Potential Role of the Formation of Tunneling Nanotubes in HIV-1 Spread in Macrophages

A previously unrecognized form of cell-to-cell communication based on F-actin–containing membrane extension that connects distant cells was reported in 2004 (1). These structures, referred to as tunneling nanotubes (TNTs), were first discovered in rat neuronal PC12 cells, but subsequently were found in many types of cells including immune cells both in vitro and in vivo (2–7). TNTs are dynamic structures with a lifetime ranging from minutes to several hours and can be several times longer than the diameter of the cell that forms them (8, 9). One important functional feature of TNTs is that they facilitate the intercellular transfer of a wide array of cellular signaling molecules and components including calcium ions, cytoplasmic and cell-surface proteins, and organelles such as mitochondria (10–13). It also has been demonstrated that prions hijack TNTs for their intercellular spread within neurons in the CNS (14).

Recently, viruses such as HTLV-1 and HIV-1 have been proposed to use TNTs to move from cell to cell (15–21). This pathway might allow more efficient viral spread and does not require the viruses to enter the extracellular compartment, and hence reduces their exposure to antiviral factors, cytotoxic T lymphocytes, neutralizing Abs, and/or drugs. Interestingly, TNTs that form in the two major targets of HIV-1, CD4⁺ T cells and macrophages, exhibit functional and structural differences. TNTs formed in T cells are almost impermeable to calcium ions, unlike those in macrophages, and contain junctions that are structurally distinct from the open-ended membranous tethers found in other cells including macrophages (16). Another noteworthy finding is that HIV-1 increases the number of TNTs in macrophages, but not in T cells (16, 20). Indeed, it was demonstrated that HIV-1 infection promotes the formation of short and long TNTs in macrophages, and the time course of this process is correlated with that of viral production (20).

More importantly, HIV-1–infected macrophages have been proposed to transfer viruses and/or the HIV-1 pathogenic protein Nef to B cells via TNTs, which ultimately results in the suppression of viral-specific IgG2 and IgA production (22). This finding might explain why Ab responses are ineffective in the majority of HIV-1–infected patients (23). Mechanistically, the formation of B cell–targeting TNTs in HIV-1–infected macrophages is likely to be dependent on the expression of the viral protein Nef (22), which is consistent with the finding that the p8 protein of HTLV-1, the features of which are often reminiscent of those of HIV-1 Nef (24), promotes the formation of TNT-like structures in a T cell line Jurkat (15).

Despite the presumed role of TNTs in the pathogenesis of HIV-1, it is not well understood how HIV-1 promotes TNT formation in macrophages. Although the viral protein Nef was reported to promote TNT formation when expressed in a macrophage-like cell line THP-1 (22), the cellular factor(s) involved in this process have not been identified. Most importantly, it is unclear to what extent TNTs contribute to intercellular spread of HIV-1 among macrophages.

Human peripheral blood monocyte-derived macrophages (MDMs) were prepared as described previously (25, 26). Approval for this study was obtained from the Kumamoto University medical ethics committee. Heparinized venous blood was collected from healthy donors after informed consent had been obtained in accordance with the Declaration of Helsinki. In brief, PBMCs were suspended in RPMI 1640 medium containing a low concentration of FCS (1%) to facilitate the adherence of monocytes in the subsequent step and then seeded in multiwell plates or dishes. Monocytes were enriched by allowing attachment to plates, chamber slides, or dishes for 1 h at 37°C, and nonadherent cells were removed by extensive wash with PBS. The adherent monocytes were differentiated into macrophages by culturing with RPMI 1640/10% FCS containing 100 ng/ml recombinant human (rh)M-CSF (a gift from Morinaga Milk Industry, Kanagawa, Japan). After 3 d, the cultures were replaced with fresh complete media after extensive wash with PBS to further remove nonadherent cells and were incubated for another 2 d. At day 5, the purity of MDMs prepared by this method was routinely >95% according to flow cytometric analysis of CD14 expression (data not shown). MDMs were used in the experiments described later or stimulated with rhIFN-γ (100 ng/ml; Prospec) followed by Western blotting in a selected experiment (Fig. 2C).

