Linkage of the CCR5Δ32 Mutation with a Functional Polymorphism of CD45RA1

A 32-bp deletion in CCR5 (CCR5Δ32) confers to PBMC resistance to HIV-1 isolates that use CCR5 as a coreceptor. To study this mutation in T cell development, we have screened 571 human thymus tissues for the mutation. We identified 72 thymuses (12.6%) that were heterozygous and 2 (0.35%) that were homozygous for the CCR5Δ32 mutation. We found that thymocyte development was normal in both CCR5Δ32 heterozygous and homozygous thymuses. In 3% of thymuses we identified a functional polymorphism of CD45RA, in which cortical and medullary thymocytes failed to down-regulate the 200- and 220-kDa CD45RA isoforms during T cell development. Moreover, we found an association of this CD45 functional polymorphism in thymuses with the CCR5Δ32 mutation (p = 0.00258). In vitro HIV-1 infection assays with CCR5-using primary isolates demonstrated that thymocytes with the heterozygous CCR5Δ32 mutation produced less p24 than did CCR5 wild-type thymocytes. However, the functional CD45RA polymorphism did not alter the susceptibility of thymocytes to HIV-1 infection. Taken together, these data demonstrate association of the CCR5Δ32 mutation with a polymorphism in an as yet unknown gene that is responsible for the ability to down-regulate the expression of high m.w. CD45RA isoforms. Although the presence of the CCR5Δ32 mutation down-regulates HIV-1 infection of thymocytes, the functional CD45RA polymorphism does not alter the susceptibility of thymocytes to HIV-1 infection in vitro.

In this study we have screened a bank of human thymus tissues for the CCR5⌬32 mutation by PCR and studied the effect of the CCR5⌬32 mutation on thymocyte development. During the course of this study, we identified a functional polymorphism of CD45RA expression in which thymocytes do not normally down-regulate expression of CD45RA high m.w. isoforms during thymocyte development. Further, we demonstrated linkage of the functional CD45RA polymorphism with the CCR5⌬32 mutation.

Thymus tissue
Human thymuses were obtained from the Department of Pathology at Duke Medical Center as discarded tissues taken in the course of corrective cardiovascular surgery or therapeutic thymectomy for myasthenia gravis using a Duke institutional review board-approved protocol. No tissue was removed that was not clinically indicated by the surgical procedure being performed. Fresh thymus tissue was either teased with forceps and scissors, and thymocyte suspensions were prepared, or a portion (ϳ0.5 ϫ 0.5 cm) was placed in RPMI 1640 medium supplemented with 7.5% DMSO and 15% FCS and snap-frozen in liquid nitrogen.

Cells and tissue culture conditions
Frozen thymocytes were thawed by incubation at 37°C for 1 h in RPMI 1640 medium containing 10% FCS and 10 g/ml of DNase I (Sigma, St. Louis, MO). After thawing, thymocytes were passed through a Ficoll-Hypaque gradient by centrifugation at 1500 rpm for 30 min and washed twice with RPMI 1640 containing 10% FCS. Thymocytes were cultured in RPMI 1640 medium supplemented with 10% FCS and 10 ng/ml of IL-2 and were maintained at 37°C in a humidified 5% CO 2 incubator. Human PBMCs were prepared from buffy coats of healthy, HIV-1-seronegative individuals obtained through the laboratory services of the American Red Cross, Carolina region (Charlotte, NC). PBMC were isolated by Ficoll-Hypaque gradient centrifugation and were used as positive control cells for in vitro HIV-1 infection assays. PBMC were washed twice in RPMI 1640 medium containing 20% heat-inactivated FCS, resuspended at a density of 2.5 ϫ 10 7 cells/ml in the same medium containing 10% DMSO, and frozen in 1-ml aliquots in liquid nitrogen. PBMC were prescreened for the ability to support the replication of syncytium-inducing (SI) and non-SI (NSI) primary isolates of HIV-1 to confirm the expression of appropriate coreceptors, including CCR5.

