Characterization of a CD4-independent clinical HIV-1 that can efficiently infect human hepatocytes through chemokine (C-X-C motif) receptor 4

Objective: HIV-1 isolates are prominently CD4-dependent and, to date, only a few laboratory-adapted CD4-independent strains have been reported. Therefore, whether CD4-independent viruses may exist in HIV-1-infected patients has remained unclear. Here, we report the successful isolation of a CD4-independent clinical HIV-1 strain, designated SDA-1, from the viral quasispecies of a therapy-naive HIV-1 and Pneumocystis jirovecii pneumonia patient in the late-stage of AIDS with extremely low CD4 cell count (CD4 = 1/μl). We characterized this virus and further explored whether it could infect or induce pathological effects in human hepatocytes.

Design and methods: To determine coreceptor usage and CD4-independent infection, the HIV-1 envelope (Env)-pseudotypes and Env-chimeric viruses were used.

Results: SDA-1 was able to infect CD4^- cell lines through either chemokine (C-X-C motif) receptor 4 or CCR5. It still maintained the ability to infect CD4⁺ cells through multiple coreceptors of chemokine (C-X-C motif) receptor 4, chemokine (C-C motif) receptor 5, chemokine (C-C motif) receptor 3 and chemokine (C-C motif) receptor 8. Productive infection by SDA-1 was noted in both CD4-negative hepatoma cells and primary cultured human hepatocytes. Moreover, we demonstrated that SDA-1 could efficiently infect human hepatocytes on both static and mitotic phases through chemokine (C-X-C motif) receptor 4, without inducing apoptotic cell death.

Conclusion: The present study provides evidence that emergence of CD4-independent HIV-1 virus in vivo may occur in HIV-1-infected patients. In addition, these results shed light on the mechanisms involved in liver damage in HIV-1-infected individuals, which could have important implications concerning the range of mutability and the pathogenesis of AIDS.

Introduction

The entry of HIV-1 into target cells requires interaction of the viral envelope (Env) with CD4 and a chemokine coreceptor ^[1,2]. Macrophage-tropic HIV-1 viruses primarily use chemokine (C-C motif) receptor 5 (CCR5) (R5) as a coreceptor, whereas T-cell-tropic viruses use chemokine (C-X-C motif) receptor 4 (CXCR4) (X4). Dual-tropic viruses (R5X4) use both coreceptors ^[3]. A few rare viruses can also use alternative coreceptors such as chemokine (C-C motif) receptor 1 (CCR1), chemokine (C-C motif) receptor 2b (CCR2b), chemokine (C-C motif) receptor 8 (CCR8), chemokine (C-X-C motif) receptor 6 (CXCR6), G protein-coupled receptor 1 (GPR1) or GPR15/Bob for entry into coreceptor-transfected CD4⁺ cell lines ^[4]. Whatever the coreceptor specificity of an HIV-1 isolate, an interaction with CD4 is always the first step in a chain of events leading to fusion of the viral envelope with the cellular membrane. However, previous studies have shown that SIV ^[5] and HIV-2 ^[6] can also infect cells independently of CD4.

In contrast to SIV and HIV-2, HIV-1 CD4-independent viruses are rarely isolated. To date, only a few laboratory CD4-independent HIV-1 variants ^[7-10] have been reported. Therefore, whether such viruses may exist in HIV-1-infected patients has remained unclear. However, several studies ^[11-14] have shown that HIV-1-DNA and p24, a core HIV-1 antigen, were detected in CD4-negative cells or tissues such as brain, kidney and liver in HIV-infected individuals, suggesting the possibility that low levels of CD4-independent variants exist in vivo. Among such CD4^- cells or tissues, liver is an important organ in determining the prognosis of HIV-1-infected patients. End-stage liver disease is becoming a frequent cause of death in HIV-1-infected hospitalized patients ^[15-17]. Although the cause of liver injury in HIV-1 patients might be multifactorial, such as hepatitis B virus (HCV) and hepatitis B virus (HBV) coinfection and the side effects of antiretroviral therapy, a number of reports have documented that histological liver abnormalities occurred solely as a result of HIV-1 infection ^[13,18,19]. Nonetheless, few attempts have been made to elucidate the mechanisms of the liver damage in HIV-1-infected individuals.

