The 2019 Conference on Retroviruses and Opportunistic Infections (CROI) took place in Seattle from March 3rd-7th. The major news on the cure research front was the possibility of two additional cases similar to Timothy Ray Brown, who for the past 12 years has been the only individual considered cured of HIV infection (TAG issued a statement offering our perspective). Other notable reports included the first results from a study using CRISPR/Cas9 to target latent SIV in macaques and a novel insight into the source of residual viral load in people on ART. Immediately preceding the conference, TAG joined with multiple other organizations to co-sponsor the annual pre-CROI community HIV cure research workshop; slides and video from the event are available online.
Stem Cell Transplant Recipients as Possible Cure Cases
Eleven years ago at CROI, a little-noticed poster by Gero Hütter and colleagues provided the first description of the case of Timothy Ray Brown (among the few people to draw attention to it at the time was activist Martin Delaney of Project Inform). As part of a series of heavy-duty treatments for acute myelogeous leukemia (AML), Brown received two stem cell transplants from a donor Hütter identified who was homozygous for the CCR5Δ32 mutation, meaning they lacked the functional CCR5 co-receptor used by most strains of HIV to infect target cells. At the time of the 2008 CROI poster presentation, Brown’s AML was in remission and he’d been off antiretroviral therapy (ART) for 285 days with no sign of HIV rebound.
Over the ensuing years, Brown has remained free of detectable HIV and there is broad consensus that he has been cured. However, multiple attempts to repeat the outcome by finding donors homozygous for the CCR5Δ32 mutation for additional people with HIV needing stem cell transplants for cancers have met with disappointing failure; the individuals have all died due to graft-versus-host disease (GVHD) or the underlying cancer, and in one case death was preceded by a rebound of HIV capable of using an alternate co-receptor, CXCR4. At least one researcher expressed concern that the CCR5Δ32 mutation might be negatively affecting the success of stem cell transplantation in some way.
Despite the setbacks, efforts to study people with HIV and cancers who receive stem cell transplants from donors homozygous for the CCR5Δ32 mutation have continued, and two headline-making presentations at CROI 2019 showed that the work has been worthwhile.
Garnering the most attention was Ravindra Gupta’s oral abstract describing a person with HIV in London who received a stem cell transplant from a CCR5Δ32 homozygote donor to treat Hodgkin’s Lymphoma that had not responded to standard therapies. Details on the case were provided in a Nature paper that was published online immediately after Gupta’s talk.
The individual was diagnosed with HIV in 2003, and ART was initiated due to a relatively low CD4 T cell count of 290 cells and viral load of 180,000 copies/ml. The diagnosis of stage 4B Hodgkin’s Lymphoma occurred in December 2012, with the cancer proving refractory to first-line chemotherapies and several salvage regimens (among them was the anti-CD30 monoclonal antibody Brentuximab, which has been reported to have potential activity against HIV-infected cells, but it seems unlikely that this played a part in the eventual outcome).
Ultimately the decision was made to undertake a stem cell transplant, and an international registry was searched to identify an appropriate donor. Fortuitously, the most closely matched stem cell donor was homozygous for the CCR5Δ32 mutation.
The transplantation process involved administration of an anti-CD52 antibody (alemtuzumab) to deplete T cells, a conditioning regimen—Lomustine, Cyclophosphamide, Ara-C and Etoposide (LACE)—and then infusion of donor stem cells. Treatments to prevent GVHD were a short course of methotrexate and cyclosporine until 180 days post-transplant. ART was maintained throughout.
There were some complications: both Epstein-Barr virus and cytomegalovirus reactivated around three months after the transplant, necessitating treatment with Rituximab and ganciclovir respectively. Mild (grade 1) GVHD transiently manifested in the gut, but resolved without any intervention.
Tests showed that donor cells homozygous for the CCR5Δ32 mutation had fully taken over by day 30 post-transplant, replacing the individual’s original cells (which had wild type CCR5 genes and expressed the normal functional co-receptor). Complete remission of the Hodgkin’s Lymphoma was confirmed at days 120 and 365.
