Monkey Models: Far from Extinct
A recent meeting on nonhuman primate models underscores their central role in AIDS research
By Andreas von Bubnoff
Nonhuman primate (NHP) studies have been and will be a central part of HIV/AIDS research, National Institute of Allergy and Infectious Diseases (NIAID) Director Anthony Fauci said at the opening of the 27th Annual Symposium on Nonhuman Primate Models for AIDS, which took place from October 28-31 in Boston. “The advances over the last 28 years are really astounding, [and] a lot of the things that have opened up doors in pathogenesis and vaccine [development] particularly come from NHP studies,” he said, ending his remarks to the about 250 attendees of the meeting by saying, “I think you are here to stay.”
The meeting featured research updates on a wide range of topics, including the characterization of existing challenge stocks, the development of new challenge stocks to better mimic human HIV infection, and studies of the role of bacterial translocation in the gut in the development of AIDS.
Optimizing challenge strains
One goal when developing NHP models for AIDS is selection of a challenge virus that mimics HIV transmission in humans. At the meeting, researchers discussed some efforts that focus on developing challenge stocks that mimic recently transmitted and not chronically replicating HIV, because the two appear to differ biologically.
Fauci addressed these differences between early and late viruses when he discussed the recent results from RV144, a 16,000-person efficacy trial of a prime-boost regimen in Thailand that showed the first hint of vaccine-induced protection against HIV infection. He said the effect the RV144 prime-boost regimen had on acquisition, but not on viral load (see Raft of Results Energizes Researchers, IAVI Report, Sep.-Oct. 2009), suggests a difference between the immune response that blocks HIV acquisition and the one that blocks virus replication. This is consistent with biological differences between the transmitted virus and the late, chronically replicating virus. For example, compared to the chronically replicating virus, the transmitted virus is less glycosylated and therefore easier to neutralize. “It may be when we are looking at the effect of a vaccine we are matching it against a virus that it doesn’t have to neutralize because the transmitting virus is very different from that,” Fauci added.
Ruth Ruprecht, a professor of medicine at Harvard Medical School, welcomed Fauci’s remarks. “That’s what I have been saying for years,” she said. Her lab has been developing SIV/HIV hybrid or SHIV strains that containenv from early, recently transmitted HIV clade C strains to better reflect the biology of the virus that is transmitted. “There is something special about the biology of the recently transmitted [clade C] virus,” Ruprecht said, adding that it is more sensitive to neutralizing antibodies and has shorter variable loops of gp120 than the other strains that predominate in the infected donor.
At the meeting, Nagadenahalli Siddappa from Ruprecht’s lab reported progress in the development of a SHIV strain with a clade C env gene that was isolated from a six-month-old child from a mother-infant cohort in Zambia (see Looking for the Perfect Challenge, IAVI Report, July-Aug. 2009). Clade C viruses are responsible for about 56%-60% of all HIV/AIDS cases in the world, according to Ruprecht. In unpublished work, Siddappa and colleagues have used this env to develop a SHIV challenge strain that is easy to neutralize, replicates reproducibly, and causes high peak viremia levels. This is the first R5 clade C SHIV that is tier 1, which means that it has a high sensitivity to neutralization, according to Ruprecht. “We have already titrated it so that we can do multiple low-dose [challenges],” Ruprecht said.
According to Ruprecht, having challenge viruses that are easy to neutralize is important in initial challenge studies, because otherwise the bar may be too high to see any protective effect of a candidate vaccine. “If you are going to need to enlist the humoral arm of the immune system, you have to have a challenge system that lets you score for that,” she said.
Some of the current SIV challenge strains such as SIVmac251 are thought to be hard to neutralize, suggesting they may not be ideal for testing vaccine candidates, said Wendy Yeh, an instructor in medicine at Harvard Medical School who works at Beth Israel Deaconess Medical Center in Boston. But just how hard it is to neutralize SIVmac251 challenge strains compared to HIV-1 has not been rigorously tested. So Yeh and colleagues determined if serum of macaques infected previously with SIVmac251 in a repeat low-dose rectal infection experiment (1) could neutralize the exact virus variants that caused the infection.
