ava i l ab l e a t www.sc i enced i r ec t . com

www.e l sev i e r. com/ l oca te /yc l im

Clinical Immunology (2008) 128, 228–237

The number and function of circulating dendritic cellsmay limit effector memory CD4+ T-cell responses inHIV patients responding to antiretroviral therapy☆

Sonia Fernandez ⁎, Shelley F. Stone, Patricia Price, Martyn A. French

a Department of Clinical Immunology and Immunogenetics, Royal Perth Hospital, Perth, Australiab School of Pathology and Laboratory Science, University of Western Australia, Perth, Australia

Received 22 February 2008; accepted with revision 25 March 2008Available online 29 May 2008

1521-6616/$ – see front matter © 200doi:10.1016/j.clim.2008.03.517

☆ The work was supported by granNational Health and Medical Researcsented (abstract WEPEA047) in part atPathogenesis, Treatment and PreveSydney, NSW, Australia.⁎ Corresponding author. School of

Medicine, University of Western AustRear 50 Murray Street, Perth 6001, Au

E-mail address: sonia.fernandez@u

Abstract Some HIV patients who previously experienced severe immunodeficiency retain lowpathogen-specific T-cell responses despite a virological response to antiretroviral therapy (ART). Toidentify correlates with dysfunction in accessory cell populations, HIV patients were stratified intogroupsmaintaining high or low CD4+ T-cell IFN-γ responses to cytomegalovirus (CMV) over 4–8 yearson ART. Myeloid dendritic cells (mDC), plasmacytoid (p) DC, M-DC8+ cells and monocytes wereenumerated and mRNA of cytokines and activation molecules were quantitated in purifiedsubpopulations. Proportions of pDC were lower (p=0.043) and mDC were higher (p=0.043) in lowresponders. TRAIL receptor 2 (DR5)mRNA levels in pDC (p=0.0008) andmDC (p=0.0062) were lowerin high responders compared to controls. Levels of IL-15 mRNA were higher in mDC from highresponders (p=0.015) and levels of IL-10 mRNA were higher in M-DC8+ cells from low responders(p=0.036). Hence CMV-specific CD4+ T-cell IFN-γ responses may be affected by numbers andfunction of circulating DC.© 2008 Elsevier Inc. All rights reserved.

KEYWORDSAntigen presentation;CD4+ T-cell;Dendritic cell;Effector memoryresponse;HIV

Introduction

Some human immunodeficiency virus (HIV)-infectedpatients who achieve a virological response to combinationantiretroviral therapy (ART) do not recover normal CD4+ T-

8 Elsevier Inc. All rights reserved

t number 404028 from theh Council of Australia. Pre-the 4th IAS Conference on HIVntion (22–25 July 2007) in

Pathology and Laboratoryralia, Level 2, MRF Building,stralia. Fax: 61 8 9224 0204.wa.edu.au (S. Fernandez).

cell responses to the antigens of opportunistic pathogens.This includes CD4+ T-cell responses to cytomegalovirus(CMV) assessed by lymphoproliferation and interferon(IFN)-γ production [1,2]. It is also evident with Candidaand mycobacterial antigens [3,4]. Low effector memoryCD4+ T-cell responses in this situation are associated withvery low nadir CD4+ T-cell counts [5,6], an increasedsusceptibility of peripheral blood mononuclear cells(PBMC) to apoptosis [7] and markers of persistent immunedysfunction including increased serum levels of IgA and IgE[8].

Antigen presenting cell (APC) dysfunction might limitCD4+ T-cell responses. Interleukin (IL)-12, IL-23 and IL-18from APC induce IFN-γ production by and proliferation of T-cells [9,10] and APC produce cytokines that promote the

.

229Dendritic cells and effector memory CD4+ T-cell responses in HIV patients responding to ART

survival of effector T-cells. For example, IL-15 from cells ofthe monocyte/macrophage lineage and blood dendriticcells (DC) [11,12] can enhance the proliferation of PBMC inresponse to mitogen and antigens [13]. Immature DC mayinhibit effector T-cell function by inducing regulatory T-cells [14] or via immunosuppressive mediators such as IL-10[15].

APC in blood include DC, monocytes and B-cells. DCcomprise 1–2% of PBMC, determine the magnitude, polarityand effector function of the T-cell response and primenaïve CD4+ T-helper cells and CD8+ cytotoxic T-cells [16].They are classified as plasmacytoid DC (pDC), myeloid DC

Figure 1 Enumeration of pDC (BDCA-4+), mDC (BDCA-1+), M-DC8+

were distinguished by forward and side scatter. (B) For the enumeaccording to expression of CD19 and CD14, respectively. (C) Expressiorespectively, within the CD19−/CD14− population. The proportionsevents in quadrant 1 (q1) or quadrant 2 (q2), respectively, divided b(D) M-DC8+ cells were identified on the basis of M-DC8 expression. (Eon the basis of CD14 expression.

(mDC) or M-DC8+ cells [17–20]. Numbers of circulating pDCand mDC are reduced in untreated HIV infection [17,18] andpDC are poorly reconstituted during ART [19,20]. The abilityof pDC and mDC from untreated HIV-infected individuals tostimulate allogeneic T-cell proliferation in a 6-day mixedleukocyte reaction is impaired [21] and viral-induced IFN-αproduction from pDC may not return to pre-infection levelswhen patients receive ART [19,20,22]. Persistently low CD4+

T-cell counts may influence this defect [23] and explain whyother studies show that DC from HIV-infected individuals canrespond normally to toll like receptor agonists [24]. PersistentHIV replication may be important as HIV induces pDC to

cells and CD14+ monocytes by flow cytometry. (A) Viable PBMCration of pDC and mDC, B-cells and monocytes were excludedn of BDCA-4 or BDCA-1 indicated the proportion of pDC and mDC,of pDC and mDC in PBMC were calculated using the number ofy the total number of events gated in A and multiplied by 100.) M-DC8+ cells co-expressed CD16. (F) Monocytes were identified

Figure 2 CD4+ T-cell IFN-γ responses to CMV varied in HIVpatients receiving long-term effective ART butwere not correlatedwith CD4+ T-cell counts. (A) CD4+ T-cell responses to CMV wereassessed by ELISPOT. A comparison of the CMV-specific CD4+ T-cellresponses used to select the patient group (prior study) and thepresent study showed that responses were stable over time. Lowand high responder groups were distinguished according to themedian IFN-γ response produced by controls of each respectivestudies [45 spots in theprior study and202 spots in thepresent study(−)]. (B) Low CMV responders had fewer CMV IFN-γ spots than highresponders (pb0.0001) and controls (p=0.0003). (C) CD4+ T-cellIFN-γ responses to CMVwere not associatedwith CD4+ T-cell countsin HIV patients (r=0.212, p=0.399).

