September 29, 2021 Functional Consequences of Increased DUSP4 Expression in Elderly was slightly lower in elderly CD4 T cells. Between 48 and 72 h, CD4 Memory T-Cell Responses. If increased activation-induced ex- expression of CD69 and CD40L continued to decline and did so pression of DUSP4 accounts for immune defects in the elderly, faster in elderly CD4 T cells. In contrast, expression of CD25 was similar patterns in elderly CD4 T-cell responses should be evi- not influenced by age. The results with elderly CD4 T cells dent in memory T cells from young adults that were manipulated mirrored CD4 memory T-cell responses of young adults with for their DUSP4 expression. Because DUSP4 expression is transfected DUSP4 (Fig. 3) with the exception of ICOS, where minimal in resting T cells and peaks at 48 h and later, we would a difference in transcript numbers at 48 h but only a minor trend expect that CD4 T cells from young and elderly adults differ in in cell surface expression at 72 h were seen. their ability to sustain the expression of activation markers and To address the question of whether inhibition of DUSP4 differentiate. The Affymetrix probes for ICOS and CD40L and transcription improved the functional activity of elderly CD4 the cytokines described in Fig. 3 failed to identify activation- memory T cells, the activation-induced transcription of DUSP4 induced changes, and we, therefore, used a quantitative PCR was silenced (Fig. 4C). The repression of DUSP4 had the ex- (qPCR) approach to examine the influence of age on the sus- pected functional consequences on the MAPK signaling path- tained expression of activation markers in a sample of 16 young ways; memory CD4 T cells from elderly individuals that were and 16 elderly adults. At 48 h after activation, transcript numbers transfected with the DUSP4-specific siRNA and activated for 48 of ICOS, CD40L, IL-4, IL-17A, and IL-21 were all reduced in h had increased ERK and JNK phosphorylation on restimulation the elderly CD4 T cells (Fig. 4A). Of the genes identified to be compared with control transfected T cells (Fig. S2). The differ- suppressed in DUSP4-transfected cells, only CD69 did not show ences were significant but not pronounced; they are likely an a difference. Similar results were obtained at the protein level underestimate of the silencing effect, because DUSP4 is strictly (Fig. 4B). CD4 memory T cells from 11 20- to 35-y-old and 11 65- restricted to the nucleus, and cytoplasmic pERK and pJNK to 85-y-old individuals were activated by culture on immobilized should, therefore, not be affected by DUSP4 silencing. As also anti-CD3/anti-CD28 antibodies; the expression of activation expected, the silencing of DUSP4 did not impact p38 phosphor- markers after 48 and 72 h was monitored by flow cytometry (Fig. ylation. Functional consequences of DUSP4 silencing are shown 4B). At 48 h, cell surface expression was essentially not different in Fig. 4 D–F. In these experiments, the influence of DUSP4 si- between young and elderly adults, consistent with the in- lencing on the expression of activation markers and the pro- terpretation that initial T-cell activation is intact. Only CD40L duction of cytokines was determined by comparing the responses

Fig. 4. DUSP4 silencing improves T-cell activity in the elderly. CD4 memory T cells were activated with plate-immobilized anti-CD3/anti-CD28 Ab. (A) Transcripts of indicated genes were quantified by qPCR after 48 h of culture. Results are shown as mean ± SEM of 16 20- to 35-y-old (open bars) and 16 65- to 85-y-old (closed bars) healthy individuals. (B) Expression of activation markers were monitored by flow cytometry 48 (Left) and 72 (Right) h after stimulation. Results from 20- to 35-y-old (open bars) and 65- to 85-y-old (closed bars) healthy individuals are shown as mean ± SEM of 11–14 experiments. (C) CD4 T cells were transfected with DUSP4-specific (closed bar) or control siRNA (open bar) and stimulated for 48 h. DUSP4 transcript numbers were quantified by qPCR, and DUSP4 protein was quantified by Western blot. (D) CD4 T cells were transfected with siRNA and activated with plate-immobilized anti-CD3/anti-CD28 Ab. Expression of activation markers after 72 h is shown as the percent increase after DUSP4 silencing in 11 20- to 35-y-old (open bars) and 11 65- to 85-y-old (closed bars) healthy individuals. (E) Cell cultures described in D were restimulated on day 2 for 4 h, and cytokine production was determined by flow cytometry. Results are again shown as the percent increased in DUSP4-silenced CD4 memory T cells. (F) IL-4 in supernatants from cultures as described in E was measured by ELISA.