Because SIVmac251 is a swarm of many different virus variants, Yeh and colleagues used single genome amplification (SGA) to isolate the transmitted founder viruses that caused productive infection. They found that a single transmitted founder virus was responsible for infection in each animal, and then made an SIV pseudovirion with the Env protein from these transmitted founder viruses. They then tested if sera taken from the animals at different time points after infection could neutralize the pseudovirion with the Env protein from the matching (autologous) transmitted founder virus.
Yeh reported that the sera from the animals could only neutralize the autologous pseudovirions between five and eight months after infection. This suggests that SIVmac251 induced neutralizing antibody (NAb) responses later than in HIV-infected humans, which develop responses two to three months after infection, Yeh said. In addition, the animal serum contained lower NAb titers than what is observed in humans. Even though the NAb response had a low titer and developed late, Yeh still found that it exerted selective pressure on the virus, causing the virus to develop escape mutations. It took several months before the infected animals developed NAbs to these escape mutants after they appeared.
“Compared to HIV-1, SIVmac251 appears to be even more resistant to antibody neutralization in that the antibodies did not appear until later in the course of infection, and even when they were present they were present at very low titers compared to what has been observed with HIV-1,” Yeh concluded. This suggests that SIVmac251 may not be an ideal model to test candidate vaccines for human clinical trials. “Our data suggest that SIVmac251 may not accurately estimate the ability of vaccine candidates to elicit neutralizing antibodies that can protect against infection in a nonhuman primate model,” Yeh said.
Brandon Keele, a senior scientist at the National Cancer Institute and Science Applications International Corporation (SAIC) in Frederick, Maryland, used SGA of full length env genes to analyze the sequence diversity of SIVmac251 and SIVsmE660, another challenge stock that is a swarm of many different variants. He found that these two challenge stocks had an overall sequence diversity of about 1%-3%, which is similar to the approximately 1% diversity of HIV variants in a single individual who has been infected for months to years. The same analysis showed there is approximately 20% maximum diversity between SIVmac251 and SIVsmE660 envsequences, compared with at most a 15% difference of HIV subtype B taken from different infected individuals. This suggests that a 251/660 heterologous challenge in animals is a good approximation of intra-subtype variation of HIV, Keele said.
Keele also used SGA to analyze the inoculum that was used to infect rhesus macaques intrarectally with the SIVmac251 stock generated by Ron Desrosiers at the New England Primate Research Center. He found that the transmitted founder virus often reflected the most common variants present in the inoculum. However, in some animals, rare variants from the inoculum established infection in the host. Keele is currently cloning these rare variants to identify the biological characteristics that enable them to get preferentially transmitted.
Keele also compared the effects of different doses of virus used to infect the animals via different transmission routes. When SIVmac251 was transmitted rectally, the number of transmitted founder viruses was typically lower when the challenge stock was more diluted. However, this dose dependence was less obvious when animals were infected vaginally. With vaginal infection, Keele said, “you have these huge differences where ten viruses get through once [and] one virus gets through in the next animal at the exact same dilution.” One possible reason for this variation might be different availability of local target cells due to anatomical differences in the vagina, Keele said, perhaps because female rhesus macaques may be in different stages of their vaginal menstrual cycle when they are infected with the challenge virus.
Theodora Hatziioannou, an assistant professor at the Aaron Diamond AIDS Research Center, reported progress in using a modified version of HIV to infect pigtail macaques. Directly infecting them with HIV is not possible because of host restriction factors including the APOBEC3 proteins. But Hatziioannou and colleagues are developing an HIV-1 derived challenge virus called simian tropic HIV-1 (stHIV-1) that can infect pigtail macaques. stHIV-1 differs from HIV-1 in that its vif gene comes from SIVmac239 or HIV-2. This vif gene encodes a Vif protein that can destroy the APOBEC3 proteins in pigtail macaques, enabling stHIV-1 to replicate (2; see Looking for the Perfect Challenge, IAVI Report, July-Aug. 2009).
