Virginia Peng, Contributing Illustrator

Vaccines do not offer the same protection to everyone. 

But what explains variability in protection across vaccines and between individuals? Researchers, including a team of bioinformaticians led by School of Medicine professor Steven Kleinstein, have been investigating vaccine success, an effort which began before the COVID-19 pandemic. The latest findings from the Signatures Project of the NIH Human Immunology Project Consortium, or HIPC, were published in three Nature papers which reveal insights into the immune system before and after vaccination. 

“You don’t want to only be able to show data that are only relevant to your little corner of the world,” said Rafick Sékaly, a professor at Emory University and one of the senior scientists on the HIPC team. “You want to show data which are generalizable, which are applicable to many regions of the world.”

The first goal of the HIPC Signatures Project was to compile and standardize data across many vaccine studies. The first paper, published on Oct. 20, describes ImmuneSpace  — a resource that contains both pre-processed data and the computer code used to conduct the normalization across source datasets. This project required extensive work to account for differences in design between clinical studies from across the world. 

ImmuneSpace took five years and a nationwide collaboration of over 25 bioinformaticians to complete. The work represents a critical step towards data accessibility. Equipped with data from ImmuneSpace, scientists drew two main conclusions from their analysis. Results were published in two papers on Oct. 31. 

First, if the immune system is more activated prior to receiving a vaccination, it will develop stronger antibody protection in response. Second, successful antibody production can be predicted based on gene expression patterns after vaccination. 

The ImmuneSpace team was led by Joann Diray-Arce, an instructor of pediatrics at Harvard Medical School and lead within the Precision Vaccines Program at Boston Children’s Hospital, and senior author Mayte Suárez-Fariñas, assistant director of the Center of Biostatistics at the Icahn School of Medicine at Mount Sinai. The researchers compiled immunological data across 1405 participants from 53 cohorts profiling the response to 24 different vaccines.

“Bringing the data together is a huge task,”Suárez-Fariñas said. “Now we have this beautiful resource which can become a gold standard. It is beyond any individual team’s ability to bring over 25 studies together and do what we did.”

Partnerships between institutions were key to the success of ImmuneSpace. Bali Pulendran, senior scientist on the post-vaccination study and Stanford professor, feels that research projects should not be limited to a university’s brand. 

“Institutional boundaries, although for administrative purposes they serve a role, in my psychology, they are irrelevant,” Pulendran said. “Just get the best people, wherever they might be, and get them to work in a team. That’s all.” 

Diray-Arce agreed, acknowledging that although different institutions were involved with the research, they “have one goal … to help the scientific community to standardize this dataset.” 

Without this extensive effort to consolidate and standardize the data, the studies that followed “would have taken more than double the time,” described Slim Fourati, first author of one of the latter two Signatures studies published, and researcher at the Case Western School of Medicine.

Due to the extensive data acquisition and processing achieved through ImmuneSpace, the HIPC Signatures team could conduct more thorough analyses and come to more robust conclusions. According to Sékaly, Fourati’s findings are “not confounded by geographic location.” 

Additionally, the inclusion of a wide array of vaccine types allowed for the results to be generalized across vaccinations. Although past studies have previously identified pre-vaccination immune signatures, Sékaly emphasized that no study held as much power as Fourati’s metadata analysis.

“We know there are differences from person to person,” Fourati said when describing the HIPC Signatures study he authored. “The idea of this paper was to look at whether or not, even before you get vaccinated, there is any mark that distinguishes someone who’s going to be fully protected by a vaccine from someone who’s not fully protected.” 

Fourati’s paper explored immune system data prior to vaccination. Characterizing immune “endotypes” based on markers of inflammation, Fourati and his colleagues found an association between innate immune activation and antibody production.

Individuals with more active immune systems prior to vaccination mount greater antibody responses, as measured by the presence of inflammatory markers. Although inflammation is often characterized as detrimental to health, this finding reveals that immune activation might help the body more robustly respond to a pathogen.

Vaccines are often designed to include an adjuvant, a substance that enhances the immune response. Fourati’s findings open up the possibility of individualizing vaccine formulations or otherwise activating the innate immune system before immunization. A simple blood test enables immune profiling that could inform a more personalized approach. 

The second of the papers published on Oct. 31 explored the immune system markers post-vaccination across 13 vaccines. When adjusting for differing timings of antibody responses, the researchers found common genes expressed across all vaccines studied. Their time-adjusted signature predicts whether the vaccine conferred a protective response. These findings have the potential to be incorporated into clinical trials to help “make the evaluation of candidate vaccines more efficient,” Kleinstein described. 

“What these studies have shown collectively is the incredible power of using vaccines as probes on the immune system to get a deep, mechanistic understanding of how this happens,” Pulendran said.

In addition to advancing the field of systems immunology, the HIPC Signatures papers highlight the future of big data in biology. Pulendran noted how there exists “an incredibly bright future” in fields related to vaccines, human immunology, and big data. According to him, it’s crucial that “bright young minds” enter this field and “take it to the next level.” 

According to Thomas Hagan, first author on the post-vaccination paper, the key to the impact of computational biology is free, public access to data.

“The work that we did could be done by any college student with a computer and an internet connection,” Hagan said.

High quality data is also essential to stronger analyses. When compiling data for ImmuneSpace, Diray-Arce and Suárez-Fariñas stressed that missing demographic information severely limits the applications of a dataset. Faced with a lack of patient sex data, the ImmuneSpace team had to become creative; the team analyzed patient genomes for the presence of a Y chromosome and added this attribute to the dataset.

Gaps in some demographic data became insurmountable, as they compiled many clinical datasets each designed differently. In the future, the team hopes that clinical studies pay extra attention to collecting data on sex, age, ethnicity, and other key determinants of health. 

“At the end of the day, these studies are not simply about an accumulation of data. Really, it’s from data to knowledge to understanding and then actionability – eventually,” said Pulendran.

The Human Immunology Project Consortium was founded in 2010.