E882 | www.pnas.org/cgi/doi/10.1073/pnas.1109797109 Yu et al. Downloaded by guest on September 29, 2021 of CD4 memory T cells silenced for DUSP4 with control trans- scription in CD4 T cells contributes to defective vaccine responses PNAS PLUS fected cells in 11 20- to 35-y-old and 11 65- to 85-y-old healthy in the elderly, we compared the gene expression in activated CD4 individuals. DUSP4 silencing did not significantly affect the ex- T cells of 10 18- to 35-y-old and 10 65- to 80-y-old healthy indi- pression of CD25 in the CD4 memory T cells of young adults or viduals on day 7 after influenza vaccination. Tetramer studies in the elderly individuals. In contrast, the expressions of CD69, vaccination studies have shown that antigen-specific CD8 T cells CD40L, and ICOS on day 3 after stimulation were increased by peak on days 7–14 and are contained within HLA-DR+CD38+ silencing DUSP4 (Fig. 4D). This improvement was relatively T cells (27, 28). In pilot studies, this marker profile was also best- minor for young individuals and averaged 10–20% for all three suited for identifying activated CD4 T cells. Frequencies of activation markers tested. In contrast, elderly CD4 memory T-cell CD4+HLA-DR+CD38+ cells after vaccination were less than 1% responses benefitted more by silencing; particularly, the expres- of all CD4 T cells and only insignificantly higher in younger than sion of CD40L increased by close to 50% (P < 0.001 compared older adults. Frequency estimates of influenza hemagglutinin and with the improvement seen with young CD4 memory T cells). A neuramidase-specific T cells generally have to rely on IFN-γ similar pattern was seen for cytokine expression (Fig. 4E). In production and range from 0.1% to 1% (29). CD4+HLA- these experiments, CD4 memory T cells were restimulated 48 h DR+CD38+ cells were sorted from positively selected peripheral after the initial stimulation and assayed for the presence of cy- blood T cells and analyzed for transcript expression by qPCR. toplasmic cytokines by flow cytometry. DUSP4 silencing did not T cells from the elderly adults expressed more DUSP4 (P = 0.002) majorly influence the production of IL-2 or IFN-γ in CD4 and less CD40L (P < 0.001), ICOS (P = 0.01), and IL-4 transcripts memory T cells from young or elderly individuals. In contrast, the (P = 0.03) (Fig. 5A). DUSP4 transcripts inversely correlated with frequencies of cells producing IL-4, IL-17A, and IL-21 were in- all three T-cell activation markers significantly (Fig. 5B), sug- creased by DUSP4 silencing; the amount of cytokines produced gesting that DUSP4 expression in the elderly activated CD4 T cells per cell as determined from the fluorescence intensity was not accounts for the reduced helper activity for B-cell responses. significantly different. For all three cytokines, the increase was most pronounced in CD4 memory T-cell responses from the el- Improved T Cell-Dependent B-Cell Responses After DUSP4 Silencing. derly; particularly, DUSP4 silencing caused a higher increase in Based on the finding that overexpression of DUSP4 preferentially the frequencies of IL-4 (P = 0.007) and IL-21 (P = 0.04) pro- impairs CD40L expression and the production of IL-4 and IL-21, ducing T cells compared with the improvement that was seen in we hypothesized that DUSP4 expression is of particular impor- CD4 memory T cells of young adults. The flow cytometric anal- tance in controlling T-helper function for B-cell differentiation. To yses were confirmed by ELISA (Fig. 4F). Concentrations of IL-4 examine the influence of age on the ability of memory CD4 T cells in culture supernatants harvested 48 h after activation were lower to provide B-cell help, we used a coculture system using T cells with T cells from 65- to 85-y-old individuals compared with young from 20- to 35-y-old and 65- to 85-y-old individuals. T cells were adults. This impaired production was, in part, restored by the treated with mitomycin C to prevent proliferation, and they were silencing of DUSP4 (P < 0.001). activated with anti-CD3/anti-CD28 Ab and cocultured with B cells. B cells were derived from an unrelated young donor to exclude any Correlation of DUSP4 Expression and T-Cell Function in Vivo After contribution of age-dependent defects in the B cells. T and B cells Influenza Vaccination. To determine whether these in vitro obser- were, therefore, equally MHC mismatched in all assays. Cells vations hold up for in vivo response and increased DUSP4 tran- concentrations were chosen so that, in the absence of anti-CD3 Ab IMMUNOLOGY

Fig. 5. DUSP4 expression in in vivo-activated T cells after influenza vaccination inversely correlates with T-helper markers. Peripheral blood was obtained from healthy individuals on day 7 after influenza vaccination. CD3+CD4+HLA-DR+CD38+ T cells were purified by cell sorting and analyzed for the expression of DUSP4, CD40L, ICOS, and IL-4 transcripts by qPCR. (A) Transcript numbers normalized to 18s RNA from 10 18- to 35-y-old (open bars) and 10 65- to 80-y-old (closed bars) individuals are shown as mean ± SEM. (B) Data from young (open squares) and elderly (closed squares) individuals are shown as 2D regression blots.

Yu et al. PNAS | Published online March 20, 2012 | E883 Downloaded by guest on September 29, 2021 to activate T cells, B cells stayed quiescent despite the MHC mis- CD40L-induced CD40 stimulation and activation of the p38 match. Successful B-cell differentiation was defined as the gener- pathway (30). E47 expression was significantly lower in B cells that ation of CD19+CD38high IgDlow or CD19+CD27+ cells (Fig. 6A). were cocultured with memory CD4 T cells from 65- to 85-y-old In the presence of T cells activated with anti-CD3 Ab, a pop- individuals (P = 0.003). DUSP4 silencing in the T-cell population low high ulation of IgD CD38 B cells emerged that was more fre- improved the ability of T cells to up-regulate E47 expression in B– quent when B cells were cocultured with CD4 memory T cells T cell cultures. from young adults compared with elderly adults (P = 0.004) (Fig. 6 A and B). Also, reduced expression of CD27 on B cells depending Effect of DUSP4 Expression in T Cells on Humoral Responses After on the age of the T-cell donor was consistent with defective T-cell Immunization. Data so far showed that increased DUSP4 in ac- help. Silencing of DUSP4 in the CD4 memory T-cell population tivated CD4 memory T cells impairs their ability to express only marginally improved B-cell differentiation supported by T molecular mediators important in providing B-cell help and that cells from young individuals, but it restored the B-cell response in DUSP4 overexpression is, at least in part, responsible for the the coculture system with elderly CD4 T cells to a level similar to impaired T cell-dependent B-cell responses in the elderly. To young individuals. Results from coculture systems with T cells examine whether DUSP4 expression in T cells controls an im- from 10 20- to 35-y-old and 10 65- to 85-y-old healthy individuals munization response in vivo, T cells from TCR transgenic OT-II are summarized in Fig. 6C. All B cells were derived from young mice were transduced with a DUSP4-expressing or a control adults unrelated to the T-cell donors. Results are expressed as percent increase in the cultures with DUSP4 silenced compared retroviral vector and adoptively transferred into CD4 KO mice. with control transfected T cells. In cultures with T cells from young Mice were immunized intraperitoneally with NP-OVA in alum, adults, only 10–20% improvement was seen with DUSP4 silenc- and cellular and humoral immune responses to the immuniza- ing. In contrast, in the cultures with memory CD4 T cells from the tion were assessed on day 14. Frequency of adoptively trans- elderly, a much more striking improvement was seen; the fre- ferred CD4 T cells in the spleens of the host was not different quency of CD27+ B cells increased by 30–40%, and in particular, irrespective of whether the T cells were transfected with the the frequency of CD38high IgDlow cells nearly doubled. This im- control or the DUSP4-expressing vector (Fig. 7A). Enumeration provement in elderly T-cell help was significantly more pro- of splenic cell populations showed equal numbers of ∼40–50 nounced compared with the effect of DUSP4 silencing on the B- million B cells and 1.5 million T cells irrespective of whether the cell help provided by young CD4 memory T cells. As an additional T cells overexpressed DUSP4. However, CD40L and ICOS ex- marker of B-cell differentiation, we quantified the transcription pression was significantly reduced by DUSP4 expression (Fig. factor E47 (Fig. 6D). Expression of E47 is dependent on p38 ac- 7B). The frequency of NP-specific B cells was significantly lower tivity in B cells, and in a T–B cell coculture system, it reflects when DUSP4-transduced T cells were adoptively transferred