However, while SIVmac239 can cause disease in pigtail macaques, stHIV-1 cannot. So Hatziioannou reasoned that the pigtail macaque host cells must have additional factors that inhibit stHIV-1 replication more than SIVmac239 replication. At the meeting, she reported that peripheral blood mononuclear cells (PBMCs) cultured from pigtail macaques only inhibited stHIV-1 replication more than SIVmac239 replication in the presence of interferon (IFN)-α. This suggests that IFN-α induces factors in the pigtail macaque PBMCs that keep stHIV-1 from replicating.
Hatziioannou found that one of these factors is tetherin, a host restriction factor that inhibits the release of HIV particles from infected host cells. Tetherin could indeed be one of the reasons why stHIV-1 does not make pigtail macaques sick. Although stHIV-1 contains a gene encoding human Vpu which can overcome human tetherin, it is inactive against pigtail macaque tetherin.
Therefore, one way to alter stHIV-1 so that it can cause disease in pigtail macaques is to enable it to overcome tetherin. Hatziioannou said this could be achieved by introducing the nef gene from SIVmac239 into stHIV-1, because SIVmac239, which does not have a vpu gene, uses nef instead to overcome tetherin in macaque host cells (3). Another possible approach is to replace the stHIV-1 vpu gene with vpu genes from certain NHPs that can also overcome macaque tetherin, such as SIVgsn, the natural host of which are greater spot-nosed monkeys (4). However, that still leaves another limitation of stHIV-1—its Env protein is primarily X4 tropic, and therefore infects different target cells than most HIV-1 currently in circulation. But Hatziioannou said she will deal with that later. “I don’t want to change too many things [at once] in my virus,” she said.
Jim Smith, a senior service fellow in the Division of HIV/AIDS Prevention at the Centers for Disease Control and Prevention in Atlanta, also reported on the development of a new challenge strain, called RT-Env SHIV, which contains the genes for HIV-1 reverse transcriptase (RT) as well as Env. To construct the strain, Smith and colleagues combined two SHIV strains that are known to replicate well in macaques: RT-SHIV, which contains the HIV-1 RT gene, and SHIV162P3, which encodes an R5-tropic clade B HIV-1 Env protein. “We took two things that had a very good track record of replicating in macaques and we put those two together,” Smith said.
Chinese rhesus macaques infected with RT-Env SHIV showed reasonably high viral loads of about 600,000 particles/ml, which is close to the peak viral loads typically observed in pigtail macaques challenged with RT-SHIV, Smith said. The new RT-Env SHIV strain will make it possible to test topical vaginal or rectal microbicides that not only contain RT inhibitors or fusion/entry inhibitors separately, but also combinations of them. Testing such combinations is important, Smith said, because of the anticipated resistance to these drugs. “We are trying to stay ahead of the curve,” he said. Next, Smith wants to start infecting pigtail macaques with RT-Env SHIV to be able to do challenge studies of topical microbicides in pigtail macaques because they are more similar to humans than rhesus macaques in that they have a monthly menstrual cycle. He has also given the strain to the National Institutes of Health (NIH), where it is being propagated in Indian rhesus macaques so that it can be used in these animals as well.
MHC: Heterozygous is better
One hypothesis as to why the adenovirus serotype-5 based vaccine candidate tested in the Phase IIb STEP trial did not have any protective effect is the poor breadth of the cellular immune responses it induced, said Shelby O’Connor, an associate scientist at the University of Wisconsin-Madison. That raises the question as to whether the breadth of the CD8+ T-cell response correlates with control of viral load, she said. She presented evidence from Mauritian cynomolgus macaques that indeed, the breadth of the cellular CD8+ T-lymphocyte response appears to play a key role in suppressing viremia.
She found that SIVmac239-infected Mauritian cynomolgus macaques that are heterozygous for major histocompatibility complex (MHC) class I genes had about an 80-times lower average chronic phase viral load than those that are homozygous. These genes encode MHC class I receptors used by infected, antigen-presenting cells to present peptides to CD8+ T cells, which are then activated and kill the infected cells. She also showed sequences of viral escape variants from the cellular immune response in these animals, which suggested that the CD8+ T-lymphocyte responses in the heterozygous animals were a composite of the CD8+ T-lymphocyte responses in each of the homozygous groups of animals. In addition, MHC heterozygous animals had about twice as many distinct MHC molecules than MHC homozygous animals, and therefore have the potential to present twice as many different sets of peptides to CD8+ T cells. This means that in heterozygous animals, about twice as many different CD8+ T-lymphocyte responses might be induced as in homozygous animals.
