News & Views | Published:


How to escape treatment

Nature volume 477, pages 3637 (01 September 2011) | Download Citation

Even during effective treatment with antiretroviral drugs, low levels of HIV persist. In part, this could be due to cell-to-cell transfer of multiple virions and the drugs' inability to inhibit replication when virus levels are high. See Letter p.95

One of the great triumphs of modern medicine is the development of combination antiretroviral drug therapy for the management of HIV infection. For those with access to these drugs, modern regimens can reduce the amount of circulating virus to very low levels. Despite their inherent potency, however, antiretroviral drugs are not curative, and HIV persists indefinitely. Consequently, patients have to adhere to these expensive and potentially toxic drugs for life. On page 95 of this issue, Sigal and colleagues1 provide insights into why antiretroviral drugs cannot fully inhibit HIV replication. Their data have notable implications for the emerging efforts aimed at curing HIV infection2.

Several mechanisms might contribute to HIV persistence during antiretroviral therapy. These include maintenance of a transcriptionally silent (latent) HIV genome in long-lived, resting target cells such as CD4+ memory T cells; proliferation of these latently infected cells; inadequate anti-HIV clearance mechanisms; and ongoing de novo infection of susceptible target cells through continued viral replication3,4. Although most researchers agree on the involvement of the first three mechanisms, the degree to which HIV can effectively replicate during therapy is a highly contentious issue.

The arguments against persistent replication are that the virus does not evolve in peripheral blood, and that following intensification — the process whereby potent drugs are added to a stable regimen — the steady-state levels of HIV RNA in the plasma do not change5.

Nevertheless, several lines of evidence support ongoing replication. Preliminary studies6,7 suggest that intensification may affect levels of the virus in tissues known to be enriched in HIV-susceptible target cells; the rapid reduction in virus levels in response to an antiretroviral drug can most easily be explained by the inhibition of the ongoing complete cycles of replication. Also, cells of many treated individuals contain unintegrated episomal HIV DNA (a potential marker of recent cell infection)8. Finally, activated cells contain higher levels of HIV DNA than resting cells do9 — an observation that is more readily explained by active cycles of infection than by preferential activation of cells that harbour latent virus.

Sigal et al.1 provide a compelling and intuitive mechanism that might account for the failure of potent drugs to completely inhibit transfer of replication-competent virus. Using a complex mathematical model, they show that the drug concentration required to prevent a single transmitted virion from successfully infecting a target cell is much lower than that needed to stop multiple transmitted virus particles from infecting the same cell. They then show in an in vitro system that physiologically relevant drug concentrations can readily inhibit cell-free transmission — that is, transmission of a free virus to a target cell — but not cell-to-cell transfer of the virus (Fig. 1). These modelling and experimental findings are consistent, because cell-to-cell transmission is known10,11 to be associated with the transfer of multiple virions to the target cell.

Figure 1: HIV infection of target cells.
Figure 1

HIV infects CD4+ memory T cells, which are scattered throughout the body. a, Where these cells are sparse (for instance, at effector sites in the mucosa) successful infection is likely to involve cell-free transfer of single virions (blue pyramids) to distant target cells (arrow). b, But in areas of high T-cell density (such as inductive tissues in lymph nodes), transmission is likely to involve direct transfer of multiple virions into adjacent cells. Sigal and colleagues' findings1 suggest that antiretroviral drugs (small pink circles) readily inhibit cell-free transfer of single virions, but may not completely stop cell-to-cell transfer of virions in target-rich areas.

Sigal and co-workers further show that under conditions of maximal drug exposure, the replication ratio following cell-to-cell transfer was less than 1, yet greater than that predicted for complete suppression. This suggests that although replication may not fully account for HIV persistence, it is likely to be a contributing factor. Finally, the authors argue through another mathematical model that low-level replication of the virus can replenish the virus reservoir even in the absence of obvious evolution, because localized tissue-based chains of infection are independent, intermittent and unlikely to be linked temporally. This is yet another intuitive conclusion that is easy to place in the context of existing evidence, but like all the predictions in this paper it is exceedingly difficult to prove in vivo.

I should emphasize that these observations1 are not definitive. An important matter for the field is to define whether cell-to-cell virus transmission does indeed occur during effective therapy. Given the logistics of accessing lymphoid tissues and their T cells in humans, it may be necessary to develop a robust non-human primate model in which animals are treated for prolonged periods. Such work could complement the ongoing studies aimed at defining the size and distribution of the HIV reservoir in humans.

One might argue that the existing antiretroviral regimens will probably be effective as long as a patient harbours drug-susceptible virus and adheres to the correct dosage of the drugs. So why should we care whether there is a small amount of difficult-to-prove de novo infection of certain cells in lymphoid tissues?

There is growing recognition that delivering antiretroviral drugs for life to everyone who might benefit from them is not going to be possible. Despite the unprecedented global investment in providing drugs, the number of newly diagnosed infections remains far greater than the number of people with access to therapy. As resources are likely to become more limited over time, we will not be able to treat our way out of this epidemic. And even if therapy is delivered, many people cannot fully adhere to it over timescales of years to decades. These public-health and individual failures result in significant harm to the individuals (owing to untreated progressive HIV infection) and to the society (because untreated people are far more likely than treated people to transmit the virus to others12).

The only way to fully address individual and public needs is to cure those infected with HIV. Given the existence of long-lived reservoirs of HIV, even complete inhibition of this virus's replication is unlikely to clear it completely from the body. But achieving complete inhibition is almost certainly going to be necessary for the many other curative strategies now being considered2. Addressing the questions Sigal and colleagues1 raise in a more definitive in vivo experiment will be a challenge, but it is probably among the most crucial tasks for those who conduct translational research in this area.


  1. 1.

    et al. Nature 477, 95–98 (2011).

  2. 2.

    Science 332, 784–789 (2011).

  3. 3.

    et al. Science 323, 1304–1307 (2009).

  4. 4.

    et al. Science 329, 174–180 (2010).

  5. 5.

    et al. Proc. Natl Acad. Sci. USA 106, 9403–9408 (2009).

  6. 6.

    et al. Nature Med. 16, 460–465 (2010).

  7. 7.

    et al. AIDS 24, 2451–2460 (2010).

  8. 8.

    et al. PLoS Pathog. 7, e1001303 (2011).

  9. 9.

    et al. J. Clin. Invest. 115, 3250–3255 (2005).

  10. 10.

    et al. Nature 418, 144 (2002).

  11. 11.

    et al. J. Virol. 85, 7169–7176 (2011).

  12. 12.

    et al. N. Engl. J. Med. 365, 493–505 (2011).

Download references

Author information


  1. Steven G. Deeks is in the Department of Medicine, University of California, San Francisco, San Francisco, California 94110, USA.

    • Steven G. Deeks


  1. Search for Steven G. Deeks in:

Corresponding author

Correspondence to Steven G. Deeks.

About this article

Publication history



Further reading


By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.

Newsletter Get the most important science stories of the day, free in your inbox. Sign up for Nature Briefing