Richard Jefferys, TAG
CROI 2017 offered a dizzying parade of new data. Webcasts of presentations and PDF files of posters were rapidly placed online and are accessible via the CROI website.
A fillip for kick & kill
On the cure research front, the results that drew the most attention related to a small trial combining a latency-reversing agent (the HDAC inhibitor romidepsin) with therapeutic vaccination – a strategy commonly referred to as “kick & kill.” 
Presented by Beatriz Mothe from IrsiCaixa in Barcelona, the crux was that five out of 13 recipients of the interventions have since interrupted ART and displayed control of viral load to low levels for several months (the longest a little over six months). None of the five have yet met the study criteria for restarting ART, which is a viral load over 2,000 copies/mL; the other eight participants quickly rebounded to levels above this cutoff and resumed ART.
Contrary to a slew of erroneous headlines in the mainstream media,  none of the five individuals are “virus-free”; based on the slide presentation, three appear to have viral load below the limit of detection of the assay used (20 copies/mL) whereas the other two are oscillating between the limit of detection and approximately 2,000 copies/mL.
The study represented a rollover from a prior trial that administered two therapeutic vaccines to 24 people who had started ART within three months of HIV infection. The vaccine vectors were based on a chimpanzee adenovirus (ChAdV63) and modified Vaccinia Ankara strain (MVA), both encoding antigens designed to focus T cell responses on highly conserved parts of HIV, including elements from the Gag, Pol, Env and Vif proteins. In a poster presented at last year’s CROI, Mothe reported that receipt of these vaccines shifted HIV-specific T cell responses toward the intended conserved targets but did not have a measurable effect on the size of the HIV reservoir. 
A total of 15 participants from this original trial then agreed to enrol in a follow up study, which provided booster immunisations with the MVA vector before and after three infusions of romidepsin. Eight weeks after the final MVA dose, all participants interrupt ART, and so far 13 individuals have reached this stage and contributed data to Mothe’s report at CROI 2017.
Although the numbers are small and follow up still relatively short, Mothe noted that the frequency of viral load containment in the cohort (~38%) is higher than has been observed in any studies involving early initiation of ART (where rates have varied from 0-15%). Ongoing analyses are exploring potential correlates of control, with Mothe suggesting there are hints of links between the induction of T cell responses to the conserved HIV antigens, lower HIV DNA levels, and the achievement of post-ART viral load control.
The contribution of romidepsin is unclear due to the lack of any control group, but the drug did not have a measurable effect on the size of the HIV reservoir when levels before and after administration were compared. There was evidence of blips in HIV viral load after each romidepsin dose, consistent with a latency-reversing effect. Mothe pointed out that blips also occurred after the MVA immunisations in 60% of the participants, indicating that the vaccine may have activated latently infected CD4 T cells specific for HIV antigens (a number of studies have reported that HIV-specific CD4 T cells can contain a substantial proportion of the latent HIV reservoir). 
Romidepsin infusions were associated with an array of side effects known to be caused by HDAC inhibitors—primarily grade 1 and grade 2 headaches, fatigue and nausea—and the drug also caused precipitous but transient declines in peripheral blood CD4 T cell counts of around 300 cells. One participant developed the serious complication of sepsis after the final romidepsin dose.
Additional follow up will be required before the significance of the study can be fully assessed, but it represents the first time that any kick & kill strategy has been associated with an increased frequency viral load control after ART interruption. There is an important caveat that applies to all studies reporting maintenance of low viral load in the absence of ART: while most news coverage assumes that the health benefits of viral load suppression will be the same regardless of whether the suppression is being mediated by immune responses or ART, that assumption remains unproven. 
Based on studies of elite controllers, it is possible that even low level viral load may be associated with a slight increase in the risk of morbidity and mortality compared to the stricter control of HIV replication imposed by ART. If post-ART control of viral load can eventually be induced in more significant numbers of people, there will be opportunity to more carefully investigate this issue by comparing clinical outcomes between post-ART controllers and study participants who restart or continue ART.
Complete suppression of HIV replication by ART
The question of whether low-level HIV replication persists despite ART has been a major issue of debate in the cure research field. The balance of evidence has favoured the conclusion that ART typically completely prevents HIV from reproducing in adherent individuals, but some studies have challenged this view, including a paper published last year in Nature which argued that cryptic replication occurs in lymphoid tissues. 
