Supportive regulatory guidance for cure research requires a clear understanding of all possible outcomes, including remission
By Richard Jefferys
The term remission is increasingly being invoked in the context of cure research and, by extension, is an issue for regulatory authorities such as the U.S. Food and Drug Administration considering measurements of safety and efficacy in clinical trials. Remission, as an outcome, has been applied in a number of cases where people with HIV have interrupted antiretroviral therapy (ART) and maintained low or undetectable viral loads for some period. It is also being assessed as a possible endpoint in a clinical trial (IMPAACT P1115) aiming to test whether starting ART immediately in perinatally infected newborns might later allow for temporary or even long-term treatment interruption.
The hope is that achieving remission will represent a first step toward the discovery of a permanent cure. While this idea may seem relatively straightforward, there are important differences among reported cases in terms of how remission was achieved and significant challenges in assessing whether post-ART control of viral load leads to a state of health equivalent to that of an individual on effective ART or a comparable HIV-negative person.
Current evidence indicates that the examples of possible remission that have been reported recently (see table) fall into two categories. In the widely publicized case of the so-called Mississippi baby, in whom HIV remained undetectable for 27 months after stopping ART, and the two adults known as the Boston patients, HIV appears to have been totally inactive during the period off ART. This was likely because the reservoir of latently infected CD4+ T cells in their bodies was extremely small, lowering the probability of one of the cells becoming activated and awakening the latent HIV within it (CD4+ T cells can activate if they encounter an antigen they recognize or respond to signals from immune system proteins such as cytokines and chemokines). However, the probability was not zero, and it is thought that eventually one or more of the latently infected CD4+ T cells became activated, allowing it to generate new viruses that went on to infect new cells and cause the viral load to rebound.
Importantly, no immune responses against HIV were detectable in these three individuals until after viral load became detectable, arguing against any role of the immune system in containing the virus during the remission. In the Mississippi child, very early initiation of ART suppressed the virus before HIV-specific immunity developed, whereas in the Boston patients the maintenance of ART during their stem cell transplants (given as treatment for cancer) meant that the new immune system that developed from the donated stem cells did not encounter HIV antigens, so no virus-specific immunity was generated.
A different scenario applies in individuals referred to as posttreatment controllers, who are sometimes described as being in virological remission (the term functional cure has also been used, but is falling out of favor). The most famous examples are the VISCONTI cohort, a group of individuals in France who started ART soon after infection, remained on treatment for several years, then interrupted and maintained viral loads around or below the limit of detection (typically <20 copies/mL). At the time of the last detailed published report in March 2013, the cohort comprised 14 participants who had been off ART for an average of around 7.5 years. A brief update in a scientific review article published in January 2015 stated that this number had increased to 20, with the average time off ART at just over nine years.
Another instance of virological remission that was in the news recently involves a perinatally infected French teenager in whom ART was interrupted at around age six; with the exception of two low-level detectable readings, viral load has since been maintained below the limit of detection for over 12 years.
A unifying factor that distinguishes these individuals from the Mississippi child and Boston patients is that HIV-specific immunity is present, and while the mechanisms of viral-load control are under investigation, the preponderance of opinion is that immunologic factors are most likely involved (whether adaptive HIV-specific immune responses, innate immune responses, or some combination of both).
Amid these possible examples of remission, the question of whether there are implications for long-term health that may differ from those associated with ART-mediated HIV suppression has gone largely unasked. But it is critically important, both for the individuals concerned and for future regulatory assessments of interventions that might promote remission.
There is reason to be optimistic that in cases where HIV is completely inactive, there would be little or no possibility of the virus causing immunologic or health problems. Nevertheless, it would still be desirable to formally evaluate the question in clinical trials, which may be possible if IMPAACT P1115 is successful in recapitulating the remission experienced by the Mississippi child in some participants.
In posttreatment controllers, however, there is already some evidence to suggest that immune-mediated containment of viral load could come at a cost to long-term health. The evidence derives from studies of elite controllers (ECs), who naturally suppress HIV to undetectable levels without ART. While ECs are at a massively reduced risk of disease progression compared with untreated HIV-positive individuals with higher viral loads, it has become evident over long-term follow-up that ECs can experience a slow loss of CD4+ T cells, gradual progression to AIDS, and increases in biomarkers of cardiovascular disease. The driving factor appears to be immune activation, which, on average, is higher among ECs than in comparable HIV-negative individuals. There is also some evidence that ECs may be hospitalized more often than similar HIV-positive individuals on ART, due primarily to cardiovascular disease, but this has been reported in only one study, and it’s possible that confounding factors—such as smoking—contributed to the difference.
The potential relevance of these observations to posttreatment controllers is highlighted by a recent update on the VISCONTI cohort at the IAS Towards an HIV Cure Symposium in July. Of the 14 individuals described in the 2013 publication, one has experienced a viral-load rebound reaching close to 100,000 copies/mL after six years off ART, necessitating reinstitution of treatment. Another has a persistently detectable viral load in the range of 100–1,000 copies/mL and a declining CD4+ T-cell count that is now below 500 cells/mm3. A third is reported to have developed a head and neck cancer and has resumed treatment. One of the original 14 is now lost to follow-up. Of the remaining 10 still being followed, nine have viral loads less than 20 copies/mL, while one had a viral-load level of 211 copies/mL at the time of last measurement. The presentation also notes that six posttreatment controllers have been added to the cohort, explaining the reference to a total of 20 members from earlier this year. However, data are shown for only one of these individuals, who is controlling viral load but has a CD4+ T-cell count below 400/mm3.
