Making Safe and Effective CAR T Cells: How Droplet Digital PCR Can Improve Their Quality Control

Ping Jin

December 17, 2020

9 Min Read

18-11-12-Jin-P1-300x200.jpgChimeric antigen receptor (CAR) T cells first entered US clinics in 2017 (1), and this therapeutic modality holds tremendous potential as one of the most effective forms of personalized cancer care ever to reach patients. The revolutionary impact of CAR T-cell therapy comes from its ability to rewire our own immune defenses to kill cancer cells: It essentially modifies a patient’s naturally existing immune cells to boost their recognition and attack of cancer cells so that the person’s own immune system can fight the disease (2).

Although CAR T-cell therapy is incredibly promising, production of CAR T cells presents unique challenges because these cells can’t be treated as a typical drug. They are living cells with unique biology that is not completely under our control — hence the term “living drug.” Consequently, the process of manufacturing CAR T cells requires additional oversight and stringent control to ensure that the cells will be safe and effective.

Minimizing the Risk of Harm to Patients
The delicate process of CAR T cell manufacturing begins with extracting T cells from a patient’s blood. CAR genes are delivered to those cells using viral vectors, transposon systems, or direct messenger RNA (mRNA) transduction (3). The resulting CAR protein ultimately gets expressed on cell surfaces. Finally, the CAR T cells are amplified (multiplied) in a bioreactor before they are ready to be infused back into the same patient.

Two CAR T-cell therapies are approved already (4), and many more are undergoing clinical evaluation. More than 400 active or recruiting CAR T-cell therapy trials currently are registered online at (5). In fact, the global CAR T-cell therapy market could reach US$8 billion in valuation by 2028, according to one forecast (6).

Quality Control: Because T cells are alive, CAR T-cell manufacturers rely on a biological process to construct them. Therefore, developers must perform several quality control (QC) steps to ensure that these cell therapies will be functional and safe.
First, during the transfection process, CAR genes will integrate into T-cell genomes at random locations. That process must be monitored closely. If a transgene integrates into a noncoding region of the genome, it never will get expressed. The gene also could interfere with oncogenes and promote tumor growth.

Furthermore, manufacturers cannot control the transgene copy number precisely in each T cell — a variable that also affects cell function. The US Food and Drug Administration (FDA) recommends that for cell therapies using retroviral or lentiviral vectors, fewer than five copies of the transgene should be copied per genome to mitigate clinical risks derived from insertional mutagenesis (7). Higher copy numbers can cause CAR T cells to become toxic to patients. In high enough quantities, CAR proteins can induce a systemic inflammatory response (called cytokine release syndrome or cytokine storm) that can lead to organ damage and death. Therefore, to ensure patient safety and drug efficacy, it is critical for CAR T-cell manufacturers to monitor CAR gene copy numbers and the locations where transgenes integrate into patients’ T cells.

Analytical Methods: Scientists typically use quantitative polymerase chain reaction (qPCR) to enumerate CAR gene integrations. However, qPCR generates quantification results in relative terms, requiring a standard curve against which to compare. To generate that standard curve, a biomanufacturer needs to perform serial dilutions of a standard sample, which is a time-consuming, complex process that often becomes the root of experimental variability. In addition, qPCR cannot detect integrations down to a single copy per cell; its sensitivity is limited to at least two or three copies per cell.

A more sensitive method for quantifying CAR gene copies is Droplet Digital PCR (ddPCR) analysis from Bio-Rad Laboratories. Both qPCR and ddPCR methods function on similar principles using standard reagents, primers, and probes. However, whereas qPCR requires performing only one PCR reaction per tube, ddPCR technology involves partitioning a sample into tens of thousands of nanoliter-sized droplets, with PCR amplification taking place independently in each one. If a droplet contains the target sequence (in this case, the CAR transgene), then it will get amplified, and a fluorescent probe will emit a signal that can be detected. That enables collection of yes-or-no results and ultimately more precise quantification than qPCR can provide. By calculating the percentage of droplets that contain the transgene, a scientist can determine the concentration of that gene in the sample — and from that derive the average copy number of the gene in a T-cell genome.

Unlike qPCR, ddPCR requires no standard curve. It quantifies target nucleic acid sequences directly to provide an absolute measurement. Also, ddPCR can detect as little as one copy per cell, enabling biomanufacturers to determine definitively whether or not the transgene has integrated successfully into T-cell genomes.

Case Study
My group at the National Institutes of Health (NIH) manufactures CAR T cells for the NIH clinical center. Last year, we showed that ddPCR could track and quantify CAR gene sequences effectively in transduced T cells (8). We used either lentiviral or retroviral vectors to transduce the cells, then cultured them for one to three weeks. After isolating and purifying DNA from several clinical CAR-T products, we used flow cytometry to measure transfection efficiency and used ddPCR to quantify the number of transgene copies per genome.

First, we found that ddPCR detected the same CAR copy number whether a sample was tested immediately or after being frozen for three or six weeks. Also, three different technicians and two different laboratories were able to generate the same results for the same samples, indicating minimal intralaboratory and interlaboratory variability for the method.