230 S. Fernandez et al.

produce the immunosuppressive enzyme indoleamine 2,3-dioxygenase (IDO) which can limit T-cell proliferation [25], aswell as IFN-α which can upregulate the expression of TNF-related apoptosis-inducing ligand (TRAIL) on CD4+ T-cellsresulting in increased apoptosis [26]. These factors may becritical in the pathogenesis of HIV [27].

M-DC8+ cells express CD16 and a cell surface proteinrecognised by the monoclonal antibody, M-DC8 [28]. They arephagocytic, activate antigen-specific T-cells [28] and caninduce T-helper (Th) 1 immune responses by producing IL-12following lipopolysaccharide stimulation or CD40 ligation[29]. The effects of HIV upon M-DC8+ cell numbers andfunction are unknown.

This study examined APC numbers and function in ascenario that represents the optimal outcome for a HIVpatient with advanced immunodeficiency; namely a stablevirological response to ART with improved CD4+ T-cellnumbers. We examined APC directly ex vivo using asensitive technique applicable to the small numbers ofcells that can be purified from individual patients. HenceAPC dysfunction and the susceptibility of APC to apoptosiswere assessed via mRNA of Th1 accessory cytokines, IDO,TRAIL and its receptors (DR4 and DR5). We correlated theAPC data with the restoration of CD4+ T-cell responses toan index antigen (CMV). This is an opportunistic pathogento which most patients and control donors have beenexposed. Data compiled from our laboratory [3 andunpublished data] show that IFN-γ enzyme-linked immu-nosorbent spot-forming cell (ELISPOT) assay responses toCMV correlate with those to Candida antigen (n=60,r=0.81, pb0.0001), so responses to CMV are a marker ofCD4+ effector T-cell function.

Methods

Study population

Male, CMV-seropositive HIV patients were selected on thebasis that they (1) began ARTwith CD4+ T-cell counts below50/μl, (2) had b50 copies HIV RNA/ml plasma for N24 monthsafter N36 months on ART, and (3) had been assessed for CMV-specific CD4+ T-cell responses after a median (range) of 70(12–80) months on ART [30,31]. Individuals with low or highIFN-γ responses to CMV (based on the median response ofhealthy controls) [30] were invited to participate. Studygroups comprised 18 patients and 10 HIV-negative controls(also male and CMV-seropositive). Informed consent wasobtained from all patients and controls and the humanexperimentation guidelines of Royal Perth Hospital werefollowed.

T-cell subsets were quantitated using EDTA-treatedwhole blood stained with CD8-FITC/CD4-PE/CD3-PCy5(Coulter, USA). Plasma HIV RNA levels were assayed bythe Amplicor™ method, version 1.5 (Roche DiagnosticSystems, USA).

Detection of IFN-γ producing T-cells

ELISPOT assays were performed with anti-IFN-γ antibodiespurchased fromMabTech (Sweden) using cryopreserved PBMCas described previously [32]. Cells were also stimulated with

231Dendritic cells and effector memory CD4+ T-cell responses in HIV patients responding to ART

anti-CD3 (Mabtech, Sweden; 10 ng/ml) as a positive control.Spots greater than 10 units in size and 20 units of intensitywere counted using an AID ELISPOT Reader System (AID,Germany). The selective stimulation of CD4+ T-cells by wholeCMV was confirmed using PBMC depleted of CD4+ or CD8+ T-cells using CD4 or CD8 Microbeads (Miltenyi Biotec, Germany)(data not shown).

Enumeration of APC

To enumerate pDC and mDC, 106 freshly isolated PBMC wereincubated with anti-human BDCA-4-PE, BDCA-1-FITC (Milte-nyi Biotec, Germany), CD19-PCy5 and CD14-PCy5 (Coulter,USA) for 10 min on ice. For M-DC8+ cells and monocytes, 106

PBMC were incubated with CD16-PCy5, CD14-ECD (Coulter,USA) and anti-M-DC8 (donated by Dr E. Peter Rieber,Technical University of Dresden, Germany) for 30 min onice, followed by incubation with anti-mouse IgM-FITC(Coulter, USA) for a further 30 min on ice. Analyses wereperformed on a Coulter EPICS-XL flow cytometer (Coulter,USA) within 1 h of staining. 100,000 events per sample wereanalysed and gates were set using appropriate isotypecontrol antibodies (Fig. 1). APC subpopulations are pre-sented as a percentage of total PBMC.