Fig. 6. DUSP4 silencing in CD4 memory T cells improves T cell-dependent B-cell responses. (A) CD4 memory T cells from 10 20- to 35-y-old and 10 65- to 85-y- old healthy individuals were cocultured with B cells from young healthy adults on anti-CD3/anti-CD28–coated plates. A representative density plot of the expression of CD38 and IgD on CD19+ cells is shown (Upper). (B) Cultures were examined for the frequencies of CD19+CD38+IgD– and CD19+CD27+ cells. Results are shown as mean ± SEM. (C) CD4 memory T cells were transfected with DUSP4 or control siRNA and cultured as described in A. Results are expressed as – percent increase in the frequencies of CD19+CD38+IgD and CD19+CD27+ cells and the cell surface expression of CD86 in the cultures with DUSP4-silenced compared with control-transfected T cells for the young (open bars) and elderly (closed bars) adults. (D) Cells cultured as described in C were assessed for the transcription of the transcription factor E47 by qPCR.

E884 | www.pnas.org/cgi/doi/10.1073/pnas.1109797109 Yu et al. Downloaded by guest on September 29, 2021 PNAS PLUS

Fig. 7. DUSP4 expression in T cells suppresses humoral responses after immunization in vivo. T cells from TCR transgenic (OT-II) mice were transduced with a DUSP4-expressing (solid bar) or control retroviral (open bar) vectors and adoptively transferred into CD4 KO (B6.129S2-Cd4tm1Mak/J) mice. Mice were immunized i.p. with NP-OVA spleens and sera were harvested on day 14. (A) The total numbers of splenic CD4 T cells, B220 B cells, NP-specific B cells, and NP- specific germinal center B cells in reconstituted and immunized mice were enumerated. (B) Expression of CD40L and ICOS was determined on splenic CD4 T cells by flow cytometry. Results are representative of two experimental series with four mice each and are shown as mean ± SEM. (C) NP-OVA-specific IgGs

were determined by ELISA. IMMUNOLOGY

(P = 0.003). A striking difference was also found for antigen- sizable and diverse number of naïve CD4 T cells into their eighth specific B cells that expressed a germinal center phenotype; such decade of life (34, 35). Functional defects in naïve CD4 T cells have antigen-specific B cells were nearly absent in hosts adoptively been described in murine model systems (36). These defects may transferred with DUSP4-expressing T cells compared with ∼400,000 compromise immune responses to newly developing infectious in the mice adoptively transferred with the control transduced organisms, such as the H1N1 or H1N5 influenza viruses or severe T cells (P = 0.009). The detrimental effect of DUSP4 expression acute respiratory syndrome-associated corona virus. However, in T cells on the ability to support T cell-dependent B-cell most of the immune responses in the adult are recall and not responses is also supported when antibody titers to the immu- primary responses. In contrast to the CD8 compartment, which is nizing antigen ovalbumin are compared (Fig. 7C). The induction compromised by the oligoclonal expansion of CD8 effector cells, of NP-OVA-specific IgG after immunization was about fivefold CD4 memory T cells maintain a diverse and functionally balanced reduced in mice adoptively transferred with the DUSP4-trans- repertoire with age. In murine systems, memory T cells that have duced T cells. been established early in life are generally well-preserved; only memory T cells generated from primary T-cell responses late in life Discussion are impaired. In humans, diseases such as herpes zoster or failure The current study shows that CD4 memory T cells with age lose to respond to the annual influenza vaccinations clearly indicate their ability to differentiate into effective helper T cells for B-cell defective CD4 T-cell memory responses with age. The identifica- responses because of the overexpression of DUSP4. Transcrip- tion of an increased expression of DUSP4 is a molecular charac- tion of DUSP4 is activation-induced; DUSP4 protein levels peak terization of such a defect that can be therapeutically targeted. at 48–72 h after TCR triggering in vitro. Increased and sustained DUSP4 belongs to the family of dual-specific phosphatases that expression of DUSP4 in CD4 memory T cells from individuals dephosphorylate both phosphoserine/threonine and phosphotyr- older than 65 y limits the expression of CD40L and reduces the osine residues on MAPK (37). DUSPs are pivotal regulators in production of the cytokines IL-4 and IL-21, all features of fol- the dynamic regulation of MAPK activities important in T-cell licular helper cells (31, 32). As a consequence, T cell-dependent activation, differentiation, and cytokine production and there- B-cell responses are impaired. Reduction of DUSP4 transcrip- fore, are of interest as pharmacologic targets (38–41). Inhibitors tion after T-cell activation and pharmacological inhibition of for the catalytic phosphatase domain frequently lack specificity, DUSP4 activity emerge as possible interventions to restore hu- and specific phosphatase inhibitors have been difficult to develop. moral immune responses in elderly individuals after vaccination. However, DUSP4 has a unique structure with an allosteric pocket Approaches to understand age-dependent defective adaptive close to the phosphatase domain. This pocket is shared with immunity have been dominated by the recognition of thymic de- DUSP6, for which a specific inhibitor was recently described (42). mise and its implication for maintaining T-cell homeostasis (1, 2, DUSP4 is one of four phosphatases that are transcriptionally 33). Indeed, compartment sizes and TCR diversity of naïve T cells induced on stressors and located in the nucleus (43, 44). The decline with age; however, most individuals continue to have a additional members are DUSP1, DUSP2, and DUSP5. Based on

