Technology Expands Access to Translational Medicine

The Undiagnosed Diseases Program (UDP) at the National Institutes of Health sees 120-150 previously undefined diseases per year. Given that physician-scientists are considered successful if they can describe 1-2 new diseases in their lifetime, countless patients are left waiting, wondering whether the world of medical research will ever be able to help them.

The cases that make their way to the UDP are typically genetic disorders involving maladaptation of the genome. The basic process for approaching these rare and undiagnosed diseases involves looking for mutations, examining where in the genome they occur and determining whether they are causative of the phenotype that led the patient to seek treatment.

This process can take a lifetime — especially if you are working in relative isolation, with a small number of patients and limited resources.

Enter Cornelius Boerkoel III, MD, PhD. In his time at the NIH UDP, Dr. Boerkoel led the development and implementation of a software program that enables physician-scientists to leverage the power of social media and cloud technology to solve complex and undiagnosed disease cases.

After leaving the NIH last fall, Dr. Boerkoel began working with Appistry, Inc. as their Chief Medical Officer to further refine and improve the software infrastructure so that it can be made available to the wider research community.  The system is available through Appistry as CloudDx Translational.

In his presentation Integrating Translational Medicine into the Clinic given at Nationwide Children’s Hospital, in March 2015, Dr. Boerkoel, also a senior clinical scientist at the Child & Family Research Institute at British Columbia Children’s Hospital in Vancouver, Canada, shared his experiences from working with the UDP and how the tools that they developed led to individualized medicine for patients.

Building a Village

The idea behind the CloudDx Translational software infrastructure is a virtual village. By considering how humans have solved problems anthropologically, Boerkoel was inspired by the concept.

“Villages are where people of niche expertise come together,” Dr. Boerkoel said during his presentation. “Their synergy allows for survival.”

As a whole system for workflow from patient recruitment through clinical data collection, testing and research, the system enables physician-scientists to solve problems much more quickly than when they must rely on separate contacts, databases and communications to build cohorts, compare data and get feedback from peers. In essence, it enables the construction of a virtual village of rare disease investigators.

One case that came to the UDP while Dr. Boerkoel was developing this tool involved a 6-year-old boy who had uncontrolled seizures from birth. At presentation, he was experiencing many seizures a day and was developmentally arrested at the neonate stage.

Genotyping revealed a de novo GRIN2A mutation, known to be associated with seizures, activating the NMDA receptor in nerve cells in a way never before seen. Armed with this information and using the interconnected databases in the system, the team was able to find a drug that intervened in the appropriate place.

The whole process took a matter of a few months, resulting in improved health care quality and a massive reduction in health care costs. The accurate diagnosis and targeted treatment transformed a demanding inpatient disease into a manageable chronic illness requiring only outpatient care.

Increasing Communication, Reducing Bias

Dr. Boerkoel envisions communication platforms of the system as a village piazza — a place where people with different experiences and viewpoints can come together to discuss issues and problems and come up with unique solutions.

Within that same system, physician-scientists could also access the research that other teams were doing on that patient. Shared notebooks were designed to allow the ability to add to or comment on a notebook, supporting collaboration among teams.

The system was also designed to be useful in a clinic setting. It provides real-time feedback about the strength of a clinician’s description and makes suggestions on where else to look and what else to ask as the provider is entering phenotype descriptions, recording vitals and other clinical data and commenting on the patient. When physicians are with a patient, they can adjust their approach as they gather more information.

This is critical for helping physicians overcome bias.

“Doctors are not omniscient,” Dr. Boerkoel conceded. “We’re all biased, and our biases change over time.”

Such bias colored the care received by two sisters who were referred to the UDP for developmental delay and hypoglycemia. The older girl had normal development until 18 months, when delays appeared. She was also dysmorphic and obese. The younger sister had delays from birth, but she did not appear dysmorphic or obese. Both girls had hypoglycemia since 20 months of age.

The medical community’s consensus was that there must be one disease that explained all the symptoms of both girls.

At the UDP, they began by looking at the genomes of both girls and their parents.

The team first uncovered a homozygous mutation for both girls in PCK1, and this mutation was ultimately shown to be causative of hypoglycemia. They then found two different de novo mutations — RAI1, which is associated with Smith-Magenis syndrome, in the older girl and GRIN2B, which is associated with developmental delay, autism and seizures, in the younger girl.

