Effect of Latent Human Immunodeficiency Virus Infection on Cell Surface Phenotype

Highly active antiretroviral therapy has succeeded in many casesin suppressing virus production in patients infected with humanimmunodeficiency virus (HIV); however, once treatment is discontinued,virus replication is rekindled. One reservoir capable of harboringHIV in a latent state and igniting renewed infection once therapyis terminated is a resting T cell. Due to the sparsity of Tcells latently infected with HIV in vivo, it has been difficultto study viral and cellular interactions during latency. TheSCID-hu (Thy/Liv) mouse model of HIV latency, however, provideshigh percentages of latently infected cells, allowing a detailedanalysis of phenotype. Herein we show that latently infectedcells appear phenotypically normal. Following cellular stimulation,the virus completes its life cycle and induces phenotypic changes,such as CD4 and major histocompatibility complex class I down-regulation,in the infected cell. In addition, HIV expression followingactivation did not correlate with expression of the cellularactivation marker CD25. The apparently normal phenotype andlack of HIV expression in latently infected cells could preventrecognition by the immune response and contribute to the long-livednature of this reservoir.

INTRODUCTION

Top

Introduction

Infection with human immunodeficiency virus (HIV) is characterizedby high plasma viral loads, CD4⁺ T-cell depletion, and ultimatelyimmune system failure (18). The administration of highly activeantiretroviral therapy has succeeded in reducing viral loadsin many patients to undetectable levels (19, 28, 35). However,termination of therapy results in a rebound in viral load, evenin patients with previously undetectable viral levels, due tothe presence of a viral reservoir not depleted by the presentdrug regimens (9). Previous studies have indicated that onesuch reservoir able to rekindle productive infection followingcessation of highly active antiretroviral therapy is the latentlyinfected T cell (10, 40). Based on the lack of sequence evolutionin the latent population compared to virus isolated from theperiphery, it appears that HIV in the latent state remains hiddenfrom the immune response for the duration of therapy (16, 17,38, 40). Unlike other suspected cellular reservoirs, such asproductively infected T cells and persistently infected macrophages,which have a relatively short half-life (19, 27, 35), the populationof latently infected T cells is estimated to remain stable upto 60 years on therapy (15, 30).

Much of the latent infection in the periphery likely occursduring the differentiation of infected, transcriptionally activeeffector T cells into quiescent memory cells. It is thoughtthat infection of the transcriptionally active T cell allowsfor reverse transcription and integration of the viral genome;however, prior to either viral or immunologic cytopathic effects,cellular and viral transcription ceases and the virus becomesdormant. In this form HIV is resistant to the effects of antiviraldrugs and is maintained until the cell is stimulated to increaseRNA production. Environmental stimuli which trigger T cell transcriptionare then capable of reactivating the virus from latency (5,12, 13, 25, 33). Formation of the latent reservoir in vivo,although very rare ( $~$ 1 per million resting CD4⁺ T cells), occursearly during the acute phase of infection and does not seemto decrease substantially with duration of therapy, making ita formidable barrier to viral eradication (8, 11, 15, 30).

There is an emerging body of evidence demonstrating HIV infectionof naïve T cells in vivo (3, 24, 26, 29), but exactly whereand how infection of this cell type occurs is not clear. Thepresence of CCR5-tropic strains of HIV in the naive T-cell population,which is resistant to direct infection by HIV in vitro, andthe recently demonstrated presence of HIV proviral DNA in naïve,CD45RA⁺ CD8⁺ T cells suggest that infection may occur duringdifferentiation in the thymus (21, 24, 26, 29). We have previouslyestablished using the SCID-hu (Thy/Liv) mouse that some latentHIV infection can be generated during thymopoiesis and thatthe loss of cellular gene transcription during the later stagesof thymopoiesis is associated with the transition into virallatency (5). These latently infected thymocytes are stable andcan be exported into the periphery where the virus remains latentuntil the cell becomes transcriptionally activated. These studiesstrengthen the notion that latently infected naïve T cellsidentified in HIV-positive patients may have arisen in the thymusduring thymocyte differentiation.

