Cellular infection by the human immunodeficiency virus type 1 (HIV-1) requires interaction of the viral envelope protein with CD4 and at least one additional cell surface molecule, termed a “cofactor” or “coreceptor.” Recent discoveries have determined that macrophage-tropic strains of HIV-1 which are largely responsible for sexual transmission require the β-chemokine receptor CCR5 in addition to CD4, while the T cell tropic viruses that emerge later after infection use the α-chemokine receptor CXCR4. Thus, both CD4 and the appropriate chemokine receptor must be expressed on the cell surface in order for HIV-1 to enter the cell and establish an infection. The in vivo importance of CCR5 for HIV-1 is demonstrated by the finding that individuals homozygous for a 32 bp deletion (Δ32) in the CCR5 gene that renders them effectively CCR5-negative are highly resistant to virus infection. In this review, the structure-function correlates of the chemokine receptors that serve as major coreceptors for HIV-1 and simian immunodeficiency virus entry will be reviewed. Since certain chemokines have been implicated as stem cell inhibitory factors, the biological consequences of chemokine receptor expression as it relates to HIV-1-associated hematodyspoiesis will also be discussed.
CD4+ lymphopenia has been the hallmark of human immunodeficiency virus type 1 (HIV-1) disease ever since the sentinel reports of AIDS cases in the early 1980s . The central role played by the CD4 molecule in primate lentiviral infections was established soon after the discovery of HIV-1 as the causative agent for AIDS. This fact was established via two distinct approaches: A) the use of monoclonal antibodies to the CD4 antigen not only inhibited HIV-1 infection of susceptible cells in vitro [2, 3], but was also able to coimmunoprecipitate the surface envelope (env) glycoprotein (gp120) of HIV-1  and B) the ability of transiently expressed CD4 to render previously unsusceptible human cells infectible by HIV-1 . However, it quickly became apparent that CD4, while capable of supporting virus binding, was not sufficient to trigger the conformational changes in the viral env protein required for the membrane fusion event necessary for HIV-1 to enter a cell. Rather, a specific cellular cofactor or coreceptor was required in conjunction with CD4 to support virus entry. This was made manifest by the failure of CD4 expression to confer susceptibility to HIV-1 infection in many nonprimate (and even some primate) cell lines [5-8], indicating that nonhuman cells lack coreceptors competent to support HIV-1 infection. In the ensuing decade, efforts to further understand the initial stages of cellular infection by HIV-1 were stymied by a frustrating lack of progress in identifying these additional cofactors. Concurrently, evidence was beginning to mount pointing to the existence of at least two phenotypic strains of HIV-1. Macrophage-tropic (M-tropic) strains retain the ability to infect primary macrophages, whereas T-tropic strains develop the ability to infect laboratory-derived T cell lines while sometimes losing the ability to establish a productive infection in primary macrophages [9, 10]. The differential tropisms displayed by virus isolates for CD4+ cells provided additional evidence that the mere expression of CD4 was not necessarily sufficient to render cells susceptible to virus entry, even when expressed in human cells.
Two seminal discoveries made independently in late 1995 and early 1996 brought order, and ultimately, illumination, to the disparate lines of evidence pointing to the existence of different coreceptors underlying the phenotypic variation seen in HIV-1 strains. These were the identification of several CD8+ T cell suppressive factors (specific for inhibiting HIV-1 infection) as the CC chemokines RANTES (regulated on activation, normal T cell expressed and secreted), MIP (macrophage inflammatory protein)-1α and MIP-1β , and the independent identification by expression cloning of the first bona fide coreceptor for T-tropic HIV-1 strains . Berger and colleagues, in an elegant application of expression cloning, identified the coreceptor for T-tropic HIV-1 strains as LESTR, a previously cloned orphan receptor belonging to the 7 transmembrane G-protein coupled receptor (7TM GPCR) family and most closely related to the CXC chemokine receptor family . This coreceptor, termed Fusin in light of its ability to mediate viral-cell fusion in conjunction with CD4, was subsequently designated CXCR4 when it was found to bind the CXC chemokine SDF-1 [13, 14]. The discovery of a putative chemokine receptor as the T-tropic HIV-1 coreceptor immediately suggested that the corresponding coreceptor for M-tropic HIV-1 could be the chemokine receptor for the CD8+ T cell suppressive factors RANTES, MIP-1α and MIP-1β. The contemporaneous cloning of a chemokine receptor exhibiting a ligand-binding profile for the aforementioned chemokines, designated CCR5 , led five different groups to demonstrate simultaneously that CCR5 was the major coreceptor for M-tropic HIV-1 entry [16-20]. Thus, in order for HIV-1 to infect a cell, both CD4 and the appropriate chemokine receptor must be expressed. Further, the ability of chemokines such as RANTES and SDF-1 to inhibit virus infection is at the level of virus entry; these chemokines prevent HIV-1 from utilizing their cognate coreceptors.