HIV-1 virus stocks
Seven different strains of HIV-1 belonging to the genetic clade B subtype were used to infect thymocytes and PBMC in vitro. The IIIB strain is an SI, T cell line-adapted strain that uses CXCR4 as its major coreceptor (3,21,22). 89.6 is an SI primary isolate that infects macrophages and CD4 ϩ lymphocytes (23) and is capable of using multiple coreceptors (22), although CXCR4 usage dominates (24). We used the uncloned stock of 89.6 that had been passaged minimally in PBMC and was expanded in PBMC for use in this study. V67970 is an SI primary isolate that uses both CXCR4 and CCR5 (24,25) and was expanded once in PBMC from the original coculture supernatant. Ba-L is an NSI HIV primary isolate that infects macrophages and CD4 ϩ lymphocytes (26) and uses CCR5 as its sole coreceptor (22). P15 and P46 are NSI primary isolates obtained during early seroconversion (27) and use CCR5 as their sole coreceptor (D. Montefiori, unpublished observations). These latter two HIV-1 isolates were used after a single expansion in PBMC of original coculture supernatants. JR-FL is a CCR5-using NSI primary isolate obtained from frontal lobe brain tissue of a patient with AIDS dementia (28). To be consistent with current nomenclature (29), HIV-1 viruses that use CCR5 as their major coreceptor will hereafter be referred to as R5 strains (Ba-L, P15, P46, and JR-FL), viruses that use CXCR4 will be termed X4 strains (IIIB), and viruses that use both coreceptors will be termed R5X4 strains (89.6 and V67970). Ba-L and JR-FL were obtained from the National Institutes of Health AIDS Research and Reference Reagent Program (Rockville, MD).

Infection of thymocytes
Cyopreserved thymocytes (2 ϫ 10 8 cells/ml/vial) and PBMC were thawed at 37°C and suspended in 20 ml of stimulation medium consisting of RPMI 1640 containing heat-inactivated FBS (20%), gentamicin (50 g/ml), PHA-P (5 g/ml), and delectinized IL-2 (5%, v/v; Advanced Biotechnologies, Columbia, MD). The cells were incubated at 37°C for 2 days, washed twice with 10 ml of growth medium (minus IL-2), and suspended in IL-2 growth medium at a density of 5 ϫ 10 7 cells/ml. Cells were seeded into 96-well U-bottom culture plates at a density of 5 ϫ 10 6 cells/100 l/well. Fifty microliters of virus was added to four wells for each cell type and incubated at 37°C overnight; each cell type was inoculated with an equal dose of virus. Fifty microliters of IL-2 growth medium was added to an additional four wells for each cell type as a negative control. After the overnight incubation the cells were washed twice with IL-2 growth medium to remove the virus inoculum and then were resuspended in 150 l of fresh IL-2 growth medium. Culture fluids (25 l) were harvested on days 4 and 8 of incubation and mixed with 225 l of 0.5% Triton X-100 for p24 quantification. The remaining culture fluid on day 4 was removed and replaced with fresh IL-2 growth medium. Viral p24 Ag was quantified by ELISA (HIV-1 p24 Core Profile ELISA, DuPont/NEN, Boston, MA).

CCR5 genotyping
Genomic DNA was isolated from either whole thymus tissue or isolated thymocytes and was amplified for CCR5 by PCR based on the methods described by Samon et al. (7) with modification. Briefly, primer CCR5-F1 (5Ј-CAAATTGGCTAC-3Ј) and CCR5-R1 (5ЈCCTTGGAAGCTGG-3Ј) flanking the region of the 32-nucleotide deletion in CCR5 gene were used to generate wild-type and deleted DNA fragments of 326 and 294 bp, respectively. The PCR reaction mixture contained ϳ1 g of genomic DNA, 0.2 mM dNTPs, 20 pmol of each primer, and 0.25 U of Taq DNA polymerase (Life Technologies/BRL, Gaithersburg, MD) in a total volume of 25 l. PCR amplification was conducted for 4 min at 94°C followed by 30 cycles of 94°C for 40 s, 60°C for 40 s, and 72°C for 40 s. Samples of the PCR products were fractionated on 1.2% agarose and 2% wide-range agarose (Sigma, St. Louis, MO) gel containing 0.5 g/ml ethidium bromide and visualized under UV light.

