CD22-Targeted CAR T Cells Elicit Responses With Acceptable Safety in Heavily Pretreated LBCL


Matthew Frank, MD, PhD, discusses the rationale for evaluating CD22 CAR T cells in heavily pretreated patients with large B-cell lymphoma, important safety and efficacy data with this approach, next steps for this research, and how continued efforts may shed needed light on sequencing in this space.

Matthew Frank, MD, PhD

Matthew Frank, MD, PhD

Treatment with autologous CD22 CAR T cells was safe and induced high response rates in heavily pretreated patients with large B-cell lymphoma (LBCL) who relapsed following CD19-targeted CAR T-cell therapy, according to data from a single-institution, phase 1 trial (NCT04088890) presented at the 2023 EHA Congress.

A single infusion of these cells resulted in a best overall response rate (ORR) of 68% (n = 38) and complete response (CR) rate of 53%. Moreover, this approach demonstrated a safety profile that appears to be comparable to that observed with CD19-targeted CAR T-cell therapy. When given at dose level 1, no grade 3 or higher cytokine release syndrome (CRS) or immune effector cell–associated neurotoxicity syndrome (ICANS) was observed.

“The topline data [suggest] that this is an effective CAR for patients whose disease has relapsed after CD19 therapy, and it appears to be as safe as CARs come,” Matthew Frank, MD, PhD, explained in an interview with OncLive®. “This is a CAR that appears to have the potential for curative intent, where very few things in this space [do], when compared with bispecifics or allogeneic stem cell transplantation.”

A phase 2 clinical trial evaluating CD22 CAR T cells in patients whose disease has progressed following CD19-targeted CAR T-cell therapy is slated to open in the near future. “I look forward to seeing what [data will] look like in a phase 2 multicenter trial,” noted Frank, who is an assistant professor of Medicine in the Division of Blood and Marrow Transplantation and Cellular Therapy at Stanford University, as well as a lead investigator of clinical trials at Stanford Health Care, in Stanford, California.

In the interview, Frank discussed the rationale for evaluating CD22 CAR T cells in this patient population, important safety and efficacy data with this approach, next steps for this research, and how continued efforts may shed needed light on sequencing in this space.

OncLive: What is the rationale for investigating CD22-targeted CAR T cells in patients who already received CD19-targeted therapy?

Frank: It is known that CD19, which is a great target for CAR T therapy, unfortunately gets downregulated or lost in terms of expression. Therefore, seeking alternative antigens is important. However, there are a lot [of antigens] that [can be targeted], including CD22 and CD20.

CD22 is expressed in most B-cell malignancies, including lymphoma, where it is highly uniformly expressed. We previously saw impressive results from the National Cancer Institute using the same CD22-targeted CAR T therapy in children with leukemia. Given the patients’ high response rates, investigators asked whether we could test this [approach] in adults with leukemia or lymphoma.

What was the target patient population for this research?

We predominantly targeted patients whose disease had relapsed after CD19-directed CAR T-cell therapy in lymphoma. The same was true for leukemia, except that in leukemia, there was not an approved CAR at the time of trial launch; therefore, investigators allotted for anyone who met eligibility.

However, with lymphoma, we were specifically targeting those who had relapsed after CD19-targeted CAR T-cell therapy. We also allowed [enrollment of] patients who were CD19 negative and had no expression. Notably, there was just 1 patient [who met these] criteria. The other patients had [received] prior CAR therapy, and that was predominantly who we were recruiting.

What efficacy data were reported with this approach?

[At the 2023 EHA Congress,] we presented data from the completed phase 1 trial for the first time. This [was] the first time [we shared] the full dataset, with every patient hitting the primary key efficacy end point, which is 3 months. Every patient has had 6 months of follow-up at this point [and have undergone] assessment.

Overall, we’re very happy with the results. Of the 38 patients treated, 68% had a response and 53% had a CR. Overall, it’s a very nice, high response rate.

What was learned in terms of safety?

I first want to highlight how sick or heavily pretreated these patients were coming to us. We didn’t cherry pick patients. We had patients who were often [on] their final therapies. The median number of lines of [prior] therapy was 4, so this [was] often their fifth therapy. To have a better sense of tumor burden, you can look at lactate dehydrogenase [LDH] levels, and LDH was elevated in over 80% of these patients. Thirty percent [of patients] never had a CAR before, emphasizing that this is a very high-risk population.

