|Home | About | Journals | Submit | Contact Us | Français|
We recently isolated from an infant an X4-syncytium-inducing (SI) human immunodeficiency virus type 1 (HIV-1) variant (92US143-T8) that was able to infect CD8+ lymphocytes independently of CD4. Although it was CD4 independent, the 92US143-T8 isolate also maintained the ability to infect CD4+ cells. In the present study, we investigated the role of CXCR4 in the infection of CD4+ and CD8+ cells by this primary isolate. The expression of CXCR4 was down modulated in CD8+ lymphocytes after infection with the 93US143-T8 isolate. Infection of CD8+ lymphocytes by the 93US143-T8 isolate was prevented by treatment with AMD3100, a specific antagonist for CXCR4, indicating CXCR4-dependent infection. Interestingly, AMD3100 treatment had no inhibitory role in the infection of purified CD4+ lymphocytes by the same isolate. Furthermore, AMD3100 treatment failed to prevent infection of known CD4+ CXCR4+ T-cell lines (MT-2 and CEM) by the 93US143-T8 isolate. In fact, virus replication in the CD4+ cells was often enhanced in the presence of AMD3100. Viruses produced from the infected CD4+ cells in the presence of AMD3100 maintained an unchanged envelope genotype and an SI phenotype. For the first time, these results provide evidence of CXCR4-dependent infection of CD8+ lymphocytes by a primary HIV-1 isolate. This study also shows a different mode of infection for the CD4+ and CD8+ lymphocytes by the same HIV-1 variant. Finally, our findings suggest that a more careful evaluation is necessary before the random use of AMD3100 as a new entry inhibitor in patients harboring SI HIV-1 strains.
Although a large number of chemokine receptors that can play critical roles as coreceptors for human immunodeficiency virus (HIV) entry have been identified, primary HIV type 1 (HIV-1) isolates generally use either CXCR4 (X4 strains) or CCR5 (R5 strains) as a coreceptor for infection of CD4+ cells (13). X4 viruses generally appear late in HIV-1 disease, and their appearance is frequently associated with rapid progression to AIDS. However, many HIV-1 isolates that use CXCR4 are also capable of using CCR5 and thus are called dual tropic (R5X4) (6, 16, 26). X4 and R5X4 viruses are able to infect a larger proportion of the immune cells because, unlike CCR5, the CXCR4 molecules are abundantly expressed on most primary T cells, including the resting T lymphocytes (7). Furthermore, X4 and R5X4 isolates frequently form syncytia in CD4+ cells in vitro, and it is believed that these viruses are primarily responsible for the profound loss of CD4+ cells in the late stages of HIV-1 disease (17).
AMD3100, a bicyclam compound, is a specific antagonist for CXCR4 that has emerged as a potent inhibitor of infection by X4 and R5X4 strains (11, 15, 21, 22, 28, 38). Strong selection pressure against the more pathogenic X4 strains has prompted the suggestion that AMD3100 might be used even in asymptomatic HIV-1-infected patients to prevent the emergence of the more pathogenic X4 viruses (11, 15). We have recently isolated CD4-independent variants from the quasispecies of several HIV-1-infected patients using a highly stringent selection method. Although these isolates can infect independently of CD4, most of these viruses have still maintained an unchanged ability to infect CD4+ cells by using the CD4 receptor like regular HIV-1 strains (37). Here, we used AMD3100 to study the role of CXCR4 in the infection of CD4+ and CD8+ cells by 93US143-T8, a primary CD4-independent HIV-1 variant originally isolated from an infected infant (37). We found that while treatment with AMD3100 strongly inhibits infection of CD8+ cells, it has no effect on infection of CD4+ cells.
93US143-T8, the HIV-1 isolate used in this study, came from the bulk viruses of a newborn patient, 93US143 (obtained from Merlin Robb through the AIDS Research and Reagents Program, National Institutes of Health [NIH]). The CD4-independent 93US143-T8 strain was isolated from the viral quasispecies of this newborn by stringent selection of infected CD8+ lymphocytes in short-term culture as described recently (37). The X4 laboratory isolate IIIB, the R5 virus JR-FL, and the parental 93US143 isolate were used as controls for these experiments.
