Medicine’s New Frontier
iami resident Luis Rios was diagnosed with pancreatic cancer in July 2017. For two years, chemotherapy and other drugs controlled his disease, but in summer 2019, his cancer had grown resistant to all treatments, a common occurrence. But because he had a rare inherited mutation (called RAD51C), he was a candidate for immunotherapy, which boosts the body’s immune system to fight cancer.
The results surprised even Rios’s doctors at Sylvester Comprehensive Cancer Center at the University of Miami Miller School of Medicine. Within two weeks, he was completely off pain medications and his tumor markers had plummeted. After two years of treatment, he had no detectable disease. He has now been off treatment for a year, with no evidence the cancer is coming back.
In 2015, halfway around the world in Turkey, Miller School researchers from the Dr. John T. Macdonald Foundation Department of Human Genetics began studying a family in which a number of relatives had gradually lost their hearing. An examination of the family members found no known gene mutations, stumping the investigators. By 2000, however, scientists had completed the capability for whole genome sequencing, and the Miller School researchers hit a home run.
A mutation was found in the family’s MINAR2 gene, which affects the inner-ear hair cells that are critical to hearing. More recently, a second Turkish family was found to have a different mutation in the same gene. And just last year, a third mutation was found in the same gene in two Indian families, for a total of 13 individuals in the four families studied. Because the hearing loss caused by the mutations is gradual, it leaves a window for possible medical intervention with genetic therapies for the four families.
Today, back in South Florida, researchers on all three UM campuses are developing ways to measure viral levels in wastewater for early detection of COVID-19 and other pathogens. By sampling the wastewater, the researchers conducting these metagenomic studies can identify multiple microbes in a sample, as well as determining their concentrations.
If, for example, a student residence shows a large uptick in the virus, campus leadership can focus more on testing the students living there. Also, viral detection in wastewater can provide a four- to seven-day lead indicator of the number of cases clinicians are likely to see in hospitals.
Treating a single patient’s cancer. Finding a common genetic mutation in a group of families. Detecting viral outbreaks that threaten entire communities. What do they have in common? All three are medical challenges with potentially hopeful outcomes.
At the Miller School, these kinds of advances are becoming regular occurrences. Researchers and clinicians are working together to achieve results that were once unimaginable — and the pace of discovery is picking up.
This is happening because medicine is getting bigger and smaller at the same time. It’s getting bigger because we know so much more about human biology than we did even five years ago. It’s getting smaller because all that knowledge is helping us individualize care.
Oncology is a perfect example. Tumors are like malignant snowflakes — no two are exactly alike. In fact, if we sequence the same patient’s tumor six months apart, its mutational landscape will probably be quite different.
To understand tumors — and everything else — we need to grapple with data. We can produce huge amounts for both large groups and individual patients, but the key is understanding it. As we dive deeper, we’re getting better at transferring, storing and analyzing data. Most important, we’re learning how to harness information and technology to improve care.
Big Data’s Big Promise
“We need to differentiate between data and knowledge,” said Stephan Schürer, Ph.D., professor of molecular and cellular pharmacology, associate director of data science at Sylvester and director of digital drug discovery at the Institute for Data Science and Computing. “Data on its own is not particularly useful. It needs context. Once we take in the larger picture, the data becomes useful.”
To keep up with data-intensive technologies, UM is expanding its own data science capabilities. Traditionally, academia has produced the first rough ideas, then biotech and pharmaceutical companies run with them.
“We have begun trying to take these projects further,” said Claes Wahlestedt, M.D., Ph.D., a professor of psychiatry who has significant experience in drug development, “but UM needs more specialists, not only in medicine but also in data sciences — applied, for example, to contemporary therapeutic discovery and development.”
To make that happen, UM is recruiting more data experts, and sometimes in unexpected places, like the insurance industry.
“They have the deepest bench of actuaries,” said David Reis, Ph.D., chief information officer at UHealth – University of Miami Health System. “We’re looking to do more ‘what if’ analyses: What if we change this value? What does it do to the data?”