The researchers discussed the possibility of HIV remission with the individual prior to the procedures, after testing had indicated that he harbored CCR5-tropic virus. Ethical approval was obtained from the UK National Health Service for a potential eventual interruption of ART, with criteria established ahead of time:
“if viral load was consistently <50 copies/ml on treatment with ‘target not detected’ for at least six months on the two most recent consecutive visits ART would be withdrawn and thereafter viral load would be monitored weekly for the first three months, monthly for a further nine months if undetectable at all time points and three monthly between years one and four.”
ART was interrupted in September 2017, and the individual has been off HIV treatment for 18 months with no evidence of viral load rebound. Multiple blood tests for the presence of HIV have been persistently negative, with the exception of a single detection of a low-level signal in a digital droplet PCR assay for HIV DNA (among eight replicates, the remaining seven of which were negative). The researchers have not yet studied any tissues. Antibody responses to HIV are waning, echoing results in Timothy Ray Brown.
Gupta was cautious, using the term HIV remission rather than cure due to the relatively short time since ART interruption. He suggested that perhaps when follow up is beyond two years, the terminology will be revised. The individual is included in the larger amfAR-funded IciStem research consortium, which focuses on evaluating people with HIV who receive stem cell transplants for cancers.
A second participant in the IciStem consortium was the subject of a late-breaker poster presentation by Björn-Erik O. Jensen from the research group of Guido Kobbe in Dusseldorf. Kobbe and colleagues first described the case at CROI 2016, at which time he was on ART and being followed after receiving a stem cell transplant from a CCR5Δ32 homozygote donor to treat relapsing AML. HIV was undetectable by multiple measures. At CROI 2019 it was revealed that, after ethical approval, ART was interrupted in November 2018. So far there has been no HIV rebound.
IciStem are following two additional individuals with HIV who’ve received stem cell transplants from CCR5Δ32 homozygote donors to treat cancers, but they remain on ART at this time (Monique Nijhuis described the current status of IciStem in a presentation at the pre-CROI community HIV cure research workshop).
Inevitably, the possibility that Timothy Ray Brown may finally have some company inspired a deluge of media coverage. The London case had been mentioned very briefly by Ian Gabriel at a BHIVA meeting last fall, prompting a short article by Simon Collins at HIV i-Base, so there was some awareness that further news would be forthcoming at CROI. The Nature paper was also circulated to media under embargo prior to the conference. Regrettably, The Hindu broke the embargo the day before Gupta’s presentation (whether accidentally or deliberately is unclear), leading Nature to also lift it for other media, so the coverage—including a detailed New York Times article by Apoorva Mandivilli—began appearing on the evening of Monday, March 4th.
There has been debate regarding the implications of the new cases. Most clear is that there is robust justification for continued efforts to identify CCR5Δ32 homozygote donors for people with HIV who need stem cell transplants to treat cancers. Whether lessons can be learned to apply to the search for more broadly applicable curative strategies is less certain; there are differences between the treatments given to the three individuals that may be help untangle contributing factors (see the excellent summary table in the HIV i-Base article by Simon Collins).
The common thread of receiving cells lacking a functional CCR5 receptor is viewed as supporting gene therapy research aiming to ablate CCR5 expression. A research group in China is currently studying whether the CRISPR/Cas9 gene editing tool can be used to disable the CCR5 gene in stem cells being transplanted into people with HIV and cancers—if successful, this might represent a strategy that would circumvent the need to find matched CCR5Δ32 homozygote donors.
Outside of the setting of stem cell transplants for cancers, the leading approach to knocking out CCR5 from CD4 T cells has been Sangamo Therapeutics SB-728-T. In clinical trials, CD4 T cells are extracted from individuals with HIV, edited at the CCR5 gene using zinc finger nuclease technology, and then expanded and reinfused. The company was hoping to achieve control of HIV viral load after ART interruption, but so far it hasn’t proven possible to modify sufficient numbers of CD4 T cells. Further commercial development for HIV has been abandoned, but some investigator-initiated studies continue.
Pablo Tebas presented new results from a trial of SB-728-T that broadly conformed to previous research. A total of 14 individuals on ART received a single infusion of modified CD4 T cells, either with or without a preceding dose of cyclophosphamide (intended to deplete existing CD4 T cells and make more room for modified cells). In a slight wrinkle, the delivery of the zinc finger nucleases to the cells was achieved using messenger RNA instead of the adenovirus vector employed in prior studies; the proportion of CD4 T cells successfully edited at the CCR5 gene by the two approaches was similar. The protocol included an ART interruption, and Tebas noted that there was a slight delay in viral load rebound compared to historical controls, but no cases of prolonged containment of HIV.
