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/ loca te /yc l im

Low CD4+ T-cell counts in HIV patients receivingeffective antiretroviral therapy are associated withCD4+ T-cell activation and senescence but not withlower effector memory T-cell function

Sonia Fernandez a,b, Patricia Price a,b, Elizabeth J. McKinnon c,Richard C. Nolan a, Martyn A. French a,b,*

a Department of Clinical Immunology and Biochemical Genetics, Royal Perth Hospital, Perth, Australiab School of Surgery and Pathology, University of Western Australia, Perth, Australiac Centre for Clinical Immunology and Biomedical Statistics, Murdoch University, Perth, Australia

Received 23 January 2006; accepted with revision 20 April 2006Available online 9 June 2006

1521-6616/$ — see front matter D 200doi:10.1016/j.clim.2006.04.570

* Corresponding author. Clinical ImGenetics, Royal Perth Hospital, WellinAustralia. Fax: +1 61 8 9224 2920.

E-mail address: Martyn.French@hea

KEYWORDSAntiretroviral therapy;HIV-1;Immune activation;Interferon-g;Senescence;T-cell responses

Abstract The adverse effects of immune activation on CD4+ T-cell recovery and therelationship between CD4+ T-cell counts and effector T-cell function were examined in HIV-1patients receiving long-term effective ART. Patients with nadir CD4+ T-cell counts b100/Al, N 12months on ARTand N6 months with b50 HIV RNA copies/ml were stratified by current CD4+ T-cellcounts and patients from the lowest (n = 15) and highest (n = 12) tertiles were studied. Weassessed proliferation (Ki67), activation (HLA-DR, CD38) and replicative senescence (CD57) byflow cytometry and CD4+ T-cell responses to CMV by IFN-g ELISpot. Proportions of CD4+ T-cellsexpressing HLA-DR or CD57 were strong univariate predictors of total ( P = 0.0002 and P = 0.002)and naive ( P b 0.0001 and P b 0.0001, respectively) CD4+ T-cell counts, suggesting that CD4+ T-cell activation drives the depletion of naive CD4+ T-cells. This was clearest in patients with asmall/undetectable thymus. IFN-g responses to CMV were similar in patients with low or highCD4+ T-cell counts.D 2006 Elsevier Inc. All rights reserved.

6 Elsevier Inc. All rights reserve

munology and Biochemicalgton Street, Perth WA 6000,

lth.wa.gov.au (M.A. French).

Introduction

Most HIV-1-infected patients treated with combination antire-troviral therapy (ART) experience a decrease in plasma HIV RNAlevel followed by an increase in CD4+ T-cell numbers. However,8—17% of patients have poor recovery of CD4+ T-cells despite

Clinical Immunology (2006) 120, 163—170

d.

S. Fernandez et al.164

optimal viral suppression [1], with the highest risk in patientswith nadir CD4+ T-cell counts below 100/Al [2,3]. This maypotentiate opportunistic infections [4]. Poor recovery of CD4+ T-cells in HIV patients with a virological response to ART mayreflect low production of naive T-cells by thymic and/or extra-thymic pathways [5,6]. Persistent immune activation may alsoplay a critical role, because poor recovery of CD4+ T-cell countsin HIV patients on ART correlates with increased expression ofCD38 and HLA-DR on T-cells [7—9] and increased CD4+ T-cellturnover [10].

Mechanisms linking persistent immune activation withpoor CD4+ T-cell recovery may include increased suscepti-bility of T-cells to apoptosis [11] and/or increased turnoverof T-cells leading to the accumulation of senescent T-cells[10,12]. Senescent T-cells are effector memory T-cells at theend of their replicative life. They are unable to proliferate[13], but may retain the capacity to secrete interferon-g(IFN-g) [14]. They accumulate in diseases characterized bychronic antigen exposure and immune activation, includingHIV infection, and may be defined by the absence of surfaceCD28 and/or expression of CD57 [13,15].

Here, we address the relationship between the recovery ofnaive and memory CD4+ T-cell numbers and persistent immuneactivation, CD4+ T-cell senescence and effector CD4+ T-cellfunction in two groups of previously immunodeficient HIV-1-infected patients distinguished by their CD4+ T-cell gainsfollowing a stable virological response to ART. Patients werealso stratified according to thymus size to determine the effectof immune activation and CD4+ T-cell senescence on CD4+ T-cellrecovery in patients with a small or undetectable thymus.