HIV-1–susceptible T cell lines (27) such as Jurkat (provided by H. Akari, Kyoto University, Kyoto, Japan), PM1-CCR5 (CCR5-expresing PM1 cells), and MT-2 (both provided by Y. Maeda, Kumamoto University, Kumamoto, Japan) were cultured with RPMI 1640/10% FCS. Among them, Jurkat and PM1-CCR5 cells are negative for M-Sec expression, whereas HTLV-1–infected MT-2 cells are positive for M-Sec expression (Figs. 2D, 5A). To visualize M-Sec by immunofluorescence, we allowed Jurkat and MT-2 cells to attach to the surface of collagen-coated chamber slides (BioCoat; BD Biosciences).

We identified the TNT inhibitor (TNT-i) by an affinity-based chemical array screening in a manner similar to recent studies (28, 29). In brief, the solutions of 6800 compounds (2.5 mg/ml in DMSO) from the NPDepo (RIKEN Natural Products Depository, Saitama, Japan) chemical library were arrayed onto four separate photoaffinity-linker–coated glass slides. The compounds were spotted onto each slide in duplicate. The slides were incubated with the lysates of 293T cells expressing either DsRed or DsRed–M-Sec fusion proteins. The probed slides were scanned at a resolution of 10 μM per pixel with a GenePix 4100A scanner (Molecular Devices) using the Cy3 channel (with an excitation wavelength of 532 nm and an emission wavelength of 575 nm). The fluorescence signals were quantified using the GenePix 5.0 software (Molecular Devices) with local background correction. The identified TNT-i (Fig. 4A) was purchased from Pharmeks (Moscow, Russia), dissolved in DMSO, and added to cultures at the indicated concentration (0.1% v/v). The same volume of DMSO was used as a vehicle control.

rHIV-1 was prepared as described previously (30). HEK293 cells cultured with DMEM-10% FCS were used as viral producer cells. The cells were seeded onto 12-well tissue culture plates and transfected with HIV-1 proviral plasmids using Lipofectamine 2000 reagent (Invitrogen). After 6 h of transfection, the culture medium was replaced with fresh medium, and the cells were cultured for an additional 48 h. Then the supernatants containing recombinant viruses were clarified by centrifugation, analyzed for their HIV-1 Gag protein concentrations by ELISA (MBL, Nagoya, Japan), and kept at −70°C before use. The CCR5-tropic JRFL and its Nef-deficient mutant plasmids, and the CXCR4-tropic NL43 and its Nef-deficient mutant plasmids were provided by Y. Koyanagi (Kyoto University, Kyoto, Japan) and A. Adachi (Tokushima University, Tokushima, Japan), respectively.

HIV-1 infection was performed essentially as described previously (26). The MDMs cultured on 24-well tissue culture plates were incubated with 200 μl of the supernatants of 293 cells containing HIV-1 (Gag concentration: 100 ng/ml unless otherwise stated) for 2 h at 37°C. Jurkat, PM1-CCR5, or MT-2 cells (5 × 10⁵ cells) were incubated with 500 μl of the supernatants of 293 cells containing HIV-1 (Gag concentration: 100 ng/ml) for 2 h at 37°C. Then the cells were washed twice with PBS to remove any unbound viruses and cultured with media containing rhM-CSF (only for MDM), in the absence or presence of TNT-i. The initial density of T cell lines was 1 × 10⁵ cells/ml. To monitor viral replication, we determined the concentration of p24 Gag protein in the culture supernatant by ELISA. The numbers of MDMs that formed the short or long TNTs (Fig. 1, Supplemental Fig. 1A) were quantified according to the criteria of a previous report (20), in which these TNTs were easily distinguishable from filopodia based on their length.