Purified CD4+ and CD8+ cells, as well as CD4+ CXCR4+ MT-2 and CEM T-cell lines, were used for infection and inhibition studies (34, 37). For infection of primary CD4+ and CD8+ lymphocytes, cells from four different healthy donors were used in separate experiments. Briefly, purified CD4+ and CD8+ cells or T-cell lines were infected with either the 93US143-T8, 93US143, IIIB, or JR-FL isolate for 2 h. Although different multiplicities of infection of the viruses were used at various points to check the function of AMD3100, all experiments shown in the results used viruses equivalent to a multiplicity of infection of 0.02. For infection of primary cells, peripheral blood lymphocytes from a single donor were separated for each experiment by negative selection in CD4+ and CD8+ lymphocytes. Purified CD4+ and CD8+ cells were stimulated by phytohemagglutinin (1 μg/ml) for 2 to 3 days, and the CD4+ and CD8+ cells were always analyzed for purity before infection with different viruses (33). After infection, the cells were continuously followed for the levels of CD4 and CD8 expression by double-color fluorescence-activated cell sorter (FACS) analysis at regular intervals as described below. To block the CXCR4, cells were treated for 30 min with a high concentration (1.2 μg/ml) of AMD3100 (a generous gift from J. Moore, Cornell University) prior to infection, and the cultures were continued in the presence of the drug. In some experiments, blocking anti-CXCR4 antibody (12G5) was used to test the role of CXCR4, as previously described (33). The role of CCR5 in the infection of CD4+ cells was also examined in some experiments by using TAK-779, a CCR5-specific antagonist (obtained through the AIDS Research and Reference Program, NIH) either alone or in combination with AMD3100 (33).
Infection and virus replication were monitored at regular intervals for induction of syncytia and production of p24 by enzyme-linked immunosorbent assay, as well as by measurement of virion-associated RNA in the culture supernatants with an AMPLICOR kit (Roche) as described previously (34, 37). In addition, the syncytium-inducing (SI) or non-SI phenotype and the level of infectious particles of the replicating viruses in the presence or absence of AMD3100 were examined by detection of syncytia and by measuring the 50% tissue culture infective dose (TCID50), respectively, in MT-2 cells following standard protocols (33).
Apoptosis and death of the infected CD4+ and CD8+ cells were monitored by trypan blue exclusion, as well as propidium iodide staining (33). Changes in the cell surface expression of CD4, CD8, CXCR4, and CCR5 were followed at frequent intervals by using labeled monoclonal antibodies for single- or double-color FACS analysis, as previously described (30-32). Genotypic populations in the replicating viruses in the presence or absence of AMD3100 were compared using a V1-V2-specific heteroduplex tracking assay (HTA) and a V3-specific HTA, as previously described, using probes from the JR-FL molecular clone (20, 27).
Although it is CD4 independent, the 93US143-T8 isolate is known to efficiently infect both CD4+ and CD8+ cells. In addition, 93US143-T8, as well as its parental 93US143 isolate, forms syncytia in T-cell lines, indicative of the SI phenotype (33). To determine whether the 93US143-T8 isolate used CXCR4 for cellular entry, we tested the ability of AMD3100 to inhibit infection of purified CD4+ and CD8+ cells isolated from healthy donors. As shown in a representative experiment in Fig. Fig.1A1A with cells from a single healthy donor, infection of purified CD4+ cells by the X4 IIIB isolate was completely abolished by treatment with AMD3100, which had absolutely no inhibitory effect on the replication of the 93US143-T8 isolate. In contrast, infection of purified CD8+ cells from the same donor by the 93US143-T8 isolate was strongly blocked by treatment with AMD3100 (Fig. (Fig.1B).1B). As expected, the IIIB isolate did not replicate in CD8+ cells to detectable levels (Fig. (Fig.1B).1B). As reported before (37), in addition to efficient replication in CD4+ cells (Fig. (Fig.1A),1A), the parental 93US143 isolate was also able to replicate in CD8+ cells, albeit to a much lower level than its replication in CD4+ cells (Fig. (Fig.1B).1B). Replication of the 93US143 isolate in CD8+ cells is likely to be due to the reduced amounts of 93US143-T8 variants present in the parental viral strains. We have previously shown that the CD4-independent 93US143-T8 variant constitutes a significant fraction of the viral quasispecies in the parental 93US143 isolate (37). Although the replication kinetics of these viruses in CD4+ and CD8+ cells from different donors varied considerably, the same pattern of replication and AMD3100 effects were seen in the CD4+ and CD8+ cells from each donor. Blocking antibody against CXCR4 (12G5) also failed to inhibit infection of CD4+ cells by the 93US143-T8 isolate, indicating that the failure to prevent infection is not specific to AMD3100 (data not shown).