Dr. Schürer is helping lead the Miller School’s efforts to make data useful. That may mean finding a single variation in a tumor cell’s DNA, which contains about 3 billion base pairs. Sequencing an entire genome produces around 200 gigabytes of data. Now, multiply that one sequence by thousands.
These huge databases must be analyzed, stored, transferred and protected. Patient data must remain deidentified to ensure privacy. Data standards must be unified so that researchers and clinicians everywhere are speaking the same language.
However, once the necessary algorithms are written (a work in progress) and information safeguarded, big data offers a gold mine of potential health care applications, from choosing the right therapies to mixing and matching drug combinations to meet each patient’s therapeutic needs.
“Tumors are highly heterogeneous, and we can now do single-cell sequencing,” Dr. Schürer said. “We could potentially analyze every cell in a tumor and find the right combination of therapies that could destroy it entirely.”
Increasingly sophisticated machine learning models may soon match patients with the right treatments, based on the tumor’s genomics and other traits. These algorithms might even predict whether a tumor will develop resistance and how clinicians might respond.
“Everything comes down to having a lot of data, and most important, organizing and annotating that data to make it usable,” Dr. Schürer said. “If we can use the data to figure out which therapy, or combination, is better for a specific patient, or adjust therapies based on the response, that would be a huge benefit.”
Much of the high-volume, cutting-edge research is conducted through shared resources, which are centralized research facilities that share their equipment and expertise with investigators throughout the University.
“Shared resources are laboratories that provide a wide range of research support for basic, clinical and translational studies, and have high-end equipment and expertise that are far too expensive for any individual investigator to afford to place in their own laboratory,” said George Grills, associate director of shared resources at Sylvester. “Because these resources are used on a shared basis, they play a key role in many research projects for a large number of investigators.”
This is how researchers from many disciplines in the wastewater sampling project are crunching data to track COVID-19. The study is a collaborative effort between researchers at UM (including from the Miller School and the College of Engineering) and Weill Cornell Medicine in New York City. It is also part of an international research consortium. The shared resources supporting this project include the Sylvester Shared Resources, including the Behavioral and Community-Based Research Shared Resource, which facilitates environmental sample collection and facilitates access to COVID-19 data from human surveillance; the Biospecimen Shared Resource, which provides biorepository services for the environmental samples, including annotation, tracking, processing, storage and distribution; and the Onco-Genomics Shared Resource, which provides rapid viral RNA extraction and purification, rapid detection with PCR, and next-generation sequencing for viral strain variant identification and metagenomics analysis.
Other collaborating shared resources include the Miami Center for AIDS Research (CFAR) Laboratory Sciences Core, which provides rapid viral detection with a novel and fast PCR method developed by Mark Sharkey, Ph.D., an assistant research professor; and the Miami Clinical and Translational Science Institute (CTSI) Biostatistics Collaboration and Consulting Core, which is facilitating the development of COVID-19 predictive models that integrate human and environmental surveillance data. The results from this large collaborative study can be used to create “weather maps” of viral prevalence that can help predict near-term consequences — a tremendous resource to inform public health decisions.
Unraveling DNA to Conquer Cancer
Precision medicine seeks to treat each patient based on the genetic mutations and other factors that are specific to their condition. This approach is revolutionizing many areas of medicine, especially oncology.
“The future is bright,” said Carmen Calfa, M.D., a breast medical oncologist and assistant professor of clinical medicine. “The speed of scientific breakthroughs in precision medicine is leading to groundbreaking therapies. It’s exciting to be at Sylvester, where we share our passions and our discoveries — scientists and clinicians all working under one roof.”
To help unite this interdisciplinary firepower, Dr. Calfa directed Miami Precision Medicine 2022: Unraveling the DNA to Conquer Cancer. This first-ever conference, presented by Sylvester and sponsored by the Miller School and UHealth, explored genomic sequencing, artificial intelligence and other ways to personalize care and improve outcomes.
One of Dr. Calfa’s patients who personifies this approach is Niurka Morales, who was diagnosed in 2018 with breast cancer and received surgery, radiation and hormonal treatments. Unfortunately, the cancer metastasized quickly and progressed to stage 4. Dr. Calfa’s team used information from a liquid biopsy — a test that analyzes tumor cells that detach into the blood — to get Morales into a major clinical trial. Called TAPUR (Targeted Agent and Profiling Utilization Registry), it was launched in 2016 and is the first precision medicine clinical trial led by the American Society of Clinical Oncology. The trial tests the use of FDA-approved drugs that target specific tumor gene abnormalities in people with advanced-stage cancer.
