Do Pediatric Vaccines Stand a Better Chance of Protecting Against HIV?

A committee asks age-old questions about immunity as they set out to look at how the delicate and complex immune systems of infants and young children might help inform AIDS vaccine development.

By Mary Rushton

In the decades-long search for a safe and effective AIDS vaccine, scientists have developed and tested dozens of different candidates, involving many different strategies. So far these efforts have been met with limited success. In 2009, results from the RV144 trial in Thailand showed a modest 31.2 percent efficacy against a prime-boost candidate that many AIDS researchers thought had a slim chance of working at all (see Special Report: Thai Trial Results, IAVI Report, Sep. 1, 2009). But the vaccine candidates—a canarypox vector-based ALVAC-HIV in combination with a genetically engineered version of gp120—tested in 16,000 individuals at risk for HIV, did not meet the criteria to warrant licensure. No other candidates have demonstrated any efficacy at all.

There are many reasons why it is so difficult to develop a highly effective and durable AIDS vaccine that could confer immunity against the vast numbers of circulating strains of the virus. One is that scientists have not yet been able to successfully design a vaccine capable of inducing broadly neutralizing antibodies (bNAbs) against HIV’s many diverse strains because our immune systems are generally outmatched by the virus. The field also suffers from the limitations of animal models that do not perfectly mimic what occurs in humans following HIV exposure.

There are other reasons for sure, but it is an open debate whether one of those reasons is that researchers are focusing too exclusively on developing and testing vaccine candidates for, and in, adults, rather than exploring ways to study the evolving and maturing immune responses in younger individuals. That was one of the questions that percolated in September when the AIDS Vaccine Research Subcommittee (AVRS), an advisory panel that provides advice and makes recommendations to leaders at the US National Institute of Allergy and Infectious Diseases (NIAID), devoted a whole day to the topic of pediatric HIV vaccine development.

Mary Marovich, director of the Vaccine Research Program in the NIAID Division of AIDS, says the discussion at AVRS was sparked in part by recent human and animal data, including compelling evidence that HIV-infected infants generate bNAbs faster than HIV-infected adults (Nat. Med. 20, 665, 2014).

“I was trying to be provocative with our experts… to push them to begin to think about whether we are in a situation where we could look at using and testing existing immunogens, as long as they are safe in adults, in an early life immunization strategy,” says Marovich.

The International Maternal Pediatric Adolescent AIDS Clinical Trials Group (IMPAACT), supported by NIAID, and the US National Institute of Child Health and Human Development have done early stage HIV vaccine trials in much younger volunteers. And a group of NIAID-funded scientists led by investigators from the Duke University Human Vaccine Institute have been studying neonatal immunity in nonhuman primates vaccinated with HIV vaccine candidates for the past several years, as have researchers at the University of California at Davis.

But the AVRS wants to take this idea further. The September meeting spurred promises from the committee to form a working group populated by representatives from child health groups and HIV vaccine networks. Marovich says the working group, which she expects to be organized by January, will examine the immunization landscape over the next 18 months and determine which, if any, of the immunogens that have already proven to be safe in adults could reasonably be tested in children, and if so what would be the appropriate age to immunize.

“The AVRS meeting was about whether we should develop a research agenda in this area. The working group will be more about how to do this and who is going to do it,” says Marovich. “While we are not actively discussing doing pediatric efficacy trials now, we can do Phase I safety studies and the bridging studies once we have a signal in adults.”

Just what immunogens might be tested in Phase I pediatric HIV vaccine trials, however, remains an open question. Currently, there are only two HIV vaccine efficacy trials in progress. One, a Phase IIb test-of-concept study known as HVTN 702, is evaluating a prime-boost regimen that is based on the one used in the RV144 trial. The study is enrolling 5,400 high-risk men and women in South Africa. Last month another Phase IIb study began in South Africa testing a prime-boost regimen developed by the Ragon Institute of Massachusetts General Hospital, the Massachusetts Institute of Technology, and Harvard, advanced to clinical trials by Johnson & Johnson’s Janssen Pharmaceutical Companies, the Bill & Melinda Gates Foundation, and NIAID. The trial’s organizers plan to include four other countries in southern Africa pending regulatory approvals. The study, known as Imbokodo (the Zulu word for rock, part of a South African proverb that refers to the strength of women and their importance in the community), is evaluating an adenovirus serotype 26 (Ad26) viral vector vaccine candidate containing four mosaic antigens engineered to provide optimal coverage against all circulating HIV variants, in combination with a clade C gp120 vaccine candidate.

There are many other scheduled or ongoing Phase I and II studies of vaccine regimens or antibody-mediated prophylaxis, including a study in HIV-exposed infants that will test whether bNAbs administered by injection are protective. Infants and toddlers are also a major focus of HIV cure research, given some of the recent findings that show very early initiation of antiretroviral (ARV) therapy in HIV-infected infants can result in undetectable virus levels for a period of years.

