The Patient Equation

by Glen de Vries

  A wristband activity tracker can tell us your daily movements. This may show us if the tumor burden is growing or shrinking—is your activity level going up or down? Can we find ways to either see or see by proxy information about the blood vessels penetrating the tumor, about how many new ones are surviving and growing? Can we view biomarkers like this a hundred times a second instead of once every six weeks? We will never be able to predict everything about someone's course of treatment—some patients might get hit by a bus, after all—but over time we can get a higher and higher resolution view at a closer‐and‐closer‐to‐continuous time frame so that we can truly know everything possible about someone's individual cancer, in time to act on it and stay ahead of the disease.

  p53‐ologists of the Future

  p53 is a tumor suppressor protein, encoded by a gene known as TP53. Mutations of p53 have been found in more than 50% of tumors, giving it a strong linkage to cancer.5 Normally, a cell will proliferate and die on a certain schedule. Cancerous cells don't die—they have uncontrolled growth, along with the desire to distribute themselves to the rest of the body and spread. p53 is supposed to stop the uncontrolled growth; if it isn't working, the growth continues, and the cancer wins. This is new knowledge, from over the past couple of decades, and that knowledge is limited right now. But what we do know is that if someone has a p53 mutation, lots of cancers can be predicted, across all different organ systems—prostate, kidney, brain, more. And yet we don't yet have p53‐ologists, people who have a single‐minded focus on understanding from a molecular basis what can go wrong if you have a p53 mutation as part of your tumor, regardless of what organ we're talking about—and, with that single‐minded focus, can start to experiment with what other layers of our layer cake we need to understand in order to begin to use our knowledge of p53 to eradicate disease. Instead of treating cancer organ by organ, for people with p53 mutations, it may be more useful to have someone who understands the protein and not necessarily the organ.

  PP2A is another protein, a cell regulator involved not just in cancer cell growth but in all kinds of metabolic pathways, some of which are related to cancer and some which are related to other cellular functions. It is also a tumor suppressor, and, when activated, it can help stop cancer growth. What's interesting about PP2A—and I'll tie this into our larger patient equation story in a moment—is that its anticancer properties were discovered somewhat by accident. Vector, the blog of Boston Children's Hospital, wrote about a set of researchers' efforts to find new treatments for T‐cell acute lymphoblastic leukemia (T‐ALL) by examining a library of nearly 5,000 drugs, compounds, and other natural products.6 What they discovered was that perphenazine, an antipsychotic medication, seemed to have a side effect: lower cancer rates. Its mechanism? Activating PP2A. “In reactivating PP2A,” Vector writes, “the drug was reactivating a protein that the tumor cells had turned off, triggering the cells to die.”7

  This is sparking research into whether PP2A activation is limited to helping in cases of T‐ALL, or if there are wider implications for other cancers. Between p53 and PP2A, scientists are starting to realize that if we can gain a better understanding of how to manipulate proteins that are critical to cellular functions, we can think about treating cancer patients based not on symptoms but on the very core of the disease. The thing is, these are just two of the 10,000 or 100,000 possible pathways toward solving cancer—and we are just at the beginning. It is data—well‐collected, smartly‐analyzed patient data—that will tell us more, and that will wake us up to molecular targets that are in the biosamples right now but haven't been identified yet. This is how patient equations will play out to their best effect in cancer: finding patterns we have never before been able to see.

  Or Perhaps Car‐t‐ographers of the Present

  In 2017, the FDA approved Keytruda as a cancer treatment “based on the genetic profile of a tumor, rather than the tissue or tumor type,” writes PharmaVOICE.8 It targets one particular genetic sequence found in a small (but significant) number of tumors—but more and more treatments like this are on the way. Joy Carson, senior director of oncology strategy for Novella Clinical, told PharmaVOICE, “We are just beginning to build a meta‐database of cancer targets—many of which do not yet have approved treatments…. It is a data engine that will drive drug development, but it will take many years to reap the benefits of this knowledge with effective targeted therapies.”9

  Keytruda isn't about cancer based on location in the body, it's about one particular genetic abnormality—with success in one study on tumors of the pancreas, prostate, uterus, and bone.10 According to the New York Times, “[o]ne woman had a cancer so rare there were no tested treatments.”11

  Fewer than 5% of patients benefited from these kinds of therapies in 2018—but that's still 60,000 potential patients for Keytruda alone.12 And, according to one analysis, fewer than 1 in 5 biopsied tumors were able to be matched to any targeted therapy drug (of which there are now 30).13 But this is the wave of the future. And the present.

