|Home | About | Journals | Submit | Contact Us | Français|
Coreceptor specificity of human immunodeficiency virus type 1 (HIV-1) strains is generally defined in vitro in cell lines expressing CCR5 or CXCR4, but lymphocytes and macrophages are the principal targets in vivo. CCR5-using (R5) variants dominate early in infection, but strains that use CXCR4 emerge later in a substantial minority of subjects. Many or most CXCR4-using variants can use both CXCR4 and CCR5 (R5X4), but the pathways that are actually used to cause infection in primary cells and in vivo are unknown. We examined several R5X4 prototype and primary isolates and found that they all were largely or completely restricted to CXCR4-mediated entry in primary lymphocytes, even though lymphocytes are permissive for CCR5-mediated entry by R5 strains. In contrast, in primary macrophages R5X4 isolates used both CCR5 and CXCR4. The R5X4 strains were also more sensitive than R5 strains to CCR5 blocking, suggesting that interactions between the R5X4 strains and CCR5 are less efficient. These results indicate that coreceptor phenotyping in transformed cells does not necessarily predict utilization in primary cells, that variability exists among HIV-1 isolates in the ability to use CCR5 expressed on lymphocytes, and that many or most strains characterized as R5X4 are functionally X4 in primary lymphocytes. Less efficient interactions between R5X4 strains and CCR5 may be responsible for the inability to use CCR5 on lymphocytes, which express relatively low CCR5 levels. Since isolates that acquire CXCR4 utilization retain the capacity to use CCR5 on macrophages despite their inability to use it on lymphocytes, these results also raise the possibility that a CCR5-mediated macrophage reservoir is required for sustained infection in vivo.
CD4+ T cells and macrophages are the two principal targets for productive human immunodeficiency virus type 1 (HIV-1) infection in vivo. Viral entry is initiated by the viral envelope glycoprotein gp120 binding to cell surface CD4, followed by interactions with a chemokine receptor. All naturally occurring strains of HIV-1 use CCR5 or CXCR4 or both, and both coreceptors are expressed by primary CD4+ T lymphocytes and macrophages. Early in infection, viral variants are almost uniformly CCR5 dependent (R5 variants), but CXCR4-using variants emerge later in a substantial minority of infected people, and their appearance is strongly associated with accelerated disease progression (7, 23). In most of these individuals, the late-stage variants retain the capacity to use both CCR5 and CXCR4 (R5X4), although some may harbor only CXCR4-dependent (X4) variants and others may harbor a mix of R5 and X4 variants (7, 23, 28).
Chemokine receptor selectivity of HIV-1 isolates is typically characterized on the basis of cell-cell fusion or infection of CD4-expressing cell lines transfected with CCR5 or CXCR4. CXCR4 utilization can also be defined based on the syncytia-inducing (SI) capacity in MT-2 cells, which reflects CXCR4-dependent infection, although this method cannot distinguish between X4 and R5X4 variants. While viral phenotyping carried out by using cell lines with defined coreceptor profiles can indicated the pathway that a virus is capable of using, it may not necessarily predict which pathway a virus will actually use in primary cells or in vivo. Indeed, primary macrophages express both CXCR4 and CCR5, but many X4 strains are unable to utilize macrophage CXCR4 for infection (27, 31, 35, 36). In contrast, it is not known whether similar virus- and cell-specific restrictions exist for coreceptor utilization in primary lymphocytes, largely because the vast majority of HIV-1 strains are isolated by coculture in lymphocytes, and so any virus that does not use lymphocyte coreceptors at all would likely not be identified.
In this study we employed a panel of R5X4 strains to address two related questions relevant to chemokine receptor utilization in primary lymphocytes: (i) among strains that have the capacity to use CCR5, are there differences in the ability to use it on primary target cells, and (ii) are R5X4 viral variants equally dependent on these two entry pathways for infection of primary target cells? Our results show that there exists biological diversity in primary lymphocyte coreceptor utilization among CCR5-using HIV-1 isolates, that many R5X4 strains behave as single coreceptor X4 strains in primary lymphocytes, and that some R5X4 strains utilize distinct entry pathways in primary lymphocytes and macrophages.
