Keystone Symposium: Building an Immune Toolbox

Researchers showcase efforts to apply basic immunological principles to the design of antibody-based vaccine candidates

By Yegor Voronin

For many years when leading HIV vaccine researchers gathered for the annual Keystone Symposium, the most befitting analogy to the state of the field was attempting to scale the snow-covered mountain peaks that serve as the meeting’s backdrop.

At this year’s symposium, HIV Vaccines: Adaptive Immunity and Beyond, which was held in the Canadian Rockies in the picturesque town of Banff Mar. 9-14, researchers acknowledged that although there is still a steep climb ahead, employing a better understanding of basic immunology might help ease the climb.
 
The focus on basic immunology was the choice of the conference co-organizers—Galit Alter, associate professor of medicine at Harvard Medical School; Susan Barnett, director of vaccines research at Novartis; and Nicole Frahm, associate director for laboratory science at the HIV Vaccine Trials Network—and it resulted in a different type of symposium. New faces were seen at the podium, discussions circled around understanding the immediate events after vaccination, and new collaborations were formed as researchers realized they are working on the same problem from very different angles.

Antibodies were the center of attention once again, but neutralizing antibodies shared the stage with molecular signaling, germinal centers, T follicular helper (Tfh) cells, and so-called functional antibodies (non-neutralizing). Researchers still have much to learn about the intricate process of guiding antibody responses at molecular and cellular levels. Still, recent progress holds promise for development of a new generation of vaccine candidates that are more grounded in basic immunological principles than their predecessors.

Starting with neutralization

In 2009, a collaboration led by Dennis Burton, professor at The Scripps Research Institute in La Jolla, published a paper describing isolation of two highly potent antibodies, from an HIV-infected donor, which could neutralize a broad swath of HIV isolates (Science 326, 285, 2009). Compared to previously isolated broadly neutralizing antibodies (bNAbs), the newcomers were active at 10- to 100-fold lower concentrations and were capable of neutralizing a wider spectrum of viruses. Over the past five years, dozens of bNAbs have been isolated from HIV-infected individuals, enabling a whole new wave of antibody research that has led to many discoveries about how HIV-specific antibodies develop in infected individuals.

Researchers have found that these bNAbs occur more frequently in HIV-infected individuals than was previously believed and that there is a smooth continuum between HIV-infected people that develop only weak neutralizing antibody responses and people that develop bNAbs, with a large majority having responses of intermediate potency and breadth (AIDS 28, 163, 2014). Studies have also shown that antibody gene variants conducive to generation of certain kinds of bNAbs are quite common among humans (PNAS 109, E2083, 2012). And recently, several groups have started tracking the arms race between the virus and the immune response in real time, trying to understand the details of the processes that lead to bNAbs appearance in some people (Nature 496, 469, 2013).

Delving into the germinal center

Following the discovery of the improved crop of bNAbs, researchers began studying the unique sequences and structures of these antibodies to identify what makes them so potent and broadly neutralizing. They soon realized that these bNAbs often have variable regions with unusually high levels of somatic hypermutation, genetic changes that allow them to bind more strongly to the specific pathogen.

It happens like this: Naive B cells activated by an antigen travel to lymph nodes and the spleen where, together with helper T cells, they establish special structures called germinal centers. Within germinal centers, B cells multiply and undergo the process of affinity maturation that results in higher affinity antibodies. Affinity maturation is an evolution-like process that is divided into two alternating stages. During the somatic hypermutation stage, a special enzyme mutates the antibody gene in each cell. Then, during affinity selection, the cells with mutated antibodies compete with each other for binding to the pathogen that is presented on the follicular dendritic cells (FDCs), and also for activating signals from Tfh cells. The B cells that win the competition may undergo additional rounds of mutation and selection, further improving their affinity to the pathogen.

The fact that genes coding for bNAbs against HIV have extensive somatic hypermutation suggests that B cells in germinal centers went through an unusually high number of these rounds of mutation and selection, likely reflecting a long period of exposure to HIV proteins in chronically infected people. This raised concerns that such bNAbs can only develop over prolonged periods of time in HIV-infected people and that it would be difficult for a vaccine to recapitulate this process.

