Research Article

Spread of HTLV-I Between Lymphocytes by Virus-Induced Polarization of the Cytoskeleton

See allHide authors and affiliations

Science  14 Mar 2003:
Vol. 299, Issue 5613, pp. 1713-1716
DOI: 10.1126/science.1080115

Abstract

Cell contact is required for efficient transmission of human T cell leukemia virus– type 1 (HTLV-I) between cells and between individuals, because naturally infected lymphocytes produce virtually no cell-free infectious HTLV-I particles. However, the mechanism of cell-to-cell spread of HTLV-I is not understood. We show here that cell contact rapidly induces polarization of the cytoskeleton of the infected cell to the cell-cell junction. HTLV-I core (Gag protein) complexes and the HTLV-I genome accumulate at the cell-cell junction and are then transferred to the uninfected cell. Other lymphotropic viruses, such as HIV-1, may similarly subvert normal T cell physiology to allow efficient propagation between cells.

The human T cell leukemia virus–type 1 (HTLV-I) is an oncogenic exogenous retrovirus that infects between 10 and 20 million people worldwide. Of these infected individuals, 2 to 3% develop adult T cell leukemia/lymphoma (1), and a further 2 to 3% develop a variety of chronic inflammatory syndromes, most notably HTLV-I–associated myelopathy/tropical spastic paraparesis (HAM/TSP) (2,3).

HTLV-I is transmitted between individuals by transfer of infected lymphocytes in breast milk, semen, or blood (4). Transfusion with cell-free blood products appears to carry a negligible risk of HTLV-I infection (5) . In vitro, efficient spread of HTLV-I infection also requires cell contact (6, 7). Cell contact is required because lymphocytes naturally infected with HTLV-I produce very few cell-free HTLV-I virions and because, of the virions that are released, only 1 in 105 to 106 is infectious (8, 9).

The mechanism of cell-to-cell spread of HTLV-I is not understood. HTLV-I expresses a surface glycoprotein, the envelope (Env) protein, which is required for infectivity (9) and for cell-cell fusion and syncytium formation (10–12). Env is presumed to bind to a cellular receptor for HTLV-I, but the receptor has not yet been identified (13, 14). Certain integrins, including intercellular and vascular cell-adhesion molecules ICAM-1, ICAM-3, and VCAM, act as cofactors for HTLV-I–induced cell fusion (15, 16).

In this study we tested the hypothesis that HTLV-I is transmitted directly across the cell-cell junction. We used confocal microscopy to examine the distribution of HTLV-I Gag and Env proteins and the HTLV-I genome in fresh, unstimulated peripheral blood mononuclear cells (PBMCs) isolated directly from HTLV-I–infected individuals and fixed and stained within 24 hours. We conclude that HTLV-I subverts normal T cell physiology to spread efficiently between host cells, without the need to release cell-free virus particles.

Polarization of virus proteins and nucleic acids.

In isolated lymphocytes, we observed clusters of Gag p19- and p15-staining material, predominantly near the cell membrane, as previously described (17) (Fig. 1, A and B; fig. S1, a and d). Env gp46 staining appeared uniformly on the cell surface (Fig. 1C). When T cells were allowed to form conjugates with neighboring cells, within 40 min there was strong polarization to the area of cell-cell contact of both Gag protein (Fig. 1, D and E) and Env protein (Fig. 1F).

Figure 1

HTLV-I Gag and Env proteins are unpolarized in an isolated T cell, but accumulate at the cell-cell junction within 40 min of cell contact; Gag protein is transferred from HTLV-I–infected T cells to uninfected T cells within 120 min. (A to C) Single confocal sections showing isolated CD4+ T cells from a patient with HAM/TSP. (A) CD4+ T cell, tubulin-alpha (green) and Gag p19 (red). (B) CD4+ T cell, tubulin-alpha (green), Gag p15 (red). (C) CD4+ T cell, Env gp46 (red). (D to F) Confocal images showing polarization of HTLV-I Gag and Env proteins to the cell-cell junction. Conjugates were allowed to form for 40 min between fresh CD4+ T cells from a patient with HAM/TSP. (D) CD4+ T cell, Gag p15 (red). (E) CD4+ T cell, Gag p19 (red). (F) CD4+ T cell, Env gp46 (red). (G and H) Confocal images showing transfer of Gag p19 protein from HTLV-I–infected T cells to uninfected T cells. Conjugates were allowed to form for 120 min. (G) HTLV-I–infected CD4+ and normal CD4+ T cell, Gag p19 (red). (H) HTLV-I–infected CD8+ and normal CD4+ T cell, Gag p19 (red). HTLV-I–infected T cells were marked with carboxyfluorescein succinimidyl ester (CFSE) (green). The transmission picture [(B) to (H) blue] is superimposed on a 0.4-μm confocal fluorescence single section [(C) to (F) red, (B), (G), and (H) red and green]. Scale bar, 5 μm.