Experimental data have illustrated that HIV infection perturbsmany normal cellular events. In vitro evidence indicates thatthe HIV gene products nef, vpu, and env are capable of down-regulatingexpression of CD4 (1, 6, 7) and that nef can also induce down-regulationof major histocompatibility complex I (MHC-I) (14, 31) and up-regulationof FasL (39) from the cellular surface. HIV has further beenshown to alter the expression of the homing and chemotacticreceptors CD54, CD62L, and CCR5 on the surface of productivelyinfected cells (23). In vitro microarray analysis has illustratedthat nef alone is capable of triggering a cell activation programsimilar to anti-CD3 stimulation and can alter gene expressionof proteins known to be crucial to HIV replication (32). Similarly,other studies have demonstrated that the viral protein tat,in addition to increasing HIV transcription, also transactivatescellular interleukin 2 gene expression (34, 36). While it isnot known if HIV causes changes in cellular phenotype priorto the transition into latency, it is possible that infectioninduces a change in cellular physiology that is not reversedwhen the infected cell enters a quiescent state. Alternatively,the coincident loss of viral gene transcription during cellularquiescence may cause these phenotypic changes to revert andthe cell would then appear similar to an uninfected T cell.Due to the extremely low levels of latent virus in vivo, characterizationof the latent reservoir at the cellular level has been difficult.In the SCID-hu system, latent HIV infection is generated athigh frequency and the use of reporter viruses allows identificationof productively infected cells (5, 20). This system thus providesa powerful model to assess the properties of latently infectedT cells and the effects of virus reactivation on the infectedcell. Herein we demonstrate that latent infection does not significantlyalter the normal phenotype of human thymocytes or peripheralblood lymphocytes. However, upon stimulation of the latentlyinfected cell, viral proteins are produced and the characteristic,virally mediated manifestations of CD4 and MHC-I down-regulationare observed. These data suggest that the latently infectedcell will appear normal to the immune system and thus escapeimmune clearance.

MATERIALS AND METHODS

Top

Materials and Methods

Infection of SCID-hu mice. SCID-hu mice were prepared by implantation of human fetal liverand thymus (Advanced Bioscience Resource, Alameda, Calif.) aspreviously described (2). Thy/Liv implants were mock infectedwith medium or directly injected with viral stocks such thata total of 20 ng of HIV_NL-r-HSAs was introduced. In general1 ng of p24 contains approximately 100 infectious units. Viralstocks were prepared by electroporation of cloned proviral DNAinto CEM cells. p24 gag expression was assessed by enzyme-linkedimmunosorbent assay (Coulter, Hialeah, Fla.) and was used toquantitate viral titers and reactivation from latency. Tissuedonor numbers are provided in the figure legends to differentiateseparate experiments.

Cell isolation and culture. Thy/Liv implants were harvested 5 weeks postinfection, and single-cellsuspensions of thymocytes were immediately pooled in the presenceof 100 ng of indinavir (Merck, West Point, Pa.)/ml and, whereindicated, also in 10 µM zidovudine (AZT) (Sigma, St.Louis, Mo.). Thymocytes from mock-infected implants were isolatedand cultured in parallel as controls in all experiments. Whereindicated, thymocytes were stained with monoclonal antibodiesspecific for human CD4, CD45, murine CD24 (muCD24) directlyconjugated to phycoerythrin (PE), allophycocyanin (APC), andfluorescein isothiocyanate (FITC) (Coulter) and with an antibodyto human CD8 conjugated to peridinin chlorophyll protein (PerCP;Becton-Dickinson) and sorted on a FACStar^Plus flow cytometer.Postsort analysis showed that more than 98% of cells were CD4positive, CD8 negative, and muCD24 negative.