Peripheral blood mononuclear cells were obtained by Ficol-Hypaque separation from whole blood of normal volunteers. Donors were screened by PCR for the CCR5Δ32 deletion allele (22), and only those homozygous for the wild-type allele were utilized. Lymphocytes and monocytes were separated by selective adherence and maintained as previously described (6). Nonadherent peripheral blood lymphocytes (PBL) were removed and stimulated with phytohemagglutinin (PHA) for 3 days prior to infection and maintained with interleukin-2 (IL-2) thereafter. CD8-depleted cells were prepared by immunomagnetic depletion (Dynal Biotech, Brown Deer, Wis.). Adherent monocytes were maintained in culture for 7 days to allow differentiation into monocyte-derived macrophages (MDM) prior to infection. The U87/CD4, U87/CD4/CCR5, and U87/CD4/CXCR4 cell lines were obtained from D. Littman (3) through the National Institutes of Health (NIH) AIDS Reference and Reagent Program.
The prototype strains DH12, 89.6, and NL43 were generated by transfection of molecular clones. Primary HIV-1 isolates 93BR020, 92HT594, and 96USHIPS9 were obtained from the NIH AIDS Reagent Program. The R5 primary isolate BL-2 and X4 primary isolate Tybe have been described previously (22, 34). Viruses were grown in PHA- and IL-2-stimulated PBL and clarified by centrifugation, and titers were determined by p24 antigen content. For experiments involving PCR amplification of infected cells, viral stocks were treated with DNase (50 U/ml) for 30 min prior to infection.
PBL were suspended at a density of 106 cells per ml prior to infection, stimulated for 3 days with PHA, and maintained with IL-2 thereafter. MDM were plated at 1.5 × 105 cells per well in 48-well plates and maintained in culture for 7 days prior to infection as described (6). U87 cells were seeded at a density of 105 cells per well in 24-well plates and infected the following day. Cells were infected overnight by using 10 ng of p24 antigen of each virus, washed three times, and fed with fresh medium; the supernatant was sampled periodically for p24 antigen production by enzyme-linked immunosorbent assay (Coulter Corp., Miami, Fla.). For PCR detection of viral entry, cells were exposed to DNase-treated viral inoculum, and total cellular DNA was extracted 2 days later. For blocking studies, cells were incubated with blocking agent for 1 h prior to infection, and the inhibitor was maintained throughout the course of the experiment. The CXCR4 inhibitor T22 (20) was a kind gift of N. Fujii (Kyoto University, Kyoto, Japan). The CCR5 inhibitor Merck657 (M657) (10) was a kind gift of M. Miller (Merck & Co., West Point, Pa.). The CD4-blocking monoclonal antibody (MAb) 19 (36) was kindly provided by J. Hoxie (University of Pennsylvania).
Cells were washed, lysed in DNA lysis buffer (100 mM KCl, 0.1% NP-40, 20 mM Tris [pH 8.4], 0.5 mg of proteinase K per ml), incubated at 55°C for 2 h, and then boiled for 15 min. For each reaction, 1.5 μl of DNA lysate was added per 25 μl of reaction mixture containing a 0.25 mM concentration of each deoxynucleoside triphosphate, 5 mM MgCl2, 50 pm of each primer, and 10 pm of molecular beacon probe, along with 1 U of AmpliTaq Gold in PCR Buffer II (Applied Biosystems, Foster City, Calif.). Primer and probe sets detected the HIV-1 long terminal repeat (LTR) (forward primer, 5′-GCT AGC TAG GAA ACC CAC TGC TTA-3′; reverse primer, 5′-GCT AGA GAT TTT CCA CAC TGA CT-3′; probe, FAM-5′-GCG AGT CAC ACA ACA GAC GGG CAC ACA CTA CTC GC-3′-DABCYL) and the cellular GAPDH [glyceraldehyde-3-phosphate dehydrogenase] gene (forward primer, 5′-GGT GGT CTC CTC TGA CTT CAA CA-3′; reverse primer, 5′-CCA GCC ACA TAC CAG GAA ATG-3′; probe, FAM-5′-CGC AGC CTG GCA TTG CCC TCA ACG ACC ACG CTG CG-3′-DABCYL). Amplification was carried out on an ABI 7700 real-time PCR detection system with an initial incubation at 95°C for 10 min followed by 40 cycles of 95°C for 15 sec and 60°C for 1 min. For quantification, DNA from serial dilutions of ACH2 cells over the range of 101 to 105 was amplified in parallel. Data were analyzed by using 7000 SDS Software (Applied Biosystems).
