Chimeric antigen receptor (CAR) T-cells are the culmination of multiple major advances in science, from tumor sequencing to genetic engineering and immunotherapy. Unsurprisingly, they were one of the most anticipated oncological approaches since immunotherapy first entered the clinic. The approach exploits T cells’ natural ability to seek out foreign invaders in the body and destroy them, with an added boost of selectivity through targeted genetic engineering. The U.S. Food and Drug Administration has approved two CAR-T therapies already and have their eye on many more. There are over 350 active or recruiting CAR T-cell trials on ClinicalTrials.gov. According to one forecast, the global CAR-T cell therapy market could reach $8 billion by 2028.
With just a few years of commercial sales under their belts, CAR T-cell therapy manufacturers are still ironing out the kinks. They are not typical drugs and producing them presents unique challenges. CAR-T cells are living cells, which means manufacturers, in part, depend on the physiological workings of the cells to integrate and express the CAR gene in the appropriate quantity. Since this process is not completely under a manufacturer’s control, the process of generating safe and effective CAR-T cells requires meticulous quality control.
Reducing the risk of harm to patients
The process of manufacturing CAR-T cells begins with extracting T cells from a patient’s blood. A scientist then uses either viral vectors, transposon systems, or direct mRNA transduction to insert the CAR gene into the DNA of the cells. This allows them to express the CAR protein on their surface. Finally, a scientist amplifies the engineered cells in a bioreactor, and a physician infuses a patient with their own modified T cells.Manufacturers need to monitor this process to ensure the cells are functioning appropriately and won’t cause harm to patients.
One challenge presents itself during the transduction step: Once the CAR gene is delivered to the cell, the manufacturer cannot control where in the cell’s genome it will integrate. It could embed itself in a stretch of DNA that regulates oncogene expression, thereby promoting tumor growth. Or, it could integrate into a silent region of the genome and never get expressed.
In addition to location, it can be challenging to control how many copies of the CAR gene integrate into the T cell’s genome. This impacts the safety, as well as the function, of the cell. If the gene doesn’t integrate at all, the cell won’t fulfill its intended purpose. Conversely, if too many copies of the gene integrate, the cell can become toxic. The protein it produces could induce a systemic inflammatory response, called cytokine release syndrome or “cytokine storm,” that can damage organs and cause death. To regulate this, The U.S. Food and Drug Administration (FDA) recommends that when manufacturers are using retroviral/lentiviral vectors, the CAR gene copy number per cell does not exceed four. Consequently, manufacturers must monitor CAR gene copy number, as well.
Today, scientists mostly rely on quantitative polymerase chain reaction (qPCR) to quantify CAR gene copy number, but this technique can only deliver relative results. qPCR cannot directly quantify gene copies; rather, a scientist needs to compare their result to a standard curve. This, in turn, needs to be generated using serial dilutions, which is a time-consuming and unreliable process. The variability inherent to qPCR makes it nearly impossible to detect genes when they are present in only one copy per cell.
In contrast, droplet digital PCR (ddPCR) can detect gene copies down to one per cell. While ddPCR and qPCR both employ the same chemistry, ddPCR quantifies gene sequences directly, which delivers far greater sensitivity. In practice, ddPCR divides a sample into tens of thousands of nanoliter-sized droplets and performs independent PCR reactions on the nucleic acid strands contained within each one. The CAR gene will get amplified and emit a fluorescent signal, lighting up the droplets where it is present, while the remaining droplets remain dark. This provides the scientist with a binary result for each droplet to precisely quantify those that contain the target sequence.From there, he or she can calculate the concentration of the gene in the sample, and in turn, derive the average gene copy number among the T-cells in that batch.
By quantifying nucleic acid sequences directly, ddPCR removes the need for standard curves. Also, since it is sensitive enough to detect down to one CAR gene copy per cell, manufacturers can tell with certainty whether or not their transfection method was successful.
At the National Institutes of Health (NIH) Clinical Center, Ping Jin, Ph.D., demonstrated that ddPCR can reliably quantify CAR gene sequences in T cells following transduction using either lentiviral or retroviral vectors. Jin and her team transduced a batch of T cells with the CAR gene and then cultured them for one to three weeks before assessing the success of the transduction protocol. They isolated and purified the genome from several CAR T cells, measured the transfection efficiency using flow cytometry, and quantified the CAR gene copy number per genome using ddPCR.
In this study, ddPCR was consistent in its ability to quantify CAR copy number. They found the same result whether the sample was tested immediately, or after three or six weeks of freezing. The results were also consistent across three different technicians and two different labs, demonstrating the high reliability of the technique.
The researchers achieved a 30-70 percent transfection efficiency, which falls within the expected range of 30-80 percent. Next, the researchers examined the impact of two variables, 1) multiplicity of infection (the ratio of viral vectors to T cells)and 2) centrifugation, on transfection efficiency using flow cytometry and on copy number, respectively.
When the samples weren’t centrifuged, transfection efficiency and copy number per cell increased along with MOI. Centrifugation enhanced the transfection efficiency at lower MOIs, ameliorating the deleterious effects of low MOI on copy number.
Finally, they mapped the locations where the CAR transgene integrated along the genome using next-generation sequencing (NGS). In the future, Jin’s group hopes to use ddPCR for this step as well. Although transfected genes integrate randomly, they show a preference for particular sites along the genome. Once these sites are determined, Jin and her team plan to design dedicated PCR probes, enabling them to detect integrations at those sites. Overall, ddPCR offers a shorter turnaround time than NGS.
Taken together, Jin’s data shows ddPCR reliably quantifies transgene copy number in clinical CAR-T products among different labs, technicians, and time points. She also showed that ddPCR can serve as a complement to flow cytometry. While the researchers used flow cytometry to measure the relationship between MOI and transfection efficiency, ddPCR enabled them to study the relationship between MOI and copy number, and subsequently, relate copy number and transfection efficiency. With a set of tools that includes ddPCR, CAR-T cell manufacturers can be more confident that their products are safe and that they will be effective in treating cancer.
Other aspects of CAR-T quality control
Beyond its role in measuring CAR gene copy number, ddPCR can also contribute to several other aspects of the CAR-T cell manufacturing process for greater safety throughout the product’s lifespan. In turns out that the “living” nature of CAR-T cells not only makes it more challenging to control copy number and transfection efficiency, it may also cause them to grow and behave differently over time. These sorts of variables could potentially harm patients.
For example, CAR-T cells could theoretically replicate within patients and promote the rapid growth of T-cell neoplasms. Researchers have not seen this happen in humans; nevertheless, the FDA recommends that manufacturers test their clinical vector lots and manufactured cell products, and that physicians test their patient’s blood after they’ve been treated. ddPCR is able to detect replication-competent viruses, which means they could be used to screen out CAR-T cells that contain them, well before they impact patients. ddPCR can also detect contaminants, such as microbes.
Another exciting avenue to explore is using ddPCR to measure the kinetics of the CAR-T cells’ persistence in treated patients over time. These cells are designed to survive for just a few months after treatment. If the cells don’t last long enough, they won’t be effective. However, if they last too long, patients can experience side effects — even after their cancer has entered remission. ddPCR can quantify CAR-T levels in the blood, allowing physicians to track CAR-T levels over time using serial blood monitoring.
The work described here only highlights one area where ddPCR can assist in CAR-T cell manufacturing, but given its versatility, it might one day assist in all aspects of the process, both upstream and downstream.