Indirect immunofluorescence staining and flow cytometry
Serial frozen 5-m sections of thymus were cut and incubated with saturating amounts of mAbs against CD3, CD4, CD8, CD45RA, CD45RO, CD45RB, and CD45RC or control mAb P3 in PBS containing sodium azide for 30 min at room temperature, washed twice with PBS, and then incubated with saturating amounts of goat anti-mouse IgG-FITC for 30 min at room temperature. These sections were washed three times with PBS, then viewed under fluorescence microscopy. Thymocytes in suspension (0.5 ϫ 10 6 cells/tube) before and after activation with PHA were incubated for 30 min at 4°C with various mAbs against the cell surface molecules, including CD4, CD8, CD45RA, CD45RO, CD45RB, CD45RC, and CXCR4 and control mAb P3 in 100 l of PBS containing 0.2% BSA and 0.1% sodium azide. Cells were washed twice with PBS with BSA/sodium azide, incubated with goat anti-mouse IgG-FITC for 30 min at 4°C followed by washing twice with PBS with BSA/sodium azide. In some experiments in which three-color flow cytometric analysis was performed, thymocytes were incubated with PE-conjugated mouse anti-CD4, and Cy5conjugated mouse anti-CD8 after completion of incubation with the goat anti-mouse IgG-FITC. After staining, thymocytes were washed, fixed in 0.4% paraformaldehyde in PBS, and protected from light at 4°C until analysis by a flow cytometer. Three-color flow cytometry was performed using a FACScan flow cytometer (Becton Dickinson, San Jose, CA). Data acquisition and analysis were conducted with CellQuest software. For each cell sample, a total of 10 4 cell events were analyzed. Data are expressed as the mean fluorescence channel (MFC), an indication of the intensity of cells stained with specific mAbs.

Statistical analysis
The Fisher exact test was used to analyze association of the heterozygous or the homozygous CCR5⌬32 mutation with the presence of the abnormality of CD45RA expression in thymocytes.

Detection of the CCR5⌬32 mutation in a bank of human thymus tissues
The Human Thymus Bank at Duke University contains 571 normal or myasthenia gravis thymuses. These thymuses were from patients with congenital heart disease (age range, 1 day to 17 years) and patients with myasthenia gravis (age range, 3-80 years). We screened these 571 thymus tissues for the CCR5⌬32 mutation using PCR (Table I and Fig. 1). Seventy-two thymuses (12.6%) were heterozygous for the CCR5⌬32 deletion mutation (CCR5⌬32 ϩ/Ϫ ), and two thymuses (0.35%) were homozygous for CCR5⌬32 deletion mutation (CCR5⌬32 Ϫ/Ϫ ). Although both thymuses that were homozygous for CCR5⌬32 were from patients with congenital heart disease, there was no statistical difference in the frequency of CCR5⌬32 in patients with myasthenia gravis compared with that in children with congenital cardiovascular disease (not significant). Sequence analysis of thymus CCR5⌬32 Ϫ/Ϫ DNA revealed the same CCR5⌬32 deletion mutation sequence as that previously reported in PBMC (data not shown) (7,8). Because thymuses were collected as randomly acquired discarded tissues, no data were kept for patients regarding race or ethnic background, and therefore, no relationship of the frequency of CCR5⌬32 mutation with race was studied. However, the overall distribution of CCR5 genotype in the thymus tissue bank was similar to the distribution of CCR5 genotypes as reported for a randomly selected population (7,8).
In normal thymuses most immature thymocytes express CD45RO (ϳ95%) and are destined to die in vivo, while a minority of thymocytes express CD45RA (ϳ5%), indicating that they have successfully been positively selected and are poised for export to the periphery as naive T cells (32)(33)(34). During our analysis of CCR5⌬32 ϩ/Ϫ and CCR5⌬32 Ϫ/Ϫ thymuses, we found abnormal patterns of expression of CD45RA in six of the thymuses with the CCR5⌬32 mutation, in which all thymocytes in both cortex and medulla were inappropriately brightly CD45RA positive (Fig. 2). This abnormality of CD45RA expression was isoform specific, because the expression patterns of CD45RO (Fig. 2, C and E) and CD45RC isoforms (Fig. 2, E and F) were normal in thymuses with the CD45RA abnormal phenotype. Hereafter in this paper these thymuses with this abnormal expression of CD45RA on all thymocytes are referred to as CD45RA abnormal thymuses, and those thymuses with normal CD45RA expression are referred to as CD45RA normal thymuses.

