the Development and Expansion of Resistant Subclones Precedes Relapse during Ibrutinib Therapy in Patients with CLL

The Bruton's Tyrosine Kinase (BTK) inhibitor ibrutinib is very effective in chronic lymphocytic leukemia (CLL). Resistance in CLL is at least in part mediated by acquired mutations in BTK or down-stream PLCG2. Here we describe the largest institutional cohort of CLL patients treated with ibrutinib, focusing on risk factors for relapse, prevalence of known resistance mutations, and development of a monitoring strategy to identify patients at risk for relapse.

All 308 patients from The Ohio State University Comprehensive Cancer Center participating in 4 clinical trials of ibrutinib were included. Deep sequencing for BTK and PLCG2 was performed using Ion Torrent Personal Genome Machine and covered coding regions of both genes. Preclinical experiments used XLA cell lines (Coriell Institute) infected with lentiviral constructs containing empty vector, wild type BTK, or C481S BTK.

With a median follow-up of 40.5 months (range 4-71 months), 136 patients remain on study, 14 have received transplant or therapy elsewhere, and 158 have discontinued. 83 patients experienced disease progression, classified as Richter's or PLL transformation (n=28) and progressive CLL (n=55). Using multivariable models, baseline risk factors for ibrutinib discontinuation due to transformation include complex karyotype (p<0.01) and MYC abnormalities on FISH (p=0.051). Age<65, del(17p) by FISH, and complex karyotype all associated with discontinuation due to CLL progression (p<0.05). Relapse generally has a poor prognosis, with a median survival from ibrutinib discontinuation for patients with transformation and progressive CLL of 3.9 (95% CI 2.0 to 10.1) and 22.7 (95% CI 13.5 to not reached) months, respectively.

46 patients with progressive CLL had samples at relapse available for deep sequencing. Of these, 39 had mutations in BTK and/or PLCG2 (84.8%; 95% CI 71.1-93.7). Distribution of mutation included patients with BTK C481 mutation only (n=30), mutation in PLCG2 only (n=3) and both BTK/PLCG2 genes (n=6).

Given the poor prognosis at relapse and presence of acquired mutations in the majority of progressive CLL patients, we were interested to evaluate whether the emergence of small clones of mutated cells could be used as a biomarker to predict relapse. For 15 patients with BTK or PLCG2 mutations, serial samples were available prior to relapse, and a clone of resistant cells could first be detected a median of 9.3 months prior to clinical relapse (95% CI: 7.6-11.7). Based on the finding that patients who relapsed on ibrutinib often had rapid progression and poor outcomes, we initiated a clinical-grade monitoring strategy in our institutional CLIA-certified molecular lab starting in November 2014. Mutational analysis of the entire coding regions of BTK and PLCG2 was performed on a cohort of 112 patients every 3 months. To date, 8 patients have clinically relapsed and all 8 had BTK C481S mutations with expansion of the clone prior to relapse. BTK C481S mutations of over 1% allelic frequency were detected in an additional 8 patients. All patients have had expansion of the resistant clones and all but one, who discontinued therapy without clinical relapse, have had increasing circulating CLL cells in the peripheral blood. Four of these 7 patients have also had increasing lymph node size on CT scan.

In 11 patients who relapsed with BTK C481S, samples were available following discontinuation. In 5 patients treated subsequently with venetoclax, allelic fractions of C481S decreased dramatically or were eliminated, but in 6 patients treated with other agents, the mutant clone persisted. To investigate whether this was due to differences in biology between wild type and C481S BTK, we created XLA cell lines without BTK protein expression stably expressing BTK, or C481S BTK. C481S BTK showed enhanced BCR signaling as evidenced by pERK and cMYC expression, and enhanced migration compared to wild type BTK (p=0.04). This demonstrates for the first time that this acquired resistance mutation fundamentally alters the tumor cell biology.

These data confirm our initial findings that relapse on ibrutinib is primarily mediated through mutations in BTK and PLCG2 and suggest that the presence of these mutations may have utility as an early biomarker to predict clinical relapse. Early identification of relapsing patients could improve outcomes by facilitating initiation of adjunct therapies such as venetoclax to eliminate the small resistant clone.