M-Sec mRNA was quantified by PCR using the first-strand cDNA generated from human peripheral blood fractions (Clontech), which include total PBMCs, CD14⁺ monocytes, and CD4⁺ lymphocytes (Fig. 2A). In a selected experiment (Fig. 2B), mRNA were extracted from HIV-1–infected MDMs using ISOGEN II (Nippon Gene, Tokyo, Japan), and cDNA were prepared using PrimeScript RTase (TaKaRa, Kyoto, Japan). PCR was carried out using the following primer pairs: 5′-GTGCTGGGTGC-3′ and 5′-TGACACCCTGCTCCAGAACC-3′ for M-Sec, and 5′-CCACCCTGTTGCTGTAGCCAAATTCG-3′ and 5′-TCCGGGAAACTGTGGCGTGATGG-3′ for G3PDH. DNA was amplified using LA Taq (TaKaRa) for 35 cycles (for M-Sec) or 30 cycles (for G3PDH).

Western blotting was performed as described previously (26). The Abs used were as follows: M-Sec (F-6; Santa Cruz Biotechnology), Gag (#65; BioAcademia, Osaka, Japan), and actin (C-2; Santa Cruz Biotechnology, as a loading control). The detection was performed with HRP-labeled secondary Abs (GE Healthcare), the Immunostar LD Western blotting detection reagent (Wako, Japan), and an image analyzer (ImageQuant LAS 4000; GE Healthcare).

Immunofluorescence analysis was performed essentially as described previously (31). In brief, cells were directly fixed in 2% paraformaldehyde, permeabilized with 0.1% Triton X-100, and stained with anti–M-Sec Ab (F-6; Santa Cruz Biotechnology), anti-Nef Ab (#2949; National Institutes of Health AIDS Reagent Program), or anti-Gag Ab (#Kal-1; Dako) for 12 h followed by anti-mouse IgG-AlexaFluor633, anti-rabbit IgG-AlexaFluor568, or anti-mouse IgG-AlexaFluor568 (all from Molecular Probes). Phalloidin conjugated to AlexaFluor488 and DAPI (both from Molecular Probes) were also used to visualize F-actin and nuclei, respectively. Signals were visualized with an FV1200 confocal laser-scanning microscope (Olympus). Serial Z-sections from the top to the bottom of MDMs (15–25 sections) were also obtained (Figs. 1A, 3A, 6A). Image processing was performed using the FV Viewer ver. 4.1 software (Olympus).

Transient transfection and subsequent flow cytometric analysis with 293 cells stably expressing CD4 were performed as described previously (31). In brief, cells grown on 12-well plates were transfected with a total of 1.6 μg plasmid (Nef-GFP with or without HA-M-Sec) using Lipofectamine 2000 reagent. The cell-surface expression of CD4 or MHC class I (MHC I) was determined by flow cytometry on a FACSVerse (BD Biosciences) using the FlowJo software (Tree Star) as described previously (26). The following Abs were used: allophycocyanin-labeled anti–HLA-A/B/C (W6/32; BioLegend) and allophycocyanin-labeled anti-CD4 (RPA-T4; BD Biosciences).

The numbers of cells were assessed using MTT reagent (26).

The statistical significance of the intersample differences was determined using the paired Student t test. The p values <0.05 were considered significant.

It was reported that when expressed in the macrophage-like cell line THP-1, Nef induced TNT-like thin protrusions bridging THP-1 cells with each other (22). The Nef-dependent formation of TNT-like structures was demonstrated also in MDMs (22). In this study, we initially attempted to confirm the result in our system. When infected with the wild-type (WT) HIV-1 JRFL strain, MDMs formed many visible long and thin membranous tubes within 2 d postinfection (Supplemental Fig. 1A, see orange arrowheads in middle panel). Meanwhile, such TNT-like structures were rare in mock-infected (top panel) and Nef-deficient (ΔNef) virus-infected MDMs (bottom panel), although the latter cells often formed fused giant cells (see white arrowheads), which is a characteristic of HIV-1–infected MDMs. Indeed, the visible tubes, which were found in WT virus-infected cultures and often connected MDMs with each other (Supplemental Fig. 1A, middle panel), were positive for both F-actin and HIV-1 Gag (Fig. 1A, arrowhead), indicating that HIV-1–infected MDM form TNTs.