Mauritian cynomolgus macaques are ideal for such studies because their genetic background is much more homogeneous than that of rhesus macaques, O’Connor said. In 99% of Mauritian cynomolgus macaques, seven MHC class I haplotypes can explain the MHC genetic diversity, she said.
“This data provides provocative evidence for us that an MHC heterozygote advantage exists in Mauritian cynomolgus macaques infected with SIVmac239. The data supports our hypothesis that a CD8 T-lymphocyte response directed at a maximally diverse set of epitopes is advantageous.”
Mary Carrington, a senior investigator at the National Cancer Institute and SAIC in Frederick, said that the situation is similar in humans. Genome-wide association studies show that the MHC is the single most important locus in the control of viral load. Using data from the international HIV controllers cohort, which is organized by Bruce Walker, director of the Ragon Institute, she reported that individuals who control viral load to be below 2,000 copies/ml of plasma show a greater trend to be heterozygous in their HLA loci than non-controllers. But she added that the effect was not very strong, in part because in humans these loci are very polymorphic. O’Connor’s data from cynomolgus macaques are much more obvious, she said, because these animals are genetically less diverse at the MHC loci. “I think this is one case where the nonhuman primate data really is essential to determine whether these effects are actually real,” Carrington said. “I think [the cynomolgus macaque] data actually puts the nail in the coffin in terms of a heterozygous effect in a way that the [human] HLA data just cannot do.” These observations add to the previous observation that people who are homozygous in their HLA loci progress to AIDS faster than people who are heterozygous in their HLA loci (5).
Focus on NK cells
While much of the focus was on antibodies and cellular immune responses, the pillars of the adaptive immune system, some speakers also discussed the role of natural killer (NK) cells, which are considered a part of the innate immune system. However, Ulrich von Andrian, a professor of pathology at Harvard Medical School, presented data that suggest that they also carry features of the adaptive immune system.
Previously, von Andrian and colleagues had shown that in mice without T and B cells, NK cells are both required and sufficient for a contact hypersensitivity response in their skin to hapten stimuli they were exposed to for a second time (6). This response had been thought to be only B- and T-cell dependent.
Von Andrian said that this NK cell-dependent contact hypersensitivity response is similar to T- and B-cell immune responses in that it is an antigen-specific memory response. He now showed that mice without B and T cells could develop a stimulus-specific memory response to hapten stimuli to their skin as long as four months after the initial stimulus. “An adaptive response in the absence of T and B cells was very much against text books,” von Andrian said. He also said that the mice appear to be able to develop such an NK-dependent response even to novel stimuli they could not possibly have encountered before, such as HIV. While the mechanism of this adaptive memory response of NK cells isn’t known, this suggests that vaccinologists should keep an open mind as to the kind of immune responses a vaccine might induce, von Andrian said. “It’s possible that current vaccines induce some level of NK cell memory,” he said. “It’s just not really been well explored so far.”
NK cells are also involved in antibody dependent cytotoxicity (ADCC), which some researchers have suggested may be partly responsible for the modest efficacy of the vaccine regimen in the recently completed RV144 efficacy trial in Thailand. In ADCC, the Fc regions of antibodies bound to HIV-infected cells bind to Fc receptors of innate immune cells such as NK cells. This binding can activate the NK cell, which then kills the target cell with the antibody bound to it.
At the meeting, Michael Alpert, a graduate student in the lab of David Evans at Harvard Medical School, reported the development of a new rapid, standardized assay to measure the ability of antibodies from rhesus macaques to direct ADCC against SIV-infected cells. To avoid variability, the assay uses cell lines for both the target cells and effector NK cells. The target cell line is a CD4+ T-cell line with a luciferase reporter gene that is expressed when the cell is infected by SIV. This Luciferase expression gets lost as soon as these SIV-infected target cells are killed by the effector cells, an NK cell line that expresses an Fc receptor.
