|Home | About | Journals | Submit | Contact Us | Français|
HIV-1 transinfection is a process by which one cell acts as an HIV-1 “escort” to enhance infection of another. There has recently been much debate concerning (1) the types of cells that may act as escorts, (2) requirements for virus internalization by the escort, and (3) the sensitivity of transinfection to inhibition by neutralizing antibodies. To address these questions, transinfection was monitored by incubating target cells with HIV-1 in the presence or absence of mouse or human cells as candidate escorts in vitro. After a 2-day culture, target cells were tested for levels of HIV-1 infection. Results showed that a variety of murine and human cells were capable escorts for HIV-1 transinfection. Cell integrity was not required, as escorts could be freeze/thawed (or fractionated to yield purified membranes/microsomes) prior to their incubation with HIV-1. In fact, the freeze/thawed or fractionated cells were often superior to their viable counterparts as mediators of transinfection. The process was sensitive to antibody neutralization. Confirmatory experiments were conducted with more than one target cell and more than one source of HIV-1. Results demonstrated that there may be multiple cell types and mechanisms with which transinfection can be accomplished. Apparently the simple binding of fragmented escort membranes to HIV-1 may be sufficient to enhance virus fusion or endocytosis at the target cell surface. The fact that dead cells or membranes can support this activity may explain, at least in part, the high frequency of human HIV-1 infections at sites of tissue damage.
Transinfection occurs when HIV-1 is attached to the membrane of an “escort” cell (e.g., via DC-SIGN) and subsequently transferred to a T cell target. The activity can be mediated by a sophisticated process in which an escort [e.g., a dendritic cell (DC)] internalizes virus and then shuttles intact virus to a cell membrane-associated immunological T cell synapse. The target T cell is then exposed to virus and infection proceeds either by virus-membrane fusion or by pinocytosis/phagocytosis and subsequent fusion within the host. Transinfection can significantly enhance the magnitude of virus infection, and can also mediate the trafficking of intact HIV-1 from one anatomic site to another. There is currently much debate about the mechanistic requirements for transinfection.1–8 What types of cells can act as escorts? Must virus be internalized by a viable escort? Is transinfection susceptible to inhibition by neutralizing antibodies?2–7 The current studies were designed to address these questions.
We first examined the ability of cells from different mouse tissues to potentiate HIV-1 infection. Because murine cells cannot be infected with HIV-1, this approach ensured that cis-infection (transfer of virus from infected escorts) would not confound results.
HIV-1SF2 (an X4/R5 virus) was cocultured for 2 days with “host” human TZM-bl reporter cells9 in the presence or absence of mouse cell escorts, and host cell infection was quantified as tat-induced luciferase activity. It should be emphasized that for all studies described in this report, escorts were not exposed to HIV-1 prior to the initiation of virus–host cultures. Mouse cells from peritoneal exudate [PECs, harvested by washing peritoneum with phosphate-buffered saline (PBS)], spleen, kidney, and liver all mediated transinfection (Fig. 1a), while peripheral blood mononuclear cells (PBMCs) did not.
The markedly enhanced infection observed in the presence of hepatocytes, which were poorly viable after suspension, was unexpected. We therefore assessed the contribution of nonviable cells to infection by coculturing host cells and virus with PECs (a population with good viability) that had been lysed by freeze/thaw (F/T). Surprisingly, transinfection was enhanced when PECs were lysed (Fig. 1b). Transinfection was improved with increasing PEC numbers (Fig. 1c) and was inhibited by heat treatment (Fig. 1d), indicating the necessity for protein, but not membrane integrity. Results clearly addressed debates concerning mechanisms of transinfection by demonstrating that cell viability (which would support virus internalization and cycling within the escort) was not required.
To confirm that the observed transinfection results were not unique to HIV-1SF2, experiments were also conducted with HIV-1ZM53M, a subtype C pseudovirus.10,11 As with the HIV-1SF2 cultures, infection was enhanced by F/T escort cells (Fig. 1e). Results demonstrated that escort activity was effective despite differences in virus origin.
We next questioned whether transinfection in the context of viable or F/T escort cells could be inhibited by neutralizing antibodies, as neutralizing antibodies have been reported in some cases of transinfection to have no significant effect.2,12 Might escorts circumvent the requirement for viral envelope attachment to the mammalian cell membrane as an initiator of HIV-1 infection? To address this question, HIV-1SF2 and neutralizing antibodies were mixed for overnight incubation at 37°C prior to the addition of target cells and escort cells in transinfection experiments. A representative experiment with F/T PECs is shown in Fig. 1f. As demonstrated, neutralizing antibodies (which were purified from HIV-1-positive patient blood and reconstituted to their original blood volume) inhibited viral infection in the presence or absence of PEC escorts. A preliminary study also showed that neutralization was effective when F/T PEC escorts were added prior to antibody–virus incubation. Results were consistent with the suggestion that escorts could not circumvent the participation of the HIV-1 envelope in the infection process.12
Human monocyte-derived dendritic cells (MDDC) were also tested as HIV-1 escorts.13 As before, experiments were initiated by adding uninfected escorts (either viable or F/T cells) to TZM-bl reporter cells and virus for a 2-day culture period. As shown in Fig. 2a, MDDC supported transinfection with or without F/T (in this experiment, F/T enhanced transinfection). We additionally tested human cell lines including gastric adenocarcinoma cells (previously shown to mediate transinfection via gp340-HIV-1 interactions6,7) and leukemic T cells (MT214). As shown in Fig. 2b, both cells served as escorts, particularly after F/T. These results confirmed that active processes of viable cells were not required for transinfection via either mouse or human escorts.