HIV-1 coreceptor specificity is typically established in transfected cell systems, but primary lymphocytes and macrophages are the main target cells for infection in vivo. In macrophages many strains are restricted in their ability to use CXCR4, but strain- and cell-specific restrictions to coreceptor utilization have not been addressed in primary lymphocytes, despite their central role in pathogenesis. One reason, of course, is that most HIV-1 strains are isolated by culture in PBL; so for single coreceptor-specific strains, it would not be possible to find a strain that does not use that entry pathway in lymphocytes even if it were to exist in vivo. For that reason, we chose to examine dual-tropic R5X4 strains, which would have at least one pathway for lymphocyte infection but would enable evaluation of each coreceptor independently.
We first focused on two well-defined R5X4 prototypes, 89.6 and DH12, and tested the pathways they use for productive infection in primary human PBL (Fig. (Fig.1).1). To define the role of CXCR4, we used the specific antagonist T22. In parallel we used the well-defined R5 prototype Bal and X4 prototype NL43. As shown in Fig. Fig.1A,1A, PBL infection by NL43 was almost completely blocked by the CXCR4 antagonist T22, as expected for a CXCR4-dependent strain. In contrast, Bal was unaffected by CXCR4 blockade, which was expected for an R5 strain; the result also showed that the blocking agent had no nonspecific or toxic effects that might nonspecifically inhibit viral replication. Surprisingly, PBL infection by the R5X4 prototype 89.6 was also almost completely blocked by T22. DH12 was also largely blocked by CXCR4 inhibition (≥20-fold), although not quite as completely as 89.6. DH12 also showed modest variability among donors, with nearly complete blocking in about three-quarters of PBL cultures and partial blocking in the remainder (data not shown). We saw identical results when we used AMD3100 (24) to block CXCR4 (data not shown). These results suggest that 89.6 uses only CXCR4 for infection of PBL, which differs markedly from infection of cell lines and primary macrophages, where 89.6 uses both CCR5 and CXCR4 (9, 36). Furthermore, DH12 also is largely CXCR4 restricted in lymphocytes.
Since selective coreceptor use should restrict infection at entry, we then examined entry into PBL by using quantitative PCR to detect early reverse transcription products 48 h after infection (Fig. (Fig.1B).1B). T22 was used to block CXCR4, and primer and probe sets were used that detect a very early LTR reverse transcription product. In these short time course experiments, we also used the specific CCR5 inhibitor M657 to confirm that CCR5 was functional for entry by R5 strains. In addition, a CD4-blocking MAb was used as a control to inhibit both binding and entry. Entry by the X4 prototype NL43 was inhibited by CXCR4 blocking and unaffected by CCR5 inhibition. In contrast, the R5 prototype Bal was unaffected by T22 but was blocked by M657. We also examined the R5 primary isolate BL-2 and also found no effect of T22 but complete blocking by M657. Consistent with the productive infection studies, entry by the R5X4 prototypes 89.6 and DH12 was blocked by the CXCR4 antagonist as efficiently as NL43, but no restriction to entry could be detected after CCR5 blocking. Furthermore, for neither of these R5X4 viruses did the combination of both CCR5 and CXCR4 blocking result in a major or consistent reduction in entry compared with CXCR4 blocking alone. These data confirm that the R5X4 prototype 89.6 and DH12 strains use CXCR4 only to enter and infect PBL, even though CCR5 is expressed in a manner that is effectively utilized by prototype R5 strains.
We next tested whether the exclusive use of CXCR4 to infect PBL was restricted to prototype R5X4 strains or was a feature of R5X4 primary isolates as well. We chose several primary isolates from among R5X4 strains available in the NIH AIDS Reagent Repository. To ensure that the virus stocks used in these experiments exhibited the R5X4 phenotype, we first confirmed their coreceptor usage patterns in U87 cells expressing CD4 alone, CD4 plus CXCR4, or CD4 plus CCR5 (Fig. (Fig.2).2). As expected, NL43 used CXCR4 only and Bal used CCR5 only. The three primary isolates 93BR020, 92HT594, and 96USHIPS9 were able to enter U87/CD4 cells through both CCR5 and CXCR4, consistent with their reported R5X4 phenotypes.