Purification of APC

Populations of mDC (BDCA-1+), pDC (BDCA-4+), and mono-cytes (CD14+) were purified by sequential positive selec-tions on 20×106 freshly isolated PBMC using magnetic beadkits (BDCA-1 Cell Isolation Kit, BDCA-4 Cell Isolation Kit,CD14 Microbeads; Miltenyi Biotec, Germany), according tothe manufacturer's instructions. To increase the purity ofAPC subpopulations, PBMC were first depleted of T-cellsand B-cells using anti-CD3 (7×107 beads/ml) and -CD19(2×107 beads/ml) conjugated Dynabeads (Dynal Biotech,Norway). M-DC8+ were purified after incubation with 10 μlundiluted anti-M-DC8 (Dr E. Peter Rieber, Technical

Table 1 Characteristics of patients and controls at the study tim

Low CMV responders

n 8Age (years) 54 a

(42–67)Current CD4+ T-cells/μl 594

(209–961)Current CD8+ T-cells/μl 798 b

(468–1581)Nadir CD4+ T-cells/μl 8 c

(0–48)Months on ART 103

(38–111)Months HIV RNAb50 copies/ml 77

(37–101)

Note. p value N1.0 unless otherwise indicated (Mann–Whitney test).a Values expressed as median (range).b Differed from controls p≤0.005.c Differed from high responders p=0.05.d NA = not applicable.

University of Dresden, Germany) for 10 min at 4 °C fol-lowed by incubation with 20 μl rat anti-mouse IgM micro-beads (Miltenyi Biotec, Germany) for 15 min at 4 °C. All APCisolations were performed on an OctoMACS SeparationUnit using MS columns (Miltenyi Biotec, Germany).Purified cell subsets for each patient and control werewashed, resuspended in RNA lysis buffer (Qiagen, USA)and stored at −80 °C. Flow cytometric analyses revealedthat the purity of all isolated APC subpopulations was80–98%.

Quantitation of mRNA

Total RNA was extracted from all fractions using RNeasy MiniKit (Qiagen, USA) and reverse transcribed into cDNA usingSensiscript® RT Kit (Qiagen, USA). mRNA levels for β-actin(housekeeping gene), IFN-α, IL-23p19 and IL-18 were deter-mined using quantitative real-time PCR. Each 20 μl reactioncontained 5 μl cDNA, 1.25mM dNTP (Invitrogen, USA), 20 pmolprimer (β-actin, Forward: 5′-GATGACCCAGATCATGTTTGA-3′,Reverse: 5′-GACTCCATGCCCAGGAAGGAA-3′) (IFN-α, Forward:5′-TCCATGAGATGATCCAGCAG-3′, Reverse: 5′-ATTTCTGCTCT-GACAACCTCC-3′) (IL-23p19, Forward: 5′-AGCAGCTCAAG-GATGGCACTCAG, Reverse: 5′-CCCCAAATTTCCCTTCCCATCTA-3′) (IL-18, Forward: 5′-GCTTGAATCTAAATTATCAGT-3′,Reverse: 5′-GAAGATTCAAATTGCATCTTR-3′) (Proligo, Austra-lia), 0.5× SYBR Green fluorochrome (Sigma, USA) and 1.5 unitPlatinum Taq DNA polymerase (Invitrogen, USA) [95 °C,310 seconds (s), 40× (95 °C, 10 s; 62 °C for β-actin, IFN-αand IL-18, 67 °C for IL-23p19, 15 s; 72 °C 25 s)]. mRNA levelsfor IL-12p35, IL-15, IL-10, IDO, TRAIL, DR4 (TRAIL receptor 1),and DR5 (TRAIL receptor 2) were determined using Quanti-tect® sequence specific probes (Qiagen, USA).

Real-time PCR was performed on a Rotorgene™ (Corbett,Australia) or a LightCycler® 480 (Roche Diagnostics, Aus-tralia). Quantitation of specific mRNA was achieved using astandard curve generated from serial 10-fold dilutions ofpooled PCR products. For each sample, the ratio of cytokine

e point

High CMV responders Controls

10 1051 49(43–66) (32–59)652 796(238–1152) (494–1260)1088 b 414(234–1840) (208–735)31 NA d

(0–45)102 NA(77–107)57 NA(31–103)

232 S. Fernandez et al.

mRNA to β-actin mRNA was calculated and expressed asarbitrary units. The lower limit of detection for the genesanalysed was a ratio of 0.00001.

Statistical analysis

Results are presented as median (range) values. Statis-tical significance was assessed by the Mann–Whitney testfor continuous variables or Fisher's Exact test forcategorical values. Correlation coefficients were deter-mined by the Spearman's Rank Correlation test. mRNAanalyses were also tested in multinomial logistic (case–control) regression with CMV-high/low responder groupmembership as outcome and levels of each mRNAmolecule as predictors, using the SPSS statistical package.The outcomes reproduced the findings based on Mann–Whitney tests and therefore the data is not presentedhere. For all tests, pb0.05 was considered to represent asignificant difference.

Figure 3 Low CD4+ T-cell IFN-γ responses to CMV were associatedpatients on long-term effective ART. (A) The proportions of pDC c(r=0.633, p=0.005). (B) Low responders had fewer pDC (BDCA-4respectively). (C) Low responders had a higher proportion of mDC (BDrespectively). (D) The proportions of M-DC8+ cells did not differ betwebetween groups.

Results

CD4+ T-cell IFN-γ responses to CMV were stable overtime and did not correlate with CD4+ T-cell counts

CMV-specific IFN-γ ELISPOTcounts [30,31] were used to dividepatients stably treated with ART into low CMV responders(n=8) and high CMVresponders (n=10). The division was basedon the median response of control donors. The patientscategorized were invited to donate a second sample of bloodfor APC isolation and the re-assessment of CD4+ T-cellresponses. The original ELISPOT data were compared withdata from the new sample. A median (range) period of 32 (23–38) months had elapsed between the studies. IFN-γ responsestoCMVwere stable over this period as 17/18 patients remainedin the same group (Fig. 2A). However responses of all patients(Fig. 2A) and control donors (data not shown) were higher inthe second study. This probably reflects amore potent batch ofCMV antigen.

with low proportions of pDC and high proportions of mDC in HIVorrelated with IFN-γ ELISPOT responses to CMV in HIV patients+) than high responders and controls (p=0.043 and p=0.034,CA-1+) than high responders and controls (p=0.043 and p=0.002,en groups. (E) The proportions of CD14+ monocytes did not differ

Figure 4 Low expression of DR5 mRNA in pDC and mDC from patients with high IFN-γ responses to CMV. (A) High CMV responders hadlower levels of DR5 mRNA than controls in the pDC (p=0.0008) and mDC (p=0.0062) subpopulations. (B) TRAIL mRNA did not differbetween the patient and control groups in any of the APC subpopulations.