The level of infection in the presence or absence of AMD 3100 was also monitored by the measurement of cell-free viral RNA (viral load) in the CD4+ and CD8+ culture supernatants using the AMPLICOR assay. Although the HIV RNA copy number may not necessarily correlate with the level of p24 production (8), the levels of viral RNA present in the culture supernatants in these experiments by and large matched the level of p24 production (Fig. 1A and B). While replication of the 93US143-T8 isolate in CD8+ cells was strongly inhibited (>90%) by treatment with AMD3100, as evident from lower viral-RNA copy numbers, similar treatment of the CD4+ cells actually increased the viral-RNA levels by more than twofold (Table (Table1).1). It may be noted that we observed enhanced levels of p24 production by the 93US143-T8-infected CD4+ cells in the presence of AMD3100 compared to the absence of the drug in most cases (Fig. (Fig.1A1A and data not shown). The reason for this enhanced replication of the 93US143-T8 isolate in CD4+ cells in the presence of AMD3100 remains unclear at present. Furthermore, the infectivities of viruses present in the culture supernatants of the AMD3100-treated or untreated cultures, as measured by the determination of the TCID50s of the culture supernatants, also generally agreed with the p24 concentration and viral-RNA copy number (Table (Table11).
Acute infection by HIV-1 usually induces death of the infected cells. The effects of AMD3100 treatment on the viability of the infected CD4+ and CD8+ cells were also closely monitored. While both IIIB and 93US143-T8 isolates induced death of the infected CD4+ cells, treatment with AMD3100 prevented the cytopathic effects induced by the IIIB isolate (Fig. (Fig.2).2). In contrast, AMD3100 treatment had no effect on the death of CD4+ cells infected with the 93US143-T8 isolate. Indeed, the 93US143-T8 isolate often induced a slightly higher rate of death of the infected CD4+ cells in the presence of AMD3100, which is in agreement with the results of higher replication (Fig. (Fig.1A)1A) and higher viral-RNA copy numbers (Table (Table1).1). However, acute death of the infected CD8+ cells as induced by the 93US143-T8 isolate was strongly inhibited by treatment with AMD3100, indicating a block in virus infection. IIIB infection of the CD8+ cells, with or without AMD3100 treatment, had no appreciable cytopathic effects (Fig. (Fig.2).2). Taken together, these data demonstrate that while the 93US143-T8 isolate can efficiently infect CD4+ cells independently of CXCR4, CXCR4 is essential for infection of CD8+ cells by this isolate.
Primary CD4+ T lymphocytes can be infected through either the CCR5 or CXCR4 coreceptor by dual-tropic (R5X4) HIV-1 isolates (13). It has been shown that the 93US143-T8 isolate is capable of using either CCR5 or CXCR4 as a coreceptor in U87 cells expressing CD4 molecules (37). Although CXCR4 is abundantly expressed on most primary T cells, as discussed above, expression of CCR5 is limited to a very small fraction of the CD4+ T cells. Others have also reported that CCR5 expression is reduced even more in phytohemagglutinin- and interleukin-2-stimulated CD4+ cells, further affecting infection by the R5 strains (3). Since the 93US143-T8 isolate was able to infect CD4+ lymphocytes independently of CXCR4 (Fig. (Fig.1A),1A), we examined the role of the CCR5 coreceptors in the infection of CD4+ cells by using TAK779, a known antagonist for CCR5 (14). As expected, TAK779 completely blocked infection of the R5-tropic JR-FL isolate. In contrast, TAK779 had no inhibitory effects on infection by the 93US143-T8 isolate. However, a somewhat lower level of replication of the 93US143-T8 isolate was observed when TAK779 and the CXCR4 antagonist AMD3100 (as discussed above) were used together (Fig. (Fig.1C).1C). Taken together, these results indicate that even though the 93US143-T8 variant is able to use the CCR5 and CXCR4 coreceptors for infection of CD4+ cells, it appears that this isolate is probably also capable of using other molecules (coreceptors) that are expressed on the CD4+ cells for viral entry.