The viral production reached its peak 10–14 d postinfection, whereas TNT-positive MDMs were obvious 1–3 d postinfection in our system (data not shown), which was consistent with the previous report (20). Thus, we focused on the initial phase of infection in the following experiments. The WT virus replicated more efficiently in MDMs than ΔNef virus (Fig. 1B, 1C) as reported previously (30). However, even when the replication of these two viruses was adjusted to equivalent levels (25 and 100 ng/ml input Gag for WT and ΔNef viruses, respectively; see Fig. 1B), the percentage of TNT-positive MDMs in WT virus-infected cultures was higher than that in ΔNef virus-infected cultures (Fig. 1D). Moreover, when compared with mock-infected cultures, ΔNef virus-infected cultures only slightly induced TNT formation (Fig. 1D). These results indicate that Nef is required for HIV-1–induced TNT formation in MDMs, which was further confirmed by comparing the percentage of TNT-positive MDMs in Gag-positive MDMs between WT virus-infected cultures and ΔNef virus-infected cultures (Fig. 1E).

The fact that HIV-1 promotes TNT formation in MDMs (20), but not in CD4⁺ T cells (22), indicates that protein(s) that are expressed in MDMs, but not in CD4⁺ T cells, are involved in the TNT formation. Thus far, M-Sec that shares homology with a component of the exocyst complex Sec6 (32, 33) and Myo10 that is an unconventional actin-based motor protein have been identified as key regulators of TNT formation (34), although the precise mechanisms by which they induce TNT formation remain unclear. In this study, we focused on M-Sec, a 73-kDa cytosolic protein, because Myo10 is undetectable in the leukocyte fraction containing monocytes/macrophages (35).

M-Sec mRNA was abundant in CD14⁺ monocytes, but not in other PBMCs, including CD4⁺ T cells (Fig. 2A). M-Sec mRNA was also readily detected in MDMs and even in HIV-1–infected MDMs (Fig. 2B). We also analyzed M-Sec expression at protein levels using Abs that detected the upregulated expression by IFN-γ in MDMs (Fig. 2C). However, CD4⁺ T cell lines such as PM1-CCR5 and Jurkat cells did not show any M-Sec expression, and HIV-1 did not induce any M-Sec expression in these cells, despite productive viral replication as indicated by the abundant Gag protein (Fig. 2D). Likewise, HIV-1 did not induce any obvious upregulation of M-Sec expression in MDMs (Fig. 2E, 2F). Nevertheless, we found that TNT-forming MDMs in WT virus-infected cultures showed a strong signal of M-Sec (Fig. 2G, lower panels), which was the case with both the thin TNT-forming MDMs (Fig. 2G, orange arrowheads) and the thick TNT-forming MDMs (Fig. 2G, yellow arrowheads). No such signal was observed with MDMs negative for TNTs in the same cultures (Fig. 2G, white arrowheads), suggesting that the intracellular distribution of M-Sec is altered in HIV-1–infected MDMs.