Patients and methods

Study population

All adult HIV-1-infected patients (n = 78) with a nadir CD4+ T-cell count below 100/Al after over 12 months on ARTand withundetectable plasma HIV RNA (b50 copies/ml) for at least 6months prior to study were identified from a cohort of over

Table 1 Characteristics of groups of patients with low or high C

Low CD4+ T-ce

n 15Male:female 14:1Age (years) 49 (30—65)b

Nadir CD4+ T-cells/Al 38 (0—84)Current CD4+ T-cells/Al 195 (120—405)Current CD8+ T-cells/Al 988 (336—2772CD4+CD45RA+CD62L+ (cells/Al) 24 (1—118)CD8+CD45RA+CD62L+ (cells/Al) 162 (47—416)CD4+CD45RO+ (cells/Al) 142 (97—267)CD8+CD45RO+ (cells/Al) 382 (205—776)Months on ART 48 (12—82)Months to N400 CD4+ T-cells/Al —Months to b50 HIV RNA copies/ml 7 (1—49)a P value obtained by Wilcoxon rank sum test.b Values expressed as median (range).c Two patients had N300 CD4+ T-cells/Al at the study time point onlyd One patient had b400 CD4+ T-cells/Al at the study time point only.e Compared with duration of ART in the low CD4+ T-cell group.

500 patients in the HIV patient database of the Departmentof Clinical Immunology and Biochemical Genetics, RoyalPerth Hospital (Perth, Australia). ART consisted of at leastthree antiretroviral drugs including a non-nucleoside reversetranscriptase inhibitor (nNRTI) or protease inhibitor (PI).Patients were stratified into approximate tertiles accordingto current CD4+ T-cell counts and percentages and the lowestand highest tertiles were compared to achieve the greatestdiscrimination. Patients from the lowest (n = 15) and highest(n = 12) tertiles were recruited for the study based on theirwillingness to participate (see Table 1). Patients in thelowest tertile consistently had b18% CD4+ T-cells and b300CD4+ T-cells/Al and were defined as having poor CD4+ T-cellrecovery, whereas patients in the highest tertile consistentlyhad N23% CD4+ T-cells and N400 CD4+ T-cells/Al and weredefined as having good CD4+ T-cell recovery. All patientswere CMV seropositive. Informed consent was obtained andhuman experimentation guidelines of Royal Perth Hospitaland the University of Western Australia were followed.

T-cell subsets

T-cell subsets were quantitated by flow cytometry usingEDTA-treated whole blood. CD4+ and CD8+ T-cells werequantitated by staining with CYTO-STAT triCHROMEk (CD8-FITC/CD4-RD1/CD3-PCy5). Naive T-cells (CD4/CD8+,CD45RA+, CD62L+) were quantitated using CD45RA-RD1,CD62L-FITC and either CD4 or CD8 conjugated with PCy5.Memory T-cells (CD4/CD8+, CD45RO+) were assessed usingCD3-PCy5 and CD4 or CD8-RD1 and CD45RO-FITC. All anti-bodies were purchased from Coulter (Miami, FL, USA).Analyses were performed on a Coulter EPICS-XL flow cyt-ometer and results are presented as absolute number of T-cells/Al whole blood.

HIV-1 RNA viral load

Plasma HIV RNA levels were determined by a quantitativeRT-PCR assay (Amplicork Version 1.5, Ultrasensitive Proto-

D4 T-cell counts at time of study

lls High CD4+ T-cells Pa

12 —12:0 —43 (34—62) 0.47820 (0—48) 0.238

c 760 (392—1786)d b0.0001) 1278 (448—1927) 0.382

283 (42—538) b0.0001378 (81—733) 0.004385 (114—1197) 0.0001469 (109—1445) N0.579 (55—83) b0.00123 (2—63) b0.001e

4 (1—41) 0.280

.

Low CD4+ T-cell counts in HIV patients 165

col, 50—75,000 copies/ml) [Roche Diagnostic Systems,Branchburg, USA].

Blood sample collection

PBMC were separated from ACD-treated whole blood byFicoll-Paquek density centrifugation and washed twice inRPMI 1640. Viable cells were counted by trypan blueexclusion and resuspended in 10% DMSO/90% heat inacti-vated FCS at N107 cells/ml for storage in liquid nitrogen. Cellviability was always greater than 95% after thawing.

Lymphocyte phenotype analysis

Expression of Ki67, HLA-DR, CD38 and CD57 was assessed bythree-color flow cytometry on cryopreserved peripheralblood mononuclear cells (PBMC) using combinations of threemonoclonal antibodies directed against CD3, CD4, CD8,Ki67, HLA-DR, CD38 or CD57 conjugated with either FITC,PE or PCy5 (Coulter Immunotech, Marseilles, France).Lymphocytes were distinguished from monocytes by theirforward and side light scatter. A minimum of 10,000lymphocytes per sample were analyzed and gates were setusing appropriate isotype controls. Cells expressing Ki67,HLA-DR, CD38 or CD57 are presented as a percentage ofCD4+ or CD8+ lymphocytes.

Assessment of thymic volume

Thymic volume was assessed on the day of sample collectionby non-contrast spiral computerized tomography (CT) scan-ning as described previously [6] and results reported in cubiccentimeters (cm3).