Both mutations appeared in the top 10 suggested mutations when the physicians entered the phenotype terms into the clinical description portion of the system, showing how the integration of data can help clinicians and researchers develop more targeted hypotheses of where to look for genetic irregularities.

Practical Translational Medicine

Technological advancements in translational medicine are changing the way physician-scientists work and think about research in both undiagnosed and diagnosed disease. While the tools used by Boerkoel and his team at the UDP mean more answers and more treatments for the (formerly) undiagnosed, researchers in all areas of translational medicine are developing technologies that will lead to more advancements for more patents.

“Individualized medicine has always been the ideal and is the ideal, although it cannot be divorced from the concept of the individual as part of community,” says Dr. Boerkoel in a follow-up interview. “What technology changes is the ability to define individuality and community at different levels and to use the information generated in more creative and beneficial ways. This applies to all diseases, not only undiagnosed diseases.”

At Nationwide Children’s, researchers and physician-scientists use tools similar to those that Dr. Boerkoel described — shared lab notebooks and cloud technologies among them. The tools may be a bit different, but the approach and overall goals are the same.

“Dr. Boerkoel’s work is a larger scale of what we are doing. His research is more from the perspective of a unique rare disease in a single person or handful of people,” says Kim McBride, MD, MS, principle investigator at the Center for Cardiovascular and Pulmonary Research at The Research Institute at Nationwide Children’s. “While we do some very rare disease work, we are more focused on cohorts of people. There is considerable overlap in the overall process and approach.”

At Nationwide Children’s, the role of technology in translational medicine is influenced not by the undiagnosed diseases flooding the UDP, but by a more universal need to quickly and efficiently get usable genomic information.

This need to get results faster and more efficiently inspired Peter White, PhD, principal investigator and director of the Biomedical Genomics Core at Nationwide Children’s, to create Churchill, which is a computational pipeline for sequencing genomic data.

“Churchill is a piece of the systems we use here — it’s basically how we go from raw sequence data to discovery of pathogenic variants,” Dr. White says.

Identifying pathogenic variants is a crucial piece of the diagnostic and research puzzle, Dr. White says. By quickly translating genomic data into valuable medical information, his team hopes to dramatically impact the world of customized patient care. Churchill can search a person’s genome for disease-causing variations in a matter of hours.

“Genome sequencing is widely accessible, even for small research teams,” Dr. White says. “However, after a genome is sequenced, scientists are left with billions of data points to analyze before any truly useful information can be gleaned. Churchill fully automates the analytical process.”

Clinically, this fully automated analytical process enables physicians to know in record time which pathogenic variants are present in their patients to confirm or aid in diagnoses.

The Churchill algorithm is also now available commercially. It was licensed to GenomeNext, LLC, which has built upon the Churchill technology to develop a secure and automated software-as-a-service platform.

According to the physician-scientists behind Churchill and CloudDx Translational, all of the innovations in software related to translational medicine add up to one achievement: better care for patients. It is their hope that the growing trend in advanced technologies in translational medicine will enable physician-scientists to bring individualized medicine to more patients than ever before.

References:

1. Appistry, Inc. http://www.appistry.com/ (accessed 2015 Mar 31).
2. Boerkoel CF. “Integrating Translational Medicine into the Clinic.” Presented at Nationwide Children’s Hospital, Columbus, Ohio. 19 Mar 2015.
3. Adams DR, Yuan H, Holyoak T, Arajs KH, Hakimi P, et al. Three rare diseases in one Sib pair: RAI1, PCK1, GRIN2B mutations associated with Smith-Magenis Syndrome, cytosolic PEPCK deficiency and NMDA receptor glutamate insensitivity. Molecular Genetics and Metabolism. 2014 Nov, 113(3):161-170.
4. Pierson TM, Yuan H, Marsh ED, Fuentes-Fajardo K, Adams DR, et al.GRIN2A mutation and early-onset epileptic encephalopathy: personalized therapy with memantine. Annals of Clinical and Translational Neurology. 2014 Mar 1, 1(3):190-198.
5. GenomeNext. www.genomenext.com (accessed 1 Apr 2015).
6. Kelly BJ, Fitch JR, Hu Y, Corsmeier DJ, Zhong H, et al. Churchill: an ultra-fast, deterministic, highly scalable and balanced parallelization strategy for the discovery of human genetic variation in clinical and population-scale genomics. Genome Biology. 2015 Jan 20, 16(1):6.