As I was discussed before, apheresis product is the most variable raw material in CAR T-cell manufacturing. Cellular composition of this starting material is highly heterogeneous between patients, therefore, manufacturing of consistent CAR T-cell product is a big challenge. One of the most interesting questions in the field – Can T-cell composition of starting material (apheresis collection) predict clinical outcome in CAR T-cell therapy?
In attempt to answer this question, some groups have started to look at T-cell composition of starting apheresis collections, correlate it with composition of expanded CAR T-cell product and with clinical outcome. There is not much data yet available at this point, but some of it looks interesting.
Olivia Finney from Seattle Children’s Hospital shared some data (start from 12:47 min) last year on correlation between presence of T-cell subsets in apheresis and B-cell aplasia in pediatric B-ALL CD19-CAR T-cell trial. B-cell aplasia is a marker of CAR T-cell persistence and (in their trial) durability of response after infusion. Quick summary:
- The more % of CD45RA-/CD62L- T-cell effector memory (T-EM) the shorter B-cell aplasia – less likely CAR T-cells will persist.
- The same negative correlation was observed between level of PD-1/ TIM-3 expression (used as markers of T-cell exhaustion) on CD8 cells with B-cell aplasia.
She did not show any correlative data for T-cell central memory (T-CM) and naive T-cells (T-N), which unlike effector T-cell subsets, represent “early lineage” cells. Interestingly, researchers from CHOP/ Upenn recently showed that presence of “early lineage” T-cells at collection directly correlates with success of T-cell expansion ex vivo. Patients with lower “early lineage” T-cells in the blood (which translates to apheresis collection) were less likely to pass test T-cell expansion. 24% of patients (20/83) did not pass test expansion threshold and were excluded from CD19-CAR T-cell pediatric B-ALL trial. In discussion, the authors indicated that T-N and T-CM could be used as predictors of clinical response:
Thus, early lineage T cells present at time of collection not only predict in vitro expansion potential but also directly give rise to T cells that persist long term and mediate robust antitumor responses in patients.
I haven’t seen any solid data, which proves this point, but it sounds very attractive. Significance of T-CM and T-N in starting material and manufactured product as well as right combination of CD4 and CD8 cells for improvement of therapeutic potency was demonstrated in experimental work by researchers from Fred Hutchinson Cancer Center. It was a rational for using T-CM subset of CD4 and CD8 cells, directly sorted from apheresis, in their lymphoma trials.
Intriguingly, researchers from National Cancer Institute (NCI)/ Kite Pharma did not see correlation between apheresis product, CAR T-cell product composition and clinical responses in lymphoma patients (thanks to Paul Rennert for bringing this up):
Clinical responses occurred regardless of product characteristic differences such as CD4/CD8 T cell ratio and juvenile / differentiated T cells.
Since different manufacturing processes, different diseases and low number of patients were analyzed, we cannot compare Seattle Children’s (n=16) and NCI (n=14) data as of now. More data should be accumulated by different groups for solid conclusions and for confident answer to the question that I posted at the beginning. However, there are few strong hints toward significance of T-CM in starting apheresis material and final manufactured product. Some groups may be indifferent to T-cell composition in starting apheresis product, since all of them could be shifted almost any way we want via cell culture conditions. But this is subject of another post. Just to illustrate how T-cell populations could be shifted, I’d like to conclude this post by a screenshot from Crystal Mackall talk:
Please share your thoughts and experience in comments!