Preparation of whole CMV virus

Cytomegalovirus (CMV) strain AD169 was propagated inhuman skin fibroblasts infected at a high multiplicity,harvested after 7 days and sonicated immediately in 0.1 Mglycine buffer pH 9d 5, by staff of the Department ofMicrobiology, Royal Perth Hospital, Western Australia.

ELISpot assay for the detection of IFN-; producingT-cells

ELISpot assays were performed with anti-IFN-g antibodiesfrom MabTech (Stockholm, Sweden) [16]. Spots greater than10 units in size and 20 units of intensity were counted usingan AID ELISpot Reader System (AID, Germany) and numbers ofspots in unstimulated wells were subtracted from numbers inwells stimulated with whole CMV. The selective stimulationof CD4+ T-cells by whole CMV was confirmed using PBMCdepleted of CD4+ or CD8+ T-cells using CD4 or CD8 Microbeads(Miltenyi Biotec, Bergisch-Gladbach, Germany).

Statistical analyses

Results are presented as median (range). Patients werestratified on the basis of CD4+ T-cell count, with patientcharacteristics and study variables compared using non-

parametric statistics (Wilcoxon rank sum and log-ranktests). To exclude confounding effects of the duration oftherapy, log-rank comparisons were adjusted within a Coxregression framework by inclusion of months on ART as acovariate with a dummy variable indicating group mem-bership. Correlation across the combined study groupswas assessed using Pearson’s correlation coefficient.Fisher’s exact test was used to assess the relationshipbetween immune activation or senescence and CD4+ T-cell recovery in patients with an undetectable/smallthymus or with a larger thymus. All tests were 2-sidedand P V 0.05 was considered to represent a significantdifference.

Results

Patient characteristics

Most patients were male and groups were balanced for ageand nadir CD4+ T-cell counts (see Table 1). Absolute CD4+ T-cell counts differed significantly (Wilcoxon rank sum test; P b

0.0001), reflecting the mode of recruitment. Absolute CD8+

T-cell counts were similar in the two groups ( P = 0.382).Patients in the low CD4+ T-cell group had fewer naive CD4+

and CD8+ T-cells and fewer memory CD4+ T-cells. MemoryCD8+ T-cell counts did not differ between the two groups.

The proportions of patients in the low and high CD4+ T-cell groups receiving PI-based ART were 40% and 58%,respectively (Fishers exact test; P = 0.3). The medianduration of treatment at stratification was shorter inpatients with low CD4+ T-cell counts ( P b 0.001).However, the groups were matched for the time takento achieve b50 HIV RNA copies/ml and patients with lowCD4+ T-cell counts had received ART for significantlylonger than the time taken for patients with high CD4+

T-cell counts to regain N400 CD4+ T-cells/Al. Duration oftreatment was included in the multivariable analysis offactors associated with CD4+ T-cell recovery.

Low CD4+ T-cell counts were associated with highproportions of CD4+ T-cells expressing markers ofactivation and senescence

Immune activation was assessed via expression of HLA-DR andCD38. In a univariate analysis, the proportion of CD4+ T-cellsexpressing HLA-DR was higher in patients with low CD4+ T-cellcounts (log-rank test; P b 0.001, Table 2). Proportions of CD8+

T-cells expressing HLA-DR did not differ between the groups( P = 0.3). When the groups were combined, the proportion ofCD4+ T-cells expressing HLA-DR was a strong univariatepredictor of both total and naive CD4+ T-cell counts (log10scale). On average, a 50% decrease in total CD4+ T-cell countsfollowed every increase of 6.8% in CD4+HLA-DR+ T-cells (r =�0.67, P = 0.0002; Fig. 1A) and a 50% decrease in naive CD4+

T-cell counts followed every increase of 2.9% in CD4+HLA-DR+

T-cells (r =�0.78, P b 0.0001; Fig. 1B). The proportion of CD4+

T-cells expressing HLA-DR was also inversely correlated withmemory CD4+ T-cell counts (r = �0.49, P = 0.01; data notshown).

The proportion of CD4+ and CD8+ T-cells expressing CD38was similar in the two patient groups ( P = 0.2, Table 2). HLA-

Table 2 Expression of cellular markers of T-cell activation, turnover and senescence in PBMC from patients with low or high CD4+

T-cell counts

Low CD4+ T-cells (n = 15) High CD4+ T-cells (n = 12) P valuea Adjusted P valueb

HLA-DR CD4 8.9 (2.5—20.5)c 2.9 (1.8—7.2) b0.001 0.01HLA-DR CD8 8.7 (1.2—34.4) 6.4 (1.3—30.2) 0.3 0.05CD38 CD4 5.1 (1.9—19.7) 7.4 (1.9—13.6) 0.2 0.008CD38 CD8 2.3 (0.5—7.4) 1.4 (1.0—6.5) 0.2 0.1Ki67 CD4 1.8 (0.3—4.0) 0.7 (0.3—2.0) 0.02 0.01Ki67 CD8 0.6 (0.1—5.1) 0.5 (0.2—1.2) 0.1 0.06CD57 CD4 7.7 (1.4—22.4) 1.7 (0.4—9.3) b0.001 0.007CD57 CD8 37.9 (13.0—58.5) 24.8 (11.0—43.4) 0.01 0.2a Log-rank test.b Adjusted for months on ART, covariate-adjusted log-rank test (via Cox regression).c Percentage of CD4+ or CD8+ T-cells expressing HLA-DR, CD38, Ki67 or CD57 [median (range)].