Morales has been not only surviving but thriving on a combination of two targeted medications for more than a year. This approach would not have been an option before the trial, and Dr. Calfa has now assumed the role of chair of the TAPUR Steering Group.
But improving care is only one of Dr. Calfa’s many concerns. As medical co-director of survivorship programs and associate director of community outreach at Sylvester, she is always looking at the big picture: quality of life.
“You want to help your patients live, but you also want them to live well,” she said. “Some patients are cured. Others require lifelong treatments. Cancer changes over time, and resistance is developed. That is why you need to change treatments as the disease changes. Our innovation and research and clinical trials are all seeking to offer more options — new treatments that work longer, more precisely, with less toxicity.”
Longer, Better Lives
These new treatments are making an enormous impact on patient care, helping people live longer and better.
“Ten years ago, we had different types of chemotherapy, but it was all chemotherapy,” said Marilyn Huang, M.D., professor of gynecologic oncology, director of translational gynecologic oncology research and a women’s reproductive cancer expert. “Information from the Human Genome Project and The Cancer Genome Atlas has transformed our understanding of cancer, highlighted the importance of cancer genomics and provided a foundation with which to develop better targeted therapies.”
Cancer is a genetic disease, and that can generate both strengths and vulnerabilities. Mutations take the brakes off cell growth, resist the body’s immune response and therapeutic interventions, and turn off the mechanisms that protect genes from more mutations. But they can also be targeted for treatment.
“With our enhanced understanding of cancer genomics, we have been able to identify targetable mutations, tailoring medical therapies to individuals based on their molecular profile,” Dr. Huang said.
Checkpoint inhibitors, a form of immunotherapy, are a relatively new class of treatments. Because the immune system’s killer T cells are incredibly powerful, they must be carefully controlled. Normal cells send special signals that help T cells differentiate good guys from bad. Unfortunately, cancers learn to co-opt these signals to avoid immune attack. Checkpoint inhibitors take away that advantage.
Another drug class, called PARP inhibitors, takes cancer cells over the edge. PARP is a DNA repair enzyme, and inhibiting it allows more mutations to accrue — so many, in fact, that tumor cells begin to die.
“Targeted therapies such as PARP inhibitors or immunotherapies work best in patients with specific mutational landscapes,” Dr. Huang said. “That’s where molecular testing comes in, to help providers match tumor mutations to available directed treatments.”
Sometimes the new diagnostic capabilities reveal pleasant surprises. Dr. Huang told the story of a young woman who was attempting to start a family when she was diagnosed at another institution with a rare and often deadly uterine sarcoma. The woman sought a second opinion at Sylvester. Using next-generation sequencing, which can map a patient’s tumor genomic landscape, providing fundamental data to guide diagnosis and, potentially, treatment, Dr. Huang’s team found the woman didn’t have a rare uterine sarcoma at all. Rather, she had an unusual fibroid, a common benign tissue growth that occurs in the uterus. She did not need any further intervention and was cleared to start a family.
Pancreatic cancer is a different challenge, but Sylvester scientists have found that one patient subgroup can achieve complete responses, sometimes going from near death to durable recoveries.
“This is a highly selected subpopulation of patients with pancreatic cancer who have specific inherited mutations, most commonly in their BRCA gene,” said Peter Hosein, M.D., associate professor of clinical medicine and co-leader of the Gastronomical Cancers Site Disease Group. “Their cancers respond to immunotherapy when most pancreatic tumors do not.” Luis Rios, whose successful response was described earlier, is his patient.
The treatment may work because BRCA and RAD51C mutations make tumors genetically unstable, creating abnormal proteins called neoantigens that sensitize the immune system. While pancreatic tumors usually don’t generate an immune response, these mutations, which appear in about 5% of pancreatic cancers, make the tumors more likely to respond to immunotherapies.