Characterization of CD45RA abnormal thymocytes
The CD45RA abnormal phenotype in thymus was investigated by flow cytometric analysis of purified thymocytes using CD45 mAbs. The mean fluorescence intensity of CD45RA abnormal thymocytes was ϳ10-fold higher with CD45RA mAb staining (MFC, 664) than that of thymocytes isolated from thymus tissue with the normal CD45RA staining pattern (MFC, 75; Fig. 3). In contrast, there were no differences in the level of expression of CD45RO, CD45RB, and CD45RC between thymocytes isolated from thymus tissues with the CD45RA normal vs the abnormal phenotype (Fig. 3).
CD45 is comprised of at least five different isoforms, ABC, AB, BC, B, and O, with molecular masses of 220, 200 (AB and BC), 190, and 180 kDa, respectively. On the protein level, the heterogeneity of different CD45 isoforms can be resolved by mAbs that react with restricted CD45 epitopes (19). In immunoprecipitation assays, protein bands immunoprecipitated by pan-CD45 Abs are 220, 200, 190, and 180 kDa, and those immunoprecipitated by CD45RA Abs are 220 and 200 kDa. Immunoprecipitation assays were conducted to characterize the CD45 isoforms using cell lysates of thymocytes from CD45RA normal and abnormal phenotype individuals. CD45 molecules from cell lysates of thymocytes were first immunoprecipitated using the pan-CD45 mAb F10-89-4 that is directed against a common epitope shared by all CD45 isoforms. We found that CD45 mAb F10-89-4 immunoprecipitated protein bands of 200 and 180 kDa in the CD45RA normal thymocytes, whereas the pan-CD45 mAb immunoprecipitated a large 220-kDa band in addition to 200-and 180-kDa bands in CD45RA abnormal thymocytes (Fig. 4). When normal thymocytes were analyzed with the CD45RA mAb F8-11-13, there were no major protein bands identified in normal control thymocytes (because only ϳ7% of normal thymocytes were CD45RA ϩ ; Fig. 4), indicating that the 200-kDa band seen by the pan-CD45 mAb in normal thymocytes was either the 200-kDa CD45BC isoform or the 190-kDa CD45B isoform. However, in cell lysates of thymocytes of CD45RA abnormal individuals, in which ϳ90% of thymocytes are CD45RA ϩ , CD45RA mAb F8-11-13 immunoprecipitated two FIGURE 1. Analysis of the CCR5⌬32 mutation in human thymus. Genomic DNA was isolated and amplified for CCR5 by PCR using primer pairs flanking the region of the 32-nucleotide deletion in CCR5 gene as described in Materials and Methods. Shown is a representative analysis of PCR products for CCR5 genotyping. The single 326-bp band indicates the normal CCR5 genotype. The single 296-bp band indicates thymus tissue homozygous for the CCR5⌬32 mutation, and two bands with sizes of 326 and 294 bp indicate thymus tissue heterozygous for the CCR5⌬32 heterozygous mutation. a There were no significant differences in percentage of positive cells in each subset of CD4 Ϫ CD8 Ϫ CD4 ϩ CD8 ϩ , CD4 ϩ CD8 Ϫ and CD4 Ϫ CD8 ϩ thymocytes among groups of CCR5 wild-type thymocytes and the CCR5⌬32 ϩ/Ϫ ( p Ͼ 0.05) thymocytes, or among the groups of CCR5 wild-type thymocytes and the combination of the CCR5⌬32 ϩ/Ϫ or the CCR5⌬32 Ϫ/Ϫ ( p Ͼ 0.05). major isoforms of 220 and 200 kDa, demonstrating that the CD45RA reactivity in the CD45RA abnormal phenotype thymuses was due to the lack of normal down-regulation of 200-and 220-kDa CD45RA isoforms during thymocyte development.
Relationship of CD45RA abnormal phenotype in thymocytes with a previously described genetically determined lack of CD45RA-negative lymphocytes in PBMC It has been previously reported that a genetically determined lack of CD45RA-negative PB T cells is present in 8% of healthy individuals (35). No loss of CD45RA expression in PB T cells of these individuals was observed after in vitro activation with PHA due to selective lack of down-regulation of the 200-kDa CD45RA isoform (35,36). To address whether the CD45RA abnormal phenotype identified in thymus in our study was similar to the genetically determined lack of CD45RA-negative T cells found in peripheral blood lymphocytes (35,36), CD45RA normal and abnormal thymocytes were activated in vitro with 1 g/ml of PHA, and analyzed for surface expression of CD45RA and CD45RO.
As shown in Fig. 6A, most unactivated CD45RA normal thymocytes were CD45RA low or negative with an MFC of 101, and CD45RO ϩ with an MFC of 224. In contrast, most thymocytes from the CD45RA abnormal thymus were CD45RA ϩ with an MFC of 409. Activation of thymocytes in vitro with PHA downregulated CD45RA expression in thymocytes with the CD45RA normal phenotype (as reflected by a lower MFC of 50 in Fig. 6B). However, CD45RA expression in thymocytes with the CD45RA abnormal phenotype was not down-regulated after PHA activation, but, rather, was increased by almost 200% to an MFC of 832 (Fig.  6B). After PHA activation, expression of CD45RO was up-regulated in both CD45RA normal and CD45RA abnormal phenotype thymocytes, and no differences in CD45RO expression after in vitro activation between CD45RA normal phenotype and CD45RA abnormal phenotype thymocytes was seen.
Because PB lymphocytes were not available from any subjects whose thymus tissue was studied, we could not directly confirm that the abnormality described in our study of lack of CD45RA down-regulation in thymocytes was the same as that previously described in PBMC. However, given the identical lack of downregulation of CD45RA mAb reactivity with T cell activation, the two conditions probably reflect the same functional polymorphism.