The third component of the assay is the serum, which contains the antibodies that bind to SIV on the infected CD4+ target cell line. This activates the NK cells once their Fc receptors bind to the Fc regions of the antibodies. Alpert said that the assay uses serial dilutions of serum that are added to wells that each contain the same number of effector and target cells. “This will [then tell] you what percentage of Luciferase activity is still present in the various wells after an eight hour incubation in [the] presence of plasma and NK cells,” Alpert said. Previous assays looked at just one dilution, whereas the new assay looks at a titration curve over several orders of magnitude, which leads to cleaner data, according to Alpert. In addition, previous ADCC or antibody dependent cellular virus inhibition (ADCVI) assays are more variable than the new assay because they rely on fresh cell preparations every time, he said. This could lead to differences in the killing efficiency of the NK cells, for example. “You can’t set up the same assay twice,” Alpert said. “Meanwhile, we have cell lines you can just take out of an incubator and use for the assay.” He is also working on a version of the assay that could be used to measure ADCC in human samples, which might be useful for analyzing samples from the Thai trial.
Bacterial translocation through the gut is a possible cause of the chronic immune activation that leads to the development of AIDS. However, so far the evidence that this translocation causes immune activation has been indirect, coming from correlative studies that show an association of increased bacterial lipopolysaccharides (LPS) in plasma and markers of immune activation. Now, Jacob Estes, a senior scientist at the National Cancer Institute and SAIC in Frederick, and colleagues have found more direct evidence for bacterial translocation in detailed analyses of tissue sections of the large bowel and lymph nodes in rhesus macaques acutely or chronically infected with SIVmac239, SIVmac251, or SIVsmE660. They also quantified the relative amount of LPS in the lamina propria, the layer underneath the gut epithelium, and the lymph nodes.
Estes showed that chronically infected animals with end-stage AIDS have a severe breakdown of the epithelial barriers in the gut. This resulted in a large amount of bacterial products such as LPS in the mucosal tissues of the gut and in the gut-draining and peripheral lymph nodes. Chronically infected non-end-stage animals had an intermediate to severe degree of gut epithelial damage and bacterial products in these sites, whereas uninfected animals had no epithelial damage, and bacterial products remained in the gut lumen. On the other hand, SIV-infected animals showed more signs of dividing cells in the gut epithelium than uninfected animals, indicating that infected animals were trying to regenerate the damaged epithelial barrier.
In contrast, chronically SIVsmm infected sooty mangabeys did not show evidence of bacterial translocation in the large bowel or peripheral lymph nodes, which is consistent with the fact that they don’t have signs of chronic immune activation and don’t get sick from SIV infection. “This data suggested to us that the damage to the epithelial lining really plays a key role in microbes [and microbial products] getting across into the host,” Estes said.
He also showed that chronically SIV-infected animals had many bacterial products in the gut that were not associated with macrophages, even though many macrophages were present in the area, suggesting a possible defect in the ability of macrophages to clear translocated bacterial products.
In a related talk, Sieghart Sopper, a group leader at the German Primate Center, challenged the notion that microbial translocation is the major driving force of the immune activation that leads to AIDS. He said that so far there is no evidence for microbial translocation during acute viremia in humans, and that recent studies did not find a correlation between microbial translocation and disease progression.
Sopper found that in SIV-infected rhesus macaques, viral load correlated better with immune activation and disease progression than bacterial translocation did. He did a retrospective analysis of 37 Indian rhesus macaques, half of which were infected with SIVmac239, the other half with SIVmac251. When he analyzed a group of 18 rapid progressors (animals that developed AIDS earlier than 15 weeks after infection) and 19 slow progressors, he found that LPS levels did not differ between the two groups.
In addition, LPS levels correlated only weakly with viral load or survival, and only at one time point late after infection. In contrast, viral replication and immune activation correlated very well with disease progression and survival at all time points after infection. “In contrast to the model that viral replication leads to microbial seepage and activation of the immune system, we think that viral replication can directly activate the [innate] immune system which then leads to the production of cytokines or other mediators, which leads to the impairment of the mucosa and the subsequent seepage of microbial products,” Sopper concluded. “We think that the interplay between immune activation and viral replication is more important than the effect of the microbial translocation for the development of AIDS.”