What was impressive is, despite the high risk, patients didn’t have a lot of toxicity. Although, there are obviously safer drugs out there, for a CAR T therapy, this is relatively safe. We saw that with the recommended phase 2 dose, no patient had grade 3 or higher CRS or ICANS. In contrast with every other approved CAR, this [one] has shown much less toxicity. However, what did emerge was a syndrome called immune effector cell–associated hemophagocytic lymphohistiocytosis–like syndrome [IEC-HS], which required collaborative treatment between colleagues who had seen this.

IEC-HS is also called variant CRS or CAR HLH. This syndrome [includes] very high ferritin levels; low blood counts; coagulopathies, often from fibrinogen being very low; and liver abnormalities. Five patients on this trial required treatment for that toxicity. At the higher dose level, 33% of patients had that toxicity and 1 patient died as a complication of it, namely from an infection. However, when lowering the dose, we seemed to find the sweet spot in terms of dosing.

The PFS and OS curves overlay, the response rates were not statistically different, and the CR rates were nearly identical. [As such,] we didn’t seem to lose a lot of efficacy; we just lowered the toxicity to a safe level, which is always important. That’s the point of a phase 1 trial: to identify that recommended phase 2 dose, the main end point of the trial, [and] we achieved [that].

What are the next steps for this research?

This has now been licensed by a company that will open a phase 2 multicenter trial. It’s important to get this [approach] to many more patients, [and] to get true efficacy data. Although we try not to cherry pick, we certainly want to have a multicenter experience to understand what [treatment with this CAR] looks like across all demographics, at all different locations. Therefore, we are excited to get that going.

We’re also taking the same CAR and evaluating this therapy in other lymphoma subtypes. Overall, we will be opening a trial in follicular lymphoma, mantle cell lymphoma, and an assortment of other lymphoma subtypes.

As these agents are developed, how is sequencing going to be handled?

This is a question [that still needs] to be addressed in the field. It’s still early days, and we actually don’t know the answer. We know that antigen loss is a problem and it emerged in the CD22 trial, as we saw antigen being downregulated in patients who relapsed as well as [in] those who had very low antigen or no expression of CD22.

One [way to address this] could be to simply look at antigen levels, and if there is no CD19, but you have CD20 and CD22, maybe [we] should go after those antigens. Some [approaches] may be obvious, although [investigators also] thought it was obvious for CD19 CARs, but immunohistochemistry [IHC] is not very good.

Based on data from the phase 3 ZUMA-1 trial [NCT02348216], it is known that patients whose disease had CD19 expression by IHC had the same response rates and durable remission rates as everyone else. Although there are sometimes limitations in technology, once technology is optimized, maybe using flow-based methods, I think antigens are the way to go.

Moreover, clinicians may want to use combinations. We have a lot of emerging [agents targeting] CD19 and CD20, and there are even triplets now. [We] don’t know if putting everything in 1 cell is the right way to go, or whether we need to have different cells. This question will have to be addressed CAR by CAR. Certainly, we want to be mindful of resources and of toxicity. Sequential CARs have also been shown to be feasible and efficacious, as patients are cured by [this approach]. Co-infusion may be worth investigating in the future, as well.

We also want to identify earlier those who will not benefit from chemotherapy and maybe even in [the] frontline, we [can] bring combinations of CARs. We will also study mechanisms of resistance that may be more universal, emerging mechanisms with biological properties. [Lastly,] we may also want to identify something called CD58 that prevents the CAR T cell from being stimulated, or design better CARs that don’t depend on CD58 signaling to activate; we have CARs in development for that, as well.

You mentioned that 1 patient in this trial didn’t have CD22 expression. As you go forward, will you continue to include those cases?

I believe that we should find [these things] out experimentally. On this trial, we often didn’t know about CD22 expression. Notably, I have told patients who had very low expression, this is maybe not the best trial for you. I try to be up front about what we know and what we don’t know. However, I say if [a patient is] eligible, [they should] go on the trial, as long as I don’t think that it will substantially harm [them] or delay [them] from getting the care [they] want, as I could be wrong. I think we must be open minded, get the data, and then use the data to guide us.


Frank MJ, Baird J, Kramer A, et al. CD22 CAR T-cell therapy is safe and effective in patients with large B-cell lymphoma who have relapsed after CD19 CAR T-cell therapy. Presented at: 2023 EHA Congress; June 8-11, 2023; Frankfurt, Germany. Abstract S230.

Related Videos
Michael Hurwitz, MD, PhD
Adnan F. Danish, MD
Adrienne Phillips, MD, MPH
Michael Hurwitz, MD, PhD
Lori A. Leslie, MD
Adrienne Phillips, MD, MPH
Md Kamrul Hasan, PhD
Mark Leick, MD
Mark Leick, MD
Hemant S. Murthy, MD