In other experiments, total thymocytes were stained with murineantibodies to human CD8 (Becton-Dickinson, Mountain View, Calif.)and CD4 single-positive (SP) cells isolated by negative selectionvia panning with goat anti-mouse antibody (Coulter)-coated flasks.These cells were then stained with rat antibody to muCD24 (Pharmingen,San Diego, Calif.) and were panned in flasks coated with rabbitanti-rat antibody (Sigma). Purity was assessed with differentantibody clones to prevent epitope masking (Coulter).

To obtain human peripheral blood lymphocytes from SCID-hu mice,blood from multiple mice was pooled and peripheral blood mononuclearcells were isolated over a Ficoll-Paque gradient. To depletemurine monocytes/macrophages, the remaining cells were adheredfor 2 h at 37°C in medium containing 100 ng of indinavir/mland 10 µM AZT. Peripheral blood lymphocytes were stainedas described above.

All cells were cultured in RPMI 1640 supplemented with 10% humanAB serum, 100 U of penicillin/ml, 100 µg of streptomycin/ml,2 mM glutamine, and 100 nM indinavir. Where indicated cellswere cultured in the presence of 10 µM AZT (Sigma). Cellswere costimulated with ${alpha}$ CD3 monoclonal antibody bound to goatanti-mouse antibody-coated plates and 0.1 µg of soluble ${alpha}$ CD28 (5)/ml.

Flow cytometry. Cells from cultures were stained with CD4-PE, CD8-PerCP, CD45-APC,and muCD24-FITC. For phenotypic analysis, cells were stainedwith the indicated antibody conjugated to FITC, PE, or APC.Thymocytes for Fig. 4 were stained with an antibody to humanCD25 conjugated to APC. HLA-A2 expression was assessed withan antibody specific to HLA-A2 (and HLA-B17) conjugated to biotinand counterstained with streptavidin-APC (Pharmingen). Isolationof this antibody has been previously described (37). Sampleswere analyzed on a FACSCalibur flow cytometer using the CellQuest Program (Becton-Dickinson). Forward- versus side-scatterprofiles were used to define the live population. All cellswere further gated on the human CD45⁺ population to excludemurine cells. Quadrants were set based on isotype controls fromuninfected cells for each condition.

View larger version (21K):
[in this window]
[in a new window]
FIG. 4. CD25 expression on thymocytes before and after HIV reactivation from latency. Thymocytes from pooled HIV_NL-r-HSAs-infected Thy/Liv implants were negatively selected as done for Fig. 2a and were analyzed for CD45, CD25, and muCD24 expression by flow cytometry (day 0). Thymocytes were then costimulated for 3 days in the presence of a viral protease inhibitor, and the percentage of CD45-positive thymocytes expressing muCD24 was assessed by flow cytometry (day 3, left panel). Backgating on muCD24-negative (top right), muCD24-dim-positive (middle right), and muCD24-bright-positive (bottom right) thymocytes illustrated the percentage of cells expressing CD25 in each population. The mean fluorescence intensity (MFI) of CD25 expression for each panel is indicated above each histogram.

