Rearch Briefs

Researchers Employ Systems Biology Approach to Predict Adaptive Immune Response to Yellow Fever Vaccine

Researchers have for the first time used a systems biology approach to predict the immune response to a vaccine. A study led by Bali Pulendran, a professor of pathology at Emory University, used microarray analysis to measure gene expression changes in the innate immune response to the yellow fever vaccine to predict the level of the adaptive T- and B-cell immune response with up to 90% and 100% accuracy, respectively (1).

The team vaccinated a group of volunteers with yellow fever vaccine, one of the most effective vaccines ever developed, and then used microarray analysis to measure gene expression changes as an indication of the innate immune response, which occurs within hours to days after vaccination, and is believed to regulate the adaptive antibody and T-cell response, which happens days or weeks later.

There are now overwhelming data that the innate immune system programs the adaptive immune system, according to Pulendran. For example, studies in mice suggest that eliminating the early innate immune response by eliminating certain genes severely compromises the adaptive immune response. And in 2006, Pulendran’s group showed that the yellow fever vaccine induces a number of toll-like receptors that are part of the innate immune system and this in turn was essential for the later CD4+ and CD8+ T-cell responses (2).

“If the innate immune system is acting within a few hours of pathogen entry and if it is programming the adaptive immune [response], can we use this early innate signature as a biomarker to predict which vaccinee will have a strong antibody or T-cell response?” asks Pulendran.

To find out, his group vaccinated 15 volunteers who had never been exposed to yellow fever virus or vaccine and used microarrays to measure gene expression changes in almost all of their genes. This avoided any possible bias that could come from focusing on biomarkers that are already thought to be important.

The researchers measured gene expression changes at several time points up to three weeks after vaccination, and the level of the yellow fever-specific B- and T-cell responses 60 days after vaccination. The same measurements were done in a second group of volunteers vaccinated one year later in an independent trial. In each group the researchers identified genes that had expression changes that best correlated with a high or low adaptive immune response later on. The gene expression signatures identified in one of the trials could predict with up to 90% accuracy whether vaccinees in the other independent trial would go on to develop a strong T-cell response; the accuracy was up to 100% for the antibody response.

“It’s the first study I am aware of that has used genomic data in a predictive fashion in two independent human studies,” says Paul Thomas, an assistant member in the immunology department of St. Jude Children’s Research Hospital, who is not connected to the study.

For now, it remains an open question as to whether the same signatures will apply to different vaccines. Pulendran has done preliminary studies that suggest that injectable flu vaccine as well as FluMist, which is given intranasally, both induce expression changes in genes that are very different from the ones induced by the yellow fever vaccine.

Still, the situation is different with flu. Unlike yellow fever vaccine, many people have likely previously encountered the influenza virus, according to Pulendran. “We are looking at a secondary response there, whereas with yellow fever we are mostly looking at a primary response,” he says. Also, the injectable flu vaccine is a purified protein and not a live-attenuated virus like the yellow fever vaccine. And FluMist, while live-attenuated, is given intranasally, suggesting that different types of cells are likely exposed to it. So, Pulendran says, it’s still possible that other live-attenuated vaccines have a similar signature.

Vaccine manufacturers of emerging vaccines to HIV, tuberculosis, or malaria could use this approach in small trials to identify gene expression signatures that predict long-term immunogenicity, according to Pulendran. In larger trials, they could then use it to identify people who are the most likely to be protected.

Next, Pulendran plans to study the biological role of some of the genes that showed up in the signatures. One gene that’s one of the best predictors of the T-cell response is known to be involved in the response of cells to stress, Pulendran says. “Why such a gene is predicting the cytotoxic T-cell response so well is a mystery,” Pulendran says.