However, researchers are now studying ways to either accelerate or guide the process of affinity maturation as part of a vaccination strategy. To do that, they need to first understand the processes that take place inside germinal centers.

Germinal centers and their functions have been extensively studied in mice, where it has been shown, among other things, that each center is founded by three to four B cells, which remain in that particular center and don’t migrate to neighboring germinal centers.

The laboratory of Gabriel Victora, a fellow at the Whitehead Institute for Biomedical Research, has been studying, in particular, how helper T cells behave in these structures (Science 341, 673, 2013), work they presented at Keystone. After transgenic mice, deficient of their own T cells, were transplanted with T cells randomly labeled with three different fluorescent markers, Victora and colleagues observed that cells labeled with different colors are always evenly distributed among lymph nodes and among germinal centers within a lymph node. These results suggest that T cells behave differently from B cells, in that multiple T-cell clones are engaged during formation of a germinal center or that they can freely migrate between centers.

This hypothesis was confirmed in another experiment in which T cells within a single germinal center were labeled by photoactivation and then traced to other germinal centers within the same lymph node. Moreover, Victora was able to directly show that new T-cell clones may invade existing germinal centers. T-cell-deficient mice were populated with a mixture of T cells recognizing two different antigens. Injection of one of these antigens led to establishment of germinal centers with T cells specific for that antigen. After injection of the second antigen, T cells with the new specificity were observed entering these germinal centers and providing help to B cells.

This difference in behavior between B and T cells makes sense during an immune response to a chronic mutating pathogen such as HIV. Spatial isolation of B-cell clones during the affinity maturation process prevents competition for antigen between B cells producing antibodies that target overlapping epitopes and a “winner takes all” scenario, which would limit the diversity of the antibody response. Whereas T cells are not competing with each other, and therefore the T-cell immune response benefits from wide dissemination of a new clone that is able to recognize escaping viral variants and continue to assist B cells.

Other research on germinal centers presented at Keystone suggests the events that occur at these sites may have an extremely long-lasting effect on humoral immunity. Mark Slifka, a senior scientist at the Oregon Health & Science University, and colleagues are focusing on understanding the mechanisms behind decades-long production of antibodies by plasma cells observed in response to some natural infections and live-attenuated vaccines, including the yellow fever 17D vaccine.

Some researchers believe that plasma cells are short lived and may be repopulated by memory B cells either via homeostatic mechanisms or in response to reinfection, but experiments by Slifka’s group strongly suggest that some plasma cells have a very long lifespan.

In one study, researchers prevented the repopulation of plasma cells by B cells by treating rhesus macaques with an anti-CD20 antibody, which deletes more than 99% of peripheral B cells, and by removing the spleens of these animals. Despite the lack of memory B cells, the animals maintained antibody titers against tetanus for more than 10 years, indicating continuous production of these antibodies by plasma cells that were formed before B-cell depletion occurred.

But just what makes these plasma cells so long lived is unknown. Slifka proposes that specific signals received by B cells in germinal centers, such as activation by multi-meric antigens and strong Tfh help may result in imprinting of the “long lifespan program” on the resulting plasma cells. The ability of HIV vaccine candidates to employ such signals, according to Slifka, and to trigger formation of plasma cells continuously producing neutralizing antibodies against the virus will be essential for long-lasting, sterilizing immunity.

Even in the absence of high neutralizing antibody titers, the ability to elicit long-lived plasma cells could be extremely beneficial. In the RV144 trial in Thailand, the first to show any protection against infection, the level of protection achieved by the prime-boost regimen (a canarypox vector expressing HIV env, gag, and pro, followed by a boost with B/E gp120 recombinant protein) was 60% in the first six months, but then rapidly declined (N. Engl. J. Med. 361, 2209, 2009). Some policymakers have suggested that if the 60% level of protection could be extended to last a decade or more, such a vaccine could be licensable in many countries.