The adhesion molecule talin accumulated at the cell-cell junction in conjugates containing HTLV-I–infected T cells and formed a ring in about 35% of cases (Fig. 2, B and C; fig. S2, a to c), or a single patch (about 60% of cases) or multiple patches (5%). HTLV-I Gag protein accumulated either in the central talin-free domain (Fig. 2, B and C; fig. S2, a and b) or as clusters overlying or adjacent to talin patches. This accumulation of Gag protein in the center of the cell-to-cell junction was observed in conjugates between CD4+ and CD4+ T cells (Fig. 2, B and C), CD4+ and CD8+ T cells (fig. S2a), and CD8+ and CD8+ T cells. The presence in these accumulations of nucleocapsid staining (p15) (Figs. 1D and 2A) is significant because the nucleocapsid binds the retroviral genome and incorporates it into the virion (18).

Figure 2

(A to C) HTLV-I Gag protein and talin accumulate in distinct domains at the cell-cell junction. Conjugates were allowed to form for 40 min between fresh cells from a HAM/TSP patient. (A) CD4+ T cell conjugate, talin (green), Gag p15 (red). [(B) and (C)] CD4+ T cell conjugate, talin (green), and Gag p19 (red). (C) Z-axis image reconstruction from (B), talin (green), and Gag p19 (red). (D to G) The MTOC lies adjacent to the polarized HTLV-I Gag protein at the cell-cell junction; however, treatment with nocodazole, an inhibitor of tubulin polymerization, blocks both the polarization to the cell-cell junction and cell-cell transfer of Gag protein. (D) CD4+ T cell conjugate, tubulin-alpha (green), HTLV-I Gag p19 (red). (E) Conjugate between HTLV-I–infected CD4+ T cell labeled with CFSE (green) and normal CD4+ T cell, HTLV-I Gag p19 (red). (F) Autologous CD4+ T cell conjugate, tubulin-alpha (green), HTLV-I Gag p19 (red). (G) Autologous CD4+ T cell conjugate, formed in the presence of PHA-L, talin (green), HTLV-I Gag p19 (red). [(E) to (G)] Treatment with nocodazole. The transmission picture [(A) and (E) blue] is superimposed on a 0.4-μm confocal fluorescence single section (red and green). Scale bars, 5 μm.

When CD4+ or CD8+ T cells isolated from a HAM/TSP patient were allowed to form conjugates with T cells from a healthy uninfected donor for 120 min, in addition to accumulation of Gag p19 staining at the cell-cell junction, there was frequent Gag p19 staining in the cells derived from the uninfected donor (Fig. 1, G and H). We observed transfer of Gag p19 staining from CD4+ T cells and CD8+ T cells to both CD4+ and CD8+ allogeneic T cells. This process may represent the initial establishment of HTLV-I infection in a newly infected individual, which involves contact between allogeneic lymphocytes.

Polarization of Gag complexes to the cell-cell junction and transfer to the uninfected cell were also observed in conjugates between CD4+ T cells and both B cells (fig. S1, b and c) and NK cells (fig. S1, e and f).

We used an antisense peptide nucleic acid (PNA) probe in fluorescence in situ hybridization (FISH) to detect the (plus sense) HTLV-I genome, which is normally bound to the Gag polyprotein during retroviral particle formation. The PNA probe stained the HTLV-I producer cell line MT-2 (Fig. 3D). In conjugates formed between MT-2 cells, HTLV-I nucleic acid accumulated at the cell-cell junctions. Uninfected Jurkat cells (Fig. 3C) were not stained. A plus-sense PNA probe, corresponding to the same 15 nucleotides in the HTLV-I Gag gene, was also used as a negative control. In isolated T cells that were naturally infected with HTLV-I, the viral nucleic acid was not polarized (Fig. 3A). However, in two-cell conjugates, the HTLV-I genome accumulated at the cell-to-cell contact area (Fig. 3B); when there, it resembled the polarization of Gag and Env proteins (Fig. 1, D and E). After a 120-min incubation, HTLV-I RNA was transferred from HTLV-I–infected cell to uninfected cell (Fig. 3, E and F), like the Gag protein (Fig. 1, G and H).

Figure 3

The HTLV-I genome accumulates at the cell-cell junction and is then transferred to the uninfected cell. (A to D) Conventional fluorescence images showing plus-strand HTLV-I nucleic acid by PNA-FISH and transmission picture (blue). (E and F) Confocal images showing plus-strand HTLV-I nucleic acid by PNA-FISH and transmission picture (blue). [(A), (B), and (D) to (F)] HTLV-I nucleic acid (red). [(A) and (B)] CD4+ T cell from HAM/TSP patient, conjugation time 40 min. (C) Jurkat cell (negative control), conjugation time 40 min. (D) MT2 cell (positive control). [(E) and (F)] CD4+ T cell from HAM/TSP patient (marked green with CFSE) and control (uninfected) CD4+ T cell, conjugation time 120 min. Scale bars, 5 μm.

Reorientation of microtubule organizing center.