  Novartis developed Kymriah, a customized treatment for relapsed or refractory B‐cell acute lymphoblastic leukemia, using the patient's own T cells to fight the cancer. Treatments cost nearly half a million dollars. But Novartis does not charge people whose cancers do not respond. The latest reported trial showed a remission rate of 82% within three months and overall survival of 49% after 18 months, compared to survival rates between 15% and 27% for these types of recurrent cases in the past.14 Customized immunotherapy treatments are the purest form of personalized medicine—they are literally personalized for each patient, hence the hefty price tag. But they also serve to limit the damage to surrounding cells, acting as a surgical strike against the cancer specifically. “This is just the beginning,” said Miriam Merad, director of the Precision Immunology Institute at Mount Sinai School of Medicine, on a panel at Aspen Ideas: Health. Between Keytruda and Kymriah, “[t]here are two molecules on the market, but there are hundreds of them we can exploit.”15

  Personalized Immunotherapy Beyond Cancer

  These same approaches that are showing such promise in cancer aren't limited to that disease. One startup, Alector, is developing drugs to target the brain's immune system in the hope that it can enable patients to fight off Alzheimer's disease.16 “If the immune system in the brain is not operating normally, the nerve cells cannot function normally,” Arnon Rosenthal, the co‐founder and CEO of Alector told OneZero, a technology and science publication hosted by Medium. “Eventually they degenerate and die. This is what leads to the disease.”17

  There is also another immunotherapy‐type approach that has recently made headlines for saving the life of a 15‐year‐old lung transplant recipient with cystic fibrosis—and a bacterial infection that wasn't responding to antibiotics.18 The answer? Phage therapy, a form of immunotherapy for bacteria, looking for the perfect virus to target the precise bacteria colonizing the patient.

  Phage therapy actually dates back to Russia, more than a hundred years ago. The government didn't want to pay for antibiotics, so instead, they looked for viruses that could kill bacteria. The therapy was mostly abandoned for years, but with the rise of antibiotic‐resistant infections, new approaches have been needed—and phage therapy has been unearthed from the dirt, along with the viruses hanging onto it.

  Graham Hatfull, a professor of biotechnology at the University of Pittsburgh, has been studying bacteriophages (viruses that prey on bacteria) for years—and collecting them, 15,000 of them, in storage at a deep freeze (–80° Celsius).19 I spoke to Hatfull, who says he wasn't necessarily looking to find a therapeutic use for phages, but had an unexpected opportunity to intervene with a particular patient, and it has taken his lab research in a surprising new direction.20

  At their core, bacteriophages can help us understand fundamental questions of diversity and genetics. Hatfull leads a program where 5,500 students each year, from 140 institutions, isolate new phages, categorize and examine them, and use them to
better understand biology and evolution. But in October of 2017, according to an account in Wired magazine, Hatfull received a desperate email from a microbiologist in London hoping for a miracle. Two teenagers with cystic fibrosis had received lung transplants—but their bodies were fighting infections post‐transplant that antibiotics simply couldn't resolve.21 There was nothing left in the hospital's toolkit—so Hatfull's colleague imagined that phages might be the teenagers' last best hope. One of the teenagers died before Hatfull and his team could identify the right phage to fight their infection—but the other got lucky, and the team found a group of three phages able to effectively attack the bacteria that were making her sick. She has since recovered and is doing well.

  The problem, Hatfull tells me, is that this is personalized medicine at its most personal—the specificity of the phages mean that they will work for one bacterial strain and one strain only. They'll work for one patient but no others, and each potential case requires a new search, a new cocktail, a new procedure. These are not (yet) off‐the‐shelf solutions, and if there is an off‐the‐shelf solution to be found, it is way down the road.