But to date, there have been no efficacy trials for HIV vaccine candidates conducted in infants or children, and the majority of safety and immunogenicity trials of preventive vaccine candidates developed so far have also been done in people 18 years of age or older. This is partly because HIV is primarily transmitted sexually or by sharing needles with an HIV-infected person. But Marovich questions whether researchers should consider vaccinating earlier. “Does it make sense to keep testing vaccines in adults after their sexual debut or to lower the age range to establish protection before they are exposed or during high-risk times like breast feeding?” she asks.

As Marovich points out, despite the overwhelming success of preventing mother-to-child transmission (PMTCT) of HIV, children are still vulnerable to the virus during the breast-feeding period. PMTCT is credited with helping to prevent 1.6 million new infections since 1995, a 2016 Joint United Nations Programme on HIV/AIDS (UNAIDS) report says. Cases of perinatal HIV transmission have declined by more than 90 percent in the US, according to the US Centers for Disease Control and Prevention (CDC), and there are 60 percent fewer children being newly infected in sub-Saharan Africa since 2009. Yet there are still thousands of children acquiring HIV from their mothers. A recent UNAIDS report estimates that every year 110,000 children are being newly infected with HIV in 21 sub-Saharan countries, with more than half of these infections occurring during the breast-feeding period. “In areas where the epidemic is still raging on, there have been good efforts and success in preventing mother-to-child transmission, but there are still gaps,” says Marovich. “There are still approximately 200,000 infants [worldwide] that are infected each year with HIV because they are born to infected mothers. For whatever reason, they were late presenters [to care], or adherence to ARVs post-birth was poor, or it happened through breast feeding. So it is still an issue globally.”

Past studies

Early on, before the dawn of ARV-based PMTCT, researchers were interested in testing vaccines to see if they could protect infants from the virus. In the early 1990s, a dose-escalation vaccine study led by New York University was conducted in 126 HIV-uninfected infants born to HIV-infected mothers in the US. The infants received four doses of recombinant HIV gp120 vaccine candidates developed by either VaxGen in partnership with Genentech, or Chiron, now part of Novartis, to see whether these HIV antigens were immunogenic. One regimen involved giving a dose at birth and then at one, three, and five months of age, while another more accelerated regimen started with a dose at birth, and then others at two weeks, two months, and five months of age (J. Infect. Dis. 181, 890, 2000).

At the end of the study, immune responses were detected in over half the infants after the second dose and those responses persisted for 104 weeks, with stronger responses observed in infants given the accelerated regimen. But three years after this study, two Phase III trials in men who have sex with men (MSM) and injection drug users testing the VaxGen recombinant gp120 candidate failed to demonstrate any efficacy (J. Infect Dis. 192, 974, 2005). This brought an abrupt end to any discussions about studying the regimen further in infants.

Plans to test another HIV vaccine candidate in breast-fed infants were also abandoned after the candidate failed to prevent HIV infection in adults. A collaboration between the Elizabeth Glaser Pediatric AIDS Foundation, IMPAACT, the HIV Vaccine Trials Network (HVTN), and Merck was planning to test the company’s adenovirus serotype 5 (Ad5) vaccine candidate in the breast-fed infants of HIV-infected mothers in sub-Saharan Africa, but the trial was shelved after the Phase IIb test-of-concept STEP trial in MSM and high-risk heterosexual women evaluating the same vaccine candidate was stopped early for futility (see A STEP Back?, IAVI Report, Vol. 11, No. 5, 2007).

More recently, a small randomized, placebo-controlled trial known as HPTN 027 found that ALVAC-HIV vCP1521, one of the vaccine candidates used in the RV144 trial, was safe, well tolerated, and immunogenic in infants born to HIV-infected mothers in Uganda (J. Aquir. Immune Def. Syd. 65, 68, 2014). Another Phase I/II study known as PedVacc 002 was conducted in Kenya to determine the safety and immunogenicity of a Modified Vaccinia Ankara (MVA) viral vector-based vaccine candidate in infants born to HIV-infected mothers. The single, low dose was delivered intramuscularly to healthy, four-month-old infants. The study demonstrated that the vaccine was safe but not immunogenic (Vaccine 32, 5801, 2014).

Relax and make antibodies  

A recent study in mice led by scientists at the University of Colorado School of Medicine showed that a process that protects the body from autoimmune diseases also prevents the immune system from generating antibodies that can neutralize HIV. After being immunized with HIV Envelope protein, followed by the adjuvant alum, transgenic mice expressing symptoms similar to lupus, an autoimmune disease, mounted neutralizing antibodies against HIV (J. Exp. Med. 214(8), 2283, 2017).