Being Frugal with Inflammation

The search for vaccines that are safer and have fewer side effects has led to the development of subunit vaccines composed of a single protein. These vaccines, however, are not very immunogenic on their own and usually have to be supplemented by adjuvants—compounds that boost the adaptive immune responses by activating the innate inflammatory responses. At this year’s Keystone Symposium, Nick Valiante, site head of immunology/global head of immuno-therapy at Novartis Vaccines, presented the ongoing work of the Swiss pharmaceutical company to develop new small molecule adjuvants that are safer and more potent. Their hypothesis was that most of the inflammation that occurs in response to a pathogen is to non-specifically block replication of the pathogen and limit its spread in the body, while only a small portion of these inflammatory responses are needed to activate the adaptive immune system. Because subunit vaccines do not contain replicating pathogens, a large portion of the inflammation caused by adjuvants is therefore “wasted,” according to Valiante. Starting with early prototypes, Valiante and colleagues iteratively tested several variations of adjuvants that had different physicochemical properties, comparing the amount of inflammation with the resulting immune responses. The lead candidate resulted in little to no systemic inflammation, while eliciting the highest level of antibodies. This work shows that adjuvant efficacy can be uncoupled from its toxicity and may lead to adjuvants that trigger just the minimal essential inflammation required to stimulate adaptive immune responses. –YV

 

Not just neutralizing

The idea that signaling during B-cell activation may have a long-lasting impact on antibody responses also intrigues Alter. Her group studies what other antibody functions, besides neutralization, may play a role in protection against HIV, especially those that engage the innate immune responses. The potential importance of these so-called functional antibodies was highlighted by the results of the RV144 vaccine trial that showed protection against HIV in the absence of neutralizing antibodies (see Antibodies: Beyond Neutralization, IAVI Report, Jan.-Feb. 2010; Sci. Transl. Med. 6, 228ra39, 2014). New and better assays for other antibody functions, such as antibody-dependent cellular cytotoxicity (ADCC), complement activation, and phagocytosis are being developed and applied to look for correlates of protection or of viral control in animal studies and clinical trials.

Antibody binding may activate the complement cascade, a complex system composed of over 30 interacting proteins in the blood, which immobilizes the pathogen and may lead to its lysis or capture by macrophages. In addition, binding of antibodies to the surface of an HIV-infected cell attracts natural killer (NK) cells (seeRethinking the Natural Killer, IAVI Report, Winter 2013), which may kill the cell via ADCC. The so-called Fab portion of an antibody is where antigen recognition occurs, while the Fc portion triggers other functions such as ADCC (see antibody image, at right).

The Fc portion of an antibody does not undergo recombination and hypermutation and, therefore, is usually viewed as being constant, but there are distinct variants of it, which are referred to as subclasses. Alter and others have shown that antibody subclass may have a major impact on an antibody’s ability to initiate various innate immune responses.

For example, within the immunoglobulin (Ig)G isotype antibodies, the IgG3 subclass is much better at triggering ADCC than the IgG4 subclass. Alter and colleagues reported recently that the vaccine candidates in the RV144 trial elicited primarily IgG1 and IgG3 antibodies, while the vaccine evaluated in the VAX003 trial (a B/E gp120 recombinant protein) elicited primarily the IgG4 subclass of antibodies, and these responses correlated with protection in RV144 or the lack thereof in VAX003 (Sci. Transl. Med. 19, 228ra38, 2014). This evidence suggests that vaccination regimens may impact the subclass of antibodies induced, and that the subclass of antibodies could possibly account for the difference in efficacy observed in the two trials.

Of course it’s not as simple as just a difference in subclass. Even within a specific subclass of antibodies there are differences in the functionality of the antibodies, which are determined by the composition and structure of glycans attached to the Fc portion of the antibody, as Alter discussed at Keystone. These modifications can make the antibody better, or worse, at a particular type of activity.

The exact mechanisms behind selection of a particular type of Fc glycosylation pattern are not well understood, but Alter believes that early signals during B-cell activation play a major role. And once the plasma cell adopts a particular glycosylation pattern for the produced antibody, it retains it for the rest of its lifespan. Therefore, just like with imprinting long lifespan on plasma cells, appropriate signaling during vaccination may be key to the development of antibodies with the desired functions.