Indeed, the thin TNT-forming MDMs that were infected with WT virus as evidenced by multiple nuclei (Fig. 3A, see MDMs indicated by arrowhead) or the expression of HIV-1 gp120 envelope proteins (Supplemental Fig. 1B) exhibited a protuberant shape, and M-Sec was accumulated in the upper regions of the cell (e.g., Fig. 3A, second panel from the top in the right column), from which the thin TNT was elongated and connected two MDMs. The protuberant bright MDMs were easily visible in WT virus-infected cultures (see Supplemental Fig. 1A). Consistent with previous observations with TNTs formed in the macrophage-like cell line RAW264.7 (32), the image of serial Z-sections clearly demonstrated that TNTs formed by the infected protuberant MDMs did not come into contact with the substratum (Fig. 3B). These findings led us to investigate whether M-Sec is indeed involved in HIV-1–induced TNT formation (see later).

The macrophage-like cell line RAW264.7 expresses M-Sec and forms TNTs (32). Although the precise mechanisms remain unclear, M-Sec, with no known enzymatic activity, appears to initiate TNT formation by coordinating with the small GTPase RalA and the exocyst complex, the latter of which is an octameric protein complex that mediates the tethering of post-Golgi secretory vesicles (36). In this study, we found that the enforced expression of Nef significantly increased the number of TNTs in parental RAW264.7 cells, but not in a clone lacking M-Sec expression or another clone overexpressing the dominant-negative form of RalA (Supplemental Fig. 2), which is consistent with the idea that Nef uses a signaling cascade involving M-Sec to promote TNT formation in macrophages.

When transfected with WT JRFL plasmid, HeLa cells stably expressing M-Sec formed long and thin TNTs, in which Gag was detected (Supplemental Fig. 3A). As was observed in MDMs (4), parental HeLa cells cotransfected with M-Sec and JRFL plasmid also formed thick TNT (Supplemental Fig. 3B), and Gag appeared to be transferred from the Gag-expressing cell (Supplemental Fig. 3B, upper part of panel c) to the surrounding cell (lower part) through the thick TNT. Flow cytometric analysis also showed that M-Sec expression increased the amount of Gag in the cell fraction (Supplemental Fig. 4A), which was confirmed by Western blotting: M-Sec expression increased the amount of Gag in the cell fraction in a dose-dependent manner (Supplemental Fig. 4B, top panel). This change correlated with the decreased amount of Gag in the culture supernatants in an M-Sec dose-dependent manner (Supplemental Fig. 4B, middle panel). Because M-Sec expression did not affect the infectivity of viruses released into the supernatants (Supplemental Fig. 4B, bottom panel), these results suggest that M-Sec–induced TNTs alter the compartmentalization of HIV-1 by facilitating its intercellular spread without affecting other viral life cycle.

To test the possibility that M-Sec–induced TNTs are involved in HIV-1 spread among MDMs, we attempted to identify small compounds that inhibit TNT formation. To this end, we screened 6800 small-molecule compounds by the affinity-based chemical array (28, 29). As a result, we found that NPD3064 (Fig. 4A; hereinafter referred to as TNT-i) inhibited M-Sec–induced TNT formation in HeLa cells in a reversible manner (Fig. 4B), although its inhibitory mechanism is unclear. Meanwhile, TNT-i did not inhibit other well-known functions of Nef (the downregulation of cell-surface expression of CD4 or MHC I) in 293 cells stably expressing CD4, which was also observed in cells cotransfected with M-Sec (Fig. 4C). The cotransfection of M-Sec hardly affected the downregulation of CD4 or MHC I induced by Nef (Fig. 4C). Thus, these results (Fig. 4B, 4C) suggest that M-Sec is dispensable for Nef to downregulate CD4 or MHC I, and that TNT-i primarily targets M-Sec or a molecular complex containing M-Sec. Importantly, TNT-i significantly reduced the percentage of TNT⁺ MDMs in WT virus-infected cultures when used at 10 μM (Fig. 4D, 4E). Moreover, TNT-i reduced the production of WT virus in the supernatants (Fig. 4F, 4G) and the cell fraction (data not shown) of MDMs, despite no obvious cytotoxic effect of TNT-i on MDMs at the concentration (Fig. 4H). More importantly, TNT-i did not affect the production of ΔNef virus in MDMs (Fig. 4F, 4G), which replicated less efficiently in MDMs than WT virus (see Fig. 1B, 1C) and induced little TNT formation (see Fig. 1D, Supplemental Fig. 1A). These results suggest that HIV-1 promotes TNT formation in MDMs and thereby spread among the cells, and that Nef optimizes the HIV-1 spread by promoting TNT formation.

