“I think I’ve finally found my calling in immunology.”

Having come to the United States from India in 1998, Purvesh Khatri, PhD (assistant professor, Institute for Immunity, Transplantation and Infection and BMIR), originally planned a career as an electronics and communications engineer: “I wanted to develop cell phones and things like that,” he says. As a master’s and PhD student at Wayne State University in Detroit, he first began to make subtle alterations in his career trajectory in response to a variety of events.

From Engineer to Software Developer

First, his master’s thesis advisor moved into informatics, so Khatri went along. Then his advisor went on sabbatical to California for a year, leaving Khatri new to the country and a language he could neither speak nor write, although he was able to read English. Struggling to work with postdoc collaborators at Wayne State, he happened upon another event. He explains, “I watched this friend of mine copying and pasting data from one website to the next and then the next and ultimately into Excel. It was a long and tedious process. I told him I’d be glad to write a program to help him. Three days later I gave him the software that became Onto-Express.”

Onto-Express (OE), which became the topic of Khatri’s master’s thesis, is described in a 2002 article in Genomics (79:266-70) with these words: “OE is a tool designed to mine the available functional annotation data and help the researcher find relevant biological processes. Many months of tedious and inexact manual searches are substituted by a few minutes of fully automated analysis.” Khatri boils that definition down to a few choice words: “it quickly allowed people with a lot of genomic data to interpret those data.”

“So,” he says, “we put it online and people started to use it and it exploded, which I didn’t expect. There were 100 people using it every day all over the world. There was a huge need. We had 12,000 to 15,000 users worldwide.”

This turned Khatri turned into a software developer, supporting all those users until, 7 or 8 years later, he “decided it was time to jump ship again because it was getting boring.”

By this time Khatri had his PhD in computer science, having used tools that he developed to design and interpret high throughput gene expression data as his doctoral thesis. It happened that Stanford was looking for a postdoc at about that time, and Khatri took the position, fashioning it into a joint appointment between the Departments of Pediatrics and Medicine because, he says, “what I really wanted was to be able to make some predictions and have someone there to validate them, someone with a wet lab to pay 50% of my salary so they would have some skin in the game. So that’s how I moved into translational medicine. I was going to develop some methods, use them to make predictions, and see my predictions validated.”

A Move into Transplant Medicine at Stanford

In his first position at Stanford, Khatri worked in organ transplant rejection. And he began wondering why patients with different transplanted organs – heart vs kidney vs lung – are treated differently. He explains his thinking: “The doctors who treat them are different; the treatments are different; the protocols are very different from one to another. I wondered why it is that we are not just treating them all similarly since ultimately it’s the recipient’s immune system that recognizes the allograft and tries to get rid of it. How do you find solutions that are going to be broadly applicable to large populations? You find them by finding the triggers that are common across all populations.

“We were thinking that with all of the data that already exist, why can’t we combine them? What if we were to change the approach? What if we were to stop excluding all of the heterogeneity? And we found that we were not only good at diagnosing active rejection of transplanted organs, but we were now predicting clinical injury 18 months before it showed up clinically.”

This ability to recognize rejection before its clinical appearance led to more experiments. “We weren’t sure if this was just an association or the trigger. We thought, if it is the trigger, we should be able to find drugs that will allow us to counteract the trigger and therefore counteract rejection. We showed in a mouse model that if we used drugs based on this common trigger, the immune cells that infiltrate the transplanted organ were significantly reduced and graft survival increased by a large amount.”

It doesn’t matter what virus it is, if you take out the things that it needs for its replication it won’t be able to replicate itself.

Having shown this success in mice, Khatri and his collaborators returned to analyzing large amounts of data to apply their findings to humans.

“We went into medical records in Belgium and were able to show that among the patients who got one of the drugs that we had predicted would work, graft failure was reduced by 30% over eight years.  This proved our hypothesis that by leveraging heterogeneity we could find triggers that would explain other comorbidities and that would be diagnostic, prognostic, and therapeutic. And because they were therapeutic, by definition they must be mechanistic.”

A Home in Computational Immunology

At this point, Khatri’s trajectory changed once more, because, as he says, “In the process of learning about organ transplant I came to realize that there were numerous opportunities for a computational scientist in immunology. So I became a computational immunologist.”

His early work in immunology is beginning to bear fruit. One current area of interest is the infectious diseases. Recently, Khatri’s lab has found a common host response to multiple different viral infections, which he is now using to identify host-directed broad spectrum anti-virals, in collaboration with Shirit Einav’s lab. “Because viruses cannot replicate themselves,” he says, “they hijack the host’s machinery to replicate themselves. Our hypothesis is that if we target that host cell cycle machinery to disrupt the viral life cycle, the virus will die. By repurposing drugs that were developed for other diseases that target the mechanisms that are important for the viral life cycle, we will kill the virus.”

To more simply clarify the issue for non-scientists, he offers the following war metaphor: “If you are an island and the only way to get to the island is a bridge, and you take out that bridge, then no enemy will be able to get to you. It doesn’t matter what virus it is, if you take out the things that it needs for its replication it won’t be able to replicate itself.”

With such an interesting early career in science, one hopes Purvesh Khatri will not recognize his success in communicating complex matters simply and clearly as yet another opportunity to alter the current trajectory. It should take awhile to uncover all there is to know about killing viruses.