At CROI 2017, Mary Kearney from the National Cancer Institute addressed the question with an analysis of HIV evolution in ten children who started ART early (mostly within a few months of birth) and were followed for at least seven years. 
Two of the children experienced some lapses in suppression of viral load and served as positive controls, while the remaining eight showed no evidence of any viral load blips during follow up.
Kearney found that the evolution of HIV genetic sequences was readily apparent in the two individuals whose viral load was not continually suppressed. In stark contrast, no evidence of HIV evolution was detectable in the other eight participants, supporting the idea that ART completely stymies viral replication. Kearney suggested that differences in the types of tools used to analyse HIV evolution may explain some of the conflicting results that have been reported, noting that the Bayesian approach employed in last year’s Nature paper may mistakenly suggest that the virus has been evolving because the timing of sample collections influences the outcome of the analysis.
Kearney’s results were buttressed by a poster presentation from Morgane Rolland of the US Military HIV Research Program (US MHRP). 
Rolland studied eight individuals who initiated ART at Feibig I, an extremely early stage of infection estimated to represent the period 10-17 days after HIV acquisition. After an average of around three years, ART was interrupted as part of a protocol assessing whether HIV remission might occur. All eight participants experienced a viral load rebound within a median of 26 days, and Rolland compared the re-emerging HIV genetic sequences with those sampled at the pre-ART baseline. These analyses, like Kearney’s, revealed no evidence of HIV evolution during ART.
Jintanat Ananworanich also described the results of this US MHRP clinical trial in detail in a separate oral presentation. 
Another case of temporary HIV remission in a stem cell transplant recipient
To date, Timothy Ray Brown remains the only individual considered to have been cured of HIV infection, an outcome achieved as a result of a complex series of treatments for a life-threatening cancer that included stem cell transplants from a donor homozygous for the CCR5 delta-32 mutation (which renders immune cells resistant to most HIV strains). Brown was in attendance at CROI, celebrating reaching a milestone of ten years since those transplants were performed. But it has also been learned—as a result of the experience of two individuals known as “the Boston patients”—that HIV-positive people who receive stem cell transplants for cancer treatment can experience dramatic reductions in the HIV reservoir even when the stem cell donors lack the CCR5 delta-32 mutation.
In the case of the Boston patients, this shrinking of the reservoir ultimately allowed for a temporary period of remission after ART interruption;  the individuals were able to go for three and eight months without any signs of HIV activity, respectively, before viral load rebounded (Boston patient Gary Steinkohl has since gone public to discuss the experience of participating in this research). 
A poster at CROI 2017 from Nathan Cummins at the Mayo Clinic in Rochester reported another case of HIV remission with broad similarities to the Boston patients. 
The individual received a stem cell transplant from a donor without the CCR5 delta-32 mutation as part of his treatments for acute lymphoblastic leukemia. By day 56 posttransplant, HIV DNA was no longer detectable in blood and subsequent sampling of large numbers of cells by leukapheresis showed significant declines in measures of the HIV reservoir. An occurrence of graft-versus host disease (GVHD) was associated with apparent elimination of most of the individual’s original CD4 T cells, which were present at a frequency of around 1 in 10 cells at day 142, but had diminished to ~13 per million cells by day 265. HIV-specific antibody responses also waned.
Approval was obtained to interrupt ART on day 784 posttransplant. HIV viral load rebound occurred after a period of remission lasting 288 days (a little over nine months), with levels rising relatively slowly from 60 copies/mL initially to 1640 copies/mL five days later, at which point ART was restarted. No symptoms were associated with the reappearance of viral load, contrasting with the Boston patients who both experienced sharp increases in viral load and symptoms of acute retroviral syndrome at the time they rebounded.
A colleague of Nathan Cummins, Stacey Rizza, presented the poster, and in discussing the case revealed that the individual had the misfortune to experience a car accident shortly before viral load rebounded (without serious injury, thankfully). One speculative possibility under consideration is that inflammation caused by stress might have triggered the activation of a latently infected cell (or cells). Samples were collected throughout follow up and are now being evaluated to try and gain a better understanding of what occurred.
The potential for rare latently infected cells to persist in a dormant state for extended periods despite ART interruption was emphasised in a symposium talk by Louis Picker from the Vaccine & Gene Therapy Institute at Oregon Health Sciences University. 