To further test this hypothesis, we next screened M-Sec⁺ cell lines and found that an HTLV-1–infected T cell line MT-2 (37) expressed M-Sec at similar levels to MDMs (Fig. 5A) and formed TNT-like F-actin⁺ membrane protrusions (Fig. 5B, orange arrowheads). The ectopic M-Sec expression in MT-2 cells might be caused by HTLV-1 infection, because HTLV-1⁺ or ATL cell lines such as C91-PL, TAXI-1, and JuanaW expressed M-Sec mRNA (38), unlike PBLs (see Fig. 2A). Interestingly, ΔNef virus replicated less efficiently in MT-2 cells than WT virus (Fig. 5C), and TNT-i reduced the production of WT but not ΔNef virus in the cells (Fig. 5D), without exhibiting obvious cytotoxic effect (Fig. 5E). In sharp contrast, in the M-Sec⁻ T cell line Jurkat (Fig. 5A), WT and ΔNef viruses replicated at similar levels (Fig. 5F), and TNT-i did not affect the production of these viruses in the cells (Fig. 5G). TNT-i was not toxic to Jurkat cells (Fig. 5H). These results further support the idea that HIV-1 spreads efficiently by promoting TNT formation through Nef in M-Sec⁺ cells such as macrophages.

The change in the intracellular distribution of M-Sec seen in HIV-infected MDMs is likely to be a critical step for TNT formation (see Fig. 2G). Thus, we finally attempted to investigate how Nef mediates this process. In a coimmunoprecipitation experiment, in which the immunoprecipitates with anti–M-Sec Abs obtained from WT virus-infected MDMs were analyzed with anti-Nef Abs, we did not detect any obvious interaction between M-Sec and Nef (data not shown). Indeed, the intracellular distribution of Nef in WT virus-infected TNT⁺ MDMs did not necessarily overlap with that of M-Sec: Nef mainly localized at the plasma membranes, whereas M-Sec was found in the cytosol (Fig. 6A). Moreover, Nef was more clearly detected within TNTs than M-Sec. Instead, we found that the nuclei were arranged in a circular pattern in WT virus-infected multinucleated MDMs, but not in ΔNef virus-infected MDMs (Fig. 6B, 6C). We also found that WT virus-infected TNT⁺ MDMs exhibited the protuberant shape even when they were unfused mononucleated cells (Fig. 6D). Thus, it appears that Nef regulates the shapes or cytoskeletal structures of infected MDMs, resulting in the accumulation of M-Sec in the upper regions of the protuberant MDMs followed by the facilitation of TNT formation and viral spread among MDMs.

One noteworthy finding of this study is that the TNT-i reduces both TNT formation and viral production in MDMs (Fig. 4). Such inhibitory effects were also observed in MT-2 cells that were positive for M-Sec and TNT-like structures, but not in M-Sec/TNT-like structure-negative Jurkat cells (Fig. 5). The intercellular spread of HIV-1 has been estimated as two to three magnitudes more efficient when cells can form physical connections with each other (39, 40). However, it has been unclear to what extent TNTs contribute to intercellular spread of HIV-1 among MDMs. Our results suggest the importance of TNTs in HIV-1 spread among MDMs: approximately half of viral production might be mediated by intercellular viral spread via M-Sec–induced TNTs during the initial in vitro infection period (Figs. 4G, 5D).

