Chasing My Cure

by David Fajgenbaum

  Then I stopped believing in an omniscient medical system.

  I pulled a reverse Faust.

  I rejected the belief that any institution had all the answers or represented all the available knowledge in the world. I stopped hoping for the solutions. Instead of appealing to the magical and mysterious power of a higher authority to gain knowledge, I read books, reviewed studies, and investigated proteins. There was work to be done.

  I made it through several more terrible episodes of illness—and then something like a reprieve. I’m realistic about it. That’s what I have now. A reprieve.

  But there’s another aspect to the reverse Faustian bargain. I had started out trying to save my own life. Now I’m working to save so many others. It feels like my soul has stretched its borders, and it has come into contact with others’ in a way I could never have anticipated. This wasn’t short-term gain in exchange for long-term consequences, like Faust’s. In return for giving up the possibility of omniscience and enduring a living hell, I got a larger life, one more connected to others, and a shared sense of responsibility. I got more than I even could have imagined. And often still less than I have wanted.

  The very next patient to try sirolimus after me was a five-year-old named Katie. She was diagnosed at only two years old, just when she was getting into princesses and beginning to experience the world, all those firsts, like digging in the dirt and playing with her older brother. But these experiences had to be avoided to protect her fragile and flighty immune system. Her doctor told the family that he hadn’t had much experience with Castleman disease and had printed some shaky “information” off the Internet (the diagnostic criteria hadn’t been published at this point). He knew enough, though, to understand that pediatric cases of Castleman disease were even less well understood and studied than adult cases. Katie’s parents were understandably terrified. Eventually, they reached out to the CDCN. We were able to connect her with an expert in their area, but despite his best attempts, nothing seemed to be working for Katie. Over four years, she missed out on a lot of normal life. She endured multiple unsuccessful treatments and fourteen hospitalizations. She underwent a number of procedures, immunosuppressive treatments, and chemotherapies. Still no reprieve. She began receiving cytotoxic chemotherapy, which helped her symptoms somewhat but limited her energy levels, restricted her growth, and led to a serious side effect—hemorrhagic cystitis—which landed her in the hospital and required nine weeks of continuous infusions through a PICC line to treat it. All of this occurred while her dad was back and forth for military deployments to the Middle East.

  After exhausting all other options, her doctor decided to try sirolimus based on my case and data. Katie still isn’t back to 100 percent and still has her bad days, but sirolimus has significantly improved her quality of life over the last year and helped her to avoid even a single hospitalization. She has more energy, she runs, laughs, and plays more than she did in the previous four years. The drastic improvement in her health meant that she had the ability and the energy to graduate kindergarten. And Katie even learned how to ride her bike this year. That’s not a metric we’re going to track in a clinical trial, but it’s as important as any other for a parent. In short, she gets to be a kid again. Katie has also inspired her mother, Mileva, to join the CDCN fight as the volunteer leader of our patient engagement program; she does everything from consoling family members of deceased patients to motivating patients to join the fight.

  Castleman disease was very personal to me early on in my journey, because I was battling it. Now, Castleman disease is even more personal to me, because I have developed so many bonds with patients—like Katie—who have been affected by it. Seeing her doing so well brings me and my team so much joy.

  So the next time sirolimus was used as a last-ditch effort, my hopes were even greater. Lisa, a fourteen-year-old girl in Colorado, went from perfect health, horseback riding, gymnastics, and track to utter misery in the ICU in a matter of days due to iMCD. All her organs were failing. She gained fluid everywhere. She lost consciousness. Blocking IL-6 didn’t work. Carpet-bombing combination chemotherapy didn’t work. A ventilator and dialysis machine were the only things keeping her alive. Her doctors decided to try sirolimus along with everything else. She showed the slightest signs of improvement before crashing again. Despite multiple rounds of chemotherapy and full-dosage sirolimus, Lisa couldn’t survive her immune system’s attack on her body. She passed away three months after she was admitted. Idiopathic multicentric Castleman disease won again. Nothing more could have been done, but that didn’t make it any easier. Sirolimus may be my miracle for now, but it’s clearly not everyone’s. Among the first few iMCD patients treated with sirolimus that I’m aware of, some have improved and some have not. That’s not good enough. The fact that it has worked for me (at least, as of my now typing this page) isn’t enough.