Picker’s effort to develop a CMV-based vaccine for HIV has attracted considerable publicity due to unprecedented results achieved in the SIV/macaque model: when immunised with a version of the vaccine encoding SIV antigens, half the recipient macaques exert strict control over a pathogenic SIV challenge and appear to eventually clear the infection. [14, 15]
Picker outlined two possible explanations for this outcome:
- Vaccine-induced immune responses limit SIV replication to such an extent that only a very small, unstable SIV reservoir is formed, which eventually decays away over time.
- Vaccine-induced immune responses actively clear the SIV reservoir over time.
To try and discern which explanation is correct, Picker and colleagues conducted a therapeutic study in which the CMV-based vaccine was administered to SIV-infected macaques on ART (a version of the vector encoding TB antigens served as a control). The timing of ART initiation in the experiment was guided by the detection of monocyte activation after SIV challenge, because previous work suggested this coincided with the initial formation of the viral reservoir. The approach allowed the researchers to divide macaques into various groups depending on how many days after SIV challenge ART was first administered.
After 600 days of treatment, ART was interrupted in all animals. Receipt of the CMV vaccine encoding SIV antigens did not affect viral load rebound, indicating it lacked any therapeutic effect. But Picker highlighted that timing of ART showed a major influence: all six macaques started earliest, on day 4 or 5 post-challenge, did not experience any viral load rebound. Necropsy studies were eventually conducted and only rare traces of SIV genetic material could be detected in tissues. Large volumes of cells sampled from these animals were unable to transmit SIV to uninfected macaques. CD8 T cells were depleted to assess if SIV-specific CD8 T cells were suppressing the virus, but there was no return of viral load, arguing against immunological control.
Out of 35 animals administered ART from day 6 or later, only one (initiated on day 6) displayed a similar lack of rebound. But after eight months, shortly before a planned necropsy, this macaque experienced a rebound in SIV viral load. Picker noted the parallel between this outcome and those observed in the human remission cases of the Mississippi baby and Boston patients.
Picker drew several conclusions from this work:
- The window of opportunity between infection and the formation of a stable long-lived viral reservoir is tiny – in this experiment, a delay in the initiation of ART of just one day had a huge effect on the risk of rebound when ART was interrupted.
- As has been observed in human remission cases, latently infected cells can linger in an inactive state for a long time before causing a rebound in viral load. This observation supports the importance of efforts to reverse HIV latency and promote clearance of latently infected cells.
- The lack of a therapeutic effect of the CMV-based vaccine indicates that, in the preventive context, it likely works by limiting the formation of the SIV reservoir, rather than inducing immune responses capable of progressively clearing the reservoir.
Picker also mentioned that there does appear to be a case of ultra-early ART leading to temporary remission in an adult human; this is an individual who acquired HIV infection in a short period between screening for a pre-exposure prophylaxis (PrEP) demonstration project and starting on the TDF/FTC PrEP regimen. When HIV was detected in the sample taken on the first day of PrEP administration, ART was immediately substituted for PrEP (this occurred within a matter of days). Hiroyu Hatano from UCSF described the case at CROI in 2014,  and cited plans to eventually conduct an ART interruption. Early last year, ART was stopped, and the individual displayed no sign of HIV activity for 220 days before rebound was detected and treatment restarted. A full presentation of the data is expected at the International AIDS Society (IAS) conference in July 2017.
Sex differences in HIV persistence
At the IAS conference in 2015, Jonathan Karn reported that the biology of HIV latency is influenced by the oestrogen receptor on CD4 T cells, leading to sex differences in the activity of candidate latency-reversing agents (LRA). Specifically, the hormone estradiol significantly inhibited the effect of LRAs in women but not men, whereas drugs that antagonise the oestrogen receptor—including the breast cancer treatments tamoxifen and fulvestrant – enhanced latency reversal. 
At CROI 2017, Eileen Scully from Johns Hopkins University presented a poster describing results from a study comparing measures of HIV persistence in carefully matched cohorts of women and men. 