Another noteworthy finding of this study is that the TNT-i reduces the production of WT virus in MDMs, but not that of mutant virus lacking Nef (Fig. 4G), which is required for both TNT formation and optimal viral production in MDMs (Fig. 1). Similar results were obtained with M-Sec/TNT⁺ MT-2 cells, but not with M-Sec⁻ Jurkat cells (Fig. 5). It has been unclear why WT virus replicates more efficiently than ΔNef virus in several cells including MDMs, but not in other cells including Jurkat. Our results suggest that WT virus spreads efficiently by promoting TNT formation through Nef in M-Sec⁺ cells such as MDM and MT-2. Although future studies should investigate how M-Sec–induced TNTs in macrophages mediate viral spread from macrophages to CD4⁺ T cells, our results might answer the long-standing question how Nef promotes HIV-1 production in a cell type–specific manner.

HTLV-1 p8 protein promotes the formation of TNT-like structures in Jurkat T cells, which allows the rapid transfer of the virus and p8 itself to the surrounding cells (15). Although it is unclear why p8, but not Nef, induces M-Sec expression in T cells, M-Sec might promote TNT formation once expressed, regardless of type of cells. It would be interesting to examine whether the TNT-i also reduces the production of HTLV-1 in T cells. The precise mechanism by which Nef promotes TNT formation in MDMs in an M-Sec–dependent manner also remains to be determined. It has been demonstrated that Nef affects various aspects of host cell vesicular transport and cytoskeletal dynamics (41, 42). Indeed, WT virus-infected MDMs, but not ΔNef virus-infected MDMs, exhibited protuberant shapes (Supplemental Fig. 1A) and aligned nuclei (Fig. 6B). Thus, Nef might indirectly alter the intracellular distribution of M-Sec in MDMs (Figs. 2G, 3A) by manipulating their cytoskeletal dynamics, which is consistent with the finding that the distribution of Nef in WT virus-infected TNT⁺ MDMs did not necessarily overlap with that of M-Sec (Fig. 6A). Further studies are necessary to understand the molecular mechanisms by which Nef induces the protuberant MDMs and the altered intracellular distribution of M-Sec in MDMs. Recent studies have demonstrated that Nef associates with five components of the exocyst complex (43), and M-Sec interacts with the HLA III–encoded protein LST1, which is highly expressed in macrophages (44). Thus, studies including these molecules are essential to clarify how Nef promotes TNT formation through M-Sec in macrophages. It also will be necessary to examine whether Hck, a member of the Src family kinases, is involved in this process because Nef binds and activates Hck that is highly expressed in macrophages (45).

In this study, we observed the presence of Gag in TNTs and demonstrated the reduced level of Gag in the supernatants in the presence of TNT-i. However, it should be mentioned that the intercellular transfer of Gag does not necessarily imply the passage of infectious virus. Thus, more careful analyses will be needed to conclude that Nef/M-Sec–induced TNTs are indeed involved in HIV-1 spread among MDMs.

It was demonstrated that the Nef-dependent TNTs formed between infected macrophages and B cells suppresses NF-κB–induced class switch recombination by transferring Nef to B cells (22). Indeed, Nef was readily detected within TNTs formed by MDMs (Fig. 6A). Thus, TNT-targeting small compound such as TNT-i or more potent analogs might represent a new strategy for preventing the B cell abnormalities observed in the majority of HIV-1–infected patients. Moreover, Nef has been shown to strongly promote viral spread, and thereby disease progression in vivo (46–48). TNT-i and its analogs also might be effective at preventing disease progression. The identification of M-Sec–induced TNTs as an important route for HIV-1 spread will help to increase our understanding of the significance of TNTs in the pathogenesis of HIV-1 and might provide therapeutic opportunities to target HIV-1. Because TNT-like structures play a role in the efficient intercellular spread of other viruses including HTLV-1 and murine leukemia virus (49), the findings of this study will also increase our understanding of these retroviruses.

We thank K. Honda and N. Tokushige for technical assistance in chemical array screening and secretarial assistance, respectively.

The authors have no financial conflicts of interest.