Yu et al. PNAS | Published online March 20, 2012 | E885 Downloaded by guest on September 29, 2021 the expression of a nuclear localization sequence, they function by vated elderly CD4 memory T cells. These experiments indicated trapping their substrate in the nucleus in addition to their phos- that transcriptional control rather than epigenetic mechanisms phatase activity (43, 44). In our in vitro culture system, T-cell accounts for the increased promoter activity in elderly T cells. Very activation only up-regulated the expression of DUSP4 and little is known about the transcriptional regulation of DUSP4. DUSP5, whereas DUSP1 and DUSP2 transcripts declined. Only Transcription factors that have been implicated in DUSP4 regu- the activation-induced expression of DUSP4 was age-dependent. lation include HoxA10, CR1, NF-κB, HNF1, and E2F-1 (58, 59). The reduced ability to mount robust humoral responses with age The work by Berasi et al. (25) showed an induction of DUSP4 reflects a combination of T- and B-cell intrinsic defects as well as caused by activation of the AMP-activated kinase (AMPK) that structural changes, such as loss in germinal center architecture could be attributed to AMPK-induced production of the imme- (45–47). T-cell functions important in B-cell differentiation in- diate early transcription factor EGR1. The work by Berasi et al. clude production of IL-4 and IL-21 and sustained CD40L ex- (25) identified an EGR1 binding site at position −119 in the pression, the duration of which determines the outcome of CD40 DUSP4 promoter. EGR1 is known to be up-regulated by TCR signaling (31, 48). Reduced CD40L expression with age impairs stimulation. As shown in Fig. 1E, DUSP4 expression after TCR the ability to provide competent T-cell help in the mouse (49). In stimulation is dependent on the transcription of EGR1. TCR-in- our studies, decreased production of IL-4 and IL-21 and shortened duced EGR1 expression in CD4 memory T cells was increased expression of CD40L were related to DUSP4 activity. Forced with age, suggesting that this abnormality is primary. overexpression of DUSP4 in CD4 memory T cells from young, With the changing demographics in the aging population, the healthy donors reduced the expression of CD40L, ICOS, and the increased morbidity from infections in the elderly has become an cytokines IL-4 and IL-21, but it did not affect CD25 expression or increasing public health problem. Improved vaccination strate- the production of IFN-γ. Conversely, DUSP4 silencing selectively fi gies are needed to overcome the age-related immune defects. up-regulated IL-4, IL-21, CD40L, and ICOS and did so signi - The current approaches are focused on improving vaccine for- cantly more in elderly adults who have increased DUSP4 activity. mulations (22, 23, 60, 61). Our studies suggest that a two-pronged The functional implications of DUSPs are determined by their strategy might be advantageous by targeting T-cell defects in substrate specificities and their subcellular localizations as well as their expression kinetics in the context of activation-induced addition to enhancing vaccines. Inhibition of activation-induced MAPK activity. DUSP4 has been shown in in vitro studies to de- DUSP4 transcription or inhibition of DUSP4 activity in the days phosphorylate predominantly pERK and less efficiently pJNK subsequent to vaccination might be able to restore competency of while having only very low specific activity for pp38 (37, 50). Our CD4 T memory cells to provide help for T cell-dependent B-cell data in human T cells are consistent with these observations. responses. Such an intervention could involve a pharmacological DUSP4 overexpression, as well as silencing, modified activation- agent that binds to the allosteric pocket shared by DUSP4 and induced ERK and JNK while not affecting p38 phosphorylation. DUSP6. Because DUSP6 is known to raise the TCR activation This substrate specificity may provide an explanation for the se- threshold and reduce TCR signaling, such an agent would be lectivity of functional modifications related to DUSP4. The ERK predicted to also improve T-cell activation in addition to T-cell pathway has been implicated in Th2 differentiation in several differentiation. Alternatively, TCR-induced EGR1 transcription, studies. Inhibition of ERK activation in dominant negative Ras possibly by inhibiting AMPK activation or activity, could be tar- transgenic T cells impaired Th2 responses (51). In JNK1 KO mice, geted to reduce the transcription of DUSP4. T cells hyperproliferated, exhibited decreased activation-induced cell death, and preferentially differentiated into Th2 cells. In Materials and Methods contrast, p38 activation favors Th1 differentiation, and cytokine Cells. Peripheral blood was obtained from 101 young (aged 20–35 y) and 101 production by Th1 cells is dependent on JNK activity (52, 53). elderly (aged 65–85 y) healthy individuals. Subjects with acute diseases, cur- In the coculture system of young adult B cells with memory rent or previous history of immune-mediated diseases or cancer (except lim- CD4 T cells of different ages, we have used the frequency of ited basal cell carcinoma), or chronic diseases that were not controlled on oral CD38high IgDlow as well as the expression of the transcription medications were excluded. The study was in accordance with the Declaration factor E47 as B-cell differentiation markers. IL-4 and IL-21 have of Helsinki and approved by the Emory and Stanford Institutional Review been shown to be of functional importance in this system (54, 55). Boards, and all participants gave informed consent. CD4 T cells were nega- fl tively enriched with human CD4+ T-cell enrichment mixture (RosetteSep; Transcription of E47 re ects CD40-induced p38 activation and − + therefore, is a measure of sustained CD40L expression. Our studies StemCell Technologies) from whole blood. CD45RA CD4 memory T cells were further isolated by depleting CD45RA+ subsets with anti-CD45RA magnetic suggest that T-cell DUSP4 is an important regulator of T-helper – beads (Miltenyi Biotec). Vβ2+ T cells were purified from CD45RA CD4+ memory function in T cell-dependent B-cell responses. IL-4, IL-21, CD40L, – β and ICOS are all T-cell properties that are characteristic of fol- T cells with biotin anti-human TCR V 2 mAb (Beckman Coulter) and positive licular helper cells (32, 56). In our in vitro system, we only found selection using antibiotin magnetic beads (Miltenyi Biotec). Alternatively, peripheral blood mononuclear cells (PBMCs) were isolated using lymphocyte CXCR5, a cell surface marker associated with follicular helper separation medium. Total CD4 T or B cells were positively selected from PBMC cells, in a small population of T cells independent of the expression with anti-CD4 or anti-CD19 beads. To isolate memory CD4 T cells from PBMC, of DUSP4. Also, in our in vivo studies in the mouse model, we did naïve T cells and CD14+ monocytes were depleted by anti-CD45RA and anti- not observe an effect of DUSP4 expression on the frequencies of CD14 beads. CD4 cells were then positively isolated. Some of CD4 memory CXCR5 and signaling lymphocyte activation molecule (SLAM)- fi + fi T cells were puri ed using EasySep human memory CD4 T-cell enrichment kit expressing T cells, although frequencies of antigen-speci cgermi- (StemCell Technologies). mDCs were generated from monocytes isolated with nal center B cells were reduced. anti-CD14 magnetic beads (Miltenyi Biotec). Cells were cultured in RPMI–10% In addition to inhibiting cytokine production and CD40L ex- (vol/vol) FCS supplemented with 800 U/mL granulocyte-macrophage colony- pression, nuclear expression of DUSP4 may also function as an stimulating factor (GM-CSF) (R&D Systems) and 1,000 U/mL IL-4 (R&D Systems). important senescence gene and impair cell cycle entry and pro- On day 6, immature DCs were stimulated with 1,100 U/mL TNF-α (R&D Sys- liferative responses (57). Studies have shown that the DUSP- tems) and 1 μg/mL prostaglandin E2 (Sigma-Aldrich) for 24 h. mediated inactivation of nuclear ERK2 represents a key event in Peripheral blood was collected from an additional 10 18- to 35-y-old and 10 the establishment of replicative senescence. The restoration of 65- to 80-y-old individuals who received their annual influenza vaccination nuclear ERK activity by DUSP inhibition may bypass critical during the 2010–2011 influenza season. Heparinized whole blood obtained senescence checkpoints and improve the clonal expansion of T on day 7 after vaccination was first depleted of B cells and plasmablasts cells that is compromised with age. using the RosetteSep Human B Cell Enrichment Mixture (Stemcell Technol- The increased activation-induced expression of DUSP4 in ogies). Erythrocytes in rosetted cells were lysed with ACK lysis buffer (Invi- CD4 memory T cells in elderly individuals is caused by increased trogen), mononuclear cells were stained with APC-CD3–, PerCP/Cy5.5-CD4–, transcriptional activity. Reporter gene activity of constructs in- FITC-CD38–, and PE-HLA-DR–specific antibodies, and CD3+CD4+CD38+HLA- cluding the DUSP4 promoter was significantly increased in acti- DR+–activated T cells were sorted on a BD Aria II.