In this study, we successfully isolated a CD4-independent clinical HIV-1 strain, designated SDA-1, from the viral quasispecies of a therapy-naive HIV-1 and Pneumocystis jirovecii pneumonia (PJP) patient in the late stage of AIDS with extremely low CD4 cell numbers. We characterized the phenotype of this virus and further explored whether it could infect or induce pathological effects in human hepatocytes.

Materials and methods

Patient's information

A 53-year-old Japanese man infected with HIV-1 was admitted to Tohoku University Hospital owing to prolonged fever and severe dyspnea in 2000. His plasma viral load and CD4 cell count at the time of admission was 220 000 copies/ml and 1 cell/μl, respectively. He was diagnosed with PJP, and his clinical stage was classified as IV-C3 ^[20]. The onset and route of HIV-1 infection were unknown. No evidence of coinfection with HBV or HCV in this patient was found. The patient was treated with trimethoprim and sulfamethoxazole (TMP-SMX) and highly active antiretroviral therapy (HAART). His condition deteriorated rapidly and he died 33 days after admission. Consent for autopsy was denied by the patient's family.

Before HAART, plasma samples and peripheral blood mononuclear cells (PBMC) were collected from this patient and cryopreserved in liquid nitrogen until use. The institutional Ethics Committee approved this study and written informed consent was obtained from the patient.

Virus isolation

HIV-1 isolation was achieved by using an in-vitro short-term phytohemagglutinin (PHA)-PBMC coculture method. Briefly, cryopreserved PBMC (2 × 10⁶) from the patient were cocultivated with PHA-stimulated PBMC (5 × 10⁶) from an HIV-1 seronegative healthy donor. The culture was maintained in RPMI-1640 (Invitrogen, California, USA) containing 10% fetal calf serum and 5 U/ml of recombinant interleukin-2 (IL-2) (Sigma, St. Louis, Missouri, USA). Proliferation of HIV-1 was examined by measuring p24 antigen in the cell culture supernatant using a p24 ELISA kit (RETRO-TEK, ZeptoMetrix Corp., New York, USA). The virus stocks were kept at -80 °C until use.

Amplification of env and sequence analysis

The full-length HIV-1 env genes were amplified by limiting dilution nested PCR from proviral PBMC DNA or plasma RNA as previously described ^[21,22]. To avoid artificial recombination and resampling of the viral genomes, independent nested PCR reactions were carried out per specimen ^[23,24].

The first round PCR was conducted with a F5852-R8935 primer pair (F5852, 5′-TAGAGCCCTGGAAGCATCCAGGAAG, HIV-1 HXB2 nucleotide position 5852-5876; R8935, 5′-TTGCTACTTGTGATTGCTCCATGT, HXB2 nucleotide position 8912-8935). The second round PCR was performed with a F5957-R8903 primer pair (F5957, 5′-GATCGAATTCTAGGCATCTCCTATGGCAGGAAGAAG, HXB2 nucleotide position 5957-5982, containing an additional EcoRI site (underlined) to facilitate cloning; R8903, 5′-AGCTCTCGAGGTCTCGAGATACTGCTCCCACCC, HXB2 nucleotide position 8881-8903, containing an additional XhoI site (underlined)). The purified PCR products were subcloned into the EcoRI and XhoI sites of the pSM-HXB2 plasmid. All correctly oriented env clones were then screened for biological function ^[22] followed by sequencing and phylogenetic analysis as previously described ^[25,26].