S. Fernandez et al.166

DR and CD38 expression correlated in the CD8+ T-cell subset(r = 0.739, P b 0.0001) but not the CD4+ T-cell subset ( P =0.6), so CD38 expression was not evaluated further.

Expression of the nuclear antigen Ki67 was used toassess T-cell turnover [17,18]. The proportion of CD4+ T-cells expressing Ki67 was slightly increased in patientswith low CD4+ T-cell counts ( P = 0.02, Table 2), but Ki67expression on CD8+ T-cells was similar in the two groups( P = 0.1). When the groups were combined, Ki67expression on CD4+ T-cells correlated directly with expres-sion on CD8+ T-cells (r = 0.769, P b 0.001) and was a weakunivariate predictor of total CD4+ (r = �0.41, P = 0.06)and naive CD4+ (r = �0.42, P = 0.05) T-cell counts (datanot shown).

CD57 expression was used as a marker of immunolog-ical senescence [12,13]. The proportions of CD4+ and CD8+

T-cells expressing CD57 were substantially higher inpatients with low CD4+ T-cell counts than patients withhigh CD4+ T-cell counts ( P b 0.001 and P = 0.01,respectively). When the groups were combined, theproportion of CD4+ T-cells expressing CD57 was a strongunivariate predictor of both total and naive CD4+ T-cellcounts (log10 scale). On average, a 50% decrease in totalCD4+ T-cell counts followed every increase of 9.1% in

Figure 1 The proportions of CD4+ T-cells expressing HLA-DR and C(A: r = �0.67, P = 0.0002 and C: r = �0.58, P = 0.002, respectively)�0.72, P b 0.001, respectively).

CD4+CD57+ T-cells (r = �0.58, P = 0.002; Fig. 1C) and a50% decrease in naive CD4+ T-cell counts followed everyincrease of 3.6% in CD4+CD57+ T-cells (r = �0.72, P b

0.001; Fig. 1D). The proportion of CD4+CD57+ T-cells wasalso inversely correlated with memory CD4+ T-cell counts(r = �0.42, P = 0.03).

Within the CD4+ T-cell subset, there was a positivecorrelation between HLA-DR and CD57 expression (r =0.447, P = 0.022; data not shown). HLA-DR and Ki67expression (r = 0.375, P = 0.05; data not shown) and Ki67and CD57 expression (r = 0.330, P = 0.09; data notshown) were also moderately correlated.

Since levels of immune activation decrease with ART,we assessed whether the difference in duration of ARTbetween the two patients groups could explain thedifferences in expression of HLA-DR, Ki67 and CD57. Afteradjusting for time on ART, the proportion of CD4+ T-cellsexpressing HLA-DR ( P = 0.01), Ki67 ( P = 0.01) and CD57( P = 0.007) remained higher in patients with low CD4+ T-cell counts (see adjusted P values, Table 2). In multivar-iable regression analyses, the proportion of CD4+ T-cellsexpressing HLA-DR and CD57 remained independent pre-dictors of total CD4+ ( P = 0.01 and P = 0.05, respectively)and naive CD4+ ( P = 0.0001 and P = 0.0005, respectively)

D57 were strong univariate predictors of total CD4+ T-cell countsand naive CD4+ T-cell counts (B: r = �0.78, P b 0.0001 and D: r =

Low CD4+ T-cell counts in HIV patients 167

T-cell counts, irrespective of duration of ART ( P N 0.4;data not shown).

Immune activation correlates with poor CD4+ T-cellrecovery in HIV-1-infected patients with a small orundetectable thymus

Late-phase CD4+ T-cell increases in HIV patients on ART areaffected by thymic production of naive T-cells [5,6]. Asthymic volume was assessed in this cohort [6], we couldrelate immune activation and CD4+ T-cell recovery inpatients with an undetectable/small thymus [b2 cm3,median (range) 0 (0—1.6) cm3; n = 18] and those with alarger thymus [N2 cm3, median (range) 7.3 (3.0—15.6) cm3;n = 8].