As observed in prior trials, participants heterozygous for the CCR5Δ32 mutation appeared to respond best. Because these individuals already have one disabled CCR5 gene, the zinc finger nucleases only have to edit one of the two alleles present in each CD4 T cell in order to prevent expression of a functional CCR5 co-receptor. Tebas concluded that more efficient CCR5 modification could potentially lead to more stringent control of HIV off ART, but it appears unlikely that such an outcome can be achieved with SB-728-T.
One variation on the theme of trying to genetically protect CD4 T cells from HIV infection involves focusing on modifying cells capable of recognizing and responding to the virus (HIV-specific CD4 T cells). A seminal study by Danny Douek many years ago showed that the virus preferentially infects HIV-specific CD4 T cells, which become dysfunctional and unable to perform their task of coordinating an effective immune response to the virus. The company American Gene Technologies is pursuing a strategy involving the genetic modification of HIV-specific CD4 T cells, with trials planned soon, and results should shed light on whether this is a better approach than attempting to modify CD4 T cells in bulk.
Taking a Bite Out of the Latent Reservoir with CRISPR/Cas9
One of the more intuitively appealing ideas in cure research involves attempting to cut the integrated HIV genome out of the DNA of latently infected cells. In this scenario, gene-editing strategies are targeted against the virus itself rather than a host gene like CCR5. The goal is to perform a sort of genetic surgery, excising HIV genes from infected cells without damaging the cell’s genome.
The gene-editing tool CRISPR/Cas9 has emerged as the leading candidate in this research, and some very preliminary results in mouse models have suggested it may have potential. The laboratory of Kamel Khalili at Temple University has pioneered these studies, in tandem with Excision Biotherapeutics, a company Khalili founded to move the approach into the clinic.
At CROI 2019, Tricia Burdo from Temple University debuted the results of a study exploring whether CRISPR/Cas9 could excise latent SIV in the SIV/macaque model of HIV infection. The cutting machinery of CRISPR/Cas9 is aimed at a target by the inclusion of molecules called guide RNAs (gRNAs), and in this case the researchers created gRNAs capable of recognizing three relatively conserved sites in the SIV genome (two in the long terminal repeats present at either end of the genome, and one in the gag gene). Burdo noted that the targeting of multiple sites is necessary for both attempting to excise large chunks of the viral genome and avoiding the potential development of resistance (somewhat similar to the rationale for combination ART). The technique produces “very little to no off-target effects,” according to Burdo.
The SIV-targeted CRISPR/Cas9 was delivered using an adeno-associated virus serotype nine (AAV9) vector. AAVs are a popular gene therapy delivery vehicle that can carry their payload into a broad range of both dividing and non-dividing cells without apparent safety issues (two AAV-delivered gene therapies have been approved by regulatory agencies).
The study included three macaques, all infected with SIVmac239 and placed on a suppressive ART regimen. In initial experiments, peripheral blood mononuclear cells (PBMC) sampled from the animals were transduced with the AAV9-CRISPR construct, producing evidence of excision of SIV genes between targeted sites.
AAV9-CRISPR was then infused into two of the macaques at a dose of 1013(ten trillion) copies per kilogram, a lengthy process involving the delivery of 100ml at a rate of 1ml per minute. Three weeks after the infusion, animals were euthanized and necropsy studies conducted. The third animal served as a control and was also euthanized to facilitate comparisons with the AAV9-CRISPR recipients.
Burdo reported that prior to euthanasia, fragments of the SIV genome that had been cut at targeted sites—referred to as excision products—could be detected in PBMC from the treated animals, as was observed when PMBC were exposed to AAV9-CRISPR in a laboratory dish. Cas9 DNA could also be detected in cells, indicating uptake of the gene-editing tool.
Necropsy studies included a preliminary evaluation of SIV outgrowth from PBMC samples. The PBMC were combined with SIV-susceptible CEM cells and then SIV p27 Gag protein levels were measured over time. SIV replication could be detected in samples from the control but not those from the macaques that received AAV9-CRISPR. However, Burdo emphasized that this assessment did not involve activating the PBMC to induce virus production (as is the case with the standard virus outgrowth assays used in human studies)—those experiments are pending.