Cell lines and cell culture

All the cell lines, unless otherwise specifically mentioned, were cultured in Dulbecco's modified Eagle's medium (DMEM, Invitrogen) containing 10% fetal calf serum. Human glioma NP-2-CD4⁺ cells transfected with a variety of chemokine receptors as indicated ^[27] were maintained in medium containing 500 μg/ml of G-418 (Promega, Wisconsin, USA) and 1 μg/ml of puromycin (Sigma). Human CD4-negative osteosarcoma (HOS) cells expressing either CXCR4 or CCR5 ^[28] were cultured in medium containing 1 μg/ml of puromycin. Human hepatoma cells Huh-7 and Hep-G2 ^[29] were obtained through the Cell Resource Center for Biomedical Research, Tohoku University, Japan. Human primary cultured hepatocytes (p-hepatocytes, BD Bioscience, California, USA) were maintained on BD Matrigel with Hepato-STIM hepatocyte culture medium (BD Bioscience).

Reagents and antibodies

The CXCR4 antagonist AMD3100 ^[30], and the CCR5 antagonist TAK-779 ^[31] were provided by the NIH AIDS Research and Reference Reagent Programme and Takeda Chemical Industries, Ltd., Osaka, Japan, respectively. Recombinant human soluble CD4 (sCD4) was from ImmunoDiagnostics, Inc. (Woburn, Massachusetts, USA). Antialbumin-fluorescein isothiocyanate (FITC) antibody was from Cedarlane Laboratories Ltd. (Hornby, Ontario, Canada). Anticytokeratin-18-phycoerythrin and antialpha fetoprotein (AFP)-FITC antibodies were from Santa Cruz Biotechnology, Inc. (Santa Cruz, California, USA). Anti-HIV-1-p24 (clone KC57)-FITC antibody was from Beckman Coulter. All other antibodies were from BD Pharmingen (San Diego, California, USA).

Pseudotyped virus infection assay

The HIV-1 Env-pseudotypes were generated as previously described ^[32]. Briefly, 293T cells (5 × 10⁶ cells/10 cm-dish) were transfected with 5 μg of luciferase-expressing pNL4-3-Luc-R^-E^{- [33]} or green fluorescent protein (GFP)-expressing pNL4-3-GFP ^[34] plasmid in combination with 10 μg of one of the env-expressing plasmids, pSM-SDA-1, pSM-HXB2 (X4), pSM-ADA (R5), or pSM-89.6 (R5X4). The vesicular stomatitis virus-G pseudotypes were also prepared ^[35].

For infection assays of luciferase-pseudotypes (luc-p), 10 ng p24 of luc-p were added into each well of 24-well plates (5 × 10⁴ cells/well). After 12 h infection, the cells were washed and incubated for an additional 36 h at 37 °C. The cells were then lysed using a Luciferase Assay kit (Promega) and the luciferase activity was examined by a luminometer (Lumat 9507, Germany). To determine the effects of various reagents related to the viral receptors, target cells were preexposed for 1 h with the indicated concentration of the antagonists, or the antibodies. For GFP-pseudotypes (GFP-p) infection, target cells were infected with 10 ng p24 of GFP-p virus for 48 h and fixed by 5% paraformaldehyde. Infectivities were visualized under a Zeiss LSM510 confocal microscopy and DIC images with a 512 × 512 resolution were acquired.

Chimeric viruses

All env recombinant chimeric viruses in this study were generated in the background of pNL43, an X4-tropic HIV-1 infectious clone ^[36]. Briefly, the fragment of pNL43 containing EcoRI (nt 5743-5748) and KpnI (nt 6343-6348) was amplified by PCR with a F5671-R6472 primer pair (F5671, 5′-GGCTCCATAACTTAGGACAAC, pNL43 nucleotide position 5671-5691; R6472, 5′-TACTTCTTGTGGGTTGGGGTC, pNL43 position 6452-6472), followed by insertion into the pSM-SDA-1 using EcoRI and KpnI. The new EcoRI-XhoI fragment (3155 bp) covering the entire SDA-1 env gene was then replaced with the equivalent region of pNL43 to construct the Env-chimeric virus NL43_SDA-1. Similarly, Env-chimeras of ADA (NL43_ADA), 89.6 (NL43_89.6) or truncated env (NL43_env (-)) were created, respectively. All Env-chimeric viruses were prepared by transfecting 293T cells as described above. For infection assays, 100 ng p24 of the chimeric viruses or virus stock supernatants were added in each well of 24-well plates (5 × 10⁴ cells/well). After 2 h adsorption, the cells were washed and incubated for 48 h. Viral replication was monitored by p24 antigen production.