E886 | www.pnas.org/cgi/doi/10.1073/pnas.1109797109 Yu et al. Downloaded by guest on September 29, 2021 T-Cell Cultures. Vβ2+ memory CD4 T cells (25 × 103/well) were stimulated PE-CD154 (CD40L), and FITC or PE-CD278 (ICOS) (eBioscience). For intracellular PNAS PLUS with 0.5 × 103 mDCs pulsed with 0.04 ng/mL TSST-1 (Toxin Technology) in 96- cytokine staining, FITC-IFN-γ, FITC-IL-2, PE-IL-4, and PE-IL-21 (BD Biosciences) well round plates. DC and TSST-1 concentrations were chosen that did not and Alexa Fluor 647-IL-17A (eBioscience) were used. For phosphoepitope yield detectable alloreactive and suboptimal TSST-1–speciﬁc proliferative analysis, 1 × 106 CD4 T cells were stimulated by anti-CD3/CD28 mAb (1 μg/mL responses. The cells were harvested before and at 16, 40, and 72 h after each) cross-linking, ﬁxed with 2% formaldehyde for 10 min at room tem- stimulation for gene expression analysis. In other experiments, CD4 memory perature, permeabilized in 100% methanol at −20 °C overnight, and stained T cells were stimulated with mDCs and a combination of TSST and staphy- with antibodies to phospho-ERK1/2, phospho-JNK, and phospho-p38 (Cell lococcal enterotoxin B (0.1 ng/mL each) to activate ∼40% of all cells or with Signaling Technology). Cytometry was performed on an LSRII system (BD anti-CD3 and anti-CD28 antibodies immobilized on tissue culture plates. Biosciences), and data were analyzed using FlowJo software (Tree Star).

Gene Expression Arrays. RNA before and 16, 40, and 72 h after stimulation was DUSP4 Transient Transfection. Freshly purified CD4 T cells were transfected with fi fi puri ed using the RNeasy Mini Kit (Qiagen). RNA was ampli ed and labeled 4 μg DUSP4-pEGFP-N1 (subcloned from pCDNA3.1 DUSP4 vector; a gift from fi using a modi cation of the technique described in the work by Baugh et al. Dr. Y. Yin, Columbia University, New York, NY) or 4 μg pEGFP-N1 vector using (62). Affymetrix GeneChips (U133A) hybridization was performed at the the Amaxa Nucleofector system and the Human T-Cell Nucleofector Kit Institute for Systems Biology (Seattle, WA). To identify genes differentially (Lonza); 24 h after transfection, MAPK activation after CD3 cross-linking was expressed over time, we used the EBarrays method implemented in Bio- analyzed in gated GFP-positive cells by phosphoepitope-specific flow cytom- conductor (63). EBarrays assigns a probability to each gene at each time, etry. Alternatively, activated T cells (36 h on anti-CD3/anti-CD28–coated plates) regardless of whether gene expression levels are equivalent (64). The ap- were transfected with 2 μg DUSP4-pIRES2-AcGFP1 or 2 μg pIRES2-AcGFP1 proach uses information across genes and arrays to optimize model fit. After empty vector (Clontech Laboratories) and examined for cell surface marker model parameters were obtained, probabilities for differential expression expression. Cytoplasmic cytokines were examined after additional stimulation were calculated for every gene. To control for possible pairing effects (a with PMA plus ionomycin in the presence of Brefeldin A for the last 4 h. sample from a young and an elderly adults was always run in parallel in each experiment), each pair of subjects was compared, and four different prob- RNAi. Total CD4 or memory CD4 T cells using the Amaxa Nucleofector system abilities were obtained for each gene. The maximum or geometric averages μ of these probabilities were adopted to yield a single score summarizing the and the Human T-Cell Nucleofector Kit (Lonza) were transfected with 1.5 g fi – fi evidence of differential expression (65). Results were similar for several DUSP4-speci c siRNA or EGR-1 speci c siRNA (siGENOME SMARTpool; different approaches, and only results based on the maximum are reported. Dharmacon). Negative control siRNA came from Qiagen; 12 h after transfection, cell numbers were adjusted, and cells were stimulated on plates fi qPCR. Total RNA from cell cultures was isolated with TRIzol reagent (Invi- coated with anti-CD3/anti-CD28 antibodies. Knockdown ef ciencies were trogen), and cDNA templates were synthesized with AMV reverse tran- monitored by qPCR and Western blotting. scriptase (Roche) and random hexamer primers (Roche). In the studies of the 6 T-cell response in influenza-vaccinated individuals, 2,000–10,000 sorted cells ELISA. CD4 T cells (1 × 10 /mL) were stimulated with plate-bound anti-CD3 were directly lysed using the CellsDirect kit (Invitrogen), and cDNA was and anti-CD28 for 48 h. IL-4 in the supernatants was analyzed using the synthesized. SYBR quantitative PCR was performed using the primers de- Human IL-4 ELISA Ready-SET-Go Kit (eBioscience). scribed in Table S2. Copy numbers were calculated from standard curves using plasmids containing the relevant gene. Results are given as relative In Vitro T Cell-Dependent B-Cell Differentiation. T cells were treated with 30 transcript numbers after adjusting for 18s ribosomal RNA transcripts. μg/mL mitomycin C (Sigma-Aldrich) for 30 min at 37 °C, washed three times, and then cocultured at a concentration of 1 × 105 cells/well for 7 d with 0.5 × 5 Western Blotting. Cultured total CD4 or memory CD4 T cells were lysed in lysis 10 B cells purified from an unrelated young adult in 96-well flat-bottomed IMMUNOLOGY