Chemokine Receptors and HIV Tropism
The discovery that certain chemokine receptors were the missing keys for the entry of HIV-1 into cells largely explains viral tropism. CXCR4 serves as the major coreceptor for T-tropic HIV-1 strains while CCR5 serves as the major coreceptor for M-tropic viruses. A number of virus strains, such as 89.6 and DH12 are capable of infecting both macrophages and certain T cell lines [21-23] presumably due to their ability to use both CCR5 and CXCR4 as fusion coreceptors [18, 23, 24]. These so-called “dual-tropic” strains may represent intermediate strains caught in the act of evolving from an M-tropic to a T-tropic phenotype . However, it is unclear whether a transition through a dual-tropic stage is a necessary in vivo phenomenon, although the increasing numbers of primary HIV-1 isolates that can use both CCR5 and CXCR4 seem to indicate that transition through a dual-tropic stage is not a rare event [25, 26].
The development of T-tropic viruses is associated with more rapidly declining CD4 counts and progression to AIDS. The more ubiquitous expression of CXCR4 among both hematopoietic and nonhematopoietic tissues relative to CCR5 has attractive explicatory power for the apparently aggressive nature of T-tropic strains. Furthermore, the presence of CXCR4 and CD4 in many subsets of hematopoietic progenitor cells, including CD34+ cells ( and unpublished data), suggests that T-tropic HIV-1 may be involved in the development of multiple cytopenias that coincide with the onset of clinical AIDS. The discovery that some viruses can utilize CXCR4 in the absence of CD4 provides another possible avenue by which T-tropic viruses can exhibit expanded tropism [28-30]. The increasing number of chemokine receptors or orphan receptors shown to be permissive for HIV-1 entry in vitro may also play a role in the pathogenesis of AIDS. For example, CCR3 has been shown to be present on microglia cells and is used as an efficient coreceptor by several neurotropic strains of HIV-1 , suggesting that utilization of CCR3 may contribute to viral neuropathogenesis. However, not all neurotropic HIV-1 strains can use CCR3, indicating that additional mechanisms may underlie the ability of HIV-1 to sometimes cause significant neurologic impairment .
The differential expression of CXCR4 and CCR5 on various T cell subsets has also afforded an explicatory paradigm for the preferential transmission of M-tropic, CCR5-using viruses. Bleul et al.  determined that CCR5 was expressed preferentially on the CD45RO+ subset of activated/memory T cells, whereas CXCR4 was more highly expressed on naive CD45RA+ T cells. Since activated/memory T cells can traffic to peripheral lymphoid organs at mucosal surfaces and form conjugates with dendritic cells or macrophages at these sites, they may be the initial targets of infection by CCR5-using viruses . Another avenue of evidence corroborating this differential expression of CCR5 and CXCR4 on activated versus resting T cells comes from studies of the differential effects of the reverse transcriptase inhibitors AZT and ddI on viral replication. Since AZT and ddI are differentially phosphorylated to their active forms in activated and resting T cells, respectively [35-37], clinical studies showing the selective inhibition of M-tropic or T-tropic strains of HIV-1 in patients receiving AZT or ddI, respectively, corroborates the differential expression of CCR5 and CXCR4 on T-cell subsets . Thus, the discovery of the HIV-1 coreceptors along with determinations of their pattern of expression in vivo has allowed the clarification and refinement of many previous observations and hypotheses regarding the nature of HIV-1 tropism.
