Current models strongly suggest that chronic immune activation in T cells is a strong driver of HIV-1 disease progression. However, the mechanisms of this effect remain largely unexplained. Although HERVs are encoded in the DNA in all cells, their expression appears to be actively modulated in mature cells by methylation and small RNAs (reviewed in 
), and retrotransposition is inhibited by innate antiretroviral mechanisms such as APOBEC3 
. In certain pathological states such as HIV-1 infection, they can be expressed but not recognized as self-antigens, and therefore may elicit an immune response 
that is analogous to the antiviral response against an exogenous infection. Given that exogenous viruses can play a partial role in exacerbating immune activation 
, we had hypothesized that HERVs could also be an important driver of abnormal activation in chronic HIV-1 disease.
HERV-K family of HERVs are the most recent endogenous retroviruses to be incorporated into the genome 
and are therefore the most conserved, with several reading open frames in many of the dozens of inserts throughout the genome. One group (HERV-K(HML2)) has 91 insertions of full or near full length proviruses 
. This makes it a good candidate for expressing mRNA capable of being translated into immunogenic molecules, in addition to being able to be recognized as viral nucleic acid by the innate immune system. Env and Gag peptides are the most T cell immunogenic genes of HIV-1 
, so we tested the expression of these genes in the HERV-K family.
To address this question, we studied untreated HIV-1-infected individuals who had progressed to advanced or severe immunologic stages of infection 
, with high viral loads and low CD4+
T cell counts. As expected, these subjects had increased CD4+
T cell activation compared to healthy controls. While HERV-K gag
expression was present to a low degree in some HIV-1-negative subjects, the infected group had a significantly higher level of relative HERV-K expression, as expected and previously reported 
. Surprisingly, the HIV-1-infected subjects with the greatest immune activation were among those with the lowest HERV-K expression, suggesting that HERV expression does not contribute to pathologic immune activation in chronic HIV-1 infection.
HERV epitopes have been proposed as potential, non-mutating markers of HIV-1 infected cells that can be targeted and lysed by cytotoxic T cells. The fact that HIV-1-infected individuals with a stronger and broader T cell response against HERV are able to control HIV-1 viremia better than other patients 
and our present data showing that subjects with less advanced disease (as characterized by abnormal immune activation) included those with increased HERV expression, although many individuals with low HERV-K expression also evidenced lower abnormal immune activation and would suggest that expression of HERV genes and their protein products is associated with a more controlled form of HIV-1 disease.
As HIV-1 infection advances and escapes immune pressures from the host, the immunological state of the infected individual deteriorates. Many studies have shown that in more advanced HIV-1 infection, immune cells express pro-apoptotic markers 
, the strength and breadth of immune response to HIV-1 antigens decreases 
, immune effector cells are activated even without T-cell receptor (TCR) stimulation 
, and less APOBEC3G is expressed than in recently infected or exposed uninfected individuals 
. Our previous reports, and the data presented here, support the notion that HERV expression and the immune responses against them are part of a beneficial response that the host mounts against HIV-1 infection, but that is lost when HIV-1 has finally escaped immune pressures and the patient progresses towards AIDS if not treated with antiretroviral drugs. Of note, we have previously shown that individuals who have an undetectable HIV-1 viral load as a result of ART do not have a strong HERV-specific T cell response (in contrast to untreated HIV-1 controllers), suggesting that this response is a potential cause rather than a consequence of good viral control 
shows that many patients with low levels of immune activation also have low levels of HERV-K expression, and the current work does not allow further interpretation of this observation. However, future longitudinal studies may test whether these patients have more rapid disease progression than the ones who still express HERV-K.
Interestingly, we observed that HERV-K expression, at least at the level of transcription, was inversely associated with immune activation but not with standard clinical measures of HIV-1 disease status (CD4+ T cell blood count and plasma viral load) in patients with progressive infection. Future studies will focus on examining these relationships in a longitudinal design including earlier time-points prior to CD4+ T cell decline and later time-points to determine the effect of antiretroviral treatment on HERV-K expression and immunity.
There are several possible mechanisms that can be speculated for the inverse correlation between HERV-K expression and immune activation that we observed in this study. HERV transcription has a distinct pattern of expression in the normal state 
and certain pathological states are thought to disrupt this pattern and cause transcriptional leakage of HERVs 
. It is possible that HIV-1 causes a similar effect, leading to targeting by cytotoxic T cells 
or antibodies 
. As HIV-1 persists and mutates to escape CTL and antibody immune responses, some mutations may also leave the HERV-K inhibition intact, and therefore serve to evade the host's immune system against HIV-1.
It is possible that HERV-K expression leads to the development of HERV-specific cytotoxic T cells that lyse HIV-1-infected cells and thereby indirectly contain immune activation. However, we did not observe high levels HERV-specific CD8+
T cells in this group of untreated progressors (reported in 
), though the peptide set tested was limited and did not completely correspond to the HERV-K gag
sequences used to identify mRNA expression. It is also possible that in some individuals, HERV expression leads to the develoment of HERV-specific regulatory T cells (Tregs). Recent literature demonstrating the dependence of murine Treg expansion on endogenous retroviral antigens encoded in the genome also indicates that there may be an intriguing and important role for HERVs in the development of Tregs and their impact on the adaptive immune response to HIV-1 infection 
. In another scenario, immune activation in advanced HIV-1 disease itself might lead to less HERV-K expression via as yet undefined pathways, though older studies suggest that the opposite is true: activated cells actually express greater levels of HERV-K. Upcoming studies involving larger cohorts of HIV-1-infected subjects with a broad range of natural disease control, including elite controllers, non-controllers and ART-suppressed individuals will be necessary to clearly delineate the relationship between immune activation, HERV-K expression and immunity. We speculate that HERV-K expression could be an alternative mechanism for eliminating HIV-1-infected cells without inducing immune activation (in contrast to HIV-1-specific T cells) given that HERV-K epitopes cannot escape immune responses by mutation as HIV does, and high expression of HERV-K is actually associated with lower activation.
Further studies also should focus on determining the role each cell type in HERV expression, and a systematic analysis of the individual HERVs that are expressed in HIV-1 infection will also be of great interest. The current study forms the basis for future investigations into the complex and important interactions between exogenous and endogenous retroviruses, and the role that HERV might play in HIV-1 pathogenesis.