Long-Term Productive Human Immunodeficiency Virus Infection of CD1a-Sorted Myeloid Dendritic Cells

ABSTRACT Myeloid, CD1a-sorted dendritic cells (MDC) productively replicated human immunodeficiency virus strains encoding envelope genes of either primary X4R5 or R5 strains for up to 45 days. Cell-free supernatant collected from long-term infected MDC, which had been exposed to an X4R5 virus 45 days earlier, was still infectious when placed over activated T cells. These data imply that DC can act as a persistent reservoir of infectious virus.

We exposed the CD1a-sorted MDC to HIV-GFP, an X4R5 virus containing the envelope gene of the dual-tropic, primary HIV isolate 89.6 (9) and the gene for green fluorescent protein (GFP). Like the parental HIV89.6, HIV-GFP was able to use both CXCR4 and CCR5 for virus entry, while the reporter gene GFP placed in lieu of nef allowed detection of infection at the single-cell level. Because we wanted to show that DC could represent a reservoir of productive HIV replication during chronic infection, we chose this dual-tropic HIV strain as a representative of a primary HIV isolate that can be found during chronic infection. Throughout disease progression, some HIV strains expand their usage of coreceptors for viral entry, shifting from exclusive use of CCR5 to use of CXCR4; this shift has been described for approximately half of the HIV clade B-infected individuals with progressive disease (10,27,32).
Our purified MDC were highly positive for DC-SIGN and expressed both CCR5 and CXCR4 ( Fig. 2F; Table 1). As expected, these MDC were productively infected by the dualtropic X4R5 HIV-GFP (Fig. 3). Virus-exposed MDC were cultured for 45 days, and throughout this time period, GFPexpressing and hence productively infected MDC were detected ( Fig. 3A and B). In general, between 1.5 to 4.5% of the virus-exposed MDC were GFP ϩ and thus HIV ϩ . Supernatants from infected DC collected at various time points contained p24 Gag that peaked on day 10 and was present in culture supernatants during the next 35 days (Fig. 3C). MDC on day 30 still expressed DC-SIGN at high levels (Table 1), and we were able to colocalize GFP and DC-SIGN, indicating productive HIV infection of DC at the single-cell level (Fig. 3D). On day 45 of culture, the viability of virus-exposed MDC and that of unexposed controls was 63 and 70%, respectively ( Table 2). Both cultures were CD14 negative; no surface CD14 expression was found on GFP-expressing, HIV-infected cells (data not shown). The absolute number of virus-exposed and nonexposed control cells decreased from 3 ϫ 10 7 on day one to 1.9 ϫ 10 7 and 2.1 ϫ 10 7 on day 45 after viral exposure, respectively. These data indicate that MDC can be maintained in culture for extended periods of time with the majority of the   a Viability of cells in the long-term MDC culture was determined by staining of noninfected and HIV-exposed MDC with 0.4% trypan blue.

604
NOTES J. VIROL. cells remaining viable. Furthermore, productive HIV replication by a small fraction of GFP ϩ cells did not appreciably influence the long-term viability of the entire culture. The use of dual-tropic HIV-GFP allowed us to examine whether long-term productively infected MDC could transfer viral progeny to T cells. On day 45 after virus exposure, MDC were separated by flow cytometry into GFP ϩ and GFP Ϫ frac-tions ( Fig. 1 and Fig. 4B, C, and D); approximately 4.5% of the MDC were GFP ϩ and thus productively infected. Real-time DNA-PCR with primers specific for HIV pol was used to demonstrate the presence of proviral DNA. A strong signal was detected in GFP ϩ cells (Fig. 4E), indicating that HIV-GFP entered MDC, proceeded through the early steps of the viral life cycle, and completed reverse transcription. The oligonucleotides 5Ј-CAAGGGAAGGCCAGGGAAT-3Ј, 5Ј-CCCCAAACCTGAAGCTCTCTT-3Ј, and 5Ј-CTTCAG AGCAGACCAGAGCCAACAGCC-3Ј (Integrated DNA Technologies Inc., Coralville, Iowa) represent the sequences of forward and reverse primers for PCR and the sequence of the probe, respectively. The latter was labeled on the 5Ј end with the reporter dye FAM (6-carboxyfluorescein) and at the 3Ј end with the quencher dye TAMRA (6-carboxytetramethyl-rhodamine) to identify PCR products, respectively (curves a and b represent analysis of GFP ϩ and GFP Ϫ fractions of HIV-exposed MDC, respectively; curve c represents values of nonexposed control MDC).