buffer [20 mM Tris·HCl, pH 7.5, 150 mM NaCl, 1 mM Na2EDTA, 1 mM EGTA, wells uncoated or coated with anti-CD3 and anti-CD28 antibodies (54). 1% Triton X-100, 2.5 mM sodium pyrophosphate, 1 mM β-glycerophosphate,

1mMNa3VO4, protease inhibitors (protease inhibitor mixture for mamma- In Vivo T Cell-Dependent B-Cell Responses. CD4 KO (B6.129S2-Cd4tm1Mak/J) lian cell extracts; Sigma-Aldrich)]. Lysate was cleared by centrifugation mice were reconstituted with T cells from TCR transgenic (OT-II) mice (Jackson (12,000 × g at 4 °C for 10 min), and the supernatants were boiled in SDS Laboratory) and immunized as described (49). The experimental protocol was loading buffer. Proteins were separated on 10% SDS-polyacrylamide gels, approved by the Emory and Palo Alto Veterans Affairs Institutional Animal electroblotted to nitrocellulose membranes (Whatman), and developed with Care and Use Committees. –β –β anti-DUSP4, anti -actin (Santa Cruz Biotechnology), anti-EGR1, or anti - Mouse DUSP4 cDNA (clone ID 40092218; Open Biosystems) was subcloned tubulin antibody (Cell Signaling Technology) and ECL reagent (Pierce). Band into the retroviral expression vector MSCV PIG (Puro IRES GFP; Addgene). Ret- fi intensities were quanti ed with Quantity One 1D Analysis Software (Bio-Rad roviral supernatant was produced using the Phoenix-ECO cell line (ATCC). CD4 T β β Laboratories) and expressed relative to -actin or -tubulin. cells were isolated from OT-II mouse spleens by negative selection (Miltenyi Biotec), stimulated with 2 μg/mL Con A and 100 U/mL IL-2 for 48 h, and then Immunoprecipitation and Phosphatase Assays. To quantify DUSP4 activity, cultured with retrovirus; 48 h after infection, cells were transferred into fresh whole-cell lysates of activated CD4 memory T cells from young and elderly adults complete RPMI 1640 media with 20 U/mL IL-2 for an additional 48 h puromycin were incubated with anti-DUSP4 antibody and protein A/G plus-agarose (Santa selection. Transfection efficiency was monitored using flow cytometry. Cruz Biotechnology). Immune complexes were washed four times with ice-cold DUSP4-overexpressing CD4 T cells (2 × 106 cells/mouse) were i.v. trans- immunoprecipitation lysis buffer without phophatase inhibitors (Pierce). ferred into CD4 KO hosts; control hosts received empty vector-infected CD4 T Phosphatase activity was assessed using a pNPP phosphatase assay kit (BioAssay cells. On the next day, mice were immunized intraperitoneally with 150 μg Systems) according to the manufacturer’s protocol. Nonspecifichydrolysisof NP-OVA (Biosearch Technologies) in PBS with alum. After reimmunization pNPP was assessed in parallel samples precipitated with control antibodies. with 100 μg NP-OVA on day 12, splenocytes and serum were collected on day Results are shown as arbitrary units defined as the phosphatase activity in specific minus nonspecific precipitates divided by the nonspecific activity. 14. For each experiment, four hosts were used in each group, and each experiment was performed two times. DUSP4 Promoter Reporter Gene Assays. Purified memory CD4 T cells were Splenocytes were analyzed for cell surface marker expression using the stimulated with plate-bound anti-CD3/anti-CD28 antibodies for 36 h. Acti- following antibodies: PE-CD4, APC-CD62L, APC-CD154 (CD40L), PE-B220, and vated cells were transfected with either 4.5 μg pGL3 basic vector or 4.5 μg APC-streptavidin (eBioscience), Alexa Fluor 647-CD278 (ICOS), PerCP/Cy5.5- DUSP4-luc reporter construct (a gift from Dr. M. S. Roberson, Cornell Uni- CD150 (SLAM), and PE/Cy7-CD38 (BioLegend), PerCP-B220, PerCP/Cy5.5-CD44, versity, Ithaca, NY) together with 0.5 μg TK-pRL control vector. Cells were and PE/Cy7-CXCR5 (BD Pharmingen) as well as biotin-peanut agglutinin left unstimulated for 16 h or restimulated with PMA plus ionomycin after 12 (Vector Laboratories) and NP-PE (Biosearch Technologies). h for 4 h. Luciferase activity was determined by the dual luciferase reporter NP-OVA-specific IgG was quantified by the mouse IgG ELISA quantitation assay kit (Promega). kit (Bethyl Laboratories) using NP-OVA (10 μg/mL) as the capture antigen.

