Quantification of latent HIV induction by vorinostat

Gareth Hardy, HIV i-Base

Various compounds are being tested to induce the latent HIV reservoir, enabling its depletion by a combination of virus-induced cytopathic effects and the action of ARVs. Of these, the most studied compounds are histone deacetylase inhibitors (HDACi) such as vorinostat (also know as SAHA).

In patients on ART, vorinostat induces expression of cell-associated HIV RNA in resting memory CD4 T cells, [1]

Despite this, two cautions have recently emerged about the potential efficacy of the current leading HDACi. Firstly, the induction of HIV RNA may not lead to production of viral particles, causing the cytopathic effects that kill infected cells: the so called “kick and kill” strategy. Secondly, it is unclear what proportion of latent proviruses HDACi can reactivate.

Anthony Cillo and colleagues from the University of Pittsburgh, PA, investigated these issues by comparing the ability of vorinostat to induce viral reactivation in vitro with global T cell activation using anti-CD3 antibodies. [2]

While activating T cells through CD3 is the most effective means of reactivating latent HIV, this approach is too toxic for clinical use. In vitro, CD3 activation enables quantitation of the total proportion of proviruses that can produce infectious virions. Cillo et al treated resting CD4 T cells from 13 HIV positive patients who had received ART for an average of 8 years, with either anti-CD3 to induce maximal latency reversal, or with SAHA. Limiting serial dilutions of CD4 T cells were set up to quantify the number of cells from which HIV could be reactivated, either as whole viral particles or cell-associated unspliced HIV RNA. Virus expression was normalised to the number of cells bearing proviral DNA, to calculate the fractional provirus expression (fPVE).

The maximal fPVE was assessed by stimulating CD4 T cells with antibodies to CD3 for 7 days. The fraction of proviruses that could be reactivated by SAHA was determined be culturing CD4 T cells with 0.5 uM vorinostat for 7 days. Culture with CD3antibodies led to global T cell activation as >94% of cells expressed CD25 and HLA-DR. Treatment with vorinostat did not result in T cell activation. HIV DNA levels increased by an average of 27-fold in cells treated with CD3 antibodies, suggesting proliferation of infected cells. In cells treated with vorinostat, HIV DNA levels increased two-fold.

Expression of virions in supernatants was detected using the Roche TaqMan assay. Global T cell activation with CD3 antibodies induced virion-associated RNA in an average of 1.5% of CD4 T cells that harboured proviral DNA. Vorinostat induced virion RNA in an average of 0.12% of CD4 T cells that harboured proviral DNA. The average fPVE ratio of vorinostat to anti-CD3 treatment was 0.05, demonstrating that vorinostat induces a fraction of the proviral expression achieved with global T cell activation.

In two patients for whom sufficient cells were available, the proportion of proviruses that could produce cell-associated unspliced HIV RNA were assessed. In response to global T cell activation, 6.8% and 8.2% of proviruses produced unspliced HIV RNA from donors 4 and 5 respectively. This amounted to a 2.4 and 2.0 fold higher induction than seen for production of virions in supernatant.

In contrast, treatment with vorinostat induced unspliced HIV RNA in 0.09% and 0.19% of proviruses, which amounted to a 3.1 and 1.4-fold higher induction than for production virions.

Finally, the relationship between induction of virions and cell-associated unspliced HIV RNA was assessed, in response to vorinostat or global T cell activation. A significant correlation was found between cellular unspliced HIV RNA and virion production when cells were stimulated with anti-CD3 [rho = 0.67, p<0.001]. In contrast there was no relationship between cellular unspliced HIV RNA and virion production in cells treated with vorinostat [rho = 0.21, p=0.99]. The data distribution suggests that a relationship is not observed here because expression of cellular HIV RNA occurs without virion production. This may explain why low-levels of plasma viraemia are not consistently observed in individuals treated with vorinostat, despite three to five fold increases in cell-associated HIV RNA [1].

This data suggests that vorinostat is not a potent inducer of replication competent latent HIV. The question of why expression of cellular unspliced HIV RNA does not correlate with production of virions may have at least two answers.

First of all, mutations will arise in any part of the viral genome, which can disable replication. While unspliced RNA is produced, any of the multiple subsequent steps required for virus assembly could be disabled. Therefore expression of unspliced HIV RNA does not necessarily equate to production of virions. The disassociation between the two is probably overridden in the case of global T cell activation, as a result of maximal viral induction.

The second reason is that there are multiple mechanisms that maintain HIV latency, of which HDACs represent just one. While HDACi may enable expression of unspliced HIV RNA in a proportion of cells capable of producing virus, the remaining steps required for complete viral production may still be regulated by other cellular mechanisms. If this is the case, one solution may be to use HDACi in combination with different latency reversing agents, which act on the other cellular mechanisms, in order to achieve full reactivation of inducible proviruses allowing reservoir depletion to occur.

Despite the disappointing lack of virion production wtih the current leading HDACi compounds, there is at least one caveat worth considering. The effectiveness of latency reversing agents is assumed to depend on their ability to induce production of virions, causing viral cytopathic effects that kill infected host cells. Studies investigating the efficacy of HDACi have focused on production of virions or HIV RNA. What is not known is whether HDACi can induce expression of viral proteins. If proteins can be induced, infected cells could then become targets for components of the immune response, including cytotoxic T cells and Natural Killer cells.

These cells may need to be induced themselves, by immunotherapy. But the expression of viral proteins should be sufficient to unmask latently infected cells from the immune response. The problem here is that expression of viral proteins is a hard parameter to measure, because no assay for protein detection has the sensitivity to detect the very low frequency of reactivated latently infected cells on a single cell basis. Despite this, one recent study attempted to address this question by assessing expression of multiply spliced HIV RNA that encodes tat and rev proteins, as well as assessing the induction of new T cell responses to HIV gag as a crude indicator of renewed viral protein expression, in a patients receiving vorinostat with ART. [3]

Unfortunately, there was no evidence for increased expression of multiply spliced HIV RNA, or induction of new HIV-specific T cell responses, providing evidence that vorinostat does not induce viral protein expression. There is therefore mounting evidence that vorinostat on its own is unlikely to be effective in the proposed “kick and kill” strategy to diminish the HIV reservoir.


  1. Archin NM et al. Administration of vorinostat disrupts HIV-1 latency in patients on antiretroviral therapy. 2012. Nature. 15 (8). 482-485.
  2. Cillo AR et al. Quantification of HIV-1 latency reversal in resting CD4+ T cells from patients on suppressive antiretroviral therapy. 2014. PNAS. 111(19). 7078-7083.
  3. Elliot JH et al. Activation of HIV transcription with short-course Vorinostat in HIV-infected patients on suppressive antiretroviral therapy. PLoS Path. 2014. 10 (11).

Links to other websites are current at date of posting but not maintained.