The researchers repeated the experiment in normal, healthy mice with a drug that impairs immunological tolerance—the mechanism that prevents the generation of auto-reactive antibodies that trigger diseases like lupus—and found that these animals started to produce antibodies with weak neutralization against HIV. When alum was added, the mice produced potent broadly neutralizing antibodies (bNAbs) that were able to neutralize a range of HIV strains. The findings raise the question whether transiently relaxing immunological tolerance, without triggering detrimental autoimmune reactions, could be a reasonable strategy for eliciting HIV-specific bNAbs through vaccination. —MR


Why immunize earlier?

Most preventive vaccines are administered to infants and young children, largely before age two, because the developing immune systems of young children are more vulnerable to infection with life-threatening viruses. The sooner children are vaccinated, the less likely they are to be affected by the diseases the vaccines protect against. And when vaccine coverage is high enough, as is the case with many childhood immunizations, it can help eliminate virus transmission almost entirely.

To accomplish this though, the immune responses to childhood vaccinations must be strong and durable enough to be protective for long periods of time, which is something that has eluded HIV vaccine researchers so far.

It is possible, however, that immune responses to HIV antigens may be stronger in infants than adults, which is one reason the AVRS is discussing the topic of infant trials. Barton Haynes, director of the Human Vaccine Institute at Duke University, says that while some aspects of the neonatal immune system are immature and slower to respond, it is clear that in response to candidate HIV vaccines, infant immune systems respond as well or even better than adults.

Julie Overbaugh, a researcher at the Fred Hutchinson Cancer Research Center who has studied the mechanisms of HIV transmission and pathogenesis extensively, says there is still a ways to go in understanding why this is the case. “We know surprisingly little about the differences in the way infants develop antibody responses compared to adults and need to know this to understand the best way forward.”

Recent studies from her lab detail the speed at which an infant mounts both neutralizing and binding antibody responses. Three years ago, Overbaugh published the first findings showing that some HIV-infected infants can rapidly mount a broad neutralizing antibody response to the virus, sometimes within the first year of life (Nat. Med. 20, 655, 2014), unlike the minority (10-20 percent) of adults who take two to three years on average to develop bNAbs following infection.

In collaboration with scientists in Kenya, Overbaugh’s lab tested the ability of serum from 28 HIV-infected infants enrolled in the Nairobi breast-feeding trial to see if it could neutralize a panel of viruses. Serum samples from more than 70 percent of the infants were able to neutralize one or more viruses from a different clade than the virus with which they were infected, in some instances within just 12 months after acquiring HIV. Serum from seven infants neutralized viruses across all four clades.

Building on these findings, Overbaugh’s team next isolated 10 neutralizing antibodies exhibiting low levels of somatic hypermutation (SHM) from an infant at around a year post-infection, including one bNAb with cross-clade neutralization capabilities that targeted the glycan-dependent N332 supersite on HIV’s Envelope (Env) protein that some bNAbs isolated from adults also favor. The low levels of SHM that the infant’s neutralizing antibodies exhibited suggest that neutralization breadth can occur without the much more extensive level of SHM that is evident in all adult antibodies isolated to date (Cell 166 (1), 77, 2016).

Work led by Duke University researchers indicates that neonatal immunization may also be helpful in skewing antibody responses to subdominant HIV epitopes. Six years ago, a study led by Haynes’ lab determined that the initial antibody responses that arise in HIV-infected people within 14 days of infection, and which target HIV Env’s gp41 protein, demonstrated cross-reactivity between gp41 and the gut flora. The hypothesis was that the early response to gp41 was really a secondary response triggered by a pre-existing pool of memory B cells interacting with bacteria living in the gut (J. Exp. Med. 208: 2237, 2011; Cell Host Microbe 16, 215, 2014). Next his group studied samples from vaccinees in the Phase II and IIb HVTN 505 trials of a DNA/rAd5 vaccine candidate that contained Env gp41. This analysis showed dominance of the vaccine-induced antibody repertoire to gp41 with antibodies that cross-reacted with intestinal microbiome antigens (Science 349: aab1253-1, 2017). They went on to isolate gp41-microbiome cross reactive antibodies before vaccination, demonstrating microbiome-reactive B cells are expanded by HIV Env vaccination. Because intestinal microbiota shape the B-cell repertoire from birth, the researchers concluded that neonatal immunization with HIV envelope antigens may be able to imprint the B-cell repertoire to respond to antigenic sites on HIV Env that may otherwise be subdominant or disfavored, such as the broadly neutralizing antibody epitopes on HIV Env.