It has been reported that R5 strains may be selectively enhanced when mixed (R5- and X-tropic) HIV-1 quasispecies are cultured in the presence of AMD3100 (15). To determine whether a minor CXCR4-independent variant may have been selected during infection of the CD4+ cells by the 93US143-T8 isolate in the presence of AMD3100, we used a V1-V2-specific, as well as a V3-specific, HTA to compare the viral variants in the different supernatants of the infected cells. We showed previously that the V1-V2 region of the envelope gene had the most variation in the 93US143 and 93US143-T8 isolates (37), making this region useful for determining whether the virus produced by the infected CD4+ cells was the same as the input virus. As shown in Fig. Fig.3,3, the same dominant sequence variant and most of the minor variants found in the initial 93US143-T8 isolate replicated in CD4+ cells with or without AMD3100. In addition, the same pattern of variants was also found in the viruses produced by the infected CD8+ cells. Because alterations in the V3 region are commonly linked with coreceptor usage changes, we also used a V3-specific HTA to determine whether any sequence changes occurred in these cultures. The V3 sequences also showed no changes in any of the 93US143-T8-infected cultures under different conditions (Fig. (Fig.3).3). For both the V1-V2 and V3 HTAs, heteroduplexes have significant shifts, which increase their sensitivity to point mutations (27). Therefore, at least by these measures, it appears that there was no selection of any minor variant of the 93US143-T8 isolate for infection of CD4+ cells in the presence of AMD3100. It may also be noted that the 93US143-T8 isolate that replicated in the presence of AMD3100 maintained a high degree of infectivity, as measured by induction of syncytia and TCID50 in MT-2 cells (Table (Table1),1), suggesting that the SI phenotype of this isolate also remained unaffected even in the presence of AMD3100.
FACS analysis was performed to determine whether infection of CD4+ and CD8+ cells in the presence or absence of AMD3100 could affect cellular expression of CD4, CD8, CXCR4, or CCR5 molecules that may influence entry of the 93US143-T8 isolate. As previously observed (37), down modulation of CD4 expression was noticeable after infection of the CD4+ cells with either the IIIB or 93US143-T8 isolate in the absence of AMD3100 (data not shown). However, in the presence of AMD3100, a complete reversal of the down-modulated CD4 expression was noticeable in the IIIB-infected CD4+ cells, indicating prevention of infection. In contrast, expression of CD4 remained down modulated even in the presence of AMD3100 after infection of the CD4+ cells with the 93US143-T8 isolate (Fig. (Fig.4).4). In fact, the level of CD4 down modulation after infection with the 93US143-T8 isolate was frequently more pronounced in the presence of AMD3100 (mean fluorescence, 3.7 [Fig. [Fig.4])4]) than in its absence (mean fluorescence, 8.2 [data not shown]) at the same time point, further supporting the idea that there may have been enhanced infection of the AMD3100-treated CD4+ cells by the 93US143-T8 isolate (data not shown). Although some degree of down modulation of CXCR4 was also noticeable after infection of the CD4+ cells by either the IIIB or 93US143-T8 isolate in the absence of AMD3100, no change in the expression of CCR5 was observed in the small number of CD4+ CCR5+ cells in the presence or absence of AMD3100 (data not shown).
Also, no significant change in the level of CD8 expression was observed after infection of CD8+ cells by either the 93US143-T8 (Fig. (Fig.5A)5A) or IIIB (Fig. (Fig.5B)5B) isolate compared to the uninfected control cells (Fig. (Fig.5C).5C). Interestingly, while most CD8+ cells expressing high levels of CXCR4 (Fig. (Fig.5C)5C) did not change after IIIB infection (Fig. (Fig.5B),5B), down modulation of CXCR4 expression was noticeable in a larger proportion of the CD8+ cells after infection with the 93US143-T8 isolate (Fig. (Fig.5A).5A). Only a small number (usually <2 to 3%) of cells that coexpressed CD4 molecules were present in both the IIIB- and 93US143-T8-infected CD8+-cell cultures, indicating that the production of virus from the CD8+ lymphocytes was primarily due to the replication of the 93US143-T8 isolate in the CD8+ cells (data not shown). The down regulation of CXCR4 expression in CD8+ cells after infection with the 93US143-T8 isolate coincided with productive virus replication in these cells (Fig. (Fig.1B).1B). The effects of AMD3100 treatment on the level of CXCR4 expression could not be assessed because such treatment rendered the cells unsuitable for FACS analysis of the CXCR4 coreceptors. Only a small percentage of the CD8+ cells expressed low levels of CCR5, and we did not find any appreciable changes in the expression of CCR5 whether untreated or after treatment with AMD3100 (data not shown). Taken together, these results suggest a direct role of CXCR4 molecules in the infection of CD8+ cells, but not CD4+ cells, by the 93US143-T8 isolate.