RESULTS

Top

Results

Latently infected cells are phenotypically similar to uninfected cells. In order to determine if latently infected thymocytes and peripheralblood lymphocytes are phenotypically different from uninfectedcells, we infected SCID-hu (Thy/Liv) mice with the pathogenic,CXCR4-tropic, reporter virus HIV_NL-r-HSAs (20) and 5 weeks postinfectionharvested thymocytes and peripheral blood lymphocytes. Thisvirus contains the coding sequence for muCD24 inserted in placeof the vpr gene, and productive infection directs the expressionof muCD24 on the surface of infected cells, allowing quantitationby flow cytometry. Infection resulted in depletion of CD4/CD8double-positive thymocytes and loss of CD4⁺ T cells in the periphery(Fig. 1). We next isolated mature, CD4 SP, muCD24-negative thymocytesby negative selection, resulting in a population of cells thatexpressed less than 1% CD8 or muCD24 (Fig. 2a). These cellswere then subjected to costimulatory signals to induce latentvirus expression. Uninfected thymocytes were obtained in thesame manner from normal Thy/Liv implants and were similarlycultured for comparison (data not shown). Cells in all experimentswere cultured in the presence of a viral protease inhibitorand AZT to prevent viral spread or de novo infection in vitro(5). As can be seen in Fig. 2a, culturing induces expressionof low levels of virus. However, following 3 days of costimulationwith activating antibodies to CD3 and CD28, more than 96% ofthymocytes expressed CD25 (data not shown) and there was a substantialincrease in expression of the HIV-encoded reporter, independentof cell size or granularity (Fig. 2a), indicating the presenceof latent virus in the initial population. Since approximately9% of CD4 SP thymocytes from HIV-infected implants harboredlatent virus, virus-induced phenotypic changes should be evidentby flow cytometry of the bulk population. Comparison of infectedand uninfected mature thymocytes immediately following isolationrevealed no phenotypic differences in expression of a varietyof cell surface proteins (Fig. 2b). Expression levels of CD11a,CD38, CD40, CD40L, CD54, Fas, and HLA-DR were similarly unchanged(data not shown). Furthermore, there was no difference in themean fluorescence intensity of any of the proteins analyzedexcept MHC-I, the intensity of which increased approximatelysevenfold on total thymocytes from infected implants and twofoldon mature thymocytes (Fig. 2c). The two- to sevenfold increasein the mean fluorescence intensity of MHC-I expression on thymocytesfrom an infected thymus has been described previously and maybe due to indirect consequences of infection involving cytokinedysregulation (22). The lack of gross phenotypic alterationbetween uninfected thymocytes and thymocytes containing latentlyinfected cells implies that latent infection does not quantifiablyinduce major external changes in cell surface protein expression.Furthermore, any phenotypic changes that may have occurred inthese molecules during the initial productive infection mustbe reversed following the transition into viral latency.

View larger version (41K):
[in this window]
[in a new window]
FIG. 1. Phenotype of SCID-hu thymocytes and peripheral blood lymphocytes (PBL). Thymocytes and peripheral blood lymphocytes from uninfected and HIV_NL-r-HSAs-infected SCID-hu (Thy/Liv) mice were harvested 5 weeks postinfection and were stained with monoclonal antibodies to CD45, CD4, and CD8 on the day of biopsy. CD45⁺ (human) cells were gated and analyzed for CD4 and CD8 expression. Human thymocytes and lymphocytes from two experiments (tissue donors 67 and 30, respectively) are shown.

View larger version (119K):
[in this window]
[in a new window]
FIG. 2. Flow cytometry analysis of cells harboring latent HIV. (a) Presence of latent HIV in CD4 SP thymocytes. At 5 weeks postinfection with HIV_NL-r-HSAs Thy/Liv implants (tissue donor 67) were pooled and cultured in medium containing a viral protease inhibitor and AZT (same experiment as Fig. 1a, experiment 1). CD4 SP, muCD24-negative thymocytes (day 0) were then isolated by panning. The left panels were gated on CD45⁺ cells to exclude any murine cell contamination and show that less then 1% of the isolated thymocytes expressed CD8 or muCD24. These thymocytes were either left unstimulated or costimulated for 3 days, and HIV (muCD24) expression was quantitated by flow cytometry. The right panels show day 3 expression of HIV as a result of each condition. HIV (muCD24) expression is illustrated based on size and granularity following costimulation (upper panels). Quadrants are set according to isotype controls for each condition. (b) Phenotype of infected thymocytes. Uninfected and HIV_NL-r-HSAs-infected CD4 SP, muCD24-negative thymocytes (from Fig. 2a, day 0) were stained for the indicated proteins on day 0 and analyzed by flow cytometry. Following activation approximately 10% of these cells showed evidence of latent infection (Fig. 2a). The y axis indicates the percentage of cells expressing each protein. Similar results were observed in multiple experiments. (c) MHC-I expression on thymocytes and peripheral blood lymphocytes (PBL) in response to HIV infection. CD4 SP, muCD24-negative, CD45-positive thymocytes and total-CD45-positive peripheral lymphocytes from uninfected and HIV_NL-r-HSAs-infected mice (tissue donor 75) were stained on day 0 for MHC-I (HLA-ABC) expression and were analyzed by flow cytometry. The gate indicates MHC-I-positive cells and is based on an isotype control. The mean fluorescence intensity (MFI) of MHC-I-positive cells is shown above each gate. Similar results were observed in multiple experiments. (d) Phenotype of human peripheral blood lymphocytes. Peripheral blood lymphocytes from uninfected and HIV_NL-r-HSAs-infected mice (tissue donor 67) were stained on day 0 with an antibody to the indicated human protein and CD45. CD45-positive (human) lymphocytes were analyzed by flow cytometry for the percentage of cells expressing each protein (shown as y axis). These results were consistent over multiple experiments.