Amongst patients with a b2 cm3 thymus (Fig. 2, o),those with a proportion of CD4+ T-cells expressing HLA-DRabove the population median (i.e., N5.2%) were morelikely to have a CD4+ T-cell count below the median (i.e.,b332 CD4+ T-cells/Al) than patients with CD4+HLA-DR+ T-cells below the median (10/11 vs. 2/7; Fishers exact testP = 0.01; Fig. 2A). Similarly, patients with HLA-DR+ CD4+

T-cells above the median were more likely to have anaive CD4+ T-cell count below the median (b84 naive CD4+

T-cells/Al; 10/11 vs. 3/7; P = 0.05; Fig. 2B). Highproportions of CD57+ CD4+ T-cells (N4.2%; n = 11) werealso moderately associated with low total CD4+ T-cellcounts (9/11 vs. 3/7, P = 0.1; Fig. 2C) and naive CD4+ T-cell counts (10/11 vs. 3/7, P = 0.05; Fig. 2D) in patientswith a small/undetectable thymus.

The outcome was less clear in patients with a largerthymus (N2 cm3; Fig. 2, .). Most had low proportions ofactivated or senescent CD4+ T-cells and high total andnaive CD4+ T-cell counts [6], so no correlations wereidentified between these parameters.

Figure 2 Expression of HLA-DR and CD57 on CD4+ T-cells was correundetectable/small thymus. Patients were divided according to thothose with a larger thymus [N2 cm3 thymus (.)]. Horizontal lines onCD4+ (plots B and D) T-cell counts for the combined cohort. Vertical(plots A and B) or CD4+CD57+ (plots C and D) T-cells for the combinCD57+ CD4+ T-cells were more likely to have low total CD4+ (parespectively) T-cell counts. Most euthymic patients (.) had few HLAcounts.

CD4+ T-cell interferon-; responses to whole CMV antigenwere similar in patients with low and high CD4+ T-cellcounts

Numbers of CD4+ T-cells able to produce IFN-g uponstimulation with CMV were assessed by ELISpot assay as amarker of effector memory CD4+ T-cell function. Nodifferences were observed between the patient groups inIFN-g responses to CMV (Wilcoxon rank sum test, P = 0.268;Fig. 3A). After correction for CD4+ T-cell counts, IFN-g CD4+

T-cell responses to CMV were higher in patients with lowCD4+ T-cell counts ( P = 0.025; Fig. 3B).

Discussion

Immune activation is a characteristic feature of untreatedHIV disease and is associated with the progressive depletionof CD4+ T-cells that follows HIV infection [19,20]. T-cellactivation decreases with ART, but can persist in patientswith a stable virological response and influence the recoveryof CD4+ T-cells [7,9,21—23]. In particular, a link betweenCD8+ T-cell activation and low CD4+ T-cell counts after ARThas been established [7,9]. Although Hunt et al. [7] haveassociated CD4+ T-cell activation with low CD4+ T-cell gainsduring the first 3 months of ART, few studies have assessedthe effects of persistent CD4+ T-cell activation on CD4+ T-cellrecovery after long-term ART. Here, the relative frequencyof CD4+ T-cells expressing HLA-DR (a marker of immuneactivation) was increased in patients with poor CD4+ T-cellrecovery after a stable virological response to ART. PoorCD4+ T-cell recovery also correlated with increased propor-tions of CD57+ senescent T-cells, and CD4+ T-cell activationand senescence were directly correlated. These effectswere independent of the duration of ART and most evidentwithin the naive CD4+ T-cell subset. Therefore, our data

lated with total and naive CD4+ T-cell counts in patients with anse with an undetectable/small thymus [b2 cm3 thymus (o)] andplots represent the median total CD4+ (plots A and C) or naive

lines on plots represent the median proportions of CD4+HLA-DR+

ed cohort. Athymic patients (o) with high levels of HLA-DR+ ornels A and C, respectively) and naive CD4+ (panels B and D,-DR+ or CD57+ CD4+ T-cells and high total and naive CD4+ T-cell

Figure 3 Patients with low and high CD4+ T-cell counts had similarly low frequencies of CD4+ T-cells able to respond to CMV whencalculated per 200,000 PBMC (plot A). Responses to CMV were elevated in patients with low CD4+ T-cell counts when data wereadjusted relative to total CD4+ T-cell numbers (plot B).

S. Fernandez et al.168

suggest that CD4+ T-cell activation (and associated T-cellturnover) drives the depletion of naive CD4+ T-cells byaccelerating terminal differentiation and senescence.