To further explore the CXCR4-independent infection of the CD4+ lymphocytes by the 93US143-T8 isolate, we used MT-2 and CEM T-cell lines that are known to express CD4 and CXCR4 but not CCR5 and are thus resistant to infection with R5 viruses (25). As expected, the X4 IIIB isolate readily infected both MT-2 (Fig. (Fig.6A)6A) and CEM (Fig. (Fig.6B)6B) cells, and these infections were almost totally abolished by treatment with AMD3100. In contrast, although the 93US143-T8 isolate infected both MT-2 (Fig. (Fig.6A)6A) and CEM (Fig. (Fig.6B)6B) cells, like the primary CD4+ cells, treatment with AMD3100 had no inhibitory effects on infection of either the MT-2 or CEM cells. Infection of MT-2 and CEM cells by the 93US143-T8 isolate also induced syncytia and caused acute death of the infected cells even in the presence of AMD3100 (data not shown). The R5 virus JR-FL failed to infect these cells in the absence or presence of AMD3100, further suggesting that CCR5 played no role in the infection of the MT-2 or CEM cells (data not shown).
Changing cellular conditions may modulate expression of coreceptors and affect HIV-1 infection (13). We did not detect any induction of CCR5 in the MT-2 or CEM cells even after treatment with AMD3100 (data not shown). Furthermore, TAK779 also failed to produce any inhibitory effects on the infection of MT-2 or CEM cells by the 93US143-T8 isolate, indicating that infection of these cells by the 93US143-T8 isolate was independent of CCR5 (data not shown). Mutations in V1-V2 and V3 are commonly associated with a change in coreceptor usage by HIV-1 (13). The V1-V2- and V3-specific HTA analyses of the viral RNAs from supernatants of infected cells also showed that these highly variable regions of the replicating 93US143-T8 isolate remained unchanged by the AMD3100 treatment (Fig. (Fig.3).3). Infection of MT-2 and CEM cells by both IIIB and 93US143-T8 isolates induced down modulation of CD4 expression, but as seen with the primary CD4+ cells (Fig. (Fig.4),4), while CD4 down modulation by the IIIB isolate was reversed after treatment with AMD3100, similar treatment had no effect on the down modulation of CD4 induced by the 93US143-T8 isolate (data not shown). However, there was visible down modulation of CXCR4 expression after infection of MT-2 (Fig. (Fig.7)7) and CEM (data not shown) cells by both the IIIB and 93US143-T8 isolates, suggesting the possible use of CXCR4 by both of these isolates in the absence of AMD3100. Taking these data together, it appears that the 93US143-T8 isolate probably used an alternative route for infection of the MT-2 and CEM cells in the presence of AMD3100, although we cannot completely rule out the possibility that the isolate has an enhanced ability to use CXCR4 even in the presence of AMD3100.
We previously reported CD4-independent infection of CD8+ lymphocytes by 93US143-T8, a primary HIV-1 isolate that was selected from the bulk viral quasispecies of a newborn. In this study, we show that the CXCR4 antagonist AMD3100 blocks the infection of CD8+ cells by the 93US143-T8 isolate, suggesting that infection of CD8+ cells is mediated through CXCR4. However, infection of CD4+ cells (both primary cells and T-cell lines) by the 93US143-T8 isolate is insensitive to AMD3100, indicating that CXCR4 is not essential for infection of CD4+ cells or that the virus binds CXCR4 on CD4+ cells in a manner that is not sensitive to AMD3100 blockage. These results provide evidence for the first time of CXCR4-dependent infection of CD8+ cells by a primary HIV-1 isolate. These data also show that the CD4+ and CD8+ cells in an individual may be infected through different mechanisms by the same HIV-1 isolate.