More recently Haynes’s group tested the same DNA/rAd5 vaccine regimen in neonatal rhesus macaques that was tested in the HVTN 505 trial, which was halted after failing to demonstrate efficacy in 2,504 MSM and transgendered women who have sex with men at 21 sites in 19 US cities (see Large AIDS Vaccine Trial Shudders to a Halt, IAVI Report blog, April 26, 2013). One of the objectives of the study was to see if gp41 immunodominance could be avoided by immunizing neonatal rhesus macaques during the early stages of microbial colonization (J. Virol. 91 21, 2017). They found that colonization of neonatal macaques occurred within the first week of life, and immunization of the animals during this time also induced a dominant gp41, microbiome cross-reactive antibody response, indicating that early vaccination could not overcome gp41 dominant responses.

Yet a retrospective study led by Sallie Permar and Genevieve Fouda, also with the Duke Human Vaccine Institute, found that infants were able to mount robust, durable Env-specific IgG responses, including anti-V1/V2 IgG responses, following vaccination with a recombinant gp120 vaccine candidate coupled with the adjuvant MF59, a potent oil-in-water emulsion made of squalene found in sharks and plants (J. Infec. Dis. 211(4), 508, 2015). This study assessed the binding and functional antibody responses in HIV-exposed vaccinated infants from historical samples collected as part of the Pediatric AIDS Clinical Trials Group 230 and 326 protocols, which were conducted in the US during the 1990s. While the study conducted by Permar’s group did not establish whether Env vaccination of infants would have been equally or more effective than HIV Env vaccination of adults, it did show they were capable of mounting robust antibody responses, including against the V1/V2 epitope that was an immune correlate of risk in the RV144 trial. They also tested samples from adults immunized with the same vaccine and found that the magnitude of the V1/V2-specific IgG response was higher in infants than in adults. (J. Virol. doi: 10.1128/JVI.01070-17).

Overbaugh says the idea of developing an AIDS vaccine for children is definitely worth thinking about, but added that there is still much to learn about how different types of immune responses are formed at this age in response to the virus. “In our study of breast-feeding infants we found that they developed broader and more potent antibodies rapidly [compared to adults], and a South African study found the same thing. So the interesting possibility that infants have an ability to make these antibodies quickly is, I think, incredibly important to study,” she says. “This could mean that vaccinating infants is the way to early protection, but it could also be that what happens in an infant is fundamentally different, and shaped by the mother’s antibodies passed on through breast milk.”

Sharon Nachman, chairman of IMPAACT and professor of pediatrics at Stony Brook University School of Medicine who was present at the AVRS meeting, agrees that there are many open questions about testing HIV vaccine candidates in infants. “What do we know about infant immunity? What do we know about child responsiveness and what do we know about the vaccines themselves that will help inform the right vaccine and the right vehicle to put the vaccine in?” asks Nachman. “These are some of the key scientific questions that were the focus of the AVRS meeting.”

Clues from animal studies

Recent findings in animal studies also offer clues about early life immune responses to HIV and HIV vaccines. Inarguably, one of the biggest challenges in AIDS vaccine development has been finding immunogens that reliably induce bNAbs to HIV. But a recent study led by Devin Sok, Antibody Discovery and Development Director at IAVI, managed to do just that by immunizing four calves (Nature 548, 108, 2017). When the animals were injected with a BG505 SOSIP trimer, all the calves developed bNAbs to HIV in their blood as rapidly as 35 to 50 days following two inoculations. This was a significant improvement from previous studies in rabbits in which the BG505 SOSIP trimer, while inducing neutralizing antibody responses against an autologous strain, was unable to induce bNAbs against the harder to neutralize Tier-2 viral strains.

What wasn’t widely known, says Marovich, is that the cows were all six months old. Whether age was a factor in the robust response is unclear though. Sok has his doubts that it mattered that much. “I don’t expect a difference between infant and adult cows because both have the same antibody repertoire,” says Sok. “It’s the antibody repertoire that’s unique from other animals.” But Marovich says a comparable study being done in adult cows could shed light on this question.

Planning for a trial

“Given what we are learning about the babies’ immune systems and their antibody responses to HIV, I think they are fully capable of responding to candidate vaccine immunizations,” says Haynes. “It’s perhaps useful to consider immunization of babies with an HIV vaccine candidate to try and understand in greater depth how babies respond.”

Ultimately, Nachman thinks the way forward for AIDS vaccine development will need to be multi-pronged. “In any epidemic, thinking about one vaccination time-point is the least effective. In order for an epidemic to change course you have to target all ages at different times, and not necessarily with the same vaccine candidate.”

She also thinks the surprise of RV144 is a good lesson for people to remember when deciding whether immunizing children is a path worth exploring. “If there is anything that RV144 has taught us it is that assumptions of how we think about vaccines are not always the right ones.”

Yet Nachman also emphasizes ethical considerations of testing vaccine candidates in children. “You have to have good reasons to bring a vaccine into children,” she says. “A region with a 40 percent acquisition rate of HIV would be a good place to do a vaccine study. It has to be a population of children where you can say it will do good and also do no harm.”

Mary Rushton is a freelance writer based in Cambridge, Massachusetts.