233Dendritic cells and effector memory CD4+ T-cell responses in HIV patients responding to ART

At the time of APC isolation, the low and high responderpatient groups were matched for current CD4+ (p=0.408) andCD8+ (p=0.274) T-cell counts and time on ART (p=0.633;Table 1). Nadir CD4+ T-cell counts were marginally lower inthe low responder group [8 (0–48) vs 31 (0–45) cells/μl;p=0.054]. Low responders had significantly fewer CMV IFNγspots than high responders [47 (19–76) vs 372 (203–850)spot-forming cells; pb0.0001] and controls [47 (19–76) vs202 (63–1106); p=0.0003] (Fig. 2B).

CD4+ T-cell IFN-γ responses to CMV were not correlatedwith CD4+ T-cell counts (Spearman's Rank Correlation test,r=0.212, p=0.399; Fig. 2C) or percentage (r=0.025, p=0.922;data not shown) in HIV patients. CD4+ T-cell IFN-γ responses toCMVwere also not correlated with CD4+ T-cell count (r=0.188,p=0.607) or percentage (r=−0.425, p=0.218) in the healthycontrol group.

Low CD4+ T-cell IFN-γ responses to CMV wereassociated with low proportions of pDC and highproportions of mDC

When all HIV patients were assessed together, their propor-tions of pDC were positively correlated with IFN-γ ELISPOTresponses to CMV (r=0.633, p=0.005; Fig. 3A). A similartrend was observed in controls, but this did not reachsignificance (r=0.474, p=0.166). When HIV patients weredivided into low and high responder groups (Fig. 3B), lowresponders had significantly fewer pDC than high respondersand controls (p=0.043 and p=0.034, respectively). Theproportions of pDC did not differ between high respondersand controls (p=0.165).

The proportions of mDC were not significantly correlatedwith IFN-γ responses to CMV in patients (r=−0.266,p=0.286) or controls (r =0.043, p=0.918) (data notshown). When the patients were divided into low and highresponder groups (Fig. 3C), low responders had a signifi-cantly higher proportion of mDC than high responders andcontrols (p=0.043 and p=0.002, respectively). Proportions

of mDC were similar in high responders and controls(p=0.280).

The proportions of M-DC8+ cells were not correlated withIFN-γ responses to CMV in HIV patients (r=0.131, p=0.604;data not shown) or controls (r=0.219, p=0.537; data notshown), nor did they differ between the low and highresponder groups (p=0.515), or between the low or highresponder groups and controls (p=0.203 and p=0.529,respectively; Fig. 3D). Similarly, the proportions of CD14+

monocytes were not correlated with IFN-γ responses to CMVin HIV patients (r=0.220, p=0.380) or controls (r=0.358,p=0.313) (data not shown). Proportions were similar acrossall groups (p=0.408–1.0; Fig. 3E).

Expression of DR5 mRNA was low in pDC and mDCfrom patients with high IFN-γ responses to CMV

TRAIL, DR4 and DR5 mRNA were quantitated in each APCsubpopulation to assess susceptibility to apoptosis. Highresponders had significantly lower levels of DR5 mRNA in pDC(p=0.0008) and mDC (p=0.0062) and marginally lower levelsof DR5 mRNA in M-DC8+ cells (p=0.065) than controls(Fig. 4A). Importantly, levels of DR5 mRNA were marginallyhigher in pDC from low responders than those from highresponders (113 (27–258) vs 68 (27–112), respectively;p=0.09, Fig. 4E). TRAIL mRNA was detectable in all APCsubpopulations, with no differences between the patient andcontrol groups (Fig. 4B). DR4 mRNA was generally notdetectable (data not shown).

Levels of mRNA for Th1 accessory cytokines did notcorrelate with IFN-γ responses to CMV

IFN-α mRNA was detected in all four APC subpopulations.IFN-αmRNA levels in pDC were lower in high responders thancontrols (2293 (0–3628) vs 2776 (2464–4981), respectively;p=0.005) and intermediate in low responders (Fig. 5A).IFN-αmRNA levels in mDC, M-DC8+ cells and monocytes were

234 S. Fernandez et al.

similar in low and high responders and controls (Fig. 5A). Lowand high responders had lower levels of IL-12p35 mRNA inmonocytes than controls (p=0.068 and p=0.015; Fig. 5B).Otherwise, there were no significant differences in levels ofIL-12p35, IL-23p19 and IL-18 mRNA between patient groupsand controls (Figs. 5B–D).

IL-15 mRNA levels were elevated in mDC frompatients with high IFN-γ responses to CMV

mDC from the high responder group had higher levels of IL-15mRNA than the low responders and control donors (p=0.015 andp=0.018, respectively; Fig. 5E). IL-15 mRNA levels in pDC didnot differ between low and high responders (detectable in 5/8vs 8/10 patients, respectively). IL-15 mRNA levels in M-DC8+

cells and monocytes were similar in all groups.

IL-10 mRNA was more frequently detectable inM-DC8+ cells from HIV patients with low IFN-γresponses to CMV

Two potentially immunosuppressive molecules, IDO and IL-10,were assessed in the APC subsets. IDO mRNA was detectablein all APC subsets from all patients and controls, but levelswere similar in all groups (Fig. 5F). IL-10 mRNA wasdetectable in M-DC8+ cells from 13/18 patients and 5/10controls. Proportionally more patients with low IFN-γresponses to CMV had detectable IL-10 mRNA (8/8 lowresponders vs 5/10 high responders and 8/8 low responders vs5/10 controls, p=0.036; Fig. 5G). IL-10 mRNA levels in pDC,mDC and monocytes were similar in all groups.