Flow cytometry and apoptosis assay

We performed cell-surface staining for CD4, CXCR4 and CCR5 by flow cytometry. To determine the purification and differentiation of p-hepatocytes, we tested the specific markers using antialbumin-FITC, anti-AFP-FITC and anticytokeratin-18-phycoerythrin antibodies. Appropriate class matched antibodies were used in each experiment. To detect the proliferation and intracellular p24, p-hepatocytes were fixed and permeabilized using a Cytofix-Cytoperm kit (BD Bioscience). Subsequently, the cells were stained with anti-Ki-67-phycoerythrin and antip24-FITC antibodies. Apoptosis of the p-hepatocytes was determined using the Apoptosis Detection kit I (BD Pharmingen). Flow cytometry analysis was performed using FACSCalibur (Becton Dickinson, New Jersey, USA). All Data were acquired and analyzed using Cell Quest software (BD Bioscience).

Nucleotide sequence accession number

The GenBank accession number for the sequence determined in this study is AY902478 (SDA-1).

Results

Evaluation of SDA-1 viral quasispecies

In an attempt to isolate CD4-independent clinical HIV-1 strain(s), we performed virus isolation from a therapy-naive HIV-1 and PJP patient with extremely low CD4 cell number, and successfully isolated the virus (peak of p24, 500 ng) from this patient and designated it SDA-1. To assess the quasispecies diversity present in vivo, we analyzed the SDA-1 env clones derived from plasma RNA and PBMC. As shown in Fig. 1, SDA-1 is grouped within the HIV-1 subtype B reference sequences. Within SDA-1's sequence cluster, three phylogenetic forms were identified. Supported by a significant bootstrap value (96%), form I was the predominant quasispecies, representing 70% of all sequences. Two minor quasispecies (forms II and III) had similar structures but differed in the position of the first breakpoint. The mean distances between major and minor quasispecies did not differ significantly from the sequence heterogeneity. Furthermore, the quasispecies diversities between plasma and PBMC were similar within each form, and were all below 5.0%.

Fig. 1

Multicoreceptor usage and CD4-independent entry of SDA-1

To determine the receptor usage of SDA-1, we randomly selected five clones from the predominant quasispecies and generated Env-pseudotypes and Env-chimeric viruses as representatives. As a control, the Envs from a variety of HIV-1 subtypes with X4 (HXB2), R5 (ADA), and R5X4 (89.6) tropism were used. Utilizing luciferase-pseudotypes (luc-p), we first examined the coreceptor usage of SDA-1. We found that in the presence of CD4, all representative SDA-1 Env-pseudotypes were able to use efficiently both CXCR4 and CCR5, with additional moderate usage of CCR3 and CCR8 (Fig. 2a).

Fig. 2

We next investigated whether SDA-1 Envs are capable of inducing CD4-independent infection. We found that SDA-1 Envs mediated entry into both HOS-CXCR4 and HOS-CCR5. However, the infectivities of SDA-1 for HOS-CXCR4 were approximately 2.5-fold higher than that for HOS-CCR5 (Fig. 2b). In stark contrast, none of the other types of luc-p viruses entered either of those cells. Furthermore, we evaluated the ability of SDA-1 Envs in mediating cell-cell fusion, a dye-transfer cell-cell fusion assay ^[37] was used with HOS-CXCR4 and HOS-CCR5 cells. Only in the cells expressing SDA-1 Envs (effector cells) did cell-cell fusion with CD4-negative, CXCR4- or CCR5-positive HOS cells (target cells) occur (data not shown).