Unlike previous studies [7,9], we found no associationbetween CD4+ T-cell recovery and proportions of CD38+

CD8+ T-cells. However, Hunt et al. [7] demonstrated thatthe relative frequency of CD38+, HLA-DR+ CD8+ T-cellsdeclined with increasing duration of viral suppression(b1000 copies/ml) on ART while Benito et al. [9] onlystudied patients who had received ART for 12 months. In amore recent study, Benito et al. demonstrated that CD4+ T-cell recovery beyond the first year of complete suppressionof viral replication is not influenced by CD8+ T-cellactivation [22]. The patients in our study had receivedART for a median time of 69 months (range 12—83) and hadstable viral suppression (b50 copies/ml) for at least 6months (median (range) 31 (9—82 months). Hence therelative frequency of HLA-DR+ and/or CD57+ CD4+ T-cellsmight be a more stable marker, or reflect a different cause,of immune activation associated with poor CD4+ T-cellrecovery in aviremic patients. Causes of CD4+ T-cellactivation in aviremic patients might include reactivationof HIV in latently infected CD4+ T-cells and/or exposure tonon-infectious virions. These were associated with HLA-DRexpression on CD4+ T-cells [24,25] but not with CD38expression [25].

Proportions of CD4+ T-cells expressing CD57 increase inuntreated HIV-1 patients as disease progresses [12] and donot normalize after 6 months of ART [26]. CD4+ CD57+ T-cellsdisplay characteristics of replicative senescence including amemory phenotype, poor IL-2 production, poor proliferativecapacity, short telomeres and frequent apoptosis [26]. Thesusceptibility of CD57+ T-cells to activation-induced apopto-sis [13,26,27] may contribute to the depressed total andmemory CD4+ T-cell counts observed here in patients withmore CD57+ CD4+ T-cells. It would also explain the increasedCD4+ T-cell apoptosis reported in HIV patients with poorCD4+ T-cell recovery on ART [11].

The poor proliferation of CD57+ T-cells, even afterstimulation with interleukin (IL)-2 [13], could be of clinicalimportance in HIV patients receiving IL-2 therapy, becausepatients with low baseline CD4+ T-cell counts generallyrespond less well [28]. The frequency of CD57+CD4+ T-cellsmight therefore predict a response to IL-2 therapy.

Late-phase CD4+ T-cell increases in HIV patients receivingART in part reflect thymic production of naive T-cells [5,6].As thymic volume was assessed in this patient cohort [6], wecorrelated immune activation and senescence with CD4+ T-cell counts in patients with and without a thymus. Inpatients with a small or undetectable thymus, HLA-DRexpression on CD4+ T-cells was inversely linked with CD4+

T-cell recovery. In contrast, most patients with a detectablethymus had high CD4+ T-cell counts and low levels of immuneactivation. Further studies are required to distinguishbetween an instructive model [i.e., a larger thymus providesbetter immune reconstitution which limits opportunisticinfections and hence reduces immune activation] and amodel where the athymic patients only achieve good CD4+ T-cell recovery if (for some other reason) they have lowimmune activation.

Despite lower total CD4+ T-cell counts, HIV patients withpoor CD4+ T-cell recovery on ART produced similar CD4+ T-cell IFN-g responses to CMV when compared to patients withgood CD4+ T-cell recovery (Fig. 3). Low CD4 T-cell countsalso did not reduce IFN-g responses to Candida antigens(data not shown). Similarly, T-cell lymphoproliferativeresponses to CMV and Candidin (Candida albicans antigen)in HIV patients receiving ART are independent of CD4+ T-cellcounts [11]. These observations might explain the discor-dance between low CD4+ T-cell counts and an increased riskof opportunistic infections observed in HIV patients with avirological response to ART [3,11].

We conclude that low CD4+ T-cell counts in patients witha virological response to long-term ART are associated withincreased proportions of activated and senescent CD4+ T-cells. The effect is most apparent in patients with a small orundetectable thymus. However, CD4+ T-cell IFN-g responsesto the antigens of a common opportunistic pathogen (CMV)are similar in patients with low or high CD4+ T-cell countsafter ART.

Acknowledgments

The authors thank Mr. Romano Krueger and Ms. KristyHingston for technical assistance. This work was supportedby the National Health and Medical Research Council ofAustralia (grant number 254590) and a grant from Abbott Pty

Low CD4+ T-cell counts in HIV patients 169

Ltd. This is publication number 2005-22 for the Departmentof Clinical Immunology and Biochemical Genetics, RoyalPerth Hospital.

References

[1] C. Piketty, L. Weiss, F. Thomas, A.S. Mohamed, L. Belec, M.D.Kazatchkine, Long-term clinical outcome of human immuno-deficiency virus-infected patients with discordant immunologicand virologic responses to a protease inhibitor-containingregimen, J. Infect. Dis. 183 (2001) 1328–1335.

[2] G.R. Kaufmann, L. Perrin, G. Pantaleo, M. Opravil, H. Furrer, A.Telenti, B. Hirschel, B. Ledergerber, P. Vernazza, E. Bernas-coni, M. Rickenbach, M. Egger, M. Battegay Swiss HIV CohortStudy Group, CD4 T-lymphocyte recovery in individuals withadvanced HIV-1 infection receiving potent antiretroviral ther-apy for 4 years: the Swiss HIV Cohort Study, Arch. Intern. Med.163 (2003) 2187–2195.