Discussion

Improvements in the function of pathogen-specific CD4+ T-cells are variable in HIV patients with a stable virologicalresponse to ART and increased CD4+ T-cell counts [1–4].Although CMV-specific CD4+ T-cell responses may be used topredict susceptibility to CMV disease [33,34], here we usedthe response as a marker of effector memory CD4+ T-cellfunction because they correlated with responses to anotheropportunistic pathogen, Candida albicans. We correlatedCMV-specific CD4+ T-cell responses with the proportions andfunctional profile of APC subsets in HIV patients with very lownadir CD4+ T-cell counts responding to ART.

CMV-specific CD4+ T-cell responses assessed after a medianof 102months on ARTwere similar to thosemeasured 32monthsearlier. We had previously described a slow rise in CMV-specificCD4+ T-cell responses on ARTover 3 to 5 years [2,3] and here weshowthat the responses are stable over 4–8 years on treatment.CMV-specific CD4+ T-cell IFN-γ responses did not correlate withCD4+ T-cell count or percentage at the timeof study (Fig. 2C), as

Figure 5 APC function assessed via cytokine mRNA. (A) IFN-α mRcontrols (p=0.005). (B) IL-12p35 mRNA levels in monocytes were lowThere were no significant differences in the levels of IL-23p19 and ILsubpopulations. (E) IL-15 mRNA levels in the mDC subpopulationresponders (p=0.015) and controls (p=0.018). (F) Levels of IDO mRmRNAwas more frequently detectable in M-DC8+ cells from HIV patiehigh responders, p=0.036).

has been shown in relation to responses to CMV [2,3] and HIV[6]. However, antigen-specific CD4+ T-cell responses can remainlow in patients with nadir CD4+ T-cell counts below 50 cells/μl[5,6]. Here nadir CD4+ T-cell counts were slightly lower in theHIV patients with low IFN-γ responses to CMV, but all patientshad nadir CD4+ T-cell counts b50 cells/μl.

HIV patients with low CD4+ T-cell IFN-γ responses to CMVhad a significantly lower proportion of pDC than patientswith high CMV responses. pDC produce IFN-α in response toviral stimuli [35,36], which promotes a Th1 cytokineresponse and primes T-cells to produce IFN-γ [35,37,38].Here IFN-α mRNA levels were significantly lower in pDC frompatients with high IFN-γ responses to CMV. Hence the lownumber of pDC may limit activation of Th1 cells, but levels ofIFN-α per pDC may not be directly limiting.

In HIV infection, pDC stimulate the expression of TRAIL byCD4+ T-cells, which promotes loss of CD4+ T-cells throughapoptosis [39]. In health, TRAIL plays a role in regulating theclonal expansion of some memory CD8+ T-cells (‘helplessmemory CD8+ T-cells’) after restimulation [40], but it isunclear if memory CD4+ T-cells are regulated in this way.Although levels of TRAIL mRNA were similar in all APCsubpopulations, expression of DR5 (TRAIL receptor 2) mRNAwas lower in pDC,mDC andM-DC8+ cells fromHIV patientswithhigh IFN-γ responses to CMV when compared to healthycontrols. Levels were intermediate in patients with low IFN-γresponses to CMV. Hence low DR5 expression may protect pDCfrom apoptosis. This observation is in accord with evidencethat mDC and pDC from the lymphoid tissues of monkeys withAIDS are prone to spontaneous death in culture, so apoptosismay contribute to their loss in disease [41].

Patients with low CMV-specific CD4+ T-cell IFN-γ responsesalso had ahigher proportion ofmDC thanpatientswith highCMVresponses. mDC are considered to be the principal producers ofIL-12 [42,43]. IL-12 deficiency is a complication of untreatedHIV infection and has been linked to poor IFN-γ production by T-cells and an enhanced susceptibility to opportunistic infections[44,45]. Here low or undetectable IL-12p25mRNA levels inmDCwere not associated with low IFN-γ production in response toCMV. IL-12 production by mDC has been demonstrated afterstimulation with bacterial products or via CD40L [46–48] andmay be less important in responses to CMV.

Immature DC can mediate antigen-specific inhibition ofeffector T-cell function [14] or prime T-cells to produce IL-10[15]. Therefore, the increased proportions of mDC in thepatients with low IFN-γ responses to CMV may have animmunosuppressive effect. Patients with low IFN-γ responsesto CMV were more likely to have detectable IL-10 mRNAlevels in the M-DC8+ population, but not in CD14+ monocytes.Levels in other subpopulations were very low.

As IL-23 enhances memory rather than naïve T-cellproliferation, IL-23 may be more important than IL-12 in thegeneration of T-cell responses in HIV patients on ART [10]. Wehave demonstrated reduced expression of IL-23p19 mRNA in

NA levels in pDC from high CMV responders were lower than iner in the high responder group than controls (p=0.015). (C and D)-18 mRNA between patient groups and controls in any of the APCwere elevated in the high responder group compared to lowNA were similar across all patient and control groups. (G) IL-10nts with low IFN-γ responses to CMV (8/8 low responders vs 5/10

235Dendritic cells and effector memory CD4+ T-cell responses in HIV patients responding to ART

adherent cells and unfractionated PBMC from HIV patients.This correlated with low levels of IFN-γ in unstimulated PBMCand purified T-cells [49]. Here, IL-23p19 mRNAwas detectable

in all APC subsets from HIV patients, but levels were similar tocontrols anddid not correlatewith IFN-γ responses toCMV. Thepatients assessed previously had detectable HIV-1 RNA and a

236 S. Fernandez et al.

shorter duration of treatment. This may explain their lowlevels of both IFNγ and IL-23 mRNA (relative to controls),whereas neither parameter was depressed in the combinedpatient cohort here.