In addition to the results with HOS-CXCR4 and CCR5, preexposure of HOS cells to Leu-3A, a CD4 monoclonal antibody (mAb) that recognizes the gp120 binding site on CD4 ^[38], failed to block SDA-1 infection. In contrast, pretreatment with antagonists for CXCR4 or CCR5 effectively inhibited infection (Table 1). Furthermore, the infectivities of SDA-1 on HOS-CXCR4 and HOS-CCR5 were enhanced by preexposure of the virus to sCD4 indicating that the binding of SDA-1 Env to CD4 induces further conformational changes in gp120 to fully expose the chemokine receptor binding domain. Collectively, SDA-1 Envs mediated the CD4-independent infection via both CXCR4 and CCR5.

Table 1

Having clarified that SDA-1 is a CD4-independent isolate, we next investigated what types of CD4^- cells are able to support SDA-1's entry. We focused first on human liver-derived cell lines, as the mechanisms of the liver damage in HIV-1-infected individuals are still unclear.

Two hepatoma cell lines, Huh-7 and Hep-G2, were used as targets. We first examined the expression of the receptors on the cell surface by flow cytometry and found that both CXCR4 and CCR5 were expressed on Huh-7 and Hep-G2 cells. In contrast, CD4 was not detected on either, which was confirmed by RT-PCR (data not shown). We then evaluated whether SDA-1 can enter into hepatoma cells with luc-p viruses. We found that only SDA-1 luc-p viruses efficiently infected Huh-7; however, its infectivity was marginal in Hep-G2 (Fig. 2c). Previous studies have shown that few HIV-1 variants can infect CD8⁺ cells using CD8 as receptor ^[10,39]. Therefore, we further explored receptors used by SDA-1 for entry into hepatoma cells. As shown in Fig. 2d, preexposure of Huh-7 to anti-CD8 Leu-2A mAb, as well as the CCR5 antagonist, TAK-779, failed to block SDA-1 infection of Huh-7, whereas anti-CXCR4 with AMD3100 effectively suppressed the infectivity. These results suggested that SDA-1 enters Huh-7 cells principally via CXCR4.

Replication of SDA-1 in human hepatoma cells

Although SDA-1 luc-p viruses infected some cells independently of CD4 cells, it was necessary to determine whether SDA-1 can replicate in those CD4^- cells, particularly in hepatoma cells. For this purpose, we constructed NL43-based Env-chimeric viruses described above. We then examined whether the chimeric viruses were able to replicate in CD4^- cells. As shown in Fig. 3a, the SDA-1 Env-chimeric viruses replicated efficiently in HOS-CXCR4 and HOS-CCR5 cells to similar levels. In contrast, none of the other Env-chimeric viruses infected either of those cells lines. Furthermore, we examined whether SDA-1 Env-chimeric viruses could replicate in hepatoma cells. As shown in Fig. 3b, high levels of NL43-SDA-1 replication were observed in Huh-7 cells. However, marginal replication was detected in Hep-G2 cells. Although both Huh-7 and Hep-G2 cells are derived from human hepatoma, many potential host factors ^[40] could influence HIV replication, which for the most part remain unknown. Similarly, only Huh-7 cells, but not Hep-G2 cells, were susceptible to HCV ^[41,42]. These reasons may be related to the difference between Huh-7 and Hep-G2 regarding the level of replication by SDA-1.

Fig. 3

SDA-1 replicates in both proliferating and static hepatocytes

To investigate further whether normal human hepatocytes could sustain entry and replication of SDA-1, p-hepatocytes were used for the following experiments. Among the three specific markers of human hepatocytes, both albumin and cytokeratin-18, but not alpha-fetoprotein were detected in the p-hepatocytes suggesting that the hepatocytes we used were well differentiated (data not shown). We also found that CXCR4 was expressed on the surface of p-hepatocytes. In contrast, neither CD4 nor CCR5 was detected on the p-hepatocyte surface or by real-time PCR (RT-PCR) (data not shown).