  Why does sirolimus work in some but not others?

  What else is going wrong that could be targeted with a new treatment approach?

  To answer these questions, I continue to work. I recently pulled together the T cell–VEGF-mTOR data into a grant application and was awarded funding by the NIH—the first-ever R01 or federal grant of any kind to study iMCD. We’re applying the funds to push forward our understanding of T cells, VEGF, and mTOR in iMCD, including a clinical trial of sirolimus in patients who do not improve with siltuximab. We hope to understand the effectiveness of sirolimus in iMCD and uncover other potential novel treatment options.

  A little while ago, I was reading through one of my medical reports when I noticed a code that I didn’t recognize on the top of the page: D47.Z2. These International Classification of Disease, or ICD, codes come from the Centers for Medicare & Medicaid Services, and are part of the vast system for categorization that medicine relies on. I guess I could be forgiven for not recognizing D47.Z2, because there are some really obscure and precise codes—V91.07XD denotes burn injury by water skis on fire; V97.33XD indicates being sucked into a jet engine for a second time (a second time!); and W61.62XD means you’ve been struck by a duck, of all things.

  Confused, I googled it: D47.Z2. It was Castleman disease. Our very own code.

  When I was first diagnosed, no unique code existed for Castleman disease. It was always referred to with a miscellaneous code, one that covered a number of hard-to-categorize diseases. Not anymore. It’s not that there was more demand (or supply) for repeat water-skiing-fire burns before I became ill. It’s just that fewer people with Castleman disease were advocating for a code than individuals with water-skiing-fire burns. And it still wouldn’t exist if Frits van Rhee, colleagues from the CDCN, and I hadn’t lobbied the Centers for Medicare & Medicaid Services for it in 2014. Someone had to turn hope into action to make this a reality. If you don’t do it, oftentimes no one else will.

  We’ve come a long way.

  Before the creation of the CDCN, the fight against Castleman disease was kind of like if American colonists were fighting one another when they were taking on the British (with apologies to my Oxford colleagues and British collaborators, “the British” represent Castleman disease in this metaphor). The colonists would have never fought off the British if they weren’t unified. The CDCN pulls individual researchers together and recruits new researchers to pursue a unifying vision and game plan that achieves more than we could on our own.

  Rather than performing studies that measure the levels of one molecule (e.g., IL-6) in the blood of a few Castleman disease patients, as had been done in the past, our collaborative network enables studies of biospecimens pooled from around the world that are measuring far more data points than ever before. Take the latest, for instance: Frits and I are currently analyzing results from a proteomics study that measured the levels of 1,300 proteins in 362 blood samples from 100 iMCD patients, 60 patients with related immunological disorders, and 40 healthy controls. Our initial review of the data su
ggests that our proposed model was correct in at least one way: iMCD involves many cytokines, and there’s a lot more to it than IL-6. Remember: You can’t see what you don’t look for and you don’t know what you don’t measure. And the data also suggest that iMCD really does seem to sit at the intersection of autoimmunity and lymphoma. Frits and I now serve as cochairs of the CDCN’s Scientific Advisory Board. He leads all efforts related to Castleman disease patient care and expert guidance; I lead our translational research efforts. We still frequently disagree on our interpretations of data, but that’s so important in science. Our debates and discussions get us closer to the right answer. And sometimes we grab a beer together afterward. He’s my colleague first, my friend second, and he also happens to be my doctor.