High CD4+ T-cell IFN-γ responses to CMV in HIV patientswereassociated with elevated IL-15 mRNA levels in mDC. A recentstudy demonstrated that IL-15 administered to SIV-infectedrhesusmacaques (whowerevirologically suppressed) increasedin vivo proliferation of both CD4+ and CD8+ effector memory T-cells with little effect on naïveor centralmemoryT-cell subsets[50]. Furthermore, IL-15 can enhance activation and effectorfunction of bothCD4+ andCD8+ humaneffectormemoryT-cells,assessed by production of IFN-γ and TNF-α after anti-CD3stimulation [51]. These authors implicated IL-15 in theinhibition of Fas-induced apoptosis in HIV-infected individuals.Hence our data suggest a role for IL-15 in the survival and IFN-γproduction of effector memory CD4+ T-cells in HIV patientsreceiving ARTwith high CD4+ T-cell IFN-γ responses to CMV.

In summary, we suggest that impaired CMV-specific CD4+

T-cell IFN-γ responses in HIV patients responding to ART mayreflect deficiency of pDC, possibly arising through TRAIL/DR5-mediated apoptosis. Further investigation of the role ofTRAIL/DR5 in immune dysfunction is warranted because itmay be amenable to immunomodulatory therapy. No clearchanges in cytokine production from APC correlated withlow CMV-specific IFN-γ responses, but roles for depressed IL-15 in survival of effector memory CD4+ T-cells and for IL-10in suppressing IFN-γ production warrant furtherinvestigation.

Acknowledgments

The authors thank Mr Steven Roberts and Dr Silvia Lee fortheir expert technical assistance and Dr Elizabeth McKinnonfor her statistical advice. This is manuscript number 2007-36for the Department of Clinical Immunology and Immunoge-netics, Royal Perth Hospital.

References

[1] G. Gerna, G. Piccinini, E. Genini, E. Percivalle, M. Zavattoni, D.Lilleri, L. Testa, G. Comolli, R. Maserati, F. Baldanti, R. Maccario,A.D. Monforte, M.G. Revello, Declining levels of rescuedlymphoproliferative response to human cytomegalovirus (HCMV)inAIDS patientswith orwithoutHCMVdisease following long-termHAART, J. Acquir. Immune Defic. Syndr. 28 (2001) 320–331.

[2] N.M. Keane, P. Price, S. Lee, C.A. Almeida, S.F. Stone, I. James,M.A. French, Restoration of CD4 T-cell responses to cytomega-lovirus is short-lived in severely immunodeficient HIV-infectedpatients responding to highly active antiretroviral therapy, HIVMed. 5 (2004) 407–414.

[3] K. Burgess, P. Price, I.R. James, S.F. Stone, N.M. Keane, A.Y. Lim,J.R. Warmington, M.A. French, Interferon-gamma responses toCandida recover slowly or remain low in immunodeficient HIVpatients responding to ART, J. Clin. Immunol. 26 (2006) 160–167.

[4] R. Sutherland, H. Yang, T.J. Scriba, B. Ondondo, N. Robinson, C.Conlon, A. Suttill, H. McShane, S. Fidler, A. McMichael, L.Dorrell, Impaired IFN-gamma-secreting capacity in mycobac-terial antigen-specific CD4 Tcells during chronic HIV-1 infectiondespite long-term HAART, Aids 20 (2006) 821–829.

[5] C.G. Lange, H. Valdez, K. Medvik, R. Asaad, M.M. Lederman,CD4+ T-lymphocyte nadir and the effect of highly activeantiretroviral therapy on phenotypic and functional immune

restoration in HIV-1 infection, Clin. Immunol. 102 (2002)154–161.

[6] M.A. Siddique, K.E. Hartman, E. Dragileva, M. Dondero, T.Gebretsadik, A. Shintani, L. Peiperl, F. Valentine, S.A. Kalams,Low CD4+ Tcell nadir is an independent predictor of lower HIV-specific immune responses in chronically HIV-1-infected sub-jects receiving highly active antiretroviral therapy, J. Infect.Dis. 194 (2006) 661–665.

[7] A. Weinberg, R.D. Jesser, C.L. Edelstein, J.R. Bill, D.A. Wohl,Excess apoptosis of mononuclear cells contributes to thedepressed cytomegalovirus-specific immunity in HIV-infectedpatients on HAART, Virology 330 (2004) 313–321.

[8] M. French, N. Keane, E. McKinnon, S. Phung, P. Price,Susceptibility to opportunistic infections in HIV-infectedpatients with increased CD4 T-cell counts on antiretroviraltherapy may be predicted by markers of dysfunctional effectormemory CD4 T cells and B cells, HIV Med. 8 (2007) 148–155.

[9] C.A. Dinarello, IL-18: a TH1-inducing, proinflammatory cytokineand newmember of the IL-1 family, J. Allergy Clin. Immunol. 103(1999) 11–24.

[10] C.S. Lankford, D.M. Frucht, A unique role for IL-23 in promotingcellular immunity, J. Leukoc. Biol. 73 (2003) 49–56.

[11] W.E. Carson, M.E. Ross, R.A. Baiocchi, M.J. Marien, N. Boiani,K. Grabstein, M.A. Caligiuri, Endogenous production of inter-leukin 15 by activated human monocytes is critical for optimalproduction of interferon-gamma by natural killer cells in vitro,J. Clin. Invest. 96 (1995) 2578–2582.

[12] H. Jonuleit, K. Wiedemann, G. Muller, J. Degwert, U. Hoppe, J.Knop, A.H. Enk, Induction of IL-15 messenger RNA and proteinin human blood-derived dendritic cells: a role for IL-15 inattraction of T cells, J. Immunol. 158 (1997) 2610–2615.

[13] J. Chehimi, J.D. Marshall, O. Salvucci, I. Frank, S. Chehimi, S.Kawecki, D. Bacheller, S. Rifat, S. Chouaib, IL-15 enhances immunefunctions during HIV infection, J. Immunol. 158 (1997) 5978–5987.

[14] M.V. Dhodapkar, R.M. Steinman, J. Krasovsky, C. Munz, N.Bhardwaj, Antigen-specific inhibition of effector Tcell functionin humans after injection of immature dendritic cells, J. Exp.Med. 193 (2001) 233–238.