[3] G.R. Kaufmann, H. Furrer, B. Ledergerber, L. Perrin, M.Opravil, P. Vernazza, M. Cavassini, E. Bernasconi, M. Rick-enbach, B. Hirschel, M. Battegay Swiss HIV Cohort Study,Characteristics, determinants and clinical relevance of CD4 Tcell recovery to b500 cells/Ml in HIV type 1-infected individualsreceiving potent antiretroviral therapy, Clin. Infect. Dis. 41(2005) 361–372.

[4] F. Dronda, S. Moreno, A. Moreno, J.L. Casado, M.J. Perez-Elias,A. Antela, Long-term outcomes among antiretroviral-naRvehuman immunodeficiency virus-infected patients with smallincreases in CD4+ cell counts after successful virologic sup-pression, Clin. Infect. Dis. 35 (2002) 1005–1009.

[5] L. Teixeira, H. Valdez, J.M. McCune, R.A. Koup, A.D. Badley,M.K. Hellerstein, L.A. Napolitano, D.C. Douek, G. Mbisa, S.Deeks, J.M. Harris, J.D. Barbour, B.H. Gross, I.R. Francis, R.Halvorsen, R. Asaad, M.M. Lederman, Poor CD4 T cellrestoration after suppression of HIV-1 replication may reflectlower thymic function, AIDS 15 (2001) 1749–1756.

[6] S. Fernandez, R.C. Nolan, P. Price, R. Krueger, C. Wood, D.Cameron, A. Solomon, S.R. Lewin, M.A. French, Thymicfunction in severely immunodeficient HIV-1 infected patientsreceiving stable and effective antiretroviral therapy, AIDS Res.Hum. Retroviruses 22 (2006) 163–170.

[7] P.W. Hunt, J.N. Martin, E. Sinclair, B. Bredt, E. Hagos, H.Lampiris, S.G. Deeks, T cell activation is associated with lowerCD4+ T cell gains in human immunodeficiency virus-infectedpatients with sustained viral suppression during antiretroviraltherapy, J. Infect. Dis. 187 (2003) 1534–1543.

[8] D. Mildvan, R.J. Bosch, R.S. Kim, J. Spritzler, D.W. Haas, D.Kuritzkes, J. Kagan, M. Nokta, V. DeGruttola, M. Moreno, A.Landay, Immunophenotypic markers and antiretroviral therapy(IMART): T cell activation and maturation help predict treat-ment response, J. Infect. Dis. 189 (2004) 1811–1820.

[9] J.M. Benito, M. Lopez, S. Lozano, C. Ballesteros, P. Martinez, J.Gonzalez-Lahoz, V. Soriano, Differential upregulation of CD38on different T-cell subsets may influence the ability toreconstitute CD4+ T cells under successful highly activeantiretroviral therapy, J. Acquired Immune Defic. Syndr. 38(2005) 373–381.

[10] K.B. Anthony, C. Yoder, J.A. Metcalf, R. DerSimonian, J.M.Orenstein, R.A. Stevens, J. Falloon, M.A. Polis, H.C. Lane, I.Sereti, Incomplete CD4 T cell recovery in HIV-1 infection after12 months of highly active antiretroviral therapy is associatedwith ongoing increased CD4 T cell activation and turnover, J.Acquired Immune Defic. Syndr. 33 (2003) 125–133.

[11] O. Benveniste, A. Flahault, F. Rollot, C. Elbim, J. Estaquier, B.Pedron, X. Duval, N. Dereuddre-Bosquet, P. Clayette, G.Sterkers, A. Simon, J.C. Ameisen, C. Leport, Mechanismsinvolved in the low-level regeneration of CD4+ cells in HIV-1-

infected patients receiving highly active antiretroviral therapywho have prolonged undetectable plasma viral loads, J. Infect.Dis. 191 (2005) 1670–1679.

[12] L. Papagno, C.A. Spina, A. Marchant, M. Salio, N. Rufer, S.Little, T. Dong, G. Chesney, A. Waters, P. Easterbrook, P.R.Dunbar, D. Shepherd, V. Cerundolo, V. Emery, P. Griffiths, C.Conlon, A.J. McMichael, D.D. Richman, S.L. Rowland-Jones, V.Appay, Immune activation and CD8+ T-cell differentiationtowards senescence in HIV-1 infection, PLOS Biol. 2 (2004)0173–0185.

[13] J.M. Brenchley, N.J. Karandikar, M.R. Betts, D.R. Ambrozak,B.J. Hill, L.E. Crotty, J.P. Casazza, J. Kuruppu, S.A.Migueles, M. Connors, M. Roederer, D.C. Douek, R.A. Koup,Expression of CD57 defines replicative senescence andantigen-induced apoptotic death of CD8+ T cells, Blood 101(2003) 2711–2720.