Analysis of linkage of abnormal thymocyte expression of CD45RA with the CCR5⌬32 mutation
Four hundred and six of the 571 original thymus tissues were available to screen for the CD45RA abnormal phenotype. We found statistically significant higher numbers of the CD45RA abnormal phenotype in thymuses with either the heterozygous and homozygous CCR5⌬32 mutation than in CCR5 wild-type thymuses (Table III). We identified the abnormality of CD45RA expression in 1 of 2 thymuses homozygous for the CCR5⌬32 mutation, in 5 of 64 thymuses heterozygous for the CCR5⌬32 mutation, and in 6 of 340 thymuses with wild-type CCR5 genes with no CCR5⌬32 mutations. The CCR5⌬32 allele was found in 50% of those expressing the CD45RA abnormal phenotype compared with 15% in those not expressing it. Using two-tailed Fisher's exact test to test for association between CD45RA abnormality and CCR5⌬32 mutation, a statistically significant association between CD45RA abnormality and CCR5⌬32 mutation was demonstrated ( p ϭ 0.00258). We next performed three pairwise comparisons of the CCR5⌬32 genotypes. Pairwise comparison of the homozygous and heterozygous CCR5⌬32 genotypes with the wild-type CCR5 genotype detected statistically significant association of both the homozygous and heterozygous CCR5⌬32 genotypes with the CD45RA abnormal phenotype (Table IV).

Susceptibility of human CCR5⌬32 ϩ/Ϫ and CD45RA abnormal thymocytes to HIV infection in vitro
Differences in HIV-1 infectivity have been previously suggested in normal PBMC CD45RA ϩ naive vs CD45RO ϩ memory CD4 ϩ PB T cells (37), and the functional capacity of CD45RO ϩ CD4 ϩ T cells is affected by HIV-1 more than that of CD45RA ϩ CD4 ϩ T cells (37). However, to date, CCR5⌬32 ϩ/Ϫ thymocytes have not been studied regarding HIV-1 infectivity. Considering the importance of CCR5 in HIV-1 infection (38) taken together with the association between the CD45RA abnormal phenotype and CCR5⌬32 genotype in thymocytes described here, it was also of interest to determine whether the CD45RA abnormal phenotype could influence the infectability of thymocytes with different CCR5 genetic backgrounds. Individuals who are homozygous for the CCR5⌬32 allele are highly resistant to infection by R5 strains of HIV-1, and their PBMC cannot be infected with R5 strains in vitro (8,39). Individuals who are heterozygous for the CCR5⌬32 allele, on the other hand, are susceptible to infection with R5 strains, but progress to AIDS at a slower rate relative to infected individuals who have wild-type CCR5 alleles (12,13,40). Likewise, PBMC from heterozygous CCR5⌬32 individuals are infectable by R5 strains, although not always to the same extent as PBMC with normal CCR5 alleles (8,41). Lower levels of infection  in heterozygous CCR5⌬32 PBMC might in some cases be due to decreased amounts of CCR5 on the cell surface (41,42). Sufficient cell numbers were not available to test thymocytes homozygous for the CCR5⌬32 mutation in HIV-1 infectivity assays. However, thymocytes with the CD45RA normal and abnormal phenotypes were tested in HIV-1 infectivity assays in combination with either wild-type CCR5 alleles or alleles that were heterozygous for the CCR5⌬32 mutation. Cells in each group were tested for infectability with five primary R5 isolates, two primary R5X4 isolates, and one X4 T cell line-adapted strain of HIV-1. Infection was compared with the infectability of PBMC from a healthy, HIV-1-negative individual who had normal wild-type CCR5 alleles. Equal numbers of viable cells of each cell type were used so that direct comparisons of levels of infection could be made. Viral replication was measured by p24 production on days 4 and 8 of infection.
Thymocytes in each group were infectable by all strains of HIV-1 tested, although infection proceeded at a slow rate in all thymocyte samples compared with PBMC (Table 5). For example, HIV-1 p24 production in PBMC increased from 6.1 to 496 ng of p24/ml on day 4 of incubation, whereas infection in thymocytes was lower (0.2-2.0 ng of p24/ml)) or undetectable (Ͻ0.1 ng of p24/ml) on day 4. The medium was replaced on day 4, and p24 again was again measured on day 8. Higher levels of p24 were seen in the thymocyte cultures at this time, in some cases ap-proaching the levels detected in PBMC (Table V). Although the results varied among cell types and viruses, at least two general observations were made. First, R5 primary isolates were capable of producing a productive infection regardless of the CCR5 genotype and CD45RA phenotype. The level of infection was lower in thymocytes that were heterozygous for the CCR5⌬32 mutation compared with CCR5 wild-type thymuses ( p Ͻ 0.01). Second, the level of HIV-1 infection of thymocytes was independent of the CD45RA expression phenotype. We saw no difference in the level of p24 production on day 4 or 8 in CD45RA normal vs CD45RA abnormal thymocytes ( p Ͼ 0.5). R5 viruses replicated in thymus as well as, if not better than, X4 and R5X4 viruses in most cases. One possible exception was thymocytes from thymus T-690 that was CCR5⌬32 ϩ/Ϫ and CD45RA normal phenotype. Here, no infection with R5 strains was detectable on days 4 and 8 of incubation. However, infection with X4 and X4R5 strains in these cells was also low, suggesting that in this case the lack of detectable infection with R5 strains may be a quantitative issue related to slow replication rather than a qualitative aspect of the infectability of the thymocytes.