We then determined which pathways were used by these primary isolates for entry and infection of lymphocytes. As shown in Fig. Fig.3A,3A, infection by the R5X4 primary isolates was markedly inhibited by CXCR4 blockade. Two of the three strains resembled 89.6 and were essentially fully blocked by T22, while one was more similar to DH12 and was largely (≥20-fold), albeit not completely, blocked. We next examined entry by using quantitative PCR in conjunction with both CXCR4 and CCR5 blockade. To be sure that these strains were sensitive to the CCR5 inhibitor, we determined whether M657 would inhibit entry into U87/CD4/CCR5 cells (Fig. (Fig.2).2). M657 blocked entry by each of these R5X4 strains in CCR5-expressing cells but not in CXCR4-expressing cells, indicating that their interactions with CCR5 are sensitive to this agent. When M657 was tested in primary lymphocytes (Fig. (Fig.3B),3B), entry by all three primary isolates was completely inhibited by the CXCR4 antagonist, but CCR5 blocking had no effect even though it efficiently blocked Bal. Thus, the complete inhibition by T22 and lack of effect by M657 indicate that entry and infection of lymphocytes by these three R5X4 primary isolates are mediated exclusively through CXCR4 and cannot proceed via CCR5 even if CXCR4 is unavailable.
A potential explanation for the inability of certain isolates to utilize CCR5 on PBL might be the presence of endogenous inhibitors, such as CCR5-blocking β-chemokines that could be produced by CD8 cells within the PBL cultures (5). While the ability of HIV-1 Bal to efficiently use lymphocyte CCR5 indicates that there is no global block to CCR5 coreceptor function, it is possible that some strains might be particularly sensitive to inhibitory effects. To determine whether CD8 cells within the cultures were inhibiting CCR5-mediated entry by R5X4 variants, we evaluated early events following the infection of CD8-depleted PBL cultures. Similar to infection in whole PBL cultures, entry into CD8-depleted lymphocytes by all three R5X4 primary strains as well as the DH12 and 89.6 prototypes was blocked by the CXCR4 antagonist (data not shown). We also found that supernatant from PHA-stimulated PBL, with or without T22, did not inhibit infection by R5X4 strains of U87/CD4/CCR5 cells (data not shown). Thus, the inability of these strains to utilize CCR5 on PBL and their exclusive reliance on CXCR4 are not due to CD8-derived inhibitors that restrict CCR5 coreceptor function.
These experiments show that the R5X4 primary strains use both CCR5 and CXCR4 in transfected cells yet are restricted to CXCR4 in PBL. However, an important question is whether they can use CCR5 for entry in any primary cell type. CCR5 is the major entry pathway in primary macrophages for R5 isolates, but along with others, we have shown that some primary isolates are able to infect macrophages through CXCR4 even though prototype X4 strains cannot (27, 31, 35, 36). We also showed previously that 89.6 and DH12 use both CCR5 or CXCR4 to infect macrophages (35), which stands in contrast to the results shown here indicating exclusive use of CXCR4 on PBL. Therefore, we determined what pathways these R5X4 primary isolates use for entry into primary macrophages (Fig. (Fig.4).4). As expected with an R5 strain, entry by Bal was inhibited by CCR5 blocking but was unaffected by the CXCR4 antagonist. Since NL43 does not use macrophage CXCR4 for entry, to assess macrophage CXCR4 use we chose HIV-1 Tybe, an X4 primary isolate that we recently showed uses CXCR4 only for macrophage infection (34). HIV-1 Tybe entry was inhibited by CXCR4 but not by CCR5 blocking.
When the primary R5X4 isolates were tested in macrophages, neither CCR5 blocking alone nor CXCR4 blocking alone prevented entry (Fig. (Fig.4).4). Inhibition of both pathways together, however, blocked entry into macrophages by these isolates. Similar results were seen for the 89.6 and DH12 prototypes. Thus, primary and prototype R5X4 strains can use both CCR5 and CXCR4 for entry into primary macrophages, in contrast to their coreceptor use in primary lymphocytes.
The inability of R5X4 strains to use lymphocyte CCR5 raised the possibility that they might interact less efficiently with the coreceptor compared with R5 strains. To determine if R5X4 and R5 strains exhibit differential interactions with CCR5, we tested the sensitivity of several prototype and primary isolates of each to the CCR5 blocker M657 on U87/CD4/CCR5 cell targets. As shown in Fig. Fig.5,5, the R5X4 isolates were approximately 10-fold more sensitive to M657 blocking than the R5 strains, suggesting that even in the context of CCR5-overexpressing cell lines, R5X4 isolates are less efficient than R5 strains in their utilization of CCR5-mediated entry pathways.