Nature's Roulette: Protective Effects of Genetic Polymorphisms in Chemokine Receptors
As we make our way well into the second decade of the AIDS epidemic, it is apparent from large cohort studies and multiple anecdotal reports that some individuals remain uninfected despite multiple exposures to the virus (either through sexual or intravenous means) [39, 40]. It became immediately obvious that mutations in the coreceptor genes could offer an attractive explanation for such natural resistance to HIV. Such a finding was made simultaneously by two groups: one by studying exposed-uninfected (EU) individuals  and the other by investigating an epidemiological cohort . Both groups found the identical 32 bp deletion in the CCR5 gene which leads to a frameshift and premature termination of the CCR5 protein after the predicted second extracellular loop [41, 42]. This mutation abolishes cell-surface expression of CCR5 [41, 43] and renders the cognate cells resistant to M-tropic HIV-1 infection, although these cells remain readily susceptible to T-tropic infection [41-43]. The allelic frequency of this 32 bp deletion was surprisingly high in the Caucasoid population and was confirmed by other large epidemiological studies to be approximately 0.10 (summarized in ). The predicted Hardy-Weinberg equilibrium figures of about 16% heterozygotes and 1% homozygotes for the Δ32 mutation agreed with the epidemiologically determined frequencies, indicating that this mutation was neutral and did not impair the fitness of its bearers . Multiple studies have also since confirmed that homozygosity for the Δ32 allele is highly protective and that seropositive Δ32 heterozygotes appear to have a moderate but significantly delayed progression to clinical AIDS [45-48]. This may be due to the delayed replication kinetics of HIV-1 on peripheral blood mononuclear cells from Δ32 heterozygotes [41, 49, 50]. Although homozygosity for the Δ32 allele is highly protective, the protection is not absolute as several individuals homozygous for the allele have been reported to be infected [51-53]. Studies of coreceptor usage of the HIV-1 env from these individuals will shed light on the in vivo importance of coreceptors other than CCR5 which might be involved in the establishment of a primary infection.
A global epidemiological survey has also determined that the Δ32 allele is largely confined to the Caucasoid population with the highest allelic frequencies found in the Scandinavian and Ashkenazi Jewish populations . However, EU individuals have been described in other ethnic groups as well, and it is conceivable that other mutations in the CCR5 gene which disrupt its function as a coreceptor may account for some instances of this natural resistance to HIV-1. A preliminary genetic survey of several major ethnic groups has found several other sequence variations in the CCR5 locus, including a frame-shift mutation in the last transmembrane region that knocks out the last 54 amino acids . This polymorphism, designated “del893C,” is present at an allelic frequency of 0.04 in the Chinese and Japanese populations and could be severe enough to knock out the coreceptor function of the molecule. Obviously, EU individuals in these populations should be screened for this polymorphism, and proper in vitro studies for coreceptor function should be performed with this mutant allele.
Recently, a polymorphism consisting of a conserved valine to isoleucine change in the first transmembrane segment of CCR2b was discovered to be associated with delayed progression (two to four years) to AIDS . This polymorphism is broadly distributed among different population groups with allele frequencies ranging from approximately 0.1 to 0.25. The mechanism by which such a conservative amino acid change in CCR2b results in delayed progression to AIDS is not readily apparent since relatively few virus strains use CCR2b to infect cells. Thus, the effects are not likely to be exerted at the level of CCR2b utilization by HIV-1; rather, the mutation in CCR2b could conceivably alter cell-surface expression of more relevant HIV-1 coreceptors such as CCR5 or CXCR4.
Interactions of Chemokine Receptors with HIV-1 and Simian Immunodeficiency Virus (SIV) Env
The HIV-1 env glycoprotein is proteolytically processed from a gp160 precursor to form a mature noncovalent multimeric complex of gp120/gp41 subunits (reviewed in ). By virtue of its function, HIV-1 env is presumed to undergo a conformational change resulting in exposure of the hydrophobic fusion peptide in gp41 which mediates the mixing of the viral and cellular membrane bilayers. Binding of env to CD4 alone is not sufficient to trigger membrane fusion, suggesting that subsequent interactions between env and the appropriate chemokine receptor are responsible for the structural alterations in env that lead to membrane fusion. Although coreceptors could potentially mediate their function by altering the location or environment in which fusion occurs, dissection of key chemokine receptor components implicates their extracellular structures as the critical mediators of the triggering event.