[14] M.C. Jimenez-Martinez, M. Linares, R. Baez, L.F. Montano, S.Martinez-Cairo, P. Gorocica, R. Chavez, E. Zenteno, R.Lascurain, Intracellular expression of interleukin-4 and inter-feron-; by a Mycobacterium tuberculosis antigen-stimulatedCD4+CD57+ T-cell subpopulation with memory phenotype intuberculosis patients, Immunology 111 (2004) 100–106.

[15] D. van Baarle, A. Tsegaye, F. Miedema, A. Akbar, Significance ofsenescence for virus-specific memory T cell responses: rapidageing during chronic stimulation of the immune system,Immunol Lett. 97 (2005) 19–29.

[16] 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. Retroviruses 16 (2000) 1991–1996.

[17] N. Sawhney, P.A. Hall, Ki67-structure, function and newantibodies, J. Pathol. 168 (1992) 161–162.

[18] J.M. Orendi, A.C. Bloem, J.C.C. Borleffs, F.J. Wijnholds, N.Machiel de Vos, H.S.L.M. Nottet, M.R. Visser, H. Snippe, J.Verhoef, C.A. Boucher, Activation and cell cycle antigens inCD4+ and CD8+ T cells correlate with plasma human immuno-deficiency virus (HIV-1) RNA level in HIV-1 infection, J. Infect.Dis. 178 (1998) 1279–1287.

[19] M.D. Hazenberg, D. Hamann, H. Schuitemaker, F. Miedema, Tcell depletion in HIV-1 infection: how CD4+ T cells go out ofstock, Nat. Immunol. 1 (2000) 285–289.

[20] M.D. Hazenberg, S.A. Otto, B.H.B. van Benthem, M.T.L. Roos,R.A. Coutinho, J.M.A. Lange, D. Hamann, M. Prins, F. Miedema,Persistent immune activation in HIV-1 infection is associatedwith progression to AIDS, AIDS 17 (2003) 1881–1888.

[21] C.A. Almeida, P. Price, M.A. French, Immune activation inpatients with HIV type 1 and maintaining suppression of viralreplication by highly active antiretroviral therapy, AIDS Res.Hum. Retroviruses 18 (2002) 1351–1355.

[22] J.M. Benito, M. Lopez, S. Lozano, C. Ballesteros, L. Capa, P.Martinez, J. Gonzalez-Lahoz, V. Soriano, CD4+ T cell recoverybeyond the first year of complete suppression of viral replica-tion during highly active antiretroviral therapy is not influencedby CD8+ T cell activation, J. Infect. Dis. 192 (2005) 2142–2146.

[23] L. Al-Harthi, J. Voris, B.K. Patterson, S. Becker, J. Eron, K.Y.Smith, R. D’Amico, D. Mildvan, J. Snidow, B. Pobiner, L. Yau, A.Landay, Evaluation of the impact of highly active antiretroviraltherapy on immune recovery in antiretroviral naRve patients,HIV Med. 5 (2004) 55–65.

[24] T. Chun, D.C. Nickle, J.S. Justement, D. Large, A. Semerjian,M.E. Curlin, M.A. O’Shea, C.W. Hallahan, M. Daucher, D.J.Ward, S. Moir, J.I. Mullins, C. Kovacs, A.S. Fauci, HIV-infectedindividuals receiving effective antiviral therapy for extendedperiods of time continually replenish their viral reservoir, J.Clin. Invest. 115 (2005) 3250–3255.

[25] G.H. Holm, D. Gabuzda, Distinct mechanisms of CD4+ and CD8+

T-cell activation and bystander apoptosis induced by human

S. Fernandez et al.170

immunodeficiency virus type 1 virions, J. Virol. 79 (2005)6299–6311.

[26] B.E. Palmer, N. Blyveis, A.P. Fontenot, C.C. Wilson, Functionaland phenotypic characterization of CD57+ CD4+ Tcells and theirassociation with HIV-1-induced T cell dysfunction, J. Immunol.175 (2005) 8415–8423.

[27] N. Shinomiya, Y. Koike, H. Koyama, E. Takayama, Y. Habu, M.Fukasawa, S. Tanuma, S. Seki, Analysis of the susceptibility of

CD57+ T cells to CD3-mediated apoptosis, Clin. Exp. Immunol.139 (2005) 268–278.

[28] B. Cordwell, S.L. Pett, S. Emery, G. Collins, C. Carey, D.Courtney-Rodgers, and D. A. Cooper, on behalf of the SILCAATstudy group. SILCAAT: CD4+T-cell responses to subcutaneous(sc) recombinant interleukin-2 (rIL-2) after one year. 16thAnnual Conference of the Australasian Society for HIV Medi-cine, Canberra, Australia.