Several important concerns are underscored by this news:
- The term virological remission tends to be truncated to just remission, which most people understand to mean a state of freedom from risk of disease. But the immune activity required to contain HIV in posttreatment controllers could be associated with negative health consequences, as has been reported in some ECs. Certainly, media descriptions of the VISCONTI cohort as examples of functional cures (which included a high-profile BBC story) were mistaken, and this term should not be used in relation to posttreatment control.
- The widely reported suggestion that the VISCONTI cohort would likely not face the disease progression and health risks reported in some ECs because of lower immune activation should be viewed with skepticism. Immune activation levels in these posttreatment controllers have not been compared with those in HIV-negative individuals, and no data on inflammation levels or biomarkers of cardiovascular disease risk have been presented.
- From the regulatory perspective, the benefits and risks of the HIV suppression seen in posttreatment controllers compared with that achieved by ART are currently unknown and will need to be evaluated in randomized studies. There are planned trials of ART interruption in individuals treated very early after HIV infection that may be able to look at this question if a sufficient number of participants display posttreatment control.
Since this may sound pessimistic, it should be noted that research on ECs offers reasons for hope as well as concern. There is evidence from several studies that a subset of ECs maintains extraordinarily strong suppression of HIV and shows immune activation and inflammatory gene expression profiles that closely resemble those of similar HIV-negative counterparts. And at least one reported case suggests that similarly strict control of HIV may be achievable in some posttreatment controllers. A logical implication is that the risk of HIV-related disease progression and illness would be extremely low or absent in these individuals unless levels of virus increase. These findings also imply that gradations in viral-load levels may be important even when the levels are extremely low and undetectable by standard clinical tests.
The refinement of biomarkers of immune activation and inflammation—which have been associated with both disease progression and morbidity and mortality in population-based studies—could also aid in the understanding of how low HIV levels may or may not affect health. Currently, there is a great deal of variability in how these biomarkers are measured in different studies, and it would be helpful to achieve consensus about how they should be evaluated in cure-related trials. Early discussions around endpoints in clinical trials where remission or posttreatment control is the goal have focused on standard virological measures (there is a proposed endpoint named virus suppression off therapy, or VSOT), which may not provide sufficient information about the prognosis of an individual who appears to be controlling viral load.
The overall message from recent research is that various forms of remission and posttreatment control are possible, but need to be better understood, particularly in terms of their long-term health implications. While the type of remission observed in the Mississippi baby and Boston patients appears ideal, it is very difficult to achieve because it requires very large reductions in the size of the latent HIV reservoir. The development of reservoir-reducing interventions is a key priority for cure research, and multiple trials of potential candidates are under way, but the task is challenging.
Posttreatment control has been posited as a more realistic goal in the near term, but there is a need to ensure that it leads to a state of health that is at least comparable to that attained on ART, if not better. When encountering terms such as remission, functional cure, and posttreatment control—which all too frequently have been used interchangeably—it’s important to appreciate that there remains a lack of consensus as to how exactly to define them, which will hopefully be resolved as the science evolves.
Recent HIV Remission and Posttreatment Control/Virological Remission Cases*
|Case(s)||Treatment||Period off ART||HIV Detection off ART||HIV-Specific Immunity off ART|
|ART initiated within 48 hours of birth, interrupted at around age 18 months||27 months||No HIV DNA, RNA, or replication-competent virus detectable in blood||Not detected|
Boston patient #1
|ART, stem cell transplant, and associated immune suppressants and chemotherapies for cancer||12 weeks||No HIV DNA, RNA, or replication-competent virus detectable in blood or rectal tissue||Not detected|
Boston patient #2
|ART, stem cell transplant, and associated immune suppressants and chemotherapies for cancer||32 weeks||No HIV DNA, RNA, or replication-competent virus detectable in blood or rectal tissue||Not detected|
|ART initiated within 100 days of infection and maintained for an average of ~3 years||Average of 9.3 years at the time of last published report (January 2015)||Average HIV DNA levels in blood: ~50 copies/million cells
HIV RNA ranging from <20 to ~200 copies/mL
|HIV-specific antibodies, “robust” HIV-specific CD4+ T-cell responses, low-magnitude HIV-specific CD8+ T-cell responses (not capable of suppressing HIV replication in vitro)|
|Combination ART initiated at 3 months of age, interrupted at around age 6 years of age||>12 years||HIV DNA levels in blood ranging from 125 to 316 copies/million cells
|HIV-specific antibodies, low-magnitude HIV-specific CD8+ T-cell responses (not capable of suppressing HIV replication in vitro)|
67-year-old European male (Jan van Lunzen case report)
|ART initiated within one month of seroconversion, maintained for 5.5 years||>9 years||No HIV DNA or RNA detectable in blood, cerebrospinal fluid, and gut tissue
Replication-competent HIV detected after transfer of CD4+ T cells into humanized mouse
|HIV-specific antibodies, strong polyfunctional HIV-specific CD8+ T-cell and CD4+ T-cell responses|
23-year-old African female (Sabine Kinloch case)
|ART initiated during acute infection, maintained for ~6 years with one switch due to virological failure||>10 years||HIV DNA levels in blood ~150 copies copies/million cells
|HIV-specific antibodies, HIV-specific CD4+ T-cell responses and CD8+ T-cell responses (capable of suppressing HIV replication in vitro)|
* These are recently reported examples, but several others have been reported in the past, including very rare cases involving individuals treated during chronic HIV infection. See citations included in the cure research section of TAG’s 2015 Pipeline Report.