[15] H. Jonuleit, E. Schmitt, G. Schuler, J. Knop, A.H. Enk, Inductionof interleukin 10-producing, nonproliferating CD4(+) T cellswith regulatory properties by repetitive stimulation withallogeneic immature human dendritic cells, J. Exp. Med. 192(2000) 1213–1222.

[16] J. Banchereau, F. Briere, C. Caux, J. Davoust, S. Lebecque, Y.J.Liu, B. Pulendran, K. Palucka, Immunobiology of dendritic cells,Annu. Rev. Immunol. 18 (2000) 767–811.

[17] H. Donaghy, A. Pozniak, B. Gazzard, N. Qazi, J. Gilmour, F.Gotch, S. Patterson, Loss of blood CD11c+ myeloid and CD11c−

plasmacytoid dendritic cells in patients with HIV-1 infectioncorrelates with HIV-1 RNA virus load, Blood 13 (2001) 759–766.

[18] J. Pacanowski, S. Kahi,M. Baillet, P. Lebon,C.Deveau,C.Goujard,L. Meyer, E. Oksenhendler, M. Sinet, A. Hosmalin, Reduced bloodCD123+ (lymphoid) and CD11c+ (myeloid) dendritic cell numbers inprimary HIV-1 infection, Blood 98 (2001) 3016–3021.

[19] J.Chehimi, D.E. Campbell, L. Azzoni, D. Bacheller, E. Papasavvas,G. Jerandi, K. Mounzer, J. Kostman, G. Trinchieri, L.J. Montaner,Persistent decreases in blood plasmacytoid dendritic cell numberand function despite effective highly active antiretroviraltherapy and increased blood myeloid dendritic cells in HIV-infected individuals, J. Immunol. 168 (2002) 4796–4801.

[20] J.S. Finke, M. Shodell, K. Shah, F.P. Siegal, R.M. Steinman,Dendritic cell numbers in the blood of HIV-1 infected patientsbefore and after changes in antiretroviral therapy, J. Clin.Immunol. 24 (2004) 647–652.

[21] H. Donaghy, B. Gazzard, F. Gotch, S. Patterson, Dysfunction andinfection of freshly isolated blood myeloid and plasmacytoiddendritic cells in patients infected with HIV-1, Blood 101 (2003)4505–4511.

237Dendritic cells and effector memory CD4+ T-cell responses in HIV patients responding to ART

[22] M. Almeida, M. Cordero, J. Almeida, A. Orfao, Persistentabnormalities in peripheral blood dendritic cells andmonocytesfrom HIV-1-positive patients after 1 year of antiretroviraltherapy, J. Acquir. Immune Defic. Syndr. 41 (2006) 405–415.

[23] S.M. Hsieh, S.C. Pan, C.C. Hung, M.Y. Chen, S.C. Chang,Differential impact of late-stage HIV-1 infection on in vitro andin vivo maturation of myeloid dendritic cells, J. Acquir. ImmuneDefic. Syndr. 33 (2003) 413–419.

[24] J.A. Martinson, A.R. Tenorio, C.J. Montoya, L. Al-Harthi, C.N.Gichinga, A.M. Krieg, L.L. Baum, A.L. Landay, Impact of class A,B and C CpG-oligodeoxynucleotides on in vitro activation ofinnate immune cells in human immunodeficiency virus-1infected individuals, Immunology 120 (2007) 526–535.

[25] A. Boasso, J.P. Herbeuval, A.W. Hardy, S.A. Anderson, M.J.Dolan, D. Fuchs, G.M. Shearer, HIV inhibits CD4+ T-cellproliferation by inducing indoleamine 2,3-dioxygenase inplasmacytoid dendritic cells, Blood 109 (2007) 3351–3359.

[26] J.P. Herbeuval, A.W. Hardy, A. Boasso, S.A. Anderson, M.J.Dolan, M. Dy, G.M. Shearer, Regulation of TNF-related apopto-sis-inducing ligand on primary CD4+ Tcells by HIV-1: role of type IIFN-producing plasmacytoid dendritic cells, Proc. Natl. Acad.Sci. U. S. A. 102 (2005) 13974–13979.

[27] A. Boasso, G. Shearer, Chronic innate immune activation as a causeof HIV-1 immunopathogenesis, Clin. Immunol. 126 (2008) 235–242.

[28] K. Schakel, E. Mayer, C. Federle, M. Schmitz, G. Riethmuller, E.P.Rieber, A novel dendritic cell population in human blood: one-step immunomagnetic isolation by a specific mAb (M-DC8) and invitro priming of cytotoxic T lymphocytes, Eur. J. Immunol. 28(1998) 4084–4093.

[29] A. de Baey, I. Mende, G. Riethmueller, P.A. Baeuerle, Phenotypeand function of human dendritic cells derived from M-DC8(+)monocytes, Eur. J. Immunol. 31 (2001) 1646–1655.

[30] S.F. Stone, P. Price, N. Khan, P.A. Moss, M.A. French, HIVpatients on antiretroviral therapy have high frequencies of CD8T cells specific for Immediate Early protein-1 of cytomegalo-virus, AIDS 19 (2005) 555–562.

[31] S. Fernandez, P. Price, E.J. McKinnon, R.C. Nolan, M.A. French,Low CD4+ T-cell counts in HIV patients receiving effectiveantiretroviral therapy are associated with CD4+ T-cell activa-tion and senescence but not with lower effector memory T-cellfunction, Clin. Immunol. 120 (2006) 163–170.

[32] N.M. Keane, P. Price, S.F. Stone, M. John, R.J. Murray, M.A.French, Assessment of immune function by lymphoproliferationunderestimates lymphocyte functional capacity in HIV patientstreated with highly active antiretroviral therapy, AIDS Res.Hum. Retrovir. 16 (2000) 1991–1996.