Not every cell could serve as an escort. Human foreskin fibroblasts (ATCC 1635) and human PBMCs, whether viable or lysed, failed to enhance virus infection (data not shown). Results likely reflected the relative lack of pertinent membrane markers (e.g., DC-SIGN, gp340) on these cell populations. These experiments served as negative controls to demonstrate that the presence of F/T cells in HIV-1 cultures was not sufficient to alter luciferase assay output.
Given that certain types of damaged cells served as potent escorts, we questioned whether fractionated membranes/microsomes could also mediate transinfection. To this end, we purified membranes/microsomes from uninfected MT2 cells. When added to virus–host cell mixtures at a dose equivalent to 105 cells, these membranes/microsomes were fully functional as transinfection escorts (Fig. 2c).
The TZM-bl cells have recently been defined as carrying a gamma-retrovirus,15 although the cells have been highly recommended for use as HIV-1 targets in endpoint neutralization assays.9 To ensure that the transinfection results were not unique to the TZM-bl cells, confirmatory experiments were performed by substituting CEM.NKR-CCR5-Luc cells16 for TZM-bl targets. As shown in Fig. 2d, there was again support of transinfection by F/T MDDCs (in this experiment the live escorts were superior to the F/T escorts). HIV-1 infection was also enhanced when CEM.NKR-CCR5-Luc targets were incubated with virus in the presence of membranes/microsomes (Fig. 2e). As additional confirmation of escort activity, experiments were repeated with a third host target, MT2. To conduct these experiments, virus (either HIV-1IIIB or HIV-1SF2) was added to targets for culture in the presence or absence of purified membranes from uninfected Mt2 cells. After 1 day, media were exchanged. After an additional 2 days, p24 values were measured by ELISA. Results from a representative experiment are shown in Fig. 2f. Again, membranes/microsomes served as potent mediators of transinfection. Results were similar when input virus was either HIV-1IIIB (Fig. 2f) or HIV-1SF2 (not shown).
The enhancement of HIV-1 infection induced by membranes/microsomes was superior to that achieved by the addition of polybrene to cell cultures. Experiments were also conducted in the presence of the antiretroviral drug AZT to ensure that virus was being produced de novo in test cell cultures. Indeed, there was essentially no p24 signal in assays when the AZT inhibitor of virus production was added to cultures. In addition, virus output was routinely improved on day 2 versus day 1 in AZT-free cell cultures, reflecting continual virus production (data not shown). Together, results showed that de novo HIV-1 production was being scored in these experiments, and that virus amplification was enhanced by the presence of escort membranes.
As a composite, these experiments proved that transinfection can be supported by a variety of escort cells, that cell viability is not required, and that the mechanism is susceptible to inhibition by neutralizing antibodies. Apparently, the sophisticated process of DC-SIGN-mediated virus internalization and transfer to a T cell synapse3 defines just one of several mechanisms by which an escort can enhance HIV-1 infections.2 Our results support those of Cavrois et al.2,17 demonstrating that internalization of HIV-1 is not required. We find that not only can damaged cells support transinfection, but that isolated membranes/microsomes (from uninfected cells) can serve as potent escorts.
What are the mechanisms of transinfection in the context of damaged cells or isolated cell membranes? While the details remain to be elucidated, it is likely that membrane fragments bearing appropriate molecules (e.g., DC-SIGN, gp340) bind virus and enhance its capacity for capture (either by fusion or endocytosis) at the host cell surface. The mere presentation of virus on a particulate structure may be sufficient to potentiate this activity. Perhaps lysed cells are in some cases better escorts than their live cell counterparts, because cell lysis exposes HIV-1 to a labyrinth of internal membranes/microsomes that supplements the plasma cell membrane in virus capture and transfer. Cell lysis may additionally alter the cytokine milieu, which could upregulate virus production by its target T cell.18 Future experiments designed to define the precise mechanisms of membrane-mediated transinfection are clearly warranted.
The fact that fractionated membranes/microsomes support transinfection may help to explain, at least in part, the association of tissue damage with HIV-1 infection in humans. Clinical trials have shown that persons with injured mucosal tissues are prone to infections with HIV-1.18–21 This association has been proposed to reflect (1) intimate contact of HIV-1 with activated T cell targets and (2) exacerbation of HIV-1 infection by cytokine release. Another explanation that can now be considered is that HIV-1 may hijack dead and fragmented cells in wounded tissues to enhance virus interactions with susceptible T cell targets.
This work was funded by NIH Cancer Center Support Core Grant P30-CA21765, NIH NIAID-P01 AI45142, NIH NIAID-R01 AI078819, and the American Lebanese Syrian Associated Charities. We thank P. Freiden for helpful technical assistance and S. Naron for assistance with scientific editing. We thank R.V. Srinivas and the AIDS Research and Reference Reagent Program (NRRRP), Division of AIDS, NIAID, NIH, for HIV1IIIB (Contributor R. Gallo). We also thank NRRRP for HIV-1SF2 (Contributor J. Levy), constructs for the production of pseudotyped HIV-1ZM53M (ZM53M.PB12, SVPC11, Contributors C.A. Derdeyn and E. Hunter and pSG3env plasmids, Contributors J.C. Kappes and X. Wu), TZM-bl cells (Contributors J.C. Kappes, X. Wu, and Tranzyme Inc), MT-2 cells (Contributor D. Richman), and CEM-NKR-CCR5-Luc cells (Contributors J. Moore and C. Spenlehauer). We thank Harold Stamey of the Tennessee Blood Services (Memphis, TN) for human blood samples.
No competing financial interests exist.