Introduction: Idelalisib (idela) is a highly selective oral inhibitor of PI3Kδ that is currently FDA-approved in conjunction with rituximab for the treatment of relapsed/refractory (R/R) chronic lymphocytic leukemia (CLL). In clinical trials performed in the R/R setting, the overall response rate was 70-80%, while the frequency of significant toxicity was easily manageable (grade ≥3 transaminitis 14%, pneumonitis 3%, grade ≥3 diarrhea 14%) (Coutre EHA 2015). Given the high efficacy of idela, we are evaluating the combination of idela plus ofatumumab (ofa) as first-line therapy for CLL in a phase II study. Surprisingly, we noted much higher rates of grade 3-4 transaminitis in particular, as well as more pneumonitis and colitis, than previously reported. Preclinical data and clinical response to corticosteroids suggest that these toxicities may be autoimmune in origin.

Methods: These results describe the clinical characteristics and toxicities of the first 21 subjects enrolled in a single-arm phase II study of idela plus ofa in previously untreated CLL patients in need of therapy. Subjects received idela 150mg twice daily during a 2 month lead-in period prior to the addition of weekly ofa infusions x 8 followed by monthly infusions x 4. For the first 2 months, subjects were monitored for toxicities with weekly clinic visits and biweekly serum chemistries. Single cell mass cytometry (CyTOF) permits the simultaneous evaluation of up to 36 markers without marker emission overlap, allowing for a comprehensive phenotypic and functional analysis of T cell subsets. We used CyTOF to compare T cell subset number and function between subjects who experienced no toxicity (n=2) and a portion of subjects (n=5) who experienced grade ≥3 toxicity while on idela.

Results: After a median follow-up of 8.1 months (range 0.7-10.8 months), sixteen subjects (76%) had experienced a grade 3 or higher toxicity. The most frequent grade ≥3 adverse events were transaminitis (n=12, 57%), enterocolitis (n=3, 14%), and pneumonitis (n=2, 10%). The subjects who experienced grade ≥3 toxicities, or who experienced multiple toxicities of at least grade 2 (n =13), were younger (median age 65 vs. 75 years, p=0.047) and had higher absolute lymphocyte counts (median 71466 vs. 19250 cells/µL, p =0.017) compared to subjects who experience no or low grade toxicity (n=7). The median time to onset of transaminase elevation was 28 days (range 14-274 days), with most occurring between days 20-30. Two subjects with ongoing elevation of grade 4 transaminitis after holding idela underwent liver biopsy. These biopsies showed increased activated cytotoxic T cells within the liver parenchyma compared to normal controls with CLL. In all cases, the organ toxicities have abated with the initiation of immunosuppressive therapy. Sixteen subjects (76%) required steroids and one subject (5%) required mycophenolate mofetil.

Preclinical data suggest that PI3Kδ is critical to the function of regulatory T cells (Tregs), and inhibition of PI3Kδ leads to autoimmunity. Indeed, CyTOF analysis demonstrated that five out of six tested subjects (83%) had a decrease in the percentage of Tregs after one cycle of idela therapy. Tregs from subjects who experienced grade ≥3 toxicity had lower baseline expression level of functional markers (GITR, T-bet, TIM-3) and higher expression level of apoptotic markers (CD95) compared to subjects who experienced no toxicity. After one cycle of idela therapy, expression levels of the Treg effector markers granzyme β, HLA-DR, and PD-1 decreased in subjects who experienced toxicity, but increased in those subjects who did not.

Conclusions: The use of idela as first-line therapy in CLL results in more frequent and severe toxicities than its use in the R/R setting. Multiple lines of evidence suggest that this toxicity is immune-mediated: the delayed time to onset, an immune cell infiltrate in biopsies of affected organs, and abatement of toxicity with immunosuppressants. Affected patients had depressed Treg functionality at baseline and lost markers of Treg activation after idela therapy, suggesting that they may be particularly sensitive to PI3Kδ blockade and Treg inhibition. In addition to elucidating the mechanisms of idelalisib, these studies will hopefully allow us to better screen and select patients in whom idela therapy will be well tolerated.

Disclosures

No relevant conflicts of interest to declare.

Author notes

*

Asterisk with author names denotes non-ASH members.

Sign in via your Institution