The amino acid residues of chemokine receptors that interact with env have been mapped in an amalgam of experiments that, despite their individual complexities, offer a general consensus of functionally important regions. In the case of CCR5, all four extracellular domains participate in coreceptor activity to some degree, a result that is not surprising since the extracellular domains are apt to be in close proximity to one another as a consequence of highly conserved disulfide bonds. Of the extracellular domains, the amino terminal domain of CCR5 is particularly important, being able to confer coreceptor activity to other nonfusogenic chemokine receptors [58-61]. However, many virus strains can tolerate dramatic changes or even truncations of the N-terminal domain of CCR5, indicating that the extracellular loops, especially the second extracellular loop, are important for coreceptor activity [60, 61]. In contrast, all T-tropic and dual-tropic virus strains studied to date are dependent upon the second extracellular loop of CXCR4, although as for CCR5, other regions of this receptor are also important [62-64]. More refined analysis has localized specific amino acids in both coreceptors that can participate in the fusion reaction [59, 60, 63, 65, 66]. Further studies will delineate the role that each of these domains plays in binding of env, inducing the conformational change, and any viral entry pathways to which the chemokine receptors may contribute .
Studies in which the cytoplasmic domains of CCR5 and CXCR4 have been modified indicate that receptor internalization and G-protein coupling are not required for coreceptor activity [60, 66, 68, 69]. Thus, the natural function of chemokine receptors appears to be independent of their ability to mediate fusion. However, the ability of the chemokine receptors to mediate intracellular signaling events that could enhance viral replication in some cell types, such as quiescent macrophages, should not be excluded. In fact, it has recently been shown that some HIV-1 and SIV env proteins can induce CCR5 signaling [70, 71], raising a host of possibilities that will no doubt be objects of intense scrutiny.
Some of the complexities that arise from studying chemokine receptor structure and function are due to differential utilization of the major coreceptors by different HIV-1 strains. Although there are exceptions, it appears that each class of envs may utilize a given chemokine receptor in a similar, though not identical, way [60, 62]. Envelopes that use CCR5 but not CXCR4 as a coreceptor (M-tropic envs) can use both the amino terminus and the loops of CCR5 in diverse backgrounds for fusion. However, dual-tropic envs that can use both CCR5 and CXCR4 as a coreceptor are more dependent on the amino terminus of CCR5. By contrast, envs that use CXCR4 as a coreceptor are most dependent upon the second extracellular loop of this receptor. This suggests a mechanism that HIV-1 may use to evolve from M-tropism to T-tropism over the course of infection. An M-tropic env that can use two distinct sites on CCR5 may evolve into a T-tropic env by proceeding through a dual-tropic intermediate that uses a single structure on CXCR4 (the loops) and a single structure on CCR5 (the amino terminus) . The ability of dual-tropic env proteins to utilize somewhat different and more restricted structures on CCR5 and CXCR4 may explain their sensitivity to even slight perturbations of chemokine receptor sequence [59, 60, 62].
The emergence of T-tropic strains of HIV-1 is accompanied by residue changes within, but not limited to, the V3 loop of env [72-74]. Thus, it is not surprising that the change from M- to T-tropism is mediated by a coreceptor switch from CCR5 to CXCR4 and that this switch can be dictated by amino acid changes in the V3 loop [25, 75]. That only one or two amino acid changes in env can change which coreceptor is used  argues that there are conserved structures of chemokine receptors that are belied by strict sequence homology (only 33% between CCR5 and CXCR4). In favor of the existence of such conserved structures is the ability of a single amino acid change in a chemokine receptor to enable usage by new envs of opposite tropism (unpublished data). Given the intractable difficulties of crystallizing seven transmembrane receptors, structural information to resolve these models will depend on genetic and antigenic dissection of the chemokine receptors. Our model of how env is interacting with chemokine receptors will advance substantially when we are able to differentiate env binding from env conformational changes induced by chemokine receptors and to map specific regions in env that interact with specific regions in the chemokine receptors.
