CISH-Targeted TILs Show Promise in GI Cancer Trial

Medscape

18 hours ago

This transcript has been edited for clarity.
Hello. I'm David Kerr, professor of cancer medicine at University of Oxford. I'd like to talk a little about a phase 1 trial that has just been published in the May edition of Lancet Oncology , a beautiful paper by Emil Lou and colleagues, in which they all looked at an extraordinarily complicated regimen of gene editing and then reinfusing tumor-infiltrating lymphocytes (TILs). This just shows how remarkable modern cancer medicine can be.
I'd like to talk a little about the procedures involved, given the route of complexity. Initially, for patients in the phase 1 trial with metastatic advanced gastrointestinal cancer, the first step was to receive a non-myeloablative lymphocyte-depletion chemotherapy. This is cyclophosphamide and fludarabine. This is then followed by high-dose interleukin-2.
The process of producing the gene-edited TILs is absolutely fascinating. The gene-editing target, the gene that we wanted to knock down, was a novel, internal immune checkpoint protein gene, CISH , which encoded a cytokine-inducible SH2 domain-containing protein. This is an internal immune checkpoint target and quite an interestingly novel one. That's the target to knock down.
The process of building the therapy is completely fascinating. Autologous TILs were generated from tumor biopsy fragments from each individual patient, and samples of the TILs were then cultured with patient-derived lymphoblastic cell lines, ormonocyte-derived dendritic cells loaded with pools of synthetic 25-mer peptides containing tumor-specific mutations that were detected from whole-exome sequencing from the patient's tumor.
Think about that for a second. The patients undergo biopsy. The TILs are taken off for culture — more on that later — but they were co-cultured with a source of patient-specific tumor-associated antigens, which were generated by whole-exome sequencing. Then, using 25-mer fragments of what they felt were the key tumor-specific antigens, were plumbed into the patients' own dendritic cells so that these antigens would be presented to at the time of co-culture with a patient, so in TILs.
It boggles the mind. The TILs with demonstrated reactivity to the neoantigens in this co-culture experiment were selected and they were then subjected to the CRISPR gene editing in vitro. They knocked down CISH , the target that they were aiming for.
Then the edited TILs underwent a rapid expansion protocol. They were cryopreserved, and then via some very sophisticated molecular quality control, infused back into the patients. Goodness gracious — I mean, it's extraordinary when you think about it.
It was a phase 1 trial, so dose-escalating the number of cells. It's pretty well tolerated. As you would expect, there was some fatigue and some fever. Nobody died in the back of any of the treatment that was given in that way.
In terms of effectiveness, these were patients with advanced disease who had undergone multiple previous lines of treatment. There were no severe cytokine-release syndromes, nothing of grade 3 or worse. No neurotoxicity. Six of 12 patients had stable disease by day 28. Four (33%) had stable disease ongoing at 56 days.
One young adult who had microsatellite-unstable and mismatch repair-deficient MSI-high tumors — therefore they've already got an existing high neoantigen load — had a complete response. That was very pleasing.
I say, again, this is an extraordinary piece of work, to think about the complexity involved in every step of that process — before the patients' own cells, the autologous cells, were manipulated and reinfused with moderately acceptable toxicity, I would say.
It's a phase 1 trial, so you're not really looking for a big efficacy readout, but the one younger patient who had, if you like, a genetic predisposition to responding to immunotherapy anyway, had a complete response.
I'd be really interested in what you think about it. Who knows what the cost of that would be in terms of the complexity — I keep using that word, don't I? — of every single step. Modern cancer medicine — don't you love it?
Well done to the team for producing this phase 1 trial result. How generalizable it will be remains to be seen, given the multiple different steps that are required. It just shows you how, in my lifetime as a cancer doctor, four decades, remarkable progress has been.
I'm very interested in any comments you'd have to make. As always, thanks for listening. For the time being, Medscapers, ahoy, and over and out.