A second recent study led by Rafick-Pierre Sékaly, a professor at the University of Montreal, also used microarray analysis to find that yellow fever vaccination consistently induces the expression of a group of transcription factor genes in three independent groups of vaccinees (3). Sékaly says the signature could be used to guide the development of vaccine candidates and adjuvants. —Andreas von Bubnoff

1. Nat. Immunol. 10, 116, 2009
2. J. Exp. Med. 203, 413, 2006
3. J. Exp. Med. 205, 3119, 2008 

Elite Controllers Found to Have More Lethal CD8+ T Cells

Even after many years of research, it’s still not well understood how rare HIV-infected individuals called long-term nonprogressors (LTNPs) control HIV replication and remain healthy without antiretroviral (ARV) therapy. Now, a study of elite controllers (ECs), which are a subset of LTNPs who control their viral loads to below 50 copies/ml of blood, has found that one key may lie in an enhanced ability of their CD8+ T cells to divide and kill HIV-infected CD4+ T cells (1).

Researchers cultured CD8+ T cells taken from ECs for six days in the presence of HIV-infected CD4+ T cells. They found that during that time, the cells divided and upregulated the protein perforin, which pokes holes into target cells, and the protein granzyme B, which kills target cells. This, they believe, explains why their CD8+ T cells could kill the HIV-infected CD4+ T cells much more efficiently than CD8+ T cells taken from progressors, which divided less vigorously and made less of these proteins. “We think going through the cell cycle causes [the CD8+ T cells] to upregulate perforin and granzyme B,” says Stephen Migueles, the lead author of the study.

“[In our research] we really did not previously have an effector function of HIV-specific CD8+ T cells that so dramatically segregated with the [ECs],” says Mark Connors, chief of the HIV-specific immunity section at the National Institute of Allergy and Infectious Diseases, part of the US National Institutes of Health, and senior author of the paper.

ECs have the same number of HIV-specific CD8+ T cells in their blood as untreated progressors but Migueles and Connors showed in 2002 that these cells can divide better than CD8+ T cells from progressors. The new study shows that on a per-cell basis, the CD8+ T cells from ECs transferred granzyme B into a greater fraction of HIV-infected CD4+ target cells than CD8+ T cells taken from a progressor. “It’s not that there are more CD8+ [cells] present in the nonprogressors, it’s that each cell kills more efficiently,” Migueles says.

The study also suggests that ARV therapy cannot repair the inability of CD8+ T cells in progressors to divide and upregulate perforin and granzyme B. ARV-treated progressors had fewer HIV-specific CD8+ T cells than ECs with equally low viral loads, and their cells behaved like CD8+ T cells from untreated progressors. This is unlike other functions like CD4+ T-cell proliferation, which improve after ARV therapy, Connors says. “This is really an impairment of the HIV-specific CD8+ T cells [in progressors] that is not fixed by ARV therapy,” says Connors.

For now, it’s still unclear what it is about ECs that makes their CD8+ T cells divide and kill so much better. “That’s the million dollar question,” Migueles says.

One consistent feature observed more often in ECs than in progressors is an allele in the major histocompatibility complex called B57. This is a host protein that HIV-infected cells such as CD4+ T cells use to present small pieces of HIV proteins on their surface to activate CD8+ T cells. But B57 alone is neither sufficient nor required for people to become ECs. “Obviously B57 has something to do with it but it can’t be the whole story,” Connors says. “There is some other interaction that we don’t yet understand that allows [ECs] to do this.”

Next, the researchers want to understand what’s different about CD8+ T cells taken from ECs once they divide and gain their ability to kill target cells. Migueles says they are looking at the genes that are not working correctly in the cells that are not dividing.

The study also found that the CD8+ T cells from progressors can be converted in vitro into ones that behave like cells taken from ECs through a combination of stimulation and rest. “If we can hit them really hard with very potent stimuli and get them to divide, they actually upregulate the killing machinery and kill very efficiently just like [cells taken from] nonprogressors,” Migueles says. “It’s far away from a treatment,” Connors adds, “but it is theoretically restorable.”

“It’s a great study,” says Guido Silvestri, an associate professor of pathology at the University of Pennsylvania who was not connected to the research. “This observation advances our mechanistic understanding of how these individuals can achieve immune control of HIV and thus avoid AIDS.” —Andreas von Bubnoff

1. Immunity 29, 1009, 2008