SIV env can also be phenotypically differentiated into M-and T-tropic strains. However, unlike HIV-1, both T- and M-tropic SIV strains use CCR5 [76-78]. Furthermore, CXCR4 is either not used or is used rarely as a SIV coreceptor. SIV tropism may be explained by utilization of additional coreceptors such as GPR1, GPR15 and STRL33 [79, 80], but may also be influenced by the way in which CCR5 is utilized. For example, we have found that there is differential utilization of CCR5 by at least some M- and T-tropic SIV strains . We have also recently established that certain strains of SIV can use CCR5 in a CD4-independent manner . This may have implications for the pathogenesis of certain aspects of SIV disease. For example, it has been suggested that infection of brain capillary endothelial cells which have CCR5 but no CD4 by CD4-independent, CCR5-dependent SIV strains may be responsible for the pathogenesis of SIV-associated encephalitis . CD4-independent binding of SIV env to rhesus CCR5 has also been recently demonstrated . Studying the molecular determinants of CCR5 usage by additional strains of CD4-independent SIV envs may shed additional light on the crucial regions of the coreceptor responsible for inducing the fusogenic conformational change in env.
Coreceptor Expression on Hematopoietic Cells: A Rosetta Stone for Understanding HIV-Associated Hematodyspoiesis?
Patients infected by HIV-1 frequently exhibit a variety of different hematological abnormalities including anemia, neutropenia and thrombocytopenia, in addition to the invariable loss of CD4+ lymphocytes (reviewed in [83, 84]). The fact that HIV-1 infects T lymphocytes and macrophages led to the early assumption that during the course of the disease, the virus could spread to hematopoietic stem and progenitor cells. Since the CD4 antigen is the primary receptor for HIV-1 entry, efforts were made to evaluate the expression of CD4 on human stem/progenitor cells. CD4 is expressed on 25%-65% of CD34+ cells [85-87], predominantly in the phenotypically primitive CD34+ CD38–/low class . The occurrence of CD4 on truly primitive cells is supported by the observation that CD4+ cells isolated from bone marrow are enriched in long-term culture-initiating cells (LTC-IC) that show the colony-forming ability of megakaryocyte and granulocyte-macrophage progenitors [85, 87]. Complement-mediated cytotoxicity assays employing an anti-CD4 monoclonal antibody are consistent with this finding, resulting in a significant reduction in the number of both classes of megakaryocyte—BFU-megakaryocyte (BFU-Meg) and colony-forming units-Meg (CFU-Meg)—and granulocyte-macrophage (CFU-GM) progenitors, whereas erythroid progenitors such as BFU-E are only slightly affected .
The presence of CD4 on hematopoietic progenitors coupled with clinical reports of the loss of primitive hematopoietic progenitors in patients with HIV-1 infection supported the convenient hypothesis that direct viral infection of progenitor cells could be responsible for the hematologic abnormalities seen in AIDS patients. However, attempts to directly demonstrate the presence of the viral genome in variously purified human CD34+, CD34+CD38– and CD34+CD4+ cells from HIV-infected individuals have largely failed (reviewed in ), although a smattering of reports suggests that a low level of virus can be found in a minority of such cells [86, 89]. Moreover, infection of CD34+ cells by either M- or T-tropic virus strains in vitro does not influence expansion and proliferation of hematopoietic progenitors, although there is an undeniable loss of primitive bone marrow progenitor cells in HIV-1-infected patients [89-92]. Together, these experiments argue that despite the expression of CD4 in a substantial fraction of progenitor cells, mechanisms other than a direct infection of early hematopoietic cells must be responsible for the hematological abnormalities encountered in AIDS patients.