Flow Cytometry. For cell surface staining, the following antibodies were Statistical Analysis. Data were analyzed by paired or unpaired t test as ap- used: FITC-IgD, PerCP-CD4, APC-CD19, FITC or PE-CD25, PE-CD27, PE or APC- propriate. DUSP4 and activation marker expression in vaccinated individuals CD45RO, PE/Cy7 or PE-CD69, and PE-CD86 (BD Biosciences) and PE/Cy7-CD38, were compared by Pearson’s correlation.

Yu et al. PNAS | Published online March 20, 2012 | E887 Downloaded by guest on September 29, 2021 ACKNOWLEDGMENTS. This work was supported by a grant from the Georgia of Health Grants R01 AR042527 (to C.M.W.), R01 EY011916 (to C.M.W.), R01 Cancer Coalition and National Science Foundation Grant DMS0847234 Faculty AI044142 (to C.M.W.), P01 HL058000 (to C.M.W.), R01 AG015043 (to J.J.G.), Early Career Development Award (CAREER) (to M. Yuan); National Institutes U19 AI057266 (to J.J.G.), and U19 AI090019 (to J.J.G.).

1. Weng NP (2006) Aging of the immune system: How much can the adaptive immune 34. Czesnikiewicz-Guzik M, et al. (2008) T cell subset-specific susceptibility to aging. Clin system adapt? Immunity 24:495–499. Immunol 127:107–118. 2. Hakim FT, Gress RE (2007) Immunosenescence: Deficits in adaptive immunity in the 35. Naylor K, et al. (2005) The influence of age on T cell generation and TCR diversity. elderly. Tissue Antigens 70:179–189. J Immunol 174:7446–7452. 3. Thompson WW, et al. (2003) Mortality associated with influenza and respiratory 36. Maue AC, et al. (2009) T-cell immunosenescence: Lessons learned from mouse models syncytial virus in the United States. JAMA 289:179–186. of aging. Trends Immunol 30:301–305. 4. Rivetti D, et al. (2006) Vaccines for preventing influenza in the elderly. Cochrane 37. Guan KL, Butch E (1995) Isolation and characterization of a novel dual specific Database Syst Rev 3:CD004876. phosphatase, HVH2, which selectively dephosphorylates the mitogen-activated 5. Goodwin K, Viboud C, Simonsen L (2006) Antibody response to influenza vaccination protein kinase. J Biol Chem 270:7197–7203. in the elderly: A quantitative review. Vaccine 24:1159–1169. 38. Zhang Y, et al. (2009) MKP-1 is necessary for T cell activation and function. J Biol 6. Simonsen L, et al. (2005) Impact of influenza vaccination on seasonal mortality in the Chem 284:30815–30824. US elderly population. Arch Intern Med 165:265–272. 39. Zhang Y, et al. (2004) Regulation of innate and adaptive immune responses by MAP 7. Lee LY, et al. (2008) Memory T cells established by seasonal human influenza A kinase phosphatase 5. Nature 430:793–797. infection cross-react with avian influenza A (H5N1) in healthy individuals. J Clin Invest 40. Kovanen PE, et al. (2008) T-cell development and function are modulated by dual 118:3478–3490. specificity phosphatase DUSP5. J Biol Chem 283:17362–17369. 8. Arvin AM (2008) Humoral and cellular immunity to varicella-zoster virus: An overview. 41. Bettini ML, Kersh GJ (2007) MAP kinase phosphatase activity sets the threshold for – J Infect Dis 197(Suppl 2):S58 S60. thymocyte positive selection. Proc Natl Acad Sci USA 104:16257–16262. 9. Donahue JG, Choo PW, Manson JE, Platt R (1995) The incidence of herpes zoster. Arch 42. Molina G, et al. (2009) Zebrafish chemical screening reveals an inhibitor of Dusp6 that – Intern Med 155:1605 1609. expands cardiac cell lineages. Nat Chem Biol 5:680–687. 10. Haynes BF, Markert ML, Sempowski GD, Patel DD, Hale LP (2000) The role of the 43. Kondoh K, Nishida E (2007) Regulation of MAP kinases by MAP kinase phosphatases. thymus in immune reconstitution in aging, bone marrow transplantation, and HIV-1 Biochim Biophys Acta 1773:1227–1237. – infection. Annu Rev Immunol 18:529 560. 44. Jeffrey KL, Camps M, Rommel C, Mackay CR (2007) Targeting dual-specificity 11. Sauce D, et al. (2009) Evidence of premature immune aging in patients phosphatases: Manipulating MAP kinase signalling and immune responses. Nat Rev – thymectomized during early childhood. J Clin Invest 119:3070 3078. Drug Discov 6:391–403. 12. Goronzy JJ, Weyand CM (2005) T cell development and receptor diversity during 45. Dunn-Walters DK, Ademokun AA (2010) B cell repertoire and ageing. Curr Opin – aging. Curr Opin Immunol 17:468 475. Immunol 22:514–520. 13. Effros RB, Dagarag M, Spaulding C, Man J (2005) The role of CD8+ T-cell replicative 46. Linton PJ, Dorshkind K (2004) Age-related changes in lymphocyte development and – senescence in human aging. Immunol Rev 205:147 157. function. Nat Immunol 5:133–139. 14. Weng NP, Akbar AN, Goronzy J (2009) CD28(-) T cells: Their role in the age-associated 47. Nikolich-Zugich J, Rudd BD (2010) Immune memory and aging: An infinite or finite decline of immune function. Trends Immunol 30:306–312. resource? Curr Opin Immunol 22:535–540. 15. Akbar AN, Fletcher JM (2005) Memory T cell homeostasis and senescence during 48. Román E, et al. (2002) CD4 effector T cell subsets in the response to influenza: aging. Curr Opin Immunol 17:480–485. Heterogeneity, migration, and function. J Exp Med 196:957–968. 16. Liuzzo G, et al. (2000) Monoclonal T-cell proliferation and plaque instability in acute 49. Eaton SM, Burns EM, Kusser K, Randall TD, Haynes L (2004) Age-related defects in CD4 coronary syndromes. Circulation 101:2883–2888. T cell cognate helper function lead to reductions in humoral responses. J Exp Med 17. Waase I, Kayser C, Carlson PJ, Goronzy JJ, Weyand CM (1996) Oligoclonal T cell 200:1613–1622. proliferation in patients with rheumatoid arthritis and their unaffected siblings. 