  To ensure that the CDCN is poised to launch large-scale studies like the proteomics study, we’ve developed a multipronged strategy. We launched a biobank to continuously procure samples and patient data. These samples can be stored until the moment that our network crowdsources a brilliant new idea for a study or until enough samples are obtained for a study that was on hold. We still turn to our community of physicians and researchers to contribute samples, but we’ve learned it’s much more efficient to go directly to patients for samples, just like we do for patient data in the registry study. Coordinating sample sharing between researchers’ institutions requires immense effort and utilization of every negotiation, strategic decision making, and managerial economics course I took in business school. Fortunately, patients want to be a part of the solution and give of themselves, literally, through their blood and tissue samples. Patients often post on social media that the CDCN is their only hope for a cure and a normal life. But actually, it’s those patients, who donate samples, data, and funds, that are the only hope the CDCN has for developing a cure.

  We also partner with tech and pharmaceutical companies on large-scale studies whenever possible. One tech company, Medidata, is contributing machine learning and data science tools to help us generate clinically meaningful insights from the half a million data points in the proteomics study. Though they are often demonized because of a few notable bad actors, pharmaceutical companies have incredible power to do good through contributing funds, data, and samples for research. These are the essential building blocks for breakthroughs, and the effort needed to obtain these resources often limits the rate of progress. Moreover, they are the only players in biomedical research that can actually develop drugs to save patient lives. Saving lives is our end goal, and we don’t ever forget it.

  We’ve had several major breakthroughs for iMCD, and we have several high-impact studies in process with collaborators around the world to identify the cause, key cell types, communication lines, and possible new treatment approaches. The results of the CDCN’s first multi-institution study, spearheaded with the viral hunter at Columbia University, indicate that a viral infection is not likely to be the cause of iMCD. Now, we’ve turned our attention to genetics and identified several mutations that are being investigated as possible causes of or contributors to iMCD. These leads were generated from another high-priority study on the CDCN’s International Research Agenda: a genomic sequencing study, made possible by more than forty thousand dollars in donations from my Wharton classmates. Multiple independent researchers are currently studying the potential role and effects of these genomic alterations in iMCD.

  I’m one of those researchers studying genomic alterations in iMCD and also one of the patients in whom alterations in an immune regulatory gene have been identified. The gene of interest serves as an on-off switch for T cells. A mutation in this gene could explain my T cells spiraling out of control and my iMCD. But it’s not very straightforward or easy to tease out whether this mutation is the cause or just a red herring; we all harbor thousands of rare variations in our genomes that have no consequences at all. Indeed, distinguishing the genetic variations that actually contribute to disease from the ones that don’t is far more challenging than the proverbial needle in a haystack. This process is more like finding a single strand of hay that looks just like the other 3 billion strands of hay in the haystack—it would be great if it stood out like a needle!

  We know that I inherited one copy of the mutated gene from each of my parents. We were able to determine my mother’s DNA sequence ten years after her death because she had been in a clinical trial where a few tubes of her blood were collected but never studied. I knew about the trial and the tubes, because I held her hand while it was drawn (my mom, like my sister Lisa, didn’t do well with needles). At the time, she had given her consent to allow for future research and sharing with outside researchers. I know she never imagined it would be her son who was doing the research or her baby boy who could benefit from the information contained in her blood samples, but I don’t think either of us expected most of what has happened to our family over the last decade and a half.

  One way to try to figure out when rare variations are predisposing us to or causing a particular disease is to introduce the exact genetic change found in the human with the disease into mouse embryos. Once the mice are born, we can compare their phenotypes or particular features with those of mice that are genetically identical except for the mutation. If the mice with the particular mutation demonstrate features similar to those of the humans with the disease, but the nonmutant mice don’t, then you’ve nailed it. Ruth-Anne Langan, a PhD student in my lab, is currently studying mice with my exact mutations to investigate the gene’s possible role in iMCD. We call these mice little Daves. We’re hopeful that we’re on to something but realistic that we still have a long way to go, especially since mutations in this gene have not been found in other iMCD patients as of yet.