We show here that several R5X4 primary and prototype HIV-1 strains use CXCR4 but not CCR5 for entry and infection of PBL, even though CCR5 is expressed and functional for entry by single coreceptor R5 strains. In contrast to lymphocytes, these isolates use both CCR5 and CXCR4 in primary macrophages. Thus, these R5X4 isolates actually function as single coreceptor X4 strains in lymphocytes but as dual R5X4 variants in macrophages. Three interrelated conclusions can be drawn from our results. First, R5X4 as defined in traditional coreceptor phenotyping systems does not necessarily mean dual coreceptor use on primary cells relevant to infection in vivo. Second, CCR5 use on primary lymphocytes is both strain specific and cell specific, which is similar to the recognized strain- and cell-specific utilization of CXCR4 use on macrophages. Finally, the inability of R5X4 strains to use CCR5 on lymphocytes raises the questions of why dual pathway use would be retained by these late-stage isolates and whether this use reflects an obligate need to infect both types of target cells in vivo.
The evolution of coreceptor utilization in general and the emergence of R5X4 variants in particular play an important role in pathogenesis. New infections are almost always initiated by R5 strains. The emergence of CXCR4-using variants in a substantial minority of subjects heralds the onset of accelerated disease progression, although it remains unknown whether the acquisition of CXCR4 use is the cause or a consequence of enhanced immune destruction. These late-stage variants were initially recognized based on their SI ability in cell lines in vitro (25). However, while T-cell-line-adapted (TCLA) prototype SI strains are restricted to CXCR4, and thus X4, it is now recognized that many or perhaps most SI primary isolates have the capacity to use both CCR5 and CXCR4 and thus exhibit an R5X4 phenotype (7, 23, 28). It is not known whether R5X4 is a transitional species in the evolution from R5 to X4, whereby most individuals would eventually transition from R5X4 to X4 if they survived long enough, or whether the as yet obscure forces responsible for coreceptor evolution favor R5X4 species as the ultimate result of evolution in vivo. Therefore, our results highlight another level of complexity in trying to understand this relationship, namely the finding that many R5X4 variants are in fact restricted to X4 use in PBL. In the context of coreceptor evolution in vivo, this finding raises the question of whether R5X4 variants that are CXCR4 restricted in lymphocytes represent a later stage of evolution in the process of disease progression than true dual-coreceptor-using variants.
Although the CXCR4 selectivity in primary lymphocytes of HIV-1 R5X4 isolates has not been widely appreciated, our results are consistent with and extend observations in several previous studies. Zhang et al. noted that HIV-1 89.6-based simian/human immunodeficiency virus (SHIV)/89.6 and its in vivo passaged derivatives were blocked by CXCR4 antagonists in PBL (38), and SHIV/89.6P appears to function like an X4 virus in vivo (30). Igarashi and colleagues showed that SHIV/DH12, derived from the R5X4 HIV-1 DH12, also uses CXCR4 exclusively for infection of both rhesus lymphocytes and macrophages (15). HIV-1 89.6 also preferentially uses CXCR4 in primary human lymphoid tissue infected ex vivo (13). Indeed, it has been recently suggested that there is a barrier to efficient CCR5 utilization intrinsic to the nature of dual coreceptor viruses (21). Taken together, these data suggest that the CXCR4-restricted nature of lymphocyte infection found here is likely representative of R5X4 strains broadly and may also reflect lymphocyte coreceptor utilization in vivo. On the other hand, Ghezzi et al. reported that lymphocyte infection by several R5X4 strains was blocked by CXCR4 inhibitors, but others were not completely blocked and several were inhibited by CCR5 antagonists (12), suggesting that some R5X4 strains may be able to use both pathways and that there may even exist R5X4 strains that are dependent on CCR5. Thus, it will be important to determine the pathways used in primary lymphocytes among a broader range of R5X4 primary isolates.
Since these R5X4 strains cannot enter lymphocytes through CCR5, their isolation by PBL culture clearly results from their ability to use lymphocyte CXCR4. Therefore, an important related question that is more difficult to answer is whether there exist in vivo R5 strains that cannot use primary lymphocyte CCR5. Such strains would be restricted to macrophages in vivo but could not be isolated through standard culture methods and, thus, would not be evident among most currently available isolates.