Analyses of Cas9 DNA demonstrated widespread distribution in the tissues of the two treated macaques, ranging from around 1-10 thousand copies per million cells in the brain to over 10 million copies per million cells in the spleen and liver (presumably reflecting the presence of multiple copies in some cells). Burdo also showed evidence of SIV excision products in spleen, lung and several lymph nodes (including inguinal, submandibular, bronchial and colonic) from the animals.
The data appear very encouraging, but do not provide information on the magnitude of effect on the latent SIV reservoir (i.e. exactly how much latent SIV was successfully excised or disabled). In response to a question, Burdo reported that future plans include conducting analytical treatment interruptions in macaques treated with AAV9-CRISPR to assess whether viral load rebound is limited or prevented by the intervention.
An issue not covered in the presentation is the potential for the induction of immune responses against Cas9, which has been observed in mouse studies. Because Cas9 is derived from bacteria, it is treated as foreign by the immune system, and AAV vectors can have an adjuvant effect that seems to promote immune responses against AAV-delivered proteins (this has occurred in studies using AAV to deliver anti-HIV broadly neutralizing antibodies). Pre-existing anti-Cas9 immune responses have also been detected in humans due to infection with Staphylococcus aureusand Streptococcus pyogenes.
In a graph displaying longitudinal viral load measurements in the macaques (on slide #4 in the webcast), it looks as if administration of AAV9-CRISPR may have been temporally associated with a transient increase in SIV viral load—which could be suggestive of immune activation—although there were also viral load fluctuations in other animals. Gaining an understanding of whether AAV9-CRISPR delivery can activate the immune system and lead to the generation (or activation) of anti-Cas9 immune responses will be important prior to initiating human trials.
A theoretical concern that researchers have raised about strategies aiming to excise latent HIV relates to what might occur in cells that have more than one integrated copy of the HIV genome (this phenomenon is thought to be uncommon, but has been reported). In this situation, it’s possible that rather than just removing HIV genes, an excision approach might make cuts in each of the separate integrated virus genomes and thereby remove all of the cell’s DNA located between the different HIV integration sites. Damaging the genome of a cell in this way could potentially have untoward effects.
Overall, Burdo’s results offer significant encouragement for efforts to translate the approach into human clinical trials. Kamel Khalili and colleagues are now working toward that goal in collaboration with Jeffrey Jacobson, a highly experienced clinical HIV researcher who joined Temple University in 2016.
Attack of the Repliclones
The past few years have seen an increasing focus on the role of CD4 T cell proliferation in sustaining the latent HIV reservoir. Evidence has accumulated demonstrating that HIV proviruses can be faithfully copied into the daughter cells of latently infected CD4 T cells when they proliferate—the phenomenon can be discerned by the detection of genetically matching copies of the HIV provirus integrated into the exact same place in the genome of multiple CD4 T cells (the progeny of proliferating CD4 T cells are known as clones). Mathematical modeling suggests CD4 T cell proliferation may be the primary mechanism that allows the latent HIV reservoir to persist, and decline only very slowly over time.
Elias Halvas from the University of Pittsburgh showed at CROI 2019 that CD4 T cell clones containing integrated, intact HIV DNA are a source of low-level HIV viral load that can be detected in some individuals on ART. Essentially, some of these cells can spit out sufficient amounts of HIV RNA to be detectable, even though the virus is not actually replicating (i.e. going on to infect other cells—this is prevented by ART).
Halvas’s study involved 10 people who had been referred due to persistent low-level viral load despite ART (HIV RNA >20 copies/ml occurring for at least 6 months). The average time on treatment was 10 years, and viral load ranged from 40 to 356 copies/ml, with a median of 97.5 copies/ml.
One individual displayed evidence of ongoing HIV evolution and the development of drug resistance mutations and was considered a case of ART regimen failure, excluding them from further analysis.
Samples from the remaining nine showed the presence of genetically identical HIV RNA at multiple timepoints, and there was no sign of viral evolution or resistance mutations against current ART. The source of the HIV RNA was identified as CD4 T cell clones containing integrated, replication-competent HIV DNA (Halvas has christened them “repliclones”). In four cases the genetic sequence of the HIV RNA could be matched to viruses detected in the quantitative virus outgrowth assay (qVOA).