Discussion
In this study we have demonstrated the association of the CCR5⌬32 mutation with a functional polymorphism of CD45RA expression in thymocytes. Moreover, we have shown that thymocytes, like PB T cells, when heterozygous for the CCR5⌬32 mutation, produced lower HIV p24 in vitro after infection with HIV. However, we could determine no effect of the CD45RA polymorphism on in vitro HIV infectivity of thymocytes.
A number of mutations in HIV-1 coreceptor genes have been found that modify HIV-1 infectivity in vitro, modify the clinical course HIV-1 infection, or both. These include CCR2 (43), stem cell-derived factor-1 (44), and CXCR4 (45) mutations as well as the CCR5⌬32 mutation (7)(8)(9)(10)(11)(12)(13)(14)(15)(16), The identification of a functional polymorphism of CD45RA expression in association with the CCR5⌬32 mutation is of potentially great interest. The presumption is that whatever evolutionary advantage conferred on humans by the CCR5⌬32 mutation may have been potentially impacted in an as yet unknown way by linkage to the CD45RA abnormal phenotype.
The family of human CD45 leukocyte common Ags is comprised of five glycoprotein members with M r of 180 kDa (O), 190 kDa (B), 200 kDa (AB and BC), and 220 kDa (ABC) that are derived from alternative splicing of a single gene. The heterogeneity of CD45 isoforms can be identified by mAbs that specifically react with epitopes of the distinct isoforms. The expression of CD45RA and CD45RO isoforms has, in general, defined complementary subsets of T cells that differ in naive and memory functional properties (46,47).
Because CD45RA ϩ PB CD4 ϩ T cells when infected with HIV-1 produce lower levels of p24 (37), we postulated that abnormal expression of CD45RA isoforms in thymus would confer on thymocytes some degree of HIV-1 resistance as well. That this was not the case suggests that factors other than CD45RA expression per se in CD45RA ϩ PB native CD4 ϩ T cells are responsible for the relative resistance to HIV-1 infection of the naive T cell subset.
Given the recent appearance of HIV-1 in man and the ancient nature of the CCR5⌬32 mutation, it has been postulated that the CCR5⌬32 mutation confers a selective advantage to humans in an  It has been shown that CD4 ϩ T cells with the heterozygous CCR5⌬32 mutation are partially resistant to HIV infection with R5 HIV-1 strains (7-9). One explanation for the severe lymphodepletion seen in thymuses from patients with end-stage AIDS is infection and destruction of thymocytes by HIV-1 (48); however, in situ hybridization for HIV-1 in these thymuses did not show a large thymus burden of HIV-1 (48). Previous studies have shown that human thymocytes are susceptible in vitro to HIV-1 infection (49 -51). We have found that thymocyte cultures were less sensitive to HIV-1 infection in vitro compared with PBMC cultures. Thymocytes with the heterozygous CCR5⌬32 mutation infected with R5 strains produced lower levels of p24 than did CCR5 wild-type thymocytes infected with the same HIV-1 isolates. These results suggested that thymocytes with the heterozygous CCR5⌬32 mutation are partially resistant to R5 HIV-1. In contrast, thymocytes with the heterozygous CCR5⌬32 mutation and thymocytes with wild-type CCR5 had similar susceptibility to X4 and R5X4 strains. Although CXCR4 is expressed on fresh thymocytes and up-regulated in activated thymocytes, the levels of p24 produced in thymocyte cultures infected with the X4 strain, IIIB, and the R5X4 strain, 89.6, were lower than those in thymocyte cultures infected with R5 HIV-1 strains. Thus, as postulated by others there may be thymotropic strains of HIV-1 that are more destructive to thymuses than other less thymotropic HIV-1 strains (52).
Given the similarity of the defect of activation-induced CD45RA isoform down-regulation in thymocytes described in this report to that of the previously described lack of CD45RA downregulation in PB T cells (35), we suggest that the defects are identical, and that we have now observed the consequence of the previously described PB T cell CD45RA abnormality in thymocytes (35,36). However, this question remains unresolved until PB and thymocyte specimens are available from the same person with the CD45RA abnormality for study.
It is of interest that in the CD45RA abnormal phenotype described here, there is lack of down-regulation of both the 220-and 200-kDa CD45RA isoforms during thymocyte development. In contrast, in the previously described polymorphism of persistent CD45RA expression on memory PB T cells (34), in the transition from naive to memory T cells only the 200-kDa CD45RA isoform was not down-regulated. Therefore, if the two polymorphisms of CD45 isoform regulation are the same, they are expressed differently in thymocyte maturation (lack of down-regulation of both the 220-and 200-kDa CD45RA isoform) vs that in naive to memory  PB T cell conversion (lack of down-regulation of only the 200-kDa isoform). Thus, we have shown a new association of the CCR5⌬32 mutation with an as yet uncharacterized gene mutation responsible for the inability to appropriately down-regulate CD45RA isoforms during thymocyte development and activation. Because CD45 is encoded on chromosome 1 (53,54), and CCR5 is encoded on chromosome 3 (12), this association must be mediated by an unidentified gene product that modifies CD45 processing and is located on chromosome 3 at a distance from CCR5. It will be of interest to determine in population studies whether the presence of the CD45RA abnormal polymorphism in association with the CCR5⌬32 mutation has any effect on the clinical outcome of HIV-1 infection.