FDA Proposes New Databases to Monitor CAR T-Cell Safety Across INDs

The US Food and Drug Administration (FDA) wants to create two new databases that will allow it to look at safety and manufacturing information across multiple applications for products within a promising class of cancer immunotherapies called anti-CD19 CAR modified T-cells.

FDA officials presented their proposal to pilot the databases last week during a meeting of the Recombinant DNA Advisory Committee (RAC), which is organized by the National Institutes of Health's (NIH) Office of Biotechnology Activities, to oversee the clinical development of gene therapies.

During the meeting, Maura O'Leary, a medical officer in FDA's Office of Cellular, Tissue and Gene Therapies at the Center for Biologics Evaluation and Research (CBER), explained that by looking at safety and chemistry, manufacturing and controls (CMC) data across multiple investigational new drug applications (INDs), FDA could "establish a safety profile for [anti-CD19 CAR T-cell therapies]."

"Within CBER, the databases discussed at the meeting are unique for their ability to enable the evaluation of cross-IND safety issues," FDA spokesperson Andrea Fisher told Focus.

Through the pilot, FDA hopes to evaluate the feasibility of these databases, which could serve as a model for other types of products being reviewed by FDA.

What are CAR T-Cells?

CAR T-cell therapies work by genetically engineering a patient's own T-cells, a type of white blood cell, to produce what are called chimeric antigen receptors (CAR) that are able to stick to tumor cells based on the presence of certain proteins.

Many CAR T-cell therapies being developed today specifically target the CD19 (cluster of differentiation 19) antigen, as it is found on the surface of most B-cells, making it an attractive target for therapies to treat various B-cell lymphomas.

According to O'Leary, anti-CD19 CAR T-cells have been "widely reported in peer reviewed journals to produce remarkable remissions in refractory acute lymphoblastic leukemia in both children and adults," she said, citing recent a study published in the New England Journal of Medicine.

O'Leary also referred to research by NIH showing anti-CD19 CAR T-cells promise for treating large B-cell and indolent B-cell lymphomas.

Safety Concerns

Despite their promise as a potential cancer cure, there are serious safety concerns related to CAR T-cell therapies that FDA hopes to better understand.

In her presentation, O'Leary said that FDA's "main safety concern" regarding these therapies is a potentially life-threatening complication called cytokine release syndrome (CRS). Juno Therapeutics of Seattle, Kite Pharma of Santa Monica, Calif. and Novartis are all developing CAR T-based therapies and have disclosed the occurrence of cytokine release syndrome in some trial patients.

"While experiencing cytokine release syndrome, patients often develop renal, pulmonary and cardiac complications [that] linger after the initial phase of the cytokine release syndrome [and] deaths have been reported both in the acute phase, as well as weeks later," O'Leary said.

Additionally, CAR T-cell therapies have been associated with neurologic toxicity and an increased risk for infection.

CAR T-Cell Landscape

Anti-CD19 CAR T-cells were chosen for the pilot "based on the number of INDs available, as well as the … preliminary evidence that the products have potential for substantial benefits and substantial risks," O'Leary said, adding that the complexity of these products, both in terms of activity within the body and in terms of their manufacture, may be relevant to some of the known safety issues for CAR T-cell therapies.

FDA has so far received 105 INDs for genetically engineered T-cell therapies, 36 of which are specifically for anti-CD19 CAR T-cells from more than 15 different sponsors.

Furthermore, O'Leary said the agency has data on at least 275 patients that have been treated with anti-CD19 CAR T-cell therapies, who in total have received more than 500 individual doses.

Pilot Project: Safety and CMC Databases

One of the main reasons FDA wants to build these databases is that the sample sizes of studies supporting individual INDs for anti-CD19 CAR T-cell therapies are too small to paint a complete picture of their safety characteristics.

However, by building a central database for safety and CMC across all INDs within the class, O'Leary said that FDA will be able to "build risk-prediction and risk-mitigation models [to] advise sponsors on these safety issues."

Through the database, FDA hopes to answer some lingering questions about CAR T-cell therapies, such as how dose, CAR T-cell levels and cytokine levels might affect patient outcomes and adverse events.

"Are there interactions between demographic features [and] disease characteristics … that predispose to a specific adverse event? Do concomitant treatments, such as tocilizumab, etanercept or corticosteroids influence … the characteristics of the adverse event?" O'Leary asked.

The database would also be used to investigate the correlation between cytokine release syndrome and the neurological events that have been reported with CAR T-cell therapies.

"The FDA is in a unique position to conduct this analysis in that we are able to collect data from across sponsors, and therefore use the totality of data to make data-driven regulatory decisions," said Kimberly Shultz, a commissioner's fellow at CBER who helped develop the pilot.

On the other hand, Shultz said, "IND sponsors are often unwilling to share information on their products with one another. There are no data sharing limitations within the FDA. We recognize that the data collected are proprietary and we will maintain confidentiality."

FDA already has much of the data it needs to populate the databases, but Shultz said the agency will be asking sponsors for additional data to fill in some gaps.

While participation in the pilot is entirely voluntary, both O'Leary and Shultz said they've received strong support from sponsors.

"We've heard from quite a few sponsors that they're welcoming this analysis," Shultz said. "A better understanding between product attributes and clinical outcomes will facilitate development of [anti-CD19 CAR T-cells] as a safe and effective therapy."

CMC Specific Issues

According to Shultz, the reason FDA wants to look at cross-IND CMC data is to identify certain steps in the manufacturing process that might significantly influence a product's safety.

"We're especially interested in understanding whether we can predict the safety outcomes by profiling the incoming apheresis characteristics and how vector lot characteristics contribute to the final product," Shultz said.

"The manufacturing process for [anti-CD19 CAR T-cells] is a multi-step, complex process, which starts with a patient specific, heterogeneous apheresis product which comes in contact with many biologically active reagents throughout the process," she added.

Currently, many companies developing anti-CD19 CAR T-cell therapies are transitioning from manufacturing processes developed in academia to commercial manufacturing processes. During this transition, Shultz said, many sponsors will be refining their manufacturing controls and more closely defining their product attributes such as potency.

"We've taken care to design our database so that we can capture these intricacies [and] capture process changes that have occurred throughout the lifecycle of the IND," she said.

Confidentiality and Data Sharing

During their presentations, both O'Leary and Shultz emphasized that the information gathered in these databases would remain confidential within FDA.

However, Mildred Cho, associate director of the Stanford Center for Biomedical Ethics, asked whether FDA would share or present their analyses or aggregate data from the pilot.

O'Leary responded that the agency hasn't gotten that far in their planning yet, but said that sharing of high-level data would not be "unrealistic."

Keith Wonnacott, director of regulatory affairs at Novartis, responded by asking how FDA would protect confidential information if the agency plans to share data from the pilot.

In response, Shultz said she imagines the data would represent trends or ranges of information rather than individual data points that could be tied back to the companies that provided them, thus protecting individual sponsors' commercial interests.

Implementation and Timeline

FDA plans to roll out the pilot in three overlapping phases, the first of which is already underway.

O'Leary says the pilot will be "user-friendly" for both FDA staff and sponsors, as it relies on an existing database system, called HIVE (High performance Integrated Virtual Environment), and will use standardized data formats that sponsors are already familiar with.

In terms of a timeline for the project, O'Leary says that FDA already has some of the data it needs, and expects it will get enough additional data from sponsors to be able to at least perform some preliminary analyses in the next six to nine months.

Recombinant DNA Advisory Committee Day 2 (Pilot presentation begins at 2:38:10)