We next explored whether SDA-1 can enter p-hepatocytes by using GFP-p. As shown in Fig. 4a, only SDA-1 GFP-p viruses gave GFP-positive cells in p-hepatocytes, whereas other HIV-1 GFP-p viruses did not. The GFP-positive cells showed spindle-like shapes suggesting that the infection occurred in the p-hepatocytes but not in the contaminating lymphocytes. Furthermore, we studied whether SDA-1 can replicate in the p-hepatocytes. As shown in Fig. 4b, the p-hepatocytes were productively infected by the SDA-1 Env-chimeric viruses and SDA-1 virus stock itself but not by the other HIV-1 Env-chimeric viruses. Moreover, we found that AMD3100 inhibited the replication of SDA-1 in p-hepatocytes in a dose-dependent manner (Fig. 4c) indicating that the infection of p-hepatocytes by SDA-1 was mediated through CXCR4.

Fig. 4

A previous study ^[19] reported that the HIV-1 gp120 env directly caused hepatocyte death by signaling through CXCR4 in vitro; however, most studies were performed using the hepatoma Huh-7 cells not hepatocytes, therefore, it may not really reflect the nature of liver damage. To explore the pathological effects of HIV-1 CD4-independent infection on hepatocytes, we exposed p-hepatocytes to the SDA-1 and analyzed cell viability. We found that the viability of the p-hepatocytes in cells cultured with or without SDA-1 Env-chimeric viruses was comparable (96%, P was not significant) indicating that HIV-1 CD4-independent infection rarely induces hepatocyte death via an apoptotic process (data not shown). To further examine whether the infection or replication of SDA-1 is limited only to a certain number of p-hepatocytes or whether the infectivity or replication is influenced by the cell cycle, we studied the intracellular expression by flow cytometry of p24 and Ki-67 ^[43], a marker strictly associated with cell proliferation, in the HIV-1-infected p-hepatocytes. As shown in Fig. 4d, we found that 32.49% of p-hepatocytes were infected by SDA-1. However, there was no significant difference in percentage of p24 expression between Ki-67⁺ (31%) and Ki-67^- p-hepatocytes (33.1%), suggesting that SDA-1 efficiently enters and replicates in both proliferating and static hepatocytes.

Considering that SDA-1 can infect hepatocytes in vitro, it would have been interesting to determine whether the patient's liver was infected in vivo. However, consent for a liver biopsy was denied by the patient's family. There was no evidence of liver dysfunction. When virus was isolated from this patient; however, liver damage [an aspartate aminotransferase (AST)/alanine aminotransferase (ALT) ratio ≥1] was observed at the end of the clinical stage. Although the cause of liver injury was unclear, our present data suggest that CD4-independent HIV-1 infection may lead to hepatocellular damage.

Discussion

In this study, we characterized a quasispecies of a CD4-independent HIV-1 isolate, termed SDA-1, which was able to utilize either CXCR4 or CCR5 in the absence of CD4. Moreover, we demonstrated that SDA-1 efficiently entered and replicated in Huh-7 hepatoma cells and normal human hepatocytes, through CXCR4, without inducing apoptotic cell death.

Many SIV and HIV-2 isolates can infect cells without CD4, at least to some extent. However, CD4-independent HIV-1 viruses have been rarely isolated and, so far, only a few laboratory-adapted CD4-independent HIV-1 variants have been reported. It must be noted that CD4-independent HIV-1 variants, isolated in vitro by passage through cells lacking CD4, have been shown to be more sensitive to neutralizing antibodies than CD4-dependent viruses ^[44,45]. Therefore, we might hypothesize that the emergence of a quasispecies of HIV-1 with a reduced requirement for CD4 is likely to be at a low abundance relative to the more common CD4⁺ strains. However, with disease progression, HIV-1 variants with reduced affinity for CD4 and with increased affinity for chemokine receptor could evolve and become more robust in the viral quasispecies, disseminate in a variety of CD4^- tissues in vivo under conditions of both reduced immunological pressure and a dramatically reduced pool of target CD4⁺ cells concomitant with high levels of virus replication. It will be important to search the viral quasispecies in other patients, especially in the later stages of HIV-1 disease for the existence of similar CD4-independent HIV-1 variants and expanded cellular tropism.