[33] S.C. Johnson, C.A. Benson, D.W. Johnson, A. Weinberg,Recurrences of cytomegalovirus retinitis in a human immuno-deficiency virus-infected patient, despite potent antiretroviraltherapy and apparent immune reconstitution, Clin. Infect. Dis.32 (2001) 815–819.

[34] K.V. Komanduri, J. Feinberg, R.K. Hutchins, R.D. Frame, D.K.Schmidt, M.N. Viswanathan, J.P. Lalezari, J.M. McCune, Loss ofcytomegalovirus-specific CD4+ Tcell responses in human immuno-deficiency virus type 1-infected patients with high CD4+ T cellcounts and recurrent retinitis, J. Infect. Dis. 183 (2001) 1285–1289.

[35] M. Cella, D. Jarrossay, F. Facchetti, O. Alebardi, H. Nakajima,A. Lanzavecchia, M. Colonna, Plasmacytoid monocytes migrateto inflamed lymph nodes and produce large amounts of type Iinterferon, Nat. Med. 5 (1999) 919–923.

[36] F.P. Siegal, N. Kadowaki, M. Shodell, P.A. Fitzgerald-Bocarsly, K.Shah, S. Ho, S. Antonenko, Y.J. Liu, The nature of the principal

type 1 interferon-producing cells in human blood, Science 284(1999) 1835–1837.

[37] N. Kadowaki, S. Antonenko, J.Y. Lau, Y.J. Liu, Naturalinterferon alpha/beta-producing cells link innate and adaptiveimmunity, J. Exp. Med. 192 (2000) 219–226.

[38] S. Pichyangkul, T.P. Endy, S. Kalayanarooj, A. Nisalak, K.Yongvanitchit, S. Green, A.L. Rothman, F.A. Ennis, D.H. Libraty,A blunted blood plasmacytoid dendritic cell response to anacute systemic viral infection is associated with increaseddisease severity, J. Immunol. 171 (2003) 5571–5578.

[39] J.P. Herbeuval, G.M. Shearer, HIV-1 immunopathogenesis: howgood interferon turns bad, Clin. Immunol. 123 (2007) 121–128.

[40] E.M. Janssen, N.M. Droin, E.E. Lemmens, M.J. Pinkoski, S.J.Bensinger, B.D. Ehst, T.S. Griffith, D.R. Green, S.P. Schoenber-ger, CD4+ T-cell help controls CD8+ T-cell memory via TRAIL-mediated activation-induced cell death, Nature 434 (2005)88–93.

[41] K.N. Brown, A. Trichel, S.M. Barratt-Boyes, Parallel loss ofmyeloidandplasmacytoiddendritic cells frombloodand lymphoid tissue insimian AIDS, J. Immunol. 178 (2007) 6958–6967.

[42] M.C. Rissoan, V. Soumelis, N. Kadowaki, G. Grouard, F. Briere, R.de Waal Malefyt, Y.J. Liu, Reciprocal control of T helper cell anddendritic cell differentiation, Science 283 (1999) 1183–1186.

[43] H. Matsuda, T. Suda, H. Hashizume, K. Yokomura, K. Asada, K.Suzuki, K. Chida, H. Nakamura, Alteration of balance betweenmyeloid dendritic cells and plasmacytoid dendritic cells inperipheral blood of patients with asthma, Am. J. Respir. Crit.Care Med. 166 (2002) 1050–1054.

[44] C. Chougnet, T.A. Wynn, M. Clerici, A.L. Landay, H.A. Kessler, J.Rusnak, G.P. Melcher, A. Sher, G.M. Shearer, Molecular analysis ofdecreased interleukin-12 production in persons infected withhuman immunodeficiency virus, J. Infect. Dis. 174 (1996) 46–53.

[45] J.D. Marshall, J. Chehimi, G. Gri, J.R. Kostman, L.J. Montaner,G. Trinchieri, The interleukin-12-mediated pathway of immuneevents is dysfunctional in human immunodeficiency virus-infected individuals, Blood 94 (1999) 1003–1011.

[46] S.E. Macatonia, N.A. Hosken, M. Litton, P. Vieira, C.S. Hsieh, J.A.Culpepper, M. Wysocka, G. Trinchieri, K.M. Murphy, A. O'Garra,Dendritic cells produce IL-12 and direct the development of Th1cells from naive CD4+ Tcells, J. Immunol. 154 (1995) 5071–5079.

[47] M. Cella, D. Scheidegger, K. Palmer-Lehmann, P. Lane, A.Lanzavecchia, G. Alber, Ligation of CD40 on dendritic cellstriggers production of high levels of interleukin-12 and enhancesT cell stimulatory capacity: T–T help via APC activation, J. Exp.Med. 184 (1996) 747–752.

[48] D. Jarrossay, G. Napolitani, M. Colonna, F. Sallusto, A.Lanzavecchia, Specialization and complementarity in microbialmolecule recognition by human myeloid and plasmacytoiddendritic cells, Eur. J. Immunol. 31 (2001) 3388–3393.

[49] S. Lee, M.A. French, P. Price, IL-23 and IFN-gamma deficiency inimmunodeficient HIV patients who achieved a long-termincrease in CD4 T-cell counts on highly active antiretroviraltherapy, Aids 18 (2004) 1337–1340.

[50] L.J. Picker, E.F. Reed-Inderbitzin, S.I. Hagen, J.B. Edgar, S.G.Hansen, A. Legasse, S. Planer, M. Piatak Jr., J.D. Lifson, V.C.Maino, M.K. Axthelm, F. Villinger, IL-15 induces CD4 effectormemory T cell production and tissue emigration in nonhumanprimates, J. Clin. Invest. 116 (2006) 1514–1524.

[51] Y.M. Mueller, V. Makar, P.M. Bojczuk, J. Witek, P.D. Katsikis, IL-15enhances the function and inhibits CD95/Fas-induced apoptosisof human CD4+ and CD8+ effector-memory Tcells, Int. Immunol.15 (2003) 49–58.