Previous data tell us that lentiviruses show a 30–80% transfection efficiency (9), and our data fell within that range. We used flow cytometry and ddPCR to test the impacts of multiplicity of infection (MoI) — the ratio of viral vectors to T cells — and of centrifugation on transfection efficiency and copy number, respectively. We found that when samples weren’t centrifuged, transfection efficiency and copy number per cell increased along with MoI. Centrifugation helped improve transfection efficiency at lower MoIs, providing for copy-number stability.

Transgenes have preferred integration sites. We used next-generation sequencing (NGS) to locate those sites and to map the transgene integrations. In the future, we plan to design a ddPCR probe that will be specific to the CAR transgene. That will enable us to track such integrations in engineered T cells.

Our data have showed that during CAR T-cell manufacturing (wherein transfection efficiency and CAR copy number must be controlled tightly), ddPCR can be a reliable tool to monitor transgene copy number in clinical CAR-T products across different laboratories, technicians, and time points. We also demonstrated that this analytical method serves as a complement to flow cytometry because ddPCR can quantify copy number and help us elucidate its relationship with both MoI and transfection efficiency. By similarly incorporating ddPCR into the CAR T-cell manufacturing process, other scientists can develop efficient manufacturing protocols with increased confidence about the safety and effectiveness of their products before they reach patients.

Other QC Applications
Apart from the uses described herein, ddPCR also can be useful in other areas of CAR T-cell manufacturing and safety testing. These genetically engineered cells can multiply over time, and their activity can change after treatment has ended. Although such persistence can provide durable benefits for patients, it also raises concerns about the safety of the cells — and ddPCR can help address such questions.

For instance, the FDA is concerned that CAR T cells might be able to replicate in patients and turn into rapidly progressing T-cell neoplasms. That phenomenon has been seen only in laboratory animals so far, but to be safe, the agency recommends that scientists test clinical vector lots, manufactured cell products, and the blood of patients who have been treated with CAR T cells (10). We believe that ddPCR could be used to detect replication-competent viruses in CAR T cells so that infected cells could be screened out. The method also could detect other sources of microbial contamination that might affect patients.

Finally, ddPCR could be used to measure CAR T-cell persistence in patients after treatment. The cells are supposed to survive in a person’s body for a few months. If they don’t survive long enough, then they might not be effective; if they last too long, then they could produce adverse side effects in patients long after their cancers have gone into remission (11). Pathologists can use ddPCR to monitor CAR T-cell counts regularly in patients’ blood and track the cells’ persistence over time.

We have demonstrated just one way that ddPCR can assist in CAR T-cell manufacturing: ensuring that the T-cells contain the right number of CAR gene copies to make a therapy both safe and effective. We predict that ddPCR one day will enter more clinical laboratories, help improve many aspects of CAR T-cell therapy, and help it live up to its promise in the future of personalized cancer care.

1 FDA Approval Brings First Gene Therapy to the United States. US Food and Drug Administration: Rockville, MD, 30 August 2017;

2 CAR-T Cell Therapy. MD Anderson Cancer Center: San Antonio, TX, 2020;

3 Li Y, et al. Quality Control and Nonclinical Research on CAR-T Cell Products: General Principles and Key Issues. Engineering 5(1) 2019: 122–131;

4 Brodsky, AN. The Promise of CAR-T Cell Therapy in 2019 and Beyond. CRI Blog 18 September 2019;

5 Search: “CAR-T cell” — Recruiting, Active, Not Recruiting Studies — List Results. 24 April 2020;

6 Global CAR-T Cell Therapy Market to be Worth US$8 Billion by 2028 — Coherent Market Insights. GlobeNewswire 31 May 2018;

7 Zhao Y, Stepto H, Schneider C. Development of the First World Health Organization Lentiviral Vector Standard: Toward the Production Control and Standardization of Lentivirus-Based Gene Therapy Products. Hum. Gene Ther. Meth. 28(4) 2017: 205–214;

8 Lu AM, et al. Poster: Applications of Droplet Digital PCR for the Detection of Copy Number in CAR-T Products. NIH Research Festival: Bethesda, MD, 12–13 September 2018.

9 Zhang Z, et al. Optimized DNA Electroporation for Primary Human T Cell Engineering. BMC Biotechnol. 18(4) 2018:

10 CBER. Guidance for Industry: Supplemental Guidance on Testing for Replication Competent Retrovirus in Retroviral Vector Based Gene Therapy Products and During Follow-Up of Patients in Clinical Trials Using Retroviral Vectors. US Food and Drug Administration: Rockville, MD, November 2006;

11 Sadelain M. What’s Next in CAR-T Cell Therapy. Cancer Research Institute: New York, NY, 2019;

Dr. Ping Jin leads the Product Assurance and Characterization Testing (PACT) section in the Department of Transfusion Medicine’s (DTM’s) Center for Cellular Engineering at the National Institutes of Health Clinical Center, 10 Center Drive, Bethesda, MD 20814; [email protected].

Droplet Digital and ddPCR are registered trademarks of Bio-Rad Laboratories, Inc.

You May Also Like