Although the extent to which CD4^- cells are infected in vivo is unclear, it has been widely thought to be low. Nonetheless, recent studies ^[11,12] utilizing the novel approach of laser capture microscopy have revealed HIV-1 sequences in isolated CD4^- cells of kidney epithelium and neuronal cells, indicating that latent infection might occur in such cells or tissues in vivo. The mechanism of viral entry into CD4^- cells remains unclear, but as we show here the evidence of emergence of CD4-independent strains in vivo must be kept in mind.

End-stage liver disease is now becoming a frequent cause of death in HIV-1-infected hospitalized patients. HCV and HBV coinfection with HIV-1 has been shown to enhance the progression of liver damage ^[16]. However, little attention has been given to the direct virological interaction between HIV and HCV/HBV in the liver, as HIV has been thought not to infect hepatocytes directly. Nonetheless, a number of reports have documented that histological liver abnormalities occurred solely as a result of HIV-1 infection. In our study, we clearly demonstrated that SDA-1 efficiently enters and replicates in both proliferating and static hepatocytes through CXCR4. To our knowledge, this is the first report that HIV-1 can efficiently replicate in normal hepatocytes. Furthermore, we have shown that HIV-1 infection did not induce significant cytocidal effects in the hepatocytes. It is noteworthy that the liver is a continuously regenerating organ. Therefore, if HIV-1 enters and integrates its DNA into the host genome, liver cells containing HIV-1 DNA will be continuously generated by the division of the infected cells. Thus, the expression of HIV-1 proteins on the infected cell surface might result in chronic damage of the liver cells by inducing host immune responses. Direct virological interaction between HIV, HCV and HBV in the liver or enhanced production of HIV-1 by inflammatory cytokines produced by the HCV and HBV-activated immune cells might also exacerbate the liver injury. At present, however, we have no definite information concerning the extent to which patients' hepatocytes harbor HIV-1 and CD4-independent HIV-1 variants.

Finally, a particularly important area of vaccine research is to take advantage of gp120 structural information to guide the design of novel envelope immunogens. As has been reported, CD4-dependent viruses hide neutralizing epitopes and only CD4 binding to gp120 induces conformational changes in gp120 to fully expose epitopes for broadly neutralizing antibodies. The CD4-independent strain we isolated here seems particularly important, as it can efficiently replicate in CD4^- hepatocytes. Therefore, the gp120 structural alterations, which might expose the coreceptor binding site without binding to CD4, may also open up other sites that could yield neutralizing antibodies. Nevertheless, evidence of a clinical CD4-independent R5X4 HIV-1 virus should have important implications concerning the range of mutability and tropism of HIV-1 and the pathogenesis of AIDS.

Acknowledgements

We would like to thank Dr Shinji Okada for the clinical data of this patient. We also thank to Dr Y. Koyanagi, Dr T. Hara, Dr R. Furuta and H. Sakai for technical supports for isolation of the virus, sequences analysis, cell-cell fusion assay and chimeric viruses of SDA-1, respectively. We are indebted to Dr D.R. Littman for the gift of HIV-1 luciferase and GFP reporter plasmids, Dr C.D. Weiss for Env expression plasmids, and Dr N. Laudau for HOS.CXCR4 and HOS.CCR5 cells. We thank Division of AIDS, NIAID, NIH for providing AMD3100 through the NIH AIDS Research and Reference Reagent Programme. We also thank Takeda Pharmaceutical Company Ltd. (Osaka, Japan) for providing TAK-779. We are grateful for Dr M. Robert-Guroff for critical reading of the manuscript.

The present work was supported by Grant-in-Aid for Scientific Research B from JSPS and the Scientific Research Expenses for Health and Welfare Programme from the Ministry of Health and Welfare, Japan.

P.X., H.L., Y.S. and T.H. designed the study. P.X., O.U., Y.S., M.Z., Y.A. and H.G. performed the experiments. P.X., O.U, Y.S., H.L. and T.H. analyzed the data. N.S. and H.H. contributed to the coreceptor expressing cell lines. P.X., H.L., Y.S., O.U., N.S., H.H. and T.H. contributed to writing the paper. T.H. contributed to grant application and financial support.