We next assessed the phenotype of human cells in the peripheryof HIV_NL-r-HSAs-infected SCID-hu mice. We observed a decreasein the relative number of CD4⁺ T cells and an increase in therelative number of CD8⁺ T cells, consistent with CD4⁺ cell depletionin the thymus (Fig. 1); however, no detectable change in othercell surface proteins was seen (Fig. 2d). Likewise, expressionof the RA and RO isoforms of CD45 was unchanged on human peripheralblood lymphocytes from infected animals (data not shown). LatentHIV infection did not alter the mean fluorescence intensityof any of the proteins analyzed, including MHC-I, the intensityof which was increased on thymocytes from HIV-infected implantsof the same mice (Fig. 2c). We are unable to quantitate expressionof the viral reporter gene in the peripheral blood lymphocytesdue to the high levels of endogenous muCD24 expression on contaminatingmurine cells in the periphery (5); however, these peripheralT cells were clearly infected with latent virus, as costimulationinduced substantial release of viral p24 into the medium, whereasno p24 was detected in the medium of unstimulated cells (600pg of p24/ml for costimulated cells versus quantities belowthe 8-pg/ml detection level of the enzyme-linked immunosorbentassay for nonstimulated cultures). The immense increase in viralp24 following costimulation, coupled with the fact that morethan 90% of peripheral human cells contain full-length HIV reversetranscripts (data not shown), implies a high level of latentinfection in the periphery. Together, these data indicate thatlatently infected T cells are exported from the thymus and thatthese cells are phenotypically indistinguishable from uninfectedlymphocytes.

Reactivation of latent HIV induces down-regulation of CD4 and MHC-I. To ascertain if viral reactivation from latency results in CD4and MHC-I down-regulation, we isolated CD4 SP, HIV reporter-negativethymocytes by fluorescence-activated cell sorter. The enrichedpopulation contained less than 1% CD8 or muCD24 positive cells,and more than 98 percent of the cells expressed high levelsof CD4. Costimulation of uninfected thymocytes had no effecton CD4 expression, while costimulation of the latently infectedthymocytes resulted in a loss of CD4 cell surface expressionfrom now productively infected cells (Fig. 3a). The down-regulationof CD4 occurred only on cells that coexpressed the virally encodedmuCD24, indicating that viral replication was associated withthe loss of CD4 expression. We are unable to specifically analyzeperipheral blood lymphocytes for CD4 down-regulation by HIVbecause of the high endogenous levels of muCD24; however, Fig.1 illustrates that the fluorescence intensity of CD4 or CD8is unchanged on human peripheral blood lymphocytes in infectedanimals. To assess if viral reactivation caused a similar decreasein MHC-I expression, we utilized donor fetal tissue that expressedthe HLA-A2 antigen and an antibody that specifically recognizesthis molecule. As seen with CD4 in the previous experiment,MHC-I was expressed at high levels on latently infected thymocytes(similar to mock; not shown); however, following HIV reactivation,there was a down-regulation of MHC-I on infected cells expressingnewly reactivated virus (Fig. 3b). This down-regulation is specificfor CD4 and MHC-I as costimulation increases the expressionof CD25 on both uninfected and previously latently infectedthymocytes (Fig. 4), and levels of CD45 are not changed followingHIV reactivation (data not shown). Thus, although latent HIVinfection does not influence cell surface protein expression,following reactivation, virus protein levels are sufficientto influence, directly or indirectly, the down-regulation ofCD4 and MHC-I.