  I’m still contacted daily by physicians or patients asking what to do to treat Castleman disease and how the disease works. I say “we don’t know” a lot less than before, but I’m still forced to say it. Perhaps one of the most gratifying aspects of my work is how we’re sharing our network blueprints with other rare disease groups, so they can follow our steps toward building patient-centric, collaborative networks to crowdsource the most promising research. No more silos. We hope there will be less “we don’t know” for other rare diseases too.

  We also look to other diseases to leverage the good work that has already been done in studying them. Most directly, we review our research findings to propose candidate drugs already approved to treat other diseases as possible off-label treatments for Castleman disease. Consider that it took twenty-five years between the discovery of elevated IL-6 in iMCD and the first-ever FDA-approved treatment for iMCD targeting IL-6. Now, consider that approximately fifteen hundred drugs are already approved by the FDA for various diseases, which could be used tomorrow or even today for the first time ever to treat one of the approximately 30 million Americans with one of the approximately seven thousand rare diseases with no FDA-approved treatments. How many lifesaving drugs already exist that are waiting to be applied to deadly diseases?

  Let me tell you about one.

  When my uncle Michael was diagnosed with metastatic angiosarcoma, a rare cancer with a horrible prognosis, I accompanied him for a visit to a top sarcoma oncologist. My uncle was told that there were two treatment options and that he likely had about one year to live. I asked the doctor if he would send off my uncle’s tumor for cancer genetic testing to search for a genetic mutation that could possibly be targeted with a treatment already FDA approved for other forms of cancer.

  My uncle’s doctor told me he wouldn’t order the genetic test, because it rarely returned any useful information. Though it is informative to diagnosis and prognosis in a large proportion of cases, it impacts treatment selection in fewer than 10 percent of cases.

  What if my uncle was in that fewer than 10 percent? I thought.

  Then I asked if the doctor would perform a test of my uncle’s cancer for something called PD-L1 and if the test was positive, consider tre
ating my uncle with an FDA-approved inhibitor of PD-L1 or its receptor, PD-1. Programmed death–ligand 1 is often found on the surface of cancer cells as a result of cancer-causing genetic mutations and DNA damage. The cellular protein doesn’t just hide the cancer from the immune system, it actually induces the death of immune cells that approach to try to kill the cancer cells. So inhibiting PD-L1 or its receptor in cancer patients in whom PD-L1 is increased allows the immune system to recognize and kill those cancer cells without being killed themselves. It would be a Hail Mary as to whether PD-L1 was elevated or if blocking it would be helpful for my uncle. The doctor explained that PD-L1 had not been studied in angiosarcoma or any other form of sarcoma and no drugs blocking PD-L1 or its receptor had been used in these cancers, so he wouldn’t order the test or consider administering the drug.

  “Even if it’s positive, the drug probably wouldn’t work and it’s prohibitively expensive anyway,” he went on.

  But you don’t know if you don’t try, I thought. Someone has to be the first. And you just told my uncle that he has limited time and options. How do you know what’s prohibitively expensive to someone anyway?

  After the visit, I encouraged my uncle to find another oncologist who would order the tests, which he did. The sarcoma expert was right about one thing: The genetic test was unrevealing. There were no mutations in the genetic code of the cancer cells that could be effectively targeted with an existing treatment. However, my uncle’s cancer cells were brilliantly positive for PD-L1. They were covered in it. Could blocking PD-L1 treat his cancer? Two drugs targeting the receptor for PD-L1 were already FDA-approved for lung cancer and melanoma. Soon thereafter, my uncle became the first angiosarcoma patient that we’re aware of to take one of those drugs. He experienced a dramatic improvement in his symptoms, laboratory abnormalities, and tumor size. By the time this book is published, I hope that he’ll be crossing three years in remission. Of course, he has no guarantees for the future, but “every day is a gift” according to my uncle Michael. This case has now led to off-label use and clinical trials of this drug and others like it in angiosarcoma, which we hope will help many more patients with this disease.