VOL. 79, 2005
NOTES 605 Next, the sorted GFP ϩ cells were plated in the upper wells of transwell tissue culture units (10 3 cells/well), while the lower wells contained CEMx174 cells (5 ϫ 10 4 cells). As negative controls, nonexposed MDC were placed into the upper wells in parallel plates. GFP ϩ MDC released HIV that crossed the semipermeable membrane of the transwell unit and infected CEMx174 cells; release of p24 Gag into the culture supernatant increased as a function of time (Fig. 5A). In parallel, we also examined the MDC fraction that was GFP negative 45 days after initial exposure to HIV-GFP. Only a low level of viral replication was detected in the transwell system involving CEMx174 cells in the bottom wells (Fig. 5A), possibly due to the presence of a few MDC, in which infectious HIV-GFP had initiated the replication cycle without yet expressing GFP at sufficient levels at the time of sorting.
We then investigated whether sorted GFP ϩ MDC could transmit HIV to primary autologous T cells. GFP ϩ MDC were placed into the upper wells of transwell units, while the lower wells contained either autologous T cells alone or T cells cocultured with autologous, freshly differentiated, noninfected MDC at a ratio of 10:1. In the latter culture, p24 Gag increased as a function of time (Fig. 5B); in the absence of MDC, virus replicated at a substantially lower level (120 ng/ml versus 27 ng/ml of p24 on day 15 of incubation). This indicated that GFP ϩ MDC in the upper wells shed virus that was propagated efficiently by the T cells activated by autologous DC but not by T cells cultured alone. Together, our data clearly demonstrate that 45 days after virus exposure, MDC harbored infectious virus that could be transmitted in the absence of cell-cell contact.
To investigate whether MDC can support long-term productive infection of a primary HIV isolate, we exposed CD1a- In the upper wells, HIV-infected GFP ϩ , HIV-exposed GFP Ϫ , or noninfected MDC were cultured. (B) p24 Gag analysis of samples collected from the bottom wells of transwell units containing either primary resting T cells or T cells mixed with autologous, freshly differentiated MDC that had not been exposed to HIV-GFP. In the upper inserts, GFP ϩ HIV-exposed cells were cultured. Data are presented as averages of p24 values, which were calculated from triplicates; error bars represent standard deviations. Results of one out of two representative experiments are shown. (C) Long-term cultured MDC support replication of R5 HIV. CD1a-sorted MDC (10 5 cells per well) were plated in 24-well tissue culture plates (Costar-Falcon). Next, MDC were exposed to 0.1 50% tissue culture infectious doses per cell of the R5 HIV1084i. After washing, MDC were cultured for 45 days in 1.5 ml of DC medium. In parallel, mock-infected MDC and HIV-exposed MDC were cultured in the presence of AZT (10 M). Supernatants (200 l) were collected and assayed for p24 Gag by enzyme-linked immunosorbent assay (Beckman Coulter). sorted MDC to HIV1084i, an R5 primary HIV clade C strain from Africa (14). This virus, which had been isolated from a 4-month-old infant infected either intrapartum or by breast feeding, was molecularly cloned and tested for its tropism (14). We maintained HIV-exposed and nonexposed MDC cultures for 45 days in the presence or absence of azidothymidine (AZT) (Fig. 5C). Clearly, MDC supported the replication of this strain, indicating permissiveness to both X4R5 (Fig. 3C) and R5 viruses.
In summary, CD1a ϩ sorted MDC supported HIV replication for 45 days. Progeny virus released by GFP ϩ MDC on day 45 was fully infectious and readily replicated in the CEMx174 cell line and in primary T cells cocultured with uninfected autologous MDC. T cells upon contact with noninfected MDC undergo activation, which leads to the upregulation of HLA-DR, CD25, and CD69 (31). According to earlier data, activation of T cells is one of prerequisites for productive HIV replication (8,22,28). We found that even a low amount of HIV released by the long-term infected MDC was sufficient to rapidly spread infection in the MDC-T-cell microenvironment.
We demonstrated here for the first time that CD1a-sorted, DC-SIGN ϩ , HIV-exposed MDC survive and release infectious virus during a protracted period of time. It would be interesting to compare the life span of our long-term in vitro HIVinfected DC with the life span of human DC in vivo. Few data are available about the life span of DC. Earlier studies reported that Langerhans cells represent a long-lived cell population (15,16,18). For instance, Langerhans cells were identified throughout 9 weeks in human skin grafts, which were transplanted to BALB/c nude mice (18). MDC are motile cells circulating in peripheral blood, residing in mucosa, and easily migrating across vascular endothelium (5,11). Our data imply that MDC could form a long-lived, motile HIV reservoir with an important role in disseminating infectious virus through peripheral blood and in lymphoid and nonlymphoid tissues. Both productive infection and DC-SIGN-mediated virus capture could contribute to HIV transmission in vivo.