AMD3100 has recently emerged as a useful agent against a wide range of X4 and R5X4 viruses (15, 21, 38). In view of its extremely potent inhibition of the X4 isolates, it has been proposed that this drug may be used even in asymptomatic patients (15). Indeed, AMD3100 is already in phase II clinical trials for potential use as a new entry inhibitor against HIV-1 (18). We found that although AMD3100 was able to prevent infection of CD8+ lymphocytes by the 93US143-T8 isolate, treatment of CD4+ lymphocytes with AMD3100 had no inhibitory effect on infection by the same isolate. Indeed, we often observed some degree of enhanced replication of the 93US143-T8 isolate in CD4+ cells in the presence of AMD3100 (Fig. (Fig.1A1A and Table Table1).1). Thus, the use of AMD3100 in patients harboring similar strains of HIV-1 variants may prove to be ineffective and could even be potentially harmful. A recent study has also shown that the sensitivities of primary X4 isolates to AMD3100 inhibition can change over time (36). Therefore, careful evaluation of different aspects HIV-1 infection is essential before the start of random use of AMD3100 as a new entry inhibitor in the treatment of HIV-1 infections.
CXCR4 is expressed on most cells of the immune system and at a relatively high level on CD8+ (Fig. (Fig.5C)5C) and CD4+ (data not shown) T lymphocytes. It is well known that, along with the CD4 receptor, CXCR4 also plays a major role as a coreceptor in HIV-1 infection of primary CD4+ T cells and CD4+-T-cell lines. We have previously shown that the 93US143-T8 isolate can infect CD8+ cells independently of CD4 (37). Here, we provide evidence that infection of the CD8+ cells by the 93US143-T8 isolate is mediated through CXCR4. Expression of CXCR4 was down modulated in the CD8+ cells after infection with the 93US143-T8 isolate (Fig. (Fig.5),5), and blocking of CXCR4 prevented infection of the CD8+ lymphocytes (Fig. (Fig.1B),1B), indicating a direct role of CXCR4 molecules in the infection of these cells. In addition, cloned envelope of the 93US143-T8 isolate was able to mediate fusion with CD4-negative quail cells using CXCR4 alone (unpublished observation; R. Doms, University of Pennsylvania, personal communication), suggesting that the 93US43-T8 isolate is capable of entering CD4-negative cells by using CXCR4. CXCR4-mediated infection of CD4-negative cells by laboratory strains of HIV has been reported (10, 29). It is well recognized that stimulation of primary CD8+ cells may induce coexpression of CD4 molecules in a small percentage of cells, and some degree of HIV-1 infection may then occur through CD4 in these double-positive (CD8+ CD4+) cells (19). A small number of double-positive cells present in the purified CD8+ population have also been observed (data not shown) (33). However, it is clear that infection of the purified CD8+ cells by the 93US143-T8 isolate was not related to the small number of CD4+ CD8+ cells that also expressed CXCR4 molecules. First, the presence of a small number (usually 2 to 3%) of contaminating CD4+ cells in the CD8+-cell population is not likely to result in the high level of virus replication that was usually associated with the 93US143-T8 isolate (Fig. (Fig.1B).1B). Second, although like the 93US143-T8 isolate, the parental 93US143 or the X4-tropic IIIB isolate replicated to high levels in CD4+ cells, these isolates replicated either to a low level (for the 93US143 isolate) or not at all (for the IIIB isolate) in CD8+ cells (Fig. 1A and B), indicating that infection of the CD8+ cells by the 93US143-T8 isolate could not be due to the contaminating CD4+ CXCR4+ cells. Also, AMD3100 did not show any inhibitory action on infection of the CD4+ cells by the 93US143-T8 isolate (Fig. (Fig.1A).1A). Thus, failure of virus replication by the AMD3100-treated CD8+ cells provides further evidence that production of the 93US143-T8 isolate in the CD8+ cells could not be due to its replication in the CD4+ lymphocytes. Furthermore, treatment of the purified CD8+ cells with anti-CD4 antibodies (or soluble CD4) either alone or in combination with AMD3100 had no effect on infection with the 93US143-T8 isolate, indicating that CD4 had no role in CD8+-cell infection by the 93US143-T8 isolate (reference 37 and data not shown). Whether memory and naïve CD8+ cells may be differentially susceptible to infection in vivo with CD4-independent viruses like the 93US143-T8 isolate remains unknown. However, we did not detect any changes in virus replication between the memory and naïve CD8+ lymphocytes after in vitro infection with the 93US143-T8 isolate (data not shown). In the absence of CD4, the exact role of CXCR4 in the infection of CD8+ lymphocytes remains unclear at present. Further studies would be necessary to determine whether CXCR4 acted as a primary receptor for the 93US143-T8 isolate for infection of the CD8+ cells. Although it appears that the 93US143-T8 isolate is able to use the CXCR4 present on the CD4+ cells (Fig. (Fig.7),7), our results also indicate that CXCR4 may not be absolutely essential for infection of the CD4+ cells by this isolate.