In another exciting collaborative effort, Sylvester is joining forces with UM’s College of Engineering in a new initiative called Engineering Cancer CuresTM to improve cancer detection and treatment.
“By cross-pollinating our knowledge of cancer biology with deep engineering expertise, we hope to discover novel ways to attack tumors,” said Sylvester Director Stephen D. Nimer, M.D. “We can do this through cell engineering, better use of data, artificial intelligence and many other efforts.”
Engineering Cancer Cures has three groups: Intelligent Materials and Targeting, Cancer Tissue Engineering, and Artificial Intelligence/Machine Learning/Deep Learning-Based Imaging and Analytics.
One priority is finding better ways to medically target tumors. While chemotherapy can be quite effective, it is also systemic, meaning it can generate ill effects throughout the body. The researchers want to transform cells into drug factories that specifically target tumors and leave healthy tissues alone.
“Therapies of the future are cell injections and tissue engineering that can go inside a body,” said Ashutosh Agarwal, Ph.D., associate professor of biomedical engineering, director of engineering and applied physics for the Desai Sethi Urology Institute, and the Cancer Tissue Engineering co-lead. “We are working with Sylvester’s cancer clinicians to develop engineered cells, tissues and gene edits that can solve incurable cancers.”
In addition, the team will engineer new models to better understand cancer biology. For example, pancreatic and other tumors co-opt immune cells to protect themselves from chemotherapy and other treatments. Researchers will be able to better interrogate these relationships to end resistance.
The Intelligent Materials group will also investigate using nanoparticles to ferry therapies to tumors, leaving surrounding tissues unharmed. Beyond that, the Artificial Intelligence/Machine Learning/Deep Learning-Based Imaging and Analytics team will leverage clinical knowledge, advanced mathematics and computer science to improve diagnostic testing.
The Fountain of Health
When Spanish conquistador Ponce de Leon landed in Florida, not far from Miami, he was after the fountain of youth. He didn’t find it, and generations of alchemists, explorers and hucksters also came up empty. We may not have any choice about growing old, but we can assert some control over how we go about it. Call it the fountain of health.
That’s the thinking behind the Precision Aging Network study, which is assessing participants based on their risk of developing cognitive problems and dementia. Eventually, clinicians would like to match an individual’s risk with a customized therapeutic plan to help them maintain their cognitive and brain health.
“If you give us your family history, education level and describe your diet and medical status, we may be able to predict your cognitive health,” said Stacy Merritt, M.A., research and administrative director of the Evelyn F. McKnight Brain Institute at the University of Miami.
Part of a $60 million University of Arizona-led initiative, the study brings in researchers at the McKnight Brain Institute and collaborators from Emory, Johns Hopkins, Baylor College of Medicine, Georgia Institute of Technology and the nonprofit Translational Genomics Research Institute.
“This five-year program will significantly advance our understanding of precision medicine to predict individual brain health risks and provide personalized solutions to maximize cognitive health span,” said Tatjana Rundek, M.D., Ph.D., professor of neurology, scientific director of the McKnight Brain Institute, the EMBI Endowed Chair for Learning and Memory in Aging, and the study’s principal investigator at the Miami site.
There are many risk factors associated with cognitive decline. While family history, age and gender can’t be modified, diet, physical activity, social interaction, education level, hypertension and diabetes can be. The researchers want to better understand how these and other factors contribute to cognition, particularly in memory, executive function and processing speed, and develop resilience profiles that can predict an individual’s cognitive journey.
The study has three main components. The first is an online cognitive test called MindCrowd (mindcrowd.org). The investigators hope to recruit half a million people to take the assessment, and they’ve already reached nearly 300,000.
During the second phase, researchers will conduct in-depth physical examinations on more than 1,600 people: brain imaging, blood work, genomic and epigenomic studies. Finally, the team will use this data trove to create better computational and analytical tools to help determine an individual’s risk of developing dementia and how they might avoid that outcome.
The scientists are intrigued because previous studies have shown that cognitively spry older adults don’t always have healthy brains.
“We studied 200 people across the United States who were 85 or older and cognitively healthy,” said Rundek. “When we scanned their brains, many of them had really bad-looking images, but they were still sharp. What was preventing those people from having cognitive deficits?”