The recent discovery of chemokine receptors as coreceptors for HIV-1 entry offers new avenues for increasing our understanding of the mechanisms underlying HIV-1-associated derangements of human hematopoietic cells and perhaps progenitor cells as well. At this point, there are 10 reported chemokine or orphan receptors that function as HIV-1 coreceptors: CXCR-4, CCR5, CCR2b, CCR3, CCR8, STRL33, GPR1, V28, ChemR23 and GPR15 [18, 20, 25, 31, 79, 93, 94 and unpublished data]. CCR5 is clearly responsible for the establishment of a productive primary infection as evidenced by the protective effect of the Δ32 CCR5 allele, and CXCR4 is probably involved in the pathogenic events that herald the onset of clinical AIDS, as suggested by the development of T-tropic CXCR4-using viruses in the later stages of the disease. The in vivo relevance of the other coreceptors has yet to be determined, though their ability to support infection by more limited numbers of virus strains raises the possibility that use of receptors other than CCR5 and CXCR4 may be involved in the myriad pathologies associated with HIV-1 infection, including the hematologic abnormalities. As such, exploring the chemokine receptor expression pattern on subsets of hematopoietic progenitors may shed light on the susceptibility of various subsets to either direct infection by HIV-1 or other forms of modulation such as chemokine-induced inhibition/proliferation or perhaps env-mediated toxicity. With regard to the latter point, recent studies have shown that soluble HIV-1 and SIV env can induce G-protein-mediated signal transduction through their cognate coreceptors [70, 71]. Therefore, intracellular signaling cascades mediated through chemokine receptors by HIV-1 env may lead to hematopoietic derangements, even in the absence of productive infection of hematopoietic progenitor populations.
Only one study to date has directly looked at CXCR4 and CCR5 expression in bone marrow progenitor cells. Deichmann et al.  consistently detected CXCR4 mRNA via reverse transcriptase-polymerase chain reaction (RT-PCR) from fluorescence-activated cell sorter (FACS)-sorted CD34+ cell samples. Significantly, CXCR4 mRNA was detected in both CD34+/CD4+ populations as well as CD34+/CD4– populations. On the other hand, CCR5 mRNA was only inconsistently detected, if at all, even though a nested PCR approach was used. This is consistent with our own results looking at protein expression using monoclonal antibodies against CXCR4 and CCR5. We found coexpression of CXCR4 on CD34+ cells but no evidence of CCR5 expression on the same population (Fig. 1). Thus, the CD4+ subset of human CD34+ cells is at least formally susceptible to HIV-1 entry via the CXCR-4 coreceptor. Because this chemokine receptor mediates T-tropic HIV-1 entry, CD34+ cells expressing CD4 and CXCR4 might be infected by HIV-1 or be affected by HIV env-mediated intracellular signaling cascades only during the later stages of the disease. This is consistent with the clinical observation that the frequency of cytopenias in nonleukocytic hematopoietic lineages (anemia, neutropenia, thrombocytopenia), alone or in combination, generally increases with the stage of disease (reviewed in ). Even the incidence of thrombocytopenia, which occurs earlier than other hematologic abnormalities, tends to increase among cohorts with more advanced disease. Thus, thrombocytopenia is rarely observed in asymptomatic seropositive patients; it is seen in 5%-12% of patients with AIDS-related complex, and up to 30% of patients with clinical AIDS exhibit thrombocytopenia [83, 95]. Whether these CD4+/CD34+/CXCR4+ cells are permissive for a productive viral infection is a separate issue that can only be addressed experimentally. Therefore, further studies looking for the presence of HIV-1 in the appropriate subset of hematopoietic progenitor cells from patients in the later stages of infection could help to address this issue. It is, of course, possible that transient infection of stem cells may occur in vivo, but productive infection may occur only after differentiation into a more mature cell type, usually along the monocytic pathway of hematopoietic differentiation . In light of the recent evidence showing env-mediated intracellular signaling through chemokine receptors, it would also be important to look at the effects of soluble HIV-1 env on progenitor cell proliferation.
An added level of complexity indicates that it is not the mere physical presence of CD4 and cognate coreceptors that determines a cell's permissiveness for viral entry. Macrophages, which express low but detectable levels of CXCR4 yet are generally resistant to entry by T-tropic strains of HIV-1, are an intriguing and relevant example. Appropriate stoichiometric levels of CD4 and the cognate coreceptor are an important determinant of a functional fusion complex that leads to productive viral entry. Kabat and colleagues have determined that when CD4 levels are limiting on the cell surface, high levels of the cognate coreceptor are required. However, when CD4 levels are high, even very low levels of coreceptor can support virus entry. This stoichiometric relationship differs for different coreceptors and even different viral strains, with primary isolates being more dependent on CD4 levels than lab-adapted or passaged strains [97, 98].