Halvas concluded that the possibility of production of HIV RNA by infected CD4 T cell clones needs to be borne in mind by clinicians caring for people with HIV, who might otherwise suspect that persistently detectable low-level viral load indicated non-adherence or treatment failure.
As to the implications for HIV cure research, Halvas suggested that repliclones may contribute to rapid viral load rebound after ART interruption (when the HIV RNA they produce is able to start infecting other cells), and he stressed that they will need to be targeted for elimination or suppression. The mechanisms prompting HIV RNA production by the cells are unclear, and need to be elucidated.
Notably, the data indicate that HIV latency can be more dynamic than was initially appreciated. It’s now clear that in some CD4 T cell clones containing integrated HIV DNA, the virus is not permanently latent, because there are times when production of HIV RNA is detectable.
In an article by Jon Cohen for Science that covers the study, John Mellors points out that the data raise questions about the “kick & kill” strategy in HIV cure research. The rationale for providing a latency-reversing “kick” is that most latently infected cells do not produce HIV RNA and therefore remain invisible to the immune system. Halvas’s results demonstrate that at least some latently infected cells do intermittently generate HIV RNA, and don’t die off as a result.
One salutary possibility is that researchers developing “kill” strategies may be able to study their efficacy in individuals like those described by Havlas, who already have low-level viral load on ART without the need for administration of any latency-reversing candidate. In theory, an effective “kill” approach should be able to reduce the amount of HIV RNA detected in such cases.
Targeting the Latent HIV Reservoir with Anti-Proliferative Therapy
The recognition that the HIV reservoir is at least partly sustained by the proliferation of CD4 T cells is rekindling interest in testing the effects of anti-proliferative therapies in the context of cure research.
At the pre-CROI community HIV cure research workshop, Joshua Schiffer from the Fred Hutchinson Cancer Research Center described a small (four person) pilot trial of the anti-proliferativedrug mycophenolate mofetil (MMF) being conducted by his research group with funding from amfAR. The rationale is based on the results of mathematical modeling work suggesting that inhibiting CD4 T cell proliferation in people on ART should significantly accelerate the decay of the HIV reservoir. Results are anticipated to be available for CROI 2020. Links to the video of Joshua Schiffer’s talk are on the workshop web page, along with the slides.
Timothy Heinrich from UCSF presented results of an AIDS Clinical Trials Group (ACTG) trial of sirolimus (a drug with potent anti-proliferative activity, also known as rapamycin) in people on suppressive ART. In the 16 participants who completed 20 weeks of dosing, there was a slight but statistically significant 0.16 logs reduction in HIV DNA levels. CD4 T cell expression of the proliferation marker Ki67 was also significantly reduced.
Heinrich noted that rates of sirolimus discontinuation were high, and there were also transient increases in the inflammatory biomarkers IL-6 and sCD14 and the coagulation biomarker D-Dimer. The results appear consistent with the notion that inhibiting CD4 T cell proliferation can affect HIV reservoir size, but additional research is needed to confirm that this was the primary mechanism for the HIV DNA reduction.
One of the pioneers in this area of HIV cure research is Andrea Savarino, who reported many years ago that the gold-based anti-proliferative drug auranofin reduced the SIV reservoir in ART-treated macaques.
Since that time, Savarino and colleagues have collaborated with investigators in Brazil to conduct a small pilot trial involving auranofin (which is a licensed treatment for rheumatoid arthritis). The latest results were presented in a poster at CROI 2019. The study design is complicated, involving multiple interventions administered to six groups, each with just five participants. The researchers report that auranofin combined with several other agents led to a reduction in HIV DNA, but the contribution of the anti-proliferative effect is unclear. Administration of the drug was not associated with any serious side effects. Given the renewed interest in targeting CD4 T cell proliferation and the uncertain safety profile of some anti-proliferative drugs, additional studies of auranofin may be justified.
Additional Webcasts and Links
The CROI website offers comprehensive webcasting including every presentation. The pre-conference workshops on March 4th are a source of several excellent overview talks, such as Paula Cannon’s on gene editing techniques in HIV research. Also recommended is Irini Sereti’s insightful plenary on HIV and inflammation.
Multiple outlets provided high quality coverage from CROI 2019, including: