In the race for the most potent CAR T-cell therapy, there is a big interest to the issue of purity and composition of the final cell product. In this post, I’ll try to summarize the current knowledge about defined CAR T-cell products, based on two clinical studies, published this week.
What is defined product and why it is important?
We can roughly divide CAR T-cell products on “bulk T-cell” or “crude” and “defined composition”. Crude CART products are not purified and contain: different ratios of CD4/CD8 T-cells and their subsets, other than T- contaminating cells and non-CAR cells. Defined composition products could be the following:
- purified CD4+ or CD8+ T-cells
- defined ratio CD4/CD8 in final product
- CD25-depleted T-cell products
- purified subsets of CD4 or CD8 T-cells (naive-, central memory-, effector- T-cells)
- purified CAR+ cells
There are few good reasons for development of defined CART products:
- Reproducible and well defined process allows manufacturing of uniform product from highly variable donor material
- Significant decrease of targeted cell dose of defined product allows to shorten the time and cost of manufacturing
- Similar defined product allows easy data comparison between manufacturing processes, clinical trial protocols and across different sites
- Precise prediction and mitigation of CAR T-cell therapy-associate toxicity
- Potential superior therapeutic efficacy
The later is single most important reason, which begs the question – Will defined CAR T-cell products deliver superior therapeutic benefit? As of today, vast majority of CART developers manufacture crude cell products. Manufacturing process usually does not include sorting of T-cells on day 0 or purification of T-cells/ CAR+ cells in-process or on a harvest day. Most of developers release final CAR T-cell product with CD3+ cells >80-90%, highly variable CD4/CD8 ratio and % of CAR+ cells.
At least one group of researchers have done a lot of experimental and clinical work on defined composition of CAR T-cell products. Defined product/ process development has been done by Michael Jensen lab, initially at City of Hope and then further adapted and modified at Fred Hutchinson Cancer Center and Seattle Children’s Hospital. Stanly Riddell’s lab from Hutch did a lot of experimental work to demonstrate potential value of defined CAR T-cell product. All together they have tried all versions of defined composition CAR T-cell products, mentioned above. One of recent and the most comprehensive studies by Riddell’s lab, showed that (1) combination of both CD4 and CD8 T-cells has superior therapeutic potency and (2) naive CD4 cells and central memory (T-CM) CD8 T-cell subsets were the most potent in vivo. Long-term persistence of human memory T-cells was demonstrated by Riddell and Jensen earlier in mouse and primate models. Importance of CD4+CD8 combo rather than use of potent cytotoxic CD8+ cells alone was also demonstrated in numerous studies (check here, here and here). Therefore, experimental studies created a strong rational for favoring T-CM, naive T-cells (T-N) and CD4+CD8 combo in adoptive cell therapy trials.
Feasibility of manufacturing defined T-CM CAR T-cell product
Before I get to the first clinical results, I’d like to look at manufacturing process of defined CAR T-cell product. Jensen started with purified CD8 T-CM or CD4 + CD8 T-CM manufacturing process, which described in details here. It includes sequential 2-step CliniMACS procedure for negative selection of CD14/CD45RA/CD4 or CD14/CD45RA-negative cells, positive selection for CD62L (marker of T-CM) and culture with IL2/IL15 for up to 30 days in bags. After two CliniMACS procedures, cell recovery was as low as 0.4% (in contrast to expected 1.4%). If input cell number was usually 5 billions PBMCs, average output cell number before starting a culture was ~19 millions. T-CM recovery efficiency was 26%. Even though, they typically started culture from 7-15 millions of CD8 T-CM cells, generation of ~3 billions of cells in 3-6 weeks was feasible. This manufacturing process was used in 2 clinical trials: NHL1 and NHL2.
In the modified manufacturing process, designed for NCT01865617 trial, included CliniMACS selection of CD4+ bulk population and 2-step CD8 T-CM (see above) or CD8 bulk selection with 2 parallel 15-20 days cultures and mixing CD4:CD8 as 1:1 before infusion. Importantly, CAR+ cells were selected before infusion by a marker (EGFRt). Interestingly, either CD8 T-CM purified on day 0 or CD8 bulk cells yielded only ~40-50% of CD8+/CAR+ cells with T-CM phenotype (CD45RA-CD62L+).
Now, feasibility of manufacturing in NCT01865617. 16/30 (53%) patients have passed threshold of 20 T-CM cells/ ul in screening assay for feasibility of manufacturing. From selected products, T-CM were successfully manufactured in all, but 1 cases. 3 out of 30 infused products were not formulated as 1:1 (10%), due to lack of expansion.
I’d summarize some of my thoughts of defined CAR T-cell product manufacturing feasibility as the following:
- CD4 and CD8 bulk selection is feasible and working well.
- Due to scarcity of T-CM in starting apheresis material, depending on targeted dose, current methods of magnetic cell selection will be feasible only in about half of the patient population. As Jensen mentioned on MSKCC CART meeting this year, population of T-CM in pediatric apheresis product is exceedingly rare. It could make the whole T-CM cell isolation exercise in pediatrics unfeasible. They may look at T-N instead.
- Sequential 2-step CliniMACS procedure results in low recoveries of total PBMCs and CD8 T-CM. Having only 10-20 millions of T-CM cells at start of culture is very risky.
- Taking into account the same % of T-CM in final CAR+ product, irrespective of starting condition (CD8 bulk versus CD8 T-CM selected), it is not clear to me what is the benefit of CliniMACS selection on day 0.
- The current culture conditions do not result in uniform defined product, in terms of T-cell subsets. It is still highly variable (from patient-to-patient) mix of T-CM, T-N and effector T-cells. Unless culture conditions are improved, highly heterogeneous final product defeat the purpose of T-CM selection on day 0. Interestingly, JCI study only mentions IL2 in culture, even though during the recent MSKCC CART meeting, Rebecca Gardner from Sealltle’s Childrens said that they use IL7+IL15 for CD4 and IL2+IL15 for CD8.
- As of now, the manufacturing process does not allow to save time and money. Even though, targeted dose is much lower than in crude CART products protocols, it still requires 15-20 days of culture and very cumbersome expensive CliniMACS procedures. Thus, day 0 CliniMACS requires 4 vials of Ab reagent ($2-3k each), 2 disposable kits ($2-3k), buffers ($0.5k). You can make it shorter and cheaper, using Prodigy instead of CliniMACS Plus, but not much.
Clinical outcomes of using defined CAR T-cell products
Results of 3 clinical trials (NCT01318317, NCT01815749, NCT01318317), using defined CAR T-cell products have been published so far (here and here). However, the therapeutic benefit of using defined versus crude CAR T-cell product remains unclear. Ideally, defined CAR T-cells should be compared with crude product within one trial settings, because even for the same conditions, clinical protocols are very very different between sites. Also, ideally, CD8 bulk vs. CD T-CM or CD8 alone vs. CD8+CD4 combo should be compared within one trial in exactly the same settings. Unfortunately, none of these “ideal comparison conditions” were met in 3 published trials, mentioned above.
Even though, it seem like CD4 + CD8 T-CM combo performed better in NHL2 trial (75% progression-free survival at 1-year) than CD8 T-CM alone in NHL1 (50% progression-free survival at 1-year), the difference is not significant, due to low number of patients (n=8 in each trial). On top of it, different CAR vectors were used between these trials, culture duration was shorter and CD25+ T-regs were depleted in NHL2 trial manufacturing protocol. So, data cannot be compared. If we look at results of other CAR T-cell lymphoma trials (narrowing to DLBCL), City of Hope results are not much better than reported from other centers (for example, from Penn). CD8 T-CM persistence was not beneficial, compare to data from other centers. Two excerpts from the study, which demonstrate that assessment of defined CAR T-cell product benefit is impossible:
CD19-CAR T cell activity is difficult to assess by disease response, since 9/16 patients were in CR at start of study, and HSCT can also produce CRs.
Thus, a T cell product derived from central-memory enrichment as described in these studies, does not persist longer than what is observed in trials with conventional bulk T cells transduced with CARs bearing CD28 co-stimulatory domains.
Now, moving to B-ALL study, published this week in JCI. First of all, B-ALL is not the best condition to assess a difference by clinical outcome between crude and defined CAR T-cell products, because response rate is too high (close to 90%) across the centers no matter what. The authors about outcome:
The 93% remission rate by flow cytometry and 86% MRD-negative CR rate in our study compares very favorably to that reported by others in which CAR–T cells of undefined composition were manufactured using CD19 CARs that incorporate either a 4-1BB costimulatory domain (children and young adults, 79%) or a CD28 costimulatory domain (adults, 75%; children and young adults, 60%) (1-4)
What about persistence? Theoretically, based on experimental work, CD8 T-CM should have superior long-term persistence. But it was not the case. The study showed persistence only at 1 month time point. What about relapses? Maybe application of T-CM will reduce the rate of relapses? No, 9/30 patients in the study relapsed, half of them (5/9) received CD8 T-CM product. Clinical outcome was significantly improved in the study after implementation of different conditioning regiment (with fludarabine). This change significantly complicates and even make impossible data comparison between CD8 bulk and CD8 T-CM groups:
The high overall rate of BM remission of 93% by flow cytometry in this study and differences in lymphodepletion regimens and infused cell doses do not allow comparison of the efficacy of CAR–T cell products manufactured from CD8+ TCM cells or from bulk CD8+ T cells. Analysis of differences in long-term persistence of cell products that were selected for CD8+ TCM or bulk CD8+ T cells in our study was further complicated by our findings that immune-mediated rejection of CAR–T cells occurs in some patients, which may provide an explanation for the loss of CAR–T cells observed in a subset of patients in other studies…
To conclude: Despite the strong experimental evidence and very attractive idea behind of defined CAR T-cell products, it is too early to conclude about their therapeutic benefit and superior potency. With greater number of patients and technical improvements in manufacturing (more efficient clinical cell sorting, IL7+IL15 in culture and other), potential benefit of defined CAR T-cell product may become more obvious. Such benefits as dropping a therapeutic dose, better correlation between dose and in vivo expansion dynamics, decreasing donor variability in manufacturing, we can see today already.