50. King AG, Ozanne BW, Smythe C, Ashworth A (1995) Isolation and characterisation of Arthritis Rheum 39:904–913. a uniquely regulated threonine, tyrosine phosphatase (TYP 1) which inactivates ERK2 18. Clambey ET, van Dyk LF, Kappler JW, Marrack P (2005) Non-malignant clonal and p54jnk. Oncogene 11:2553–2563. expansions of CD8+ memory T cells in aged individuals. Immunol Rev 205:170–189. 51. Yamashita M, et al. (1999) T cell antigen receptor-mediated activation of the Ras/ 19. Haynes L, Easton SM, Burns EM, Randall TD, Swain SL (2003) CD4 T cell memory mitogen-activated protein kinase pathway controls interleukin 4 receptor function derived from young naive cells functions well into old age, but memory generated and type-2 helper T cell differentiation. Proc Natl Acad Sci USA 96:1024–1029. from aged naive cells functions poorly. Proc Natl Acad Sci USA 100:15053–15058. 52. Dong C, et al. (1998) Defective T cell differentiation in the absence of Jnk1. Science 20. Akbar AN, Beverley PC, Salmon M (2004) Will telomere erosion lead to a loss of T-cell 282:2092–2095. memory? Nat Rev Immunol 4:737–743. 53. Rincón M, et al. (1998) Interferon-gamma expression by Th1 effector T cells mediated 21. Del Giudice G, et al. (2006) An MF59-adjuvanted inactivated influenza vaccine by the p38 MAP kinase signaling pathway. EMBO J 17:2817–2829. containing A/Panama/1999 (H3N2) induced broader serological protection against 54. Kuchen S, et al. (2007) Essential role of IL-21 in B cell activation, expansion, and plasma heterovariant influenza virus strain A/Fujian/2002 than a subunit and a split influenza cell generation during CD4+ T cell-B cell collaboration. J Immunol 179:5886–5896. vaccine. Vaccine 24:3063–3065. 55. Avery DT, Bryant VL, Ma CS, de Waal Malefyt R, Tangye SG (2008) IL-21-induced 22. Couch RB, et al. (2007) Safety and immunogenicity of a high dosage trivalent influenza vaccine among elderly subjects. Vaccine 25:7656–7663. isotype switching to IgG and IgA by human naive B cells is differentially regulated by – 23. Oxman MN, et al. (2005) A vaccine to prevent herpes zoster and postherpetic IL-4. J Immunol 181:1767 1779. neuralgia in older adults. N Engl J Med 352:2271–2284. 56. King C (2009) New insights into the differentiation and function of T follicular helper – 24. Lee WW, et al. (2008) Age-dependent signature of metallothionein expression in cells. Nat Rev Immunol 9:757 766. primary CD4 T cell responses is due to sustained zinc signaling. Rejuvenation Res 11: 57. Tresini M, Lorenzini A, Torres C, Cristofalo VJ (2007) Modulation of replicative 1001–1011. senescence of diploid human cells by nuclear ERK signaling. J Biol Chem 282: – 25. Berasi SP, et al. (2006) Inhibition of gluconeogenesis through transcriptional activation 4136 4151. of EGR1 and DUSP4 by AMP-activated kinase. J Biol Chem 281:27167–27177. 58. Wang J, Shen WH, Jin YJ, Brandt-Rauf PW, Yin Y (2007) A molecular link between – 26. Chu Y, Solski PA, Khosravi-Far R, Der CJ, Kelly K (1996) The mitogen-activated protein E2F-1 and the MAPK cascade. J Biol Chem 282:18521 18531. kinase phosphatases PAC1, MKP-1, and MKP-2 have unique substrate specificities and 59. Wang H, et al. (2007) HoxA10 activates transcription of the gene encoding mitogen- reduced activity in vivo toward the ERK2 sevenmaker mutation. J Biol Chem 271: activated protein kinase phosphatase 2 (Mkp2) in myeloid cells. J Biol Chem 282: 6497–6501. 16164–16176. 27. Miller JD, et al. (2008) Human effector and memory CD8+ T cell responses to smallpox 60. Khurana S, et al. (2010) Vaccines with MF59 adjuvant expand the antibody repertoire and yellow fever vaccines. Immunity 28:710–722. to target protective sites of pandemic avian H5N1 influenza virus. Sci Transl Med 2: 28. Akondy RS, et al. (2009) The yellow fever virus vaccine induces a broad and 15ra15. polyfunctional human memory CD8+ T cell response. J Immunol 183:7919–7930. 61. Rinaudo CD, Telford JL, Rappuoli R, Seib KL (2009) Vaccinology in the genome era. 29. He XS, et al. (2006) Cellular immune responses in children and adults receiving J Clin Invest 119:2515–2525. inactivated or live attenuated influenza vaccines. J Virol 80:11756–11766. 62. Baugh LR, Hill AA, Brown EL, Hunter CP (2001) Quantitative analysis of mRNA 30. Frasca D, Blomberg BB (2009) Effects of aging on B cell function. Curr Opin Immunol amplification by in vitro transcription. Nucleic Acids Res 29:E29. 21:425–430. 63. Gentleman R, et al. (2005) Bioinformatics and Computational Biology Solutions Using 31. Cancro MP, et al. (2009) B cells and aging: Molecules and mechanisms. Trends Immunol R and Bioconductor (Springer, Berlin). 30:313–318. 64. Kendziorski CM, Newton MA, Lan H, Gould MN (2003) On parametric empirical Bayes 32. Fazilleau N, Mark L, McHeyzer-Williams LJ, McHeyzer-Williams MG (2009) Follicular methods for comparing multiple groups using replicated gene expression profiles. helper T cells: Lineage and location. Immunity 30:324–335. Stat Med 22:3899–3914. 33. Cicin-Sain L, et al. (2010) Loss of naive T cells and repertoire constriction predict poor 65. Choi H, Ghosh D (2008) A Comparison of Methods for Meta-Analysis of Gene Ex- response to vaccination in old primates. J Immunol 184:6739–6745. pression Data (Wiley, New York).

E888 | www.pnas.org/cgi/doi/10.1073/pnas.1109797109 Yu et al. Downloaded by guest on September 29, 2021