Genetic defects in host cell coreceptor expression can influence infection by different HIV strains (24). The contrasting effects of AMD3100 on the CD4+ and CD8+ cells from the same donor found in this study suggested that the differential replication of the 93US143-T8 isolate in these two T-cell subtypes was not due to faulty genetic disposition of the coreceptors. Adaptation of HIV-1 may occur under selective pressure from the blockade of specific cellular coreceptors (1). However, the blockade of CCR5 with TAK 779, a specific CCR5-blocking agent, had no effect on infection of CD4+ cells by the 93US143-T8 isolate (Fig. (Fig.1C).1C). Furthermore, the lack of any inhibitory effects of AMD3100 on the infection of CCR5-negative MT-2 (Fig. (Fig.6A)6A) and CEM (Fig. (Fig.6B)6B) cells argues against the use of CCR5 by the 93US143-T8 isolate. Even a combination of AMD3100 and TAK 779 produced little inhibitory effect on the infection of primary CD4+ cells (Fig. (Fig.1C),1C), suggesting that apart from CXCR4 and CCR5, the 93US143-T8 isolate may be able to use other molecules for viral entry.
AMD3100 at concentrations of 100 to 400 ng/ml showed strong inhibitory effects against X4 and R5X4 strains in different studies (12, 15, 22, 28, 38, 39). We found no evidence of inhibition of infection by the 93US143-T8 isolate of CD4+ cells when AMD3100 was used at a much higher concentration (1.2 μg/ml). It has been suggested that AMD3100-resistant viruses with reduced viral fitness may emerge by sequential passages of some X4 HIV-1 strains in the presence of AMD3100 (4). The 93US143-T8 isolate was able to replicate efficiently in the presence of AMD3100 without any prior sensitization with the drug. Indeed, the 93US143-T8 viruses that were produced in the presence of AMD3100 were as “fit” as the original isolate, as tested by infectivity assays, further suggesting that the 93US143-T8 isolate is naturally resistant to AMD3100 (Table (Table1).1). AMD3100 acts by selective antagonism of CXCR4, and it generally does not trigger an intracellular signal by itself (5, 11). We have not observed any evidence of an altered activation status or differential expression of CCR5 or CXCR4 after treatment with AMD3100 (data not shown). However, the possibility that expression of other as-yet-unknown receptors and coreceptors was modulated after AMD3100 treatment of the CD4+ cells that in turn may have enhanced infection by the 93US143-T8 isolate cannot be ruled out.
It is becoming increasingly apparent that the cellular immunity mediated primarily by CD8+ lymphocytes could play a critical role in the development of an effective vaccine against HIV (2, 9, 23, 35). The results of this study demonstrate that HIV-1 may induce CXCR4-dependent infection of CD8+ lymphocytes. Future studies with HIV-1 variants capable of infecting CD8+ lymphocytes may shed more light on HIV-1 pathogenesis and may help in designing better vaccine candidates. We have also shown that the SI-X4 strain of HIV-1 is capable of infecting primary CD4+ cells, as well as CD4+-T-cell lines, independently of CXCR4, indicating the possibility of a new mode for viral pathogenesis. Finally, this study provides evidence that AMD3100, a new entry inhibitor for HIV-1 that is already in advanced clinical trials, may not be useful in some patients harboring X4 isolates.
This work is supported by NIH grants AI-42715 and AI-44974 and a grant from the American Foundation for AIDS Research to K.S.
We thank Robert Doms of the University of Pennsylvania for his study of viral entry and James Mullins of the University of Washington and Manuel Caruso of Laval University for helpful suggestions. We also thank C. McAllister for FACS analyses; Tara Riddle, Jianchao Zhang, Amy Traven, Yvette Arundel, Sankosh Sidhu, David Carr, Victoria Best, and Raquel Raices for technical assistance; and Amy Dutcher and Karen Watkins for their excellent help in preparation of the manuscript.