View larger version (37K):
[in this window]
[in a new window]
FIG. 3. HIV reactivation from latency causes CD4 and MHC-I down-regulation on infected cells. (a) CD4 down-regulation. Thymocytes from a single Thy/Liv implant (tissue donor 187) were stained with monoclonal antibodies to CD4, CD8, CD45, and muCD24 and were then isolated by fluorescence-activated cell sorting at 5 weeks postinfection with HIV_NL-r-HSAs (day 0, top). Uninfected thymocytes from the same tissue donor were sorted in parallel (day 0, bottom). Thymocytes were cultured in the presence of protease inhibitor and AZT. The left panels were gated on CD45-positive cells and indicate that more than 98% of the sorted cells were CD4 SP and muCD24 negative postsort. Less than 1% of these thymocytes expressed CD8 (data not shown). Thymocytes were then costimulated for 3 days and were analyzed for CD4 and muCD24 expression by flow cytometry. The right panels are gated on CD45-postive cells and show HIV (muCD24) and CD4 expression following costimulation (Costim). Unstim, unstimulated. The level of p24 expression from costimulated and unstimulated cultures is shown at the right of the figure. (b) MHC-I down-regulation. CD4 SP thymocytes from HIV_NL-r-HSAs-infected implants (tissue donor 62) were negatively selected as done for Fig. 2a. Isolated cells contained less than 2% CD8-positive and less than 1% muCD24-expressing cells (day 0). Isolated thymocytes were cultured either unstimulated or costimulated for 2 days in the presence of a viral protease inhibitor and AZT and were analyzed for HLA-A2 and HIV (muCD24) expression by flow cytometry. The mean fluorescence intensity (MFI) of HLA-A2 expression from muCD24-positive and -negative cells on day 2 is shown. Uninfected thymocytes isolated and cultured in the same manner for 2 days are shown in the right panels for comparison. The level of p24 expression for each condition is shown below each dot plot.

${alpha}$ CD3 and ${alpha}$ CD28 costimulation of CD4 SP, muCD24-negative thymocytesdramatically increased the expression of the ${alpha}$ chain of the interleukin2 receptor (CD25) on both latently infected and uninfected thymocytes(Fig. 4). Backgating analysis of cells that expressed HIV reporterfollowing costimulation indicated that the intensities of HIVand CD25 expression were not necessarily linked, as we observedin multiple experiments similar levels of CD25 expression oncells expressing high or low levels of muCD24 (Fig. 4). Furthermore,in all experiments, virus expression could be identified inthymocytes that did not express detectable CD25. Moreover, ${alpha}$ CD3stimulation alone was able to reactivate latent virus withoutan increase in CD25 expression (5; data not shown).