We observed frequent reorientation of the microtubule organizing center (MTOC) to the area of cell-cell contact in lymphocyte conjugates (Fig. 2D; fig. S2, d and e): in each case the MTOC lay immediately adjacent to the accumulation of HTLV-I Gag protein. This close apposition of polarized Gag molecules to the MTOC suggested that the microtubule cytoskeleton affected the polarization of Gag. In CD4+-CD4+ T cell conjugates, treatment with 33 nM nocodazole for 90 min blocked the polarization and transfer of Gag protein, both in the absence (Fig. 2, E and F) and the presence (Fig. 2G) of the T cell activator phytohemagglutinin-L (leucoagglutinin, PHA-L). These results suggest that microtubules are involved in transporting Gag-containing material toward the cell-cell junction before transfer into the recipient cell.

Binding of the T cell receptor (TCR) to the major histocompatibility complex (MHC) and the antigen on the surface of another cell causes reorientation of the responding T cell's MTOC to the cell-cell junction. Surprisingly, there was a significant association between MTOC polarization and CD4 positivity in conjugates between autologous CD4+ and CD8+ T cells from an infected individual (P = 0.046, Fisher's exact test). This observation raised the possibility that the MTOC polarization was associated with HTLV-I infection of the T cell and was not triggered by antigen recognition.

To test this possibility, we counted the orientation patterns of MTOCs in 304 spontaneous two-cell conjugates formed between fresh CD4+ T cells from two HTLV-I–infected subjects and one uninfected control. The results (table S1) showed a strong association between Gag p19 positivity and MTOC polarization to the cell junction in the same cell. The odds ratio of MTOC polarization in a Gag p19+ cell, compared with a Gag p19 cell, was 4.07 (95% confidence interval, 3.07 to 5.39; χ2 = 99; P ≪ 0.001).

Thus, MTOC polarization in a CD4+ T cell was not triggered by TCR-mediated recognition of HTLV-I antigens presented by the other (infected) T cell: rather, the polarization occurred inside the infected T cell. HTLV-I infection of a T cell apparently induced the cytoskeletal rearrangement that occurred when the HTLV-I–infected T cell made contact with another cell.

Our observations do not rule out other pathways of cell-to-cell spread of HTLV-I, including a contribution from infectious cell-free virions. However, infection by cell-free HTLV-I particles in vitro is very inefficient (8, 9). HTLV-I has retained a functional envelope protein, which is required for infectivity and for HTLV-I–induced cell-cell fusion (9, 10). Clarification of the precise roles of HTLV-I Env in cell-to-cell transmission awaits identification of the cellular receptor(s) and electron microscopic studies of the membrane contact area between the cells. It is possible that the critical role of HTLV-I Env protein is to cause fusion of the two cell membranes (10).

Initiation of polarization.

Two factors appeared to be necessary to initiate the observed polarization of the cytoskeleton: HTLV-I infection of the cell and contact with another cell. It is not yet clear which molecules mediate these signals. HTLV-I Env protein is again a candidate for this function, because it is the only HTLV-I protein that is expressed intact on the outside of the infected cell. However, HTLV-I also up-regulates expression of certain adhesion molecules such as integrins (19, 20), which will increase the likelihood of cell-cell adhesion. Furthermore, Yamamoto et al. (20) found that ligation of ICAM-1 on the cell surface induces expression of HTLV-I genes, which suggests the existence of a positive feedback loop between cell-cell adhesion and HTLV-I gene expression (fig. S3).

HTLV-I Gag protein, in complex with the HTLV-I genome, appears to be transported to the MTOC by a microtubule-dependent process. Microtubules have been shown to be involved in the intracellular transport of other viruses, e.g., adenovirus and herpesvirus (21–23).

The junction formed between an HTLV-I–infected T cell and another T cell shared two similarities—ordered talin domains and MTOC polarization—with the “immunological synapse” (24). However, in the present study the MTOC polarization occurred within the HTLV-I–infected cell, not toward the infected cell. Therefore, MTOC polarization was not triggered by recognition of HTLV-I antigens presented by a neighboring T cell, and the structures we report here cannot be considered an “immunological” synapse. The term “virological synapse” may be more appropriate.

HTLV-I can infect almost any mammalian cell in vitro, but in vivo it is almost confined to T cells, for unknown reasons (25–27). It is possible that T cell–specific factors are required either for efficient HTLV-I replication or for the process of cell-to-cell transfer reported here.

We conclude that HTLV-I exploits the normal physiology of the T cell to enable efficient cell-to-cell transmission by forming a close contact with the recipient cell and using the cytoskeleton to propel viral material into the recipient cell (fig. S3). Although HTLV-I has a peculiarly strong dependence on cell contact for efficient transmission of the virus between cells, it is possible that other lymphotropic viruses, such as HIV-1 (28, 29), use a similar mechanism to spread between lymphocytes.

Supporting Online Material

www.sciencemag.org/cgi/content/full/1080115/DC1

Materials and Methods

Figs. S1 to S3

Table S1

References

  • * To whom correspondence should be addressed. E-mail: c.bangham{at}imperial.ac.uk

REFERENCES AND NOTES

View Abstract

Navigate This Article