Even if appropriate subsets of hematopoietic progenitors are found to be susceptible to HIV-1 infection, it is still likely that the pathogenesis of hematological disorders seen in AIDS patients is multifactorial and cannot be explained by one simple mechanism, just as the decline in CD4+ lymphocytes cannot be explained solely by direct infection and killing by the virus. Other mechanisms contributing to the hematologic derangements seen in AIDS patients may include the release of different inflammatory cytokines during chronic HIV-1 infection and the infection of hematopoietic accessory cells (e.g., bone marrow stromal cells), which may in turn lead to disruption of the bone marrow microenvironment that may be deleterious to the proper function and maturation of progenitor cells . Inflammatory cytokines such as tumor necrosis factor (TNF)-α, interferon (IFN)-α, or IFN-β possess a well-documented history of having inhibitory effects on the proliferation of early human hematopoietic cells .
Other chemokine receptors, especially those which bind MIP-1α, MIP-2α, IL-8 and platelet factor-4, may mediate some important negative signals which regulate the development of human stem/progenitor cells . The β-chemokines RANTES, MIP-1α and MIP-1β have been reported to be secreted in excess during HIV-1 infection , and it will be important to determine the effects of these chemokines on hematopoiesis. The general promiscuity of each chemokine receptor for more than one ligand may complicate the relationships between the chemokine-chemokine receptor axes, but some preliminary determinations have already been made. For example, MIP-1α was reported to inhibit human erythroid colony formation after interaction with the CCR1 receptor but not with CCR5 . Interestingly, although MIP-1α was the first stem cell inhibitor to be identified, mice deficient for MIP-1α do not have an expanded progenitor pool . In another report, MIP-1α together with IL-3 expanded a pool of hematopoietic progenitors from CD34+CD33– cells . In concurrence with these reports, Gewirtz et al. also did not find inhibitory effects of this chemokine on the cloning efficiency of normal human CFU-Mix, BFU-E and CFU-GM progenitors. The only visible inhibitory effect was a slight inhibition of CFU-Meg colony formation .
With the cloning of the chemokine receptors and the appropriate assignment of the chemokine ligands to their appropriate receptor, it is time to reappraise the putative inhibitory role of chemokines in human hematopoiesis. This is especially important now that molecular approaches are being developed to target the appropriate chemokine-chemokine receptor axes in an effort to block HIV infection. Indeed, small molecule inhibitors of CXCR4 have already been described [107-109] and CCR5 antagonists have also been developed . The possibility that these chemokine-chemokine receptor axes may mediate inhibitory signals on proliferation of normal human stem/progenitor cells has to be carefully reappraised in well-controlled experiments before appropriate in vivo molecular strategies which interfere with the function of these receptors can be employed for the therapeutic treatment of HIV infection.
While our current techniques for detecting transcripts and protein from chemokine receptor genes allow an approximation of cell susceptibility to HIV-1, our limited knowledge of the regulation of chemokine receptors and their interaction with HIV-1 env remains a major hurdle for understanding their relevance in vivo. Research efforts should now be redoubled to generate better reagents and techniques for molecularly dissecting the interactions between HIV-1 env and its cognate coreceptors. However, we can say that although the path has been long, there are at least now signposts to guide our journey.
We would like to thank members of the Doms Lab for intellectual stimulus and helpful discussions; J. Rucker and J. Berson were especially helpful in providing primary references. We also thank Alan Gewirtz and members of his lab for supporting the collaborative work mentioned in this review; Jasha Ratajczak was especially generous in helping to provide us with the cells used in our studies. Last but not least, we would like to thank Monica Tsang at R&D Systems for generously providing chemokine receptor antibodies that actually work! B.L. was supported with a Measey Foundation Fellowship for Clinicians (Wistar Institute), B.J.D. was supported with a predoctoral fellowship (Howard Hughes Medical Institute). Additional work was supported by grants to R.W.D. (AI-35383, AI-38225, AI-40880).