DISCUSSION

Top

Discussion

Our data suggest that latent infection does not substantiallyperturb cellular surface protein expression on thymocytes oron peripheral blood lymphocytes. Analysis of CD4 and MHC-I expressionindicates that any residual HIV protein expression during latencyis insufficient to impart the phenotypic cellular changes associatedwith productive infection. Whether this is due to a low levelof or a complete lack of viral protein expression during latencyis unclear. Following viral reactivation, both the surface expressionof both CD4 and MHC-I is down-regulated on cells that expressHIV but not on cells in the same culture that are not expressingvirus, illustrating that viral reactivation is necessary forthis phenotype. Since latently infected cells express equivalentlevels of homing (CD11a, CD44, CD54, and CD62L) and chemotacticreceptors (CXCR4 and CCR5) to those of their uninfected, restingcounterparts, they would circulate and distribute through thebody in a normal fashion, which is consistent with previousstudies that observed latently infected cells in both the peripheralblood and lymph nodes of infected individuals (8).

The low levels of viral RNA (5), the undetectable viral reporterexpression, and the lack of CD4 and MHC-I down-regulation observedduring HIV latency suggest that viral protein expression isinsufficient for adequate MHC-I presentation of HIV peptidesto cytotoxic T lymphocytes (CTL). Similarly, the normal levelsof MHC-I expressed on latently infected cells would not triggeran NK cell response. These factors could explain in part whylatently infected cells are able to persist for such a longperiod of time in the presence of an HIV-specific CTL response(15, 30). Unfortunately, this would also suggest that therapeuticattempts to eradicate HIV by boosting the CTL response wouldnot affect the latent population, as without an increase inviral protein production, these cells would remain invisibleto the immune response. On the other hand, our data indicatethat, once the latent virus is reactivated, infection proceedswith the normal phenotypic consequences, suggesting that theinfected cell could then be targeted by HIV-specific CTL.

Interestingly, it appears that the level of HIV infection isnot precisely controlled by the degree to which the cell isactivated. By analyzing costimulated thymocytes, we found thatboth CD25-positive and CD25-negative cells express HIV and thatCD25 expression, although generally signifying a more activatedcell, did not correlate with a higher level of HIV expression.Our previous studies have demonstrated that HIV gene expressioncan be induced within 24 h following reactivation from latency(5). As CD25 is a relatively late T-cell activation marker,HIV may be expressed prior to CD25 acquisition, suggesting thatimmunotherapies targeting CD25 (4) may not eliminate cells recentlyactivated to produce virus. However, our data suggest that strategiesdesigned to reactivate HIV expression from a latently infectedcell, such that it could be targeted by the immune system, mayonly need to minimally stimulate a cell, potentially avoidingthe problems associated with nonspecific immune activation.

It is interesting that destruction of thymocytes by HIV canlead to an inversion of the human CD4-to-CD8 cell ratio in theperiphery of these mice (Fig. 1). This depletion in the peripheryis correlated with the HIV-induced depletion of thymocytes inthe Thy/Liv implant, which give rise to the peripheral cells.The inversion of the CD4/CD8 cell ratio suggests that the depletionof thymocytes can cause a quantitative change in the exportof lymphocytes from the thymus. In the setting of pediatricHIV infection and the developing immune system, the impairedexport of lymphocytes to the periphery could enhance the CD4⁺T-cell deficiency associated with HIV infection. It appearsthat, if any phenotypic aberrations occur with active infectionprior to latency, they are reversed following the loss of activeviral replication. Accordingly, our data indicate that somecells that survive differentiation in the infected thymus canbe exported to the periphery and appear phenotypically normalyet harbor latent HIV. Consequently, it is important to targetthe thymus and the latent pool generated therein with noveltherapeutic approaches in order to eliminate these reservoirsof infection.

ACKNOWLEDGMENTS

We thank G. Bristol, R. Cortado, and A. Kacena for expert technicalassistance.

This work was supported by NIH grants nos. AI 36554 and AI 36059and the UCLA CFAR.

FOOTNOTES

* Corresponding author. Mailing address: UCLA School of Medicine, Department of Medicine, 11-934 Factor Bldg., Mail Code 167817, 10833 Le Conte Ave., Los Angeles, CA 90095. Phone: (310) 794-7765. Fax: (310) 825-6192. E-mail: jzack{at}ucla.edu.

REFERENCES

Top