Early research is providing some clues. Education seems to be protective against cognitive decline. Other modifiable factors are smoking (don’t), drinking (only in moderation) and exercise (move more). Social engagement is shaping up to be a particularly powerful factor.
“Many of these cognitively sharp people in their 80s and 90s are working part-time jobs,” said Dr. Rundek, “reading, writing, playing music, teaching, mentoring. They are fully engaged and feeling optimistic about their lives that have meaning and purpose. And though we cannot create new neurons, we can create new connections between them, and that may protect people as they age.”
The researchers want to understand the genetic, molecular and biochemical mechanisms that govern these protective effects, and perhaps even duplicate them pharmacologically. In addition, they may be able to detect biological markers that might indicate who is most at risk of developing dementia.
These results won’t come into play for several years, but Dr. Rundek has one particularly important piece of advice to help people keep their edge now.
“As we age, we should seek out novelty,” she said. “If you get into the same routine, your brain stagnates. You need to show it something novel, so it can form new connections.”
Solving Medical Mysteries
For decades, patients, families and doctors wrestled with mystery diseases. While the symptoms were obvious, physicians were unable to diagnose the underlying conditions. Patients often went on years-long diagnostic odysseys — specialist after specialist, test after test — to figure out what was wrong. To make matters worse, these conditions often affect children.
But next-generation sequencing changed how we manage rare and undiagnosed diseases (RUGDs), identifying the genetic variations that cause many of them.
With this incredible tool, physicians could more effectively diagnose RUGDs. Still, there was a catch: While doctors could diagnose many of these genetic diseases, they often could not treat them.
“There are around 8,000 listed RUGDs,” said Mustafa Tekin, M.D., interim chair of the Dr. John T. Macdonald Foundation Department of Human Genetics, “but there are only a few approved treatments for them.”
In 2018, the Miller School became a clinical site for the Undiagnosed Disease Network, an NIH-funded consortium of researchers. The school’s physicians and researchers combine comprehensive exams with genomic sequencing and other tools to help diagnose patients.
“We first look at known gene mutations,” Dr. Tekin said. “If we don’t find any, we conduct whole genome sequencing to potentially identify new genes.”
Illustration of a syringe filled with binary code with a drop forming a human head
This is how, in a separately funded study, Dr. Tekin’s team was able to diagnose the Turkish and Indian families with inherited hearing loss. He has been studying the genetic underpinnings of hearing loss for more than 20 years, and has amassed a database of deafness-associated gene mutations found in families all over the world.
Still, diagnosis is no longer enough. Researchers want to expand access to treatments, which is proving to be a long and challenging journey.
Bone marrow-associated conditions may be among the easier ones to treat. Patient marrow can be removed, a new gene inserted, and the marrow reinfused to permanently replace a missing or dysfunctional protein. Metabolic conditions may also be relatively low-hanging fruit.
“These are more accessible because they are often caused by gene mutations in the liver,” Dr. Tekin said. “If we can package the gene in a vector that can target the liver and give that gene back, that’s a potential treatment and possibly even a cure. Also, given the success of mRNA vaccines for COVID, that could be another treatment strategy, providing time for developing a cure in the future.”
Leveling the Diagnostic Playing Field
Medicine’s new frontier is about more than the “what and why” of disease. Increasingly, it’s about how diseases affect specific groups. For example, Black men have higher rates of prostate cancer. Increased genetic risk and poor access to care may contribute to these bad outcomes. Another problem is clinical trial disparities. If researchers only study white patients, they will only identify new drugs that work best in white patients.
Miller School researchers are exploring different ways to study population data to increase health care equity. Molecular geneticist Sophia George, Ph.D., associate director of diversity, equity, and inclusion at Sylvester, and associate professor in the Department of Obstetrics, Gynecology and Reproductive Sciences, is leading an international study by the African-Caribbean scNetwork to decode Black genomes and investigate the genetic drivers behind breast, ovarian and prostate cancers.
“We hope to provide new insights into why Black people are at higher risk for aggressive cancers and often develop them at younger ages,” said Dr. George. “We will dissect the genetic and cellular underpinnings that can lead to poor cancer outcomes.”