Contrary to a previously published retrospective analysis, levels of HIV DNA were not significantly different between the groups.  However multiple measures indicated that expression of HIV RNA from the viral reservoir was lower in women compared to men. These included cell-associated HIV RNA and low-level viraemia (measured with an assay capable of capturing a single HIV RNA molecule), as well as the ratio between the amount of integrated HIV DNA detected and the ability to induce HIV RNA production from the reservoir (using the TILDA assay ). Consistent with the lower HIV expression, markers of T cell activation were also significantly reduced in women compared to men. Scully concluded that biologic sex is an important consideration in cure research, and that the manipulation of sex hormones may have a role to play in efforts to target the HIV reservoir.
Jintanat Ananworanich drew attention to Scully’s poster in an excellent plenary overview of the cure research field, and emphasised the need to do more to facilitate increased participation by women. 
Dual bNAb combo leads to long-term SHIV control
In a symposium on broadly neutralising antibodies (bNAbs), Michel Nussenzweig premiered unpublished results from a collaborative experiment his laboratory has conducted with Malcolm Martin at the National Institute of Allergy and Infectious Disease (NIAID). 
The study challenged macaques with a pathogenic SHIV (SHIVAD8) and, starting three days post-infection, administered a combination of two bNAbs, 3BCN117 and 10-1074, thrice weekly for two weeks. Nussenzweig reported that, interestingly, the bNAbs led to prolonged preservation of CD4 T cells and control of viral load in treated animals, with many displaying what he described as an “elite controller phenotype.” The depletion of CD8 T cells from some of the macaques led to increased viral load, indicating that the short course of combined bNAbs had positively modulated virus-specific immunity. A paper by Nishimura et al describing the results is now in press at Nature. 
Macaque models in HIV/AIDS research
Jeff Lifson delivered an exceptional Bernard Fields Memorial Lecture on the role of non-human primate models in HIV/AIDS research. 
Included in the broad survey were a few nuggets of unpublished data from ongoing work: in a collaboration with Louis Picker, the anti-B cell antibody rituximab has been used to disrupt B cell follicles in elite controller macaques, and this led to reductions in SIV viral load and numbers of T-follicular helper cells in the lymph node, consistent with previous reports that B cell follicles can represent a sanctuary site for the virus. A similar study is now being conducted in SIV-infected macaques treated with ART.
As an alternative approach to the same problem, Lifson cited the work of Dave Ott whose laboratory is investigating the genetic modification of CD8 T cells to express CXCR5, a chemokine receptor that can guide the CD8 T cells into B cell follicles. 
Lifson showed imaging results demonstrating that this strategy successfully localised CD8 T cells to the follicles, and work is now underway to modify CD8 T cells to express both CXCR5 and T cell receptors targeting SIV antigens.
Cellular & gene therapy in cancer and HIV
Finally, Carl June from the University of Pennsylvania gave a plenary talk on the dramatic advances that have occurred in the cancer field involving genetically modified T cells. 
These approaches typically extract T cells from an individual, genetically modify them in the laboratory to target the desired antigen, then expand and reinfuse the cells, which are described as “chimeric antigen receptor” (CAR) T cells. Impressive results have been obtained against a variety of cancers, although serious safety issues have also emerged in some cases due to the potential for excessively vigorous immune reactions to cause inflammation and pathology. 
June highlighted that many different clinical trials of this type of approach are currently underway across the globe in cancer, but none are occurring in HIV. June advocated for the development of combination approaches that both engineer HIV resistance in CD4 T cells (such as the Sangamo approach, which June has been involved in studying) and modify CD8 T cell antigen receptors to better target HIV-infected cells for elimination. 
He also stressed the need to foster engineering innovations to make this type of therapy cheaper, scalable and globally accessible.
Jefferys R. TAG basic science blog. (10 March 2017)
Unless stated otherwise, references are to the programme and abstracts of the Conference on Retroviruses and Opportunistic Infections (CROI), 13-16 February 2017, Seattle.
- Mothe B et al. Viral control induced by HIVconsv vaccines & romidepsin in early treated individuals. CROI 2017, Seattle. Late breaker oral abstract LB119.
- Forster K. Five HIV patients left ‘virus-free’ with no need for daily drugs in early vaccine trials. The Independent. (23 February 2017).
- Mothe B et al. Shaping CTL immunodominance with conserved HIV vaccines after early treatment (BCN01). CROI 2016, Boston. Poster abstract 320.
- Jefferys R. HIV-specific CD4 T cells harbor the majority of latent virus: implications for therapeutic vaccines. TAB blog. (September 2011).