  Because of the work done to catalog and categorize his phages, Hatfull and his team were able to bootstrap their way into success in this one case—but success is not always replicable. Hatfull in fact wondered if they had gotten lucky with this patient, and if her infection would have cleared on its own, with or without the phages. So he has done further testing, trying to figure out whether and how he could effectively move forward with therapeutic intervention in more cases. He has had some failures, where the right phage could not be identified in time to save the patient, and a few minor signs of success, including another lung transplant patient with a raging infection for whom the right phage was found in just 10 days—a remarkably short turnaround time, Hatfull told me, and was showing real progress until the patient died from complications four weeks later.

  Over the next year, Hatfull hopes to take on more and more individual cases, hoping to put some process behind these one‐off interventions and figure out if there is a path forward to intervene at scale. In the old model of life sciences, no personalized therapies could possibly be scalable. But now we might have the data to start to deliver even the most custom treatments like this. As Hatfull's team has collected different bacterial strains to search for effective phages, he has found that for around one‐third of Mycobacterium abscessus strains, he can't find any good phages—but for the others, he finds at least one, and sometimes more than one. For his team, that presents two challenges: how to expand the set of known phages to get broader penetration, and how to speed up the screening process to find the ones that can be found more quickly and easily. The question still remains whether it will be economically viable to expand the program.

  In the future—keeping in mind this is not a near‐term likelihood— Hatfull imagines that we could figure out what makes the phages work and synthetically build them in the lab, tune our phages to be ideal attackers for whatever bacteria are presenting themselves, and make the phage treatment into a pharmaceutical solution. But, he says, we're not even close, and we don't yet understand why some phages infect some bacterial hosts and strains and not others. For millions of years, phages have been co‐evolving with bacteria, and it's a very complex dance that we're only beginning to comprehend. Hatfull says that the Bill & Melinda Gates Foundation has called for proposals for the use of phages to engineer the infant gut microbiome—to turn unhealthy gut microbiomes into healthy ones—but it's still very early. Hatfull is working with others to improve the data sets and better understand phages and their potential applications. He sees a possible future in combination therapy for tuberculosis—to shorten antibiotic therapy from months to weeks by using targeted phages in addition to antibiotics, and to use the phages to make the treatments we already have more robust and lasting.

  Immunotherapies for cancer and phage therapies for bacterial infections offer tremendous hope to patients for whom standard treatments have proven ineffective. But at least in these areas there have been traditional therapies to start from, and bodies of knowledge and research that have been in development for decades. In the next chapter, we'll look at Castleman disease—a rare condition that not much has been known about…until very recently. Castleman disease stands as an example of patient equations bringing us from near‐zero to a much fuller understanding of disease, and of treatment, driven largely by the work and passion of one man, Dr. David Fajgenbaum, trying to save his own life as well as thousands of others with the disease.

  Notes

  1. Mike Montgomery, “In Cancer Fight, Artificial Intelligence Is A Smart Move For Everyone,” Forbes, December 22, 2016, http://www.forbes.com/sites/mikemontgomery/2016/12/22/in-cancer-fight-artificial-intelligence-is-a-smart-move-for-everyone/.

  2. The Economist, “Understanding Cancer's Unruly Origins Helps Early Diagnosis,” Medium (The Economist), December 12, 2017, https://medium.economist.com/understanding-cancers-unruly-origins-helps-early-diagnosis-eb449e3ff466?gi=6a5cf570ac1.

  3. Jerry S. H. Lee and Danielle Carnival, “A Global Effort to End Cancer as We Know It,” Medium (The Cancer Moonshot), September 23, 2016, https://medium.com/cancer-moonshot/a-global-effort-to-end-cancer-as-we-know-it-42a9905327e8.

  4. Sharon Begley, “We Fought Cancer . . . and Cancer Won,” Newsweek 152, no. 11 (2008): 42–44, 46, 57–58 passim, https://www.ncbi.nlm.nih.gov/pubmed/18800570.

  5. Francesco Perri, Salvatore Pisconti, and Giuseppina Della Vittoria Scarpati, “P53 Mutations and Cancer: A Tight Linkage,” Annals of Translational Medicine 4, no. 24 (December 2016): 522—522, https://doi.org/10.21037/atm.2016.12.40.

  6. Tom Ulrich, “When Is an Antipsychotic Not an Antipsychotic? When It's an Antileukemic,” Vector, January 21, 2014, https://vector.childrenshospital.org/2014/01/when-is-an-antipsychotic-not-an-antipsychotic-when-its-an-antileukemic/.