Vaccines and germinal centers

While the knowledge of what signals are needed to drive the immune response in a particular direction is important, researchers are still a long way from being able to provide these signals in the right place at the right time during vaccination. Diego Farfan-Arribas, a postdoc from Shan Lu’s laboratory at University of Massachusetts Medical School, presented their efforts in collaboration with Alex Dent from Indiana University to track and tweak the events occurring in germinal centers in response to vaccination (Hum. Vaccin. Immunother., in press). In a study comparing two different types of vaccines in mice, Farfan-Arribas and colleagues found that the vaccine administered as a prime had significant effects on cellular composition of germinal centers, which correlated with elicitation of different immune responses and modified the potency of the booster vaccination.

In germinal centers of mice primed with either HIV gp120-expressing DNA or a gp120 protein subunit vaccine, researchers observed approximately three-fold more B cells after vaccination with DNA. The effect was transient but clearly indicate these two vaccines resulted in different immune signals.

Also, priming with DNA had a delayed effect on Tfh cells. The number of observed Tfh cells was similar between the two vaccines after the prime, but boosting the DNA-primed responses resulted in a higher number of effector memory Tfh cells, an important subset of CD4+ T cells indicative of established immunity and recall response. Finally, priming with DNA and boosting with protein resulted in higher antibody titers, suggesting that the differences in the numbers of cells in germinal centers have downstream consequences for the overall immune response.

This effect was further explored by supplementing the DNA gp120 vaccine with plasmids coding either for ICOS-L or BLyS, two proteins involved in B-cell stimulation signaling. Expression of these signaling proteins in the vicinity of the gp120 expression was expected to have an adjuvant-like effect and boost immune responses to the viral protein. Also, local expression at the site of injection is preferable to systemic administration, which may lead to side effects and safety concerns.

Researchers reported that co-expressing ICOS-L resulted in slightly increased numbers of Tfh cells in germinal centers, while BLyS expression increased the number of Tfh cells by approximately 50%, while reducing the number of B cells. At the moment, the ability to rationally modulate B-cell responses is very limited, so researchers were surprised by the effect supplementing the vaccine with BLyS had on B cells, but such experiments are critical to inform future attempts to tweak the immune system via vaccination.

From mice to men

An obvious limitation of many of the basic immunology studies described above is that they were done in animal models. While mouse models allow genetic manipulation as well as the invasive sampling necessary to dissect the processes occurring in germinal centers, the mouse immune system is obviously very different from the human immune system, especially with regard to innate immune responses. Therefore, studies in humans are necessary to confirm these findings. However, to do such studies in humans, researchers will likely have to rely on samples of peripheral blood.

For this reason, some researchers are developing assays that can use blood samples to help deduce what is happening in lymphoid organs. Shane Crotty, a professor at the La Jolla Institute for Allergy & Immunology, and colleagues were studying CD4+ Tfh-like cells (generally defined by the presence of CXCR5 on the cell surface) in blood when they noticed that a subset of these cells was highly reminiscent of resting memory Tfh cells obtained from germinal centers (Immunity 39, 758, 2013). Compared to other CXCR5-positive cells, gene expression profiles of cells of this subtype (defined by expression of low levels of PD-1 and the lack of CXCR3 surface marker) was much more similar to the gene expression profile of Tfh cells. And in vitro, these cells behaved just like Tfh cells do in vivo—they activated memory B cells, stimulated expression of IgG, and led to conversion of B cells into plasma cells.

Moreover, the frequency of these CD4+ Tfh-like cells in blood samples from HIV-infected volunteers was associated with the development of bNAb responses. This indicates that it may be possible in the future to track immune responses in blood samples soon after vaccination to ensure they are developing in the desired direction.

Although there is still a long way to go in the design and development of HIV vaccine immunogens, this Keystone meeting highlighted the increasing importance of the convergence of developing a basic understanding of the immune signaling pathways with the empirical testing of various vaccine candidates. Working from opposite directions, these efforts complement each other more than ever before, with vaccine candidates getting more sophisticated via employment of the latest discoveries in immunology, and clinical trials serving as hypothesis-generating factories for basic research.

Yegor Voronin is senior science officer at the Global HIV Vaccine Enterprise.