An issue raised by these results is why late-stage R5X4 isolates would retain the capacity to use CCR5 at all if they are unable to use lymphocyte CCR5. One possibility might be an obligate step involving CCR5, apart from lymphocyte infection, that is necessary for sustained replication in vivo. Although both CCR5 and CXCR4 are used by R5X4 strains to enter macrophages, our results suggest that for some CCR5 may be quantitatively more important (reference 36 and data not shown). Thus, one possible reason to retain CCR5 use might be a requirement for macrophages as a reservoir during late-stage disease. In the SHIV model, it has been suggested that macrophages are primarily responsible for viral production during extremely late-stage disease when nearly all CD4+ T cells are depleted (14), although in that model macrophage infection is mainly CXCR4 mediated (15). Alternatively, signals, perhaps distinct from those transmitted through CXCR4, may be transmitted by Env through CCR5 and may initiate cellular activation and trigger the intercellular interactions needed for sustained replication (1, 8, 11, 19, 32).
Why some R5X4 strains are unable to use CCR5 on PBL remains unclear. While it was initially believed that the simple expression of a coreceptor in conjunction with CD4 was sufficient for it to be used as an entry pathway, it has become increasingly clear that this is not the case. Evidence for strain- and primary cell-specific coreceptor utilization came with the recognition that CXCR4 is expressed on macrophages and can be used by certain primary X4 and R5X4 isolates even though it cannot be used by prototype TCLA X4 strains (27, 31, 35, 36). It has been suggested that the basis for defective macrophage CXCR4 use by TCLA X4 strains results from relatively low coreceptor expression and the differential ability of primary versus TCLA X4 strains to use low levels of CXCR4 (29). In contrast, differences between CCR5-using isolates in the levels of CCR5 expression required have not been reported. We found that R5X4 strains were more sensitive to a CCR5 antagonist than R5 strains, suggesting lower affinity or otherwise less efficient interactions, and a previous report demonstrated a higher sensitivity to RANTES inhibition (26). Furthermore, one study directly compared CCR5 binding of CD4-triggered monomeric gp120 and found that the R5X4 isolate tested exhibited lower levels of binding than R5 strains (2). Since primary CD4+ lymphocytes express relatively low levels of CCR5 (references 4, 18, and 33 and data not shown), it is possible that R5 strains are able to utilize these low levels but R5X4 strains cannot, due to low efficiency interactions. However, we have so far been unable to demonstrate differences between R5X4 and R5 Envs in their ability to use low levels of CCR5 in cell line-based assays (data not shown). Alternatively, we have shown that among a panel of closely related R5 and R5X4 Envs generated from a single primary isolate swarm, R5X4 variants were more constrained than R5 variants in their ability to fuse with a panel of CCR5 mutants, indicating a less plastic ability to tolerate conformational variability (37). Since CCR5 may be expressed in several alternate conformational states (17), it is possible that CCR5 is expressed on lymphocytes in a conformation, different from that in macrophages or transfected cells, that is able to support entry by the more flexible R5 but not by the conformationally more restricted R5X4 variants. Finally, differential association of CXCR4 with other proteins in the cell membrane has been suggested as an explanation for cell-specific use of macrophage CXCR4 (16), so similar differences between CCR5 on lymphocytes and other cells could underlie differential utilization by some strains.
In summary, this study shows that many R5X4 isolates behave as single coreceptor X4 strains in primary lymphocytes and can use distinct entry pathways in primary lymphocytes and macrophages. Thus, we extend to CCR5 and primary lymphocytes the concept developed previously, based on macrophages and CXCR4, that entry coreceptor utilization is both cell specific and strain specific. Since many CCR5 and CXCR4 blockers that inhibit HIV-1 entry are in various stages of development as clinical therapeutics, an important question in using these agents will be to determine which chemokine receptor(s) should be targeted in an infected individual. Since the profile of coreceptor use in primary lymphocytes may differ from that in primary macrophages or transfected cells in vitro, it will be important to correlate results of studies in vivo with coreceptor utilization not just in transfected cell systems but in both primary target cell types as well.
The authors thank C. Christie for technical assistance, M. Miller for M657, N. Fujii for T22, and D. Kolson and L. Margolis for valuable discussions.
This work received valuable assistance from the Immunology Core and the Viral, Cell, and Molecular Core of the Penn Center for AIDS Research and was supported by grants from the NIH to R.G.C.