  7. Ibid.

  8. Denise Myshko, “Trend: Advanced Diagnostics and Precision Medicine,” PharmaVOICE, November 2018, https://www.pharmavoice.com/article/2018-11-diagnostics/.

  9. Ibid.

  10. Gina Kolata, “Cancer Drug Proves to Be Effective Against Multiple Tumors,” New York Times, June 8, 2017, https://www.nytimes.com/2017/06/08/health/cancer-drug-keytruda-tumors.html.

  11. Ibid.

  12. Ibid.

  13. Denise Myshko, “Trend: Advanced Diagnostics and Precision Medicine.”

  14. Matthew H. Forsberg, Amritava Das, Krishanu Saha, and Christian M. Capitini, “The Potential of CAR T Therapy for Relapsed or Refractory Pediatric and Young Adult B‐Cell ALL,” Therapeutics and Clinical Risk Management 14 (September 2018): 1573–84, https://doi.org/10.2147/tcrm.s146309.

  15. Amanda Mull, “The Two Technologies Changing the Future of Cancer Treatment,” The Atlantic, June 25, 2019, https://www.theatlantic.com/health/archive/2019/06/immunotherapies-make-cancer-treatment-less-brutal/592378/.

  16. Ron Winslow, “The Future of Alzheimer's Treatment May Be Enlisting the Immune System,” Medium (OneZero), June 4, 2019, https://onezero.medium.com/the-future-of-alzheimers-treatment-may-be-enlisting-the-immune-system-d4de95ac1cff.

  17. Ibid.

  18. Sigal Samuel, “Phage Therapy: Curing Infections in the Era of Antibiotic Resistance,” Vox, May 14, 2019, https://www.vox.com/future-perfect/2019/5/14/18618618/phage-therapy-antibiotic-resistance.

  19. Megan Molteni, “Genetically Tweaked Viruses Just Saved a Very Sick Teen,” Wired, May 8, 2019, https://www.wired.com/story/genetically-tweaked-viruses-just-saved-a-very-sick-teen/.

  20. Graham Hatfull, interview for The Patient Equation, interview by Glen de Vries and Jeremy Blachman, July 2, 2019.

  21. Megan Molteni, “Genetically Tweaked Viruses Just Saved a Very Sick Teen.”

  8

  Castleman Disease—Not One Rare Disease with No Treatments, But Three Rare Diseases…with Hope, Thanks to Data

  Let's talk about the typical model of drug development. Life s
ciences companies are normally looking for the blockbuster, a treatment for the masses that can cover as many patients as possible. Of course effectiveness is important, but we only need to be more effective than the current treatment in order to go to market, and if we can potentially meet the needs of millions, even if not absolutely everyone who tries the drug will wind up benefitting…well, that's just the reality of drug development.

  Precision medicine turns that paradigm on its head. As we're able to get more precise with treatments, the number of patients for whom those treatments are relevant will necessarily decrease. This creates a huge shift in the kinds of business decisions a life sciences company will choose to make. It's a different goal, about taking all of the information we can generate and activating it to be something that really makes a difference for each and every patient.

  We saw that in the previous chapter with immunotherapy for cancer and phage therapy for bacterial infections, and it's just as salient a point here, with Castleman disease. To go from treating a rare disease with one relatively low‐effectiveness treatment to figuring out that the condition may actually be three distinct diseases, each of which may need a different treatment…we're talking about affecting lives deeply, but it's really just a few hundred or a few thousand lives at a time. This is a very different endeavor than searching for the next‐generation cholesterol drug or antidepressant.

  Finding Clusters in a Random World

  Just a few thousand patients around the world are diagnosed with Castleman disease each year. It is a deadly condition that affects the lymph nodes and can cause the collapse of multiple organ systems throughout the body. Traditionally, the best treatment has been a drug that has been shown to be effective in just over a third of patients. The rest? There has been scarcely more than hope. It's true that we've had clinical data to analyze for a long time, and patient characteristics to look at, but now we can get biospecimens, genomic and proteomic profiles, more data than we used to be able to collect or comprehend—and things are changing.