The network is building a single-cell atlas of healthy breast, fallopian tube (where most ovarian cancers originate) and prostate cells from people in Kenya, Nigeria, the Bahamas, Haiti, Jamaica and the U.S. They will combine single-cell and genotyping data to better understand what constitutes normal tissue and decipher how cells become cancerous.
“We are investigating how genes contribute to different cell types, and the distribution of cells,” Dr. George said. “We will be mapping stem cells, differentiated cells, immune cells and other types. We want to understand the full spectrum of us.”
Correcting Clinical Trial Imbalances
In another effort, surgical oncologist Neha Goel, M.D., is focused on clinical trial representation. Dr. Goel co-authored a study that analyzed data from the AACR Project GENIE Consortium — one of the largest and most diverse genomic databases — on 6,652 women with primary or metastatic breast cancer.
The study found that Black patients with metastatic cancer had fewer actionable mutations. One example is PIK3CA, an enzyme that is commonly mutated in breast cancer. Clinical trials have shown that one PIK3CA inhibitor (alpelisib) improves progression-free survival. However, this drug is less likely to be offered to Black women because the population has fewer PIK3CA mutations. And because these women are understudied, there remains a paucity of therapeutic alternatives.
Dr. Goel believes these results underscore the disparities that accrue when genomic studies consist of a homogeneous, mostly white population. “If we don’t look at these potential mutations in a diverse patient cohort, we won’t necessarily know to include them in drug development and clinical trials,” she said. “It stops being precision medicine if there are wide gaps between different racial groups.”
The imbalance of clinical trial subjects is a life sciences-wide deficiency that physician-scientists are working to correct. The Miller School, Vanderbilt University Medical Center and Meharry Medical College have received a $12.4 million grant from the National Institutes of Health to jointly develop the Southeast Collaborative for Innovative and Equitable Solutions to Chronic Disease Disparities.
Racial and ethnic minorities make up 39% of the population in the Southeast, including nearly 15 million Black people and 9 million Latinos. The Southeast Collaborative aims to reduce disparities in diabetes, cardiovascular disease and obesity care among Black and Latino populations.
“Minorities in the Southeast fare worse on many health indicators compared to other regions, in large part due to poor socioeconomic status, with more than 22% of Southeastern residents living in poverty,” said Roy Weiss, M.D., Ph.D., professor and chair of the Department of Medicine and one of the four principal investigators. “Effectively addressing pervasive chronic disease disparities will require interventions that consider the needs, priorities and lived experiences of those disproportionately impacted.”
The Miller School is also a lead partner in the National Institutes of Health’s All of Us Research Program, an ambitious 10-year precision medicine initiative that aims to collect data from more than a million people in the U.S. The Miller School, Emory University, the Morehouse School of Medicine and the University of Florida make up the Southeast Enrollment Center (SEEC) network, which is strengthening the program’s reach in underserved communities.
“All of Us is the first genomic research project of this size and scope,” said SEEC lead Stephan Züchner, M.D., Ph.D., the Miller School’s chief genomics officer. “Eighty percent or more of biomedical research is conducted among non-Hispanic white populations, which does not represent the census of the nation. That is why SEEC plays such an important role — we’re possibly the most diverse geographic region in the country, and we are currently the national leader in All of Us in recruiting participants from communities traditionally underrepresented in research studies.”
All of Us is seeking to achieve both breadth and depth, such as creating a whole genome sequence for each participant. Last fall, the program began releasing genomic data for 100,000 participants through its Researcher Workbench, with data from another 100,000 expected by the end of this year, which will make it one of the largest genome databases in the world. Soon, All of Us will begin returning personalized information on genetic risk factors to participants who requested it when they enrolled.
“There is general agreement that medicine is increasingly using genomic information for diagnostic and treatment considerations,” Dr. Züchner said. “Genomic information will drive precision and personalized medicine approaches. That’s why All of Us has great potential benefit for both providers and patients everywhere. It will take time, and we’re still in the early stages, but this is not a dash — it’s a marathon.”
Louis Greenstein, Lisette Hilton and Richard Westlund contributed to this article.