- Jefferys R. The challenge of defining HIV remission. TAGline, Autumn 2015.
- Lorenzo-Redondo R et al. Persistent HIV-1 replication maintains the tissue reservoir during therapy. Nature (2016). doi:10.1038/nature16933.
- Katusiime MGK et al. No evidence of ongoing HIV replication after 7 years on ART. CROI 2017, Seattle. Oral abstract 120.
- Roland M et al. Resurgence of HIV-1 Founder Viruses Following Antiretroviral Treatment Interruption. CROI 2017, Seattle. Late breaker poster abstract 299LB.
http://www.croiconference.org/sites/default/files/posters-2017/299LB_Rolland.pdf (poster PDF)
- Ananworanich J et al. HIV RNA rebound postinterruption in persons suppressed in Fiebig I acute HIV. CROI 2017, Seattle. Oral abstract 124.
- Henrich TJ et al. Antiretroviral-free HIV-1 remission and viral rebound following allogeneic stem cell transplantation: a report of two cases. Ann Intern Med. 2014 Sep 2; 161(5): 319-327. doi: 10.7326/M14-1027.
- Halter C. A case of two diagnoses. Poz magazine. (November 2015)
- Cummins N et al. 288 day drug-free remission from HIV rebound by allogeneic PBSCT. CROI 2017, Seattle. Poster abstract 319.
- Picker L et al. Therapeutic vaccination for HIV/SIV: what will it take for cure? CROI 2017, Seattle. Oral abstract 49.
- Hansen SG et al. Profound early control of highly pathogenic SIV by an effector-memory T cell vaccine. Nature. 2011 May 26; 473(7348): 523-527.
- Hansen SG et al. Immune clearance of highly pathogenic SIV infection. Nature. 2013 Oct 3; 502(7469): 100-104.
- Hatano H et al. Lack of Detectable HIV DNA in a PrEP Study Participant Treated During “Hyperacute” HIV Infection. CROI 2014. 3-6 March 2014, Boston. Late breaker poster 397LB.
- Karn J et al. Estrogen blocks HIV re-emergence from latency and points to gender-specific differences in HIV reservoirs. IAS 2015, Vancouver. Late breaker abstract TUAA0205LB.
- Scull EP et al. Sex based differences in HIV reservoir activity and residual immune activation. CROI 2017, Seattle. Poster abstract 281.
http://www.croiconference.org/sessions/sex-based-differences-hiv-reservoir-activity-and-residual-immune-activation (abstract and poster)
- Cuzin L et al. Levels of intracellular HIV-DNA in patients with suppressive antiretroviral therapy. AIDS 2015. 29(13);1665-1671. doi: 10.1097/QAD.0000000000000723.
- Procopio FA et al. A novel assay to measure the magnitude of the inducible viral reservoir in HIV-infected individuals. EBioMedicine. 2015 Aug; 2(8): 874-883. doi: 10.1016/j.ebiom.2015.06.019.
- Ananworanich J. The emerging potential for HIV cure for infants, children, and adults. CROI 2017, Seattle. Oral abstract 12.
- Nussenzweig M. Clinical studies with broadly neutralizing antibodies. CROI 2017, Seattle. Oral abstract 145.
- Nishimura Y et al. Early antibody therapy can induce long-lasting immunity to SHIV. Nature (2017) 543;559-563. doi:10.1038/nature21435.
- Lifson J et al. Insights into HIV prevention, pathogenesis and treatment from nonhuman primate models. CROI 2017, Seattle. Oral abstract 10.
- Ayala VI et al. CXCR5 dependent entry of CD8 T cells into rhesus macaque B-cell follicles achieved through T-Cell engineering. Journal of Virology, 15 March 2017. doi: 10.1128/JVI.02507-16.
- June C. Advances in cellular therapy in cancer and HIV. CROI 2017, Seattle. Oral abstract 13.
- Johnson LA et al. Driving gene-engineered T cell immunotherapy of cancer. Cell Research (2017) 27:38-58. doi:10.1038/cr.2016.154.
- Tebas P et al. Gene editing of CCR5 in autologous CD4 T cells of persons infected with HIV. N Engl J Med. 2014. 370(10): 901-910. doi: 10.1056/NEJMoa1300662.