Tumor adaptability is determined by genetic mutations that are different for each patient. Identifying the optimum combination of chemotherapies and non-cancer drugs for each patient requires modeling the tumor profile in a whole animal model. However, start-up biotech company My Personal Therapeutics is using proprietary fruit fly avatars in combination with artificial intelligence and high-throughput screening to provide personalized treatment recommendations.

Need for Targeted Cancer Therapies

Today, similar chemotherapy drugs are used to treat patients with different cancers. There are approximately 300 different types of cancers, each with a specific set of characteristics. In individual patients, tumor growth is affected by systemic inflammation, microbiome composition, adipose tissue and other factors, while cell death (apoptosis and senescence) is influenced by different combinations of gene mutations, including alterations of genes not previously associated with cancer.

Tumors can therefore rapidly adapt, leading to low response rates and relapse. As a consequence, the best cancer treatments comprise drugs, either single agents or more often combinations of substances, that target multiple specific nodes of the tumor network –– an approach referred to as polypharmacology.

Leveraging Genetic Drivers

Identifying the optimal solution cannot generally be achieved using current diagnostic test methods. It is not possible using the information obtained from current analyses to predict how a tumor with multiple altered genes will respond to treatment.

Recapitulation of the tumor with maximum fidelity in a whole-animal model is needed to enable unbiased testing of potential drug treatments. Mouse, zebrafish and fly avatars are some of the models currently being developed. Genetic information from a patient’s tumor is used to identify specific drivers for tumor growth in that patient, and this information is used to link patients to specific clinical trials or possible targeted treatments.

Founding of My Personal Therapeutics

London-based My Personal Therapeutics (MPT) was founded in August 2019 to leverage fruit fly avatar technology developed by Ross Cagan and his colleagues at the Icahn School of Medicine at Mount Sinai Center for Personalized Cancer Therapeutics (CPCT) and offer personalized, genomics-based cancer treatment recommendations.

MPT exclusively licensed the Personal Discovery Process (PDP) technology from Mount Sinai for global commercialization. PDP leverages Big Data to build personalized fruit fly avatars that model individual patients at an unprecedented level of complexity. Using robotics, thousands of drugs are screened in combinations to identify drug cocktails able to target the tumor while preserving the patient’s quality of life.

This approach enables us to account for each patient tumor’s genetic complexity at a level never before achieved in a clinical setting. Importantly, all recommended treatment combinations will include a cancer drug and one or more non-cancer drugs, making treatments less toxic and more affordable.

The genomic information and drug treatment data collected are being used to develop TuMatch, an artificial intelligence–powered service designed to enable rapid and affordable personalized treatment recommendations. We hope to have TuMatch available by early 2020 for colorectal cancer, followed by other cancers.

Advantages of Fruit Fly Avatars

Drosophila melanogaster fruit flies are advantageous as avatars because they are small, genetically tractable and live for a very short time, allowing the rapid screening of hundreds of drug combinations. As a result, the PDP serves as a massive clinical trial for each individual patient. Using the proprietary technology developed at Mount Sinai, we can generate fruit fly avatars incorporating multiple tumor mutations, effectively modeling the complexity of those tumors. For each patient, using robotics, 400,000 avatars are screened against up to 1,200 FDA-approved drugs to identify optimal combinations

The technology was demonstrated in a U.S. patient with advanced colon cancer resistant to many different chemotherapy agents.¹ The tumor was sequenced and found to contain at least nine cancerous mutations. Over 300,000 flies engineered with these mutations in their gut cells were screened against 121 existing drugs both singly and in combinations of up to three. The optimum combination consisted of the cancer drug trametinib and the osteoporosis treatment zoledronate — the patient was able to survive for an additional two years because of this combination therapy.

Choosing the Right Patients

A patient or oncologist only needs to provide My Personal Therapeutics with sequencing results from a tumor biopsy and a blood sample, even from biopsy material collected up to two to three years earlier. We prefer to work with oncologists, but recognize that more patients today are taking ownership of their health care. As a result, we have a website through which they can purchase the TuMatch screening service if they are eligible, which is determined through consultation with the oncologist.

Because the process takes time, we need to ensure that the patient’s cancer is not too far advanced. We are currently focused on colorectal and related cancers, lung cancer and rare cancers and may not yet be able to provide effective screening for patients with certain cancers.

If a patient is eligible, they do not need to visit the company, but only send the sequencing data derived from a biopsy and blood. We use the data to determine the mutations that act as drivers for tumor mutation and growth (usually 7–8, but can be as many as 15–20).

Once screening is completed, the identified drug cocktails are discussed with the patient’s oncologist, and the optimal drug combination is selected based on the patient’s overall medical condition. The choice may also depend on the location of the patient; the prices of drugs vary in different regions/countries, and some cocktails may be cost-prohibitive.

Just the Beginning

Since opening our doors in August, we have closed an initial funding round and are preparing for a formal second round in September. We have also formed several research collaborations, established a foundation to help patients cover the cost of our screening services and initiated discussions with the first patients to have their biopsies sequenced.

For colorectal cancer, we have established a database of drugs based on previous results that, using AI, enables us to eliminate non-relevant drugs from the screening process, reducing the time and cost of the service. As the database expands, we will hopefully reach the point where the avatars are no longer needed. Currently, we know that five specific drug cocktails work for approximately 75% of colorectal cancer patients.                                                                                                                             

For each tumor type, at least 200 patients are required to develop a commercial screening service. Going forward, we envision developing TuMatch solutions for all types of tumors and ultimately other genetic diseases. We are also interested in exploring the impact of comorbidities on cancer treatment and are already modeling diabetic fruit flies with cancer to determine how this disease might affect drug recommendations. Finally, we are using our existing data to gain an understanding of the mechanisms of action behind the effective drug combinations we have identified to apply this knowledge to new drug design and development.

References

1. Bangi, Erdem et al. “A personalized platform identifies trametinib plus zoledronate for a patient with KRAS-mutant metastatic colorectal cancer.” Science Advances.  22 May 2019. Web.