Characterization of CD20 expression loss as a mechanism of resistance to mosunetuzumab in patients with relapsed/refractory B-cell non-Hodgkin lymphomas

7526

Background: Mosunetuzumab (M) is a bispecific antibody targeting CD20 and CD3 that redirects T cells to engage and eliminate malignant B cells being developed for relapsed or refractory (R/R) B-cell non-Hodgkin lymphomas (B-NHL). CD20 is an optimal target, with uniform expression across B-NHL histologies and minimal receptor turnover. We characterized CD20 loss as a potential mechanism of resistance to M in patients (pts) on a Phase I/II trial (NCT02500407) receiving M monotherapy for the treatment (tx) of R/R B-NHL. Methods: Pts with R/R B-NHL received M intravenously in 3-week cycles, for eight to 17 cycles depending on tumor response. At baseline (BL), biomarker-evaluable (archival or fresh) biopsies were collected from 293 pts. Biopsies from 62 pts were collected at additional time points during tx with M and/or at disease progression (PD). The proportion of CD20+ and PAX5+ tumor cells was determined by immunohistochemistry (IHC) using dual-staining with anti-CD20 (clone L26, VENTANA) and anti-PAX5 (clone DAK-PAX5, DAKO) antibodies. Expression of MS4A1, the gene encoding CD20 , was measured by RNA-sequencing (RNA-seq); MS4A1 mutation profiling was performed by whole exome sequencing (WES). Levels of CD20 expression were assessed relative to response rates. Correlative analyses were performed and assessed centrally (IHC, RNA-seq, and WES) and locally (IHC). Results: CD20 levels were consistently high ( > 75% CD20+PAX5+ cells) in the majority of BL biopsies and generally comparable across histologies (FL, DLBCL, tFL, MCL, and RT). BL CD20 loss (≤5% CD20+PAX5+ cells) was seen in 16/293 pts (5.5%), more commonly in aggressive NHL, and responses to M were not seen in these pts. Among 62 pts with BL and on-tx/at-PD biopsies, BL CD20 levels were ≤5% in 7/62 pts (11%) (6/7 pts [86%] progressed before completing Cycle 2). CD20 levels were maintained in on-tx biopsies from 23/24 pts (96%). At PD, biopsies showed CD20 loss in 7/26 pts (27%). For five pts with BL, on-tx and at-PD biopsies, all pts maintained CD20 while on-tx and 1/5 pts (20%) had CD20 loss at PD. There was no clear association between CD20 reduction and histology. Data from 185 BL biopsies showed generally concordant levels of CD20 gene and protein expression (r = 0.72). In 10/185 pts (5%), MS4A1 was expressed without detectable CD20 protein expression; DNA sequencing revealed novel mutations in MS4A1, including mutations leading to truncation of the protein. CD20 transmembrane and extra-cellular domain mutations were also observed but do not block CD20 expression. Conclusions: In pts with R/R B-NHL treated with M, low BL CD20 expression is associated with lack of response to M. During M tx, loss of tumor cell expression of CD20 is one mechanism of acquired resistance; however, CD20 expression is maintained in most pts with PD, implying alternative mechanisms for acquired M resistance. Clinical trial information: NCT02500407.

Abstract 3527: Assessing signal modulation in preclinical models and tumor tissue from patients enrolled in a Phase 1 study of the HER3/EGFR dual action antibody, MEHD7945A (MEHD), as an exploratory pharmacodynamic (PD) analysis

Background

Aberrant activation of HER family members including EGFR results in unregulated proliferative signals involved in tumorigenesis. HER3 plays a unique role as both a dimerization-dependent allosteric catalyst and critical activator of PI3 kinase signaling (PI3K). MEHD is a dual-action antibody that blocks ligand binding to EGFR and HER3 to abrogate signaling downstream from these two receptors. Early clinical evidence shows PD modulation and biological activity. The two major signaling pathways activated by ligand binding to these receptors are the RAS/MAPK and PI3K/AKT pathways. Phosphorylation of these critical signaling proteins can be monitored to evaluate inhibition of ligand driven receptor activity.

Methods

Phosphorylation of signaling proteins in the RAS/MAPK and PI3K/AKT pathways was measured by Western blot analysis and reverse-phase protein arrays (RPPA) in MEHD-sensitive cell lines and xenograft models. Paired tumor biopsies from patients in the Phase 1 MEHD study were collected pre and post MEHD dosing. Half of each sample was formalin-fixed and paraffin-embedded for immunohistochemistry (IHC) using pS6 (S235/236), pPRAS40 (T246), and pMAPK (T202/Y204); the other half was frozen and embedded in Optimum Cutting Temperature compound (OCT) for analysis of multiple signaling pathway endpoints by RPPA.

Results

Inhibition of EGFR and/or HER3 activated signaling by MEHD was confirmed in sensitive cell lines and xenograft models by western blot analysis for key analytes including pMAPK and pS6. Broader analysis of signaling in the RAS/MAPK and PI3K/AKT pathways by RPPA demonstrated additional markers that were inhibited by MEHD. Signaling from EGFR driven models was blocked by EGFR inhibition alone but signaling in models in which both EGFR and HER3 were contributing factors, as assessed by tumor growth inhibition, were not inhibited by EGFR inhibition alone. However, MEHD was equally effective at blocking signaling in both models. Analytes that were inhibited by MEHD in these pre-clinical models were measured by RPPA and IHC in paired biopsy specimens from the Ph1 MEHD study as a PD assessment. Thirty-two paired biopsies were obtained. 20/32 had sufficient tissue and tumor for assessment and 11/20 showed decreased phosphorylation of pERK, pPRAS40, or pS6 by IHC and/or RPPA.

Conclusions

Preclinical evaluation of RAS/MAPK and PI3K/AKT signaling in MEHD-sensitive cell lines and xenograft models identified pERK, pPRAS40, and pS6 as measurable indicators of MEHD target engagement and biological activity. Evaluation of these markers in patient tumor biopsies demonstrated that signaling pathways consistent with HER3 and/or EGFR inhibition are downregulated in tumors, suggesting that MEHD is biologically active in patients (see Juric et al., AACR abstract).

Citation Format: Elicia M. Penuel, Luciana Burton, Howard Stern, Lackner Mark, Andrea Pirzkall, Lukas Amler, Gabriele Schaefer. Assessing signal modulation in preclinical models and tumor tissue from patients enrolled in a Phase 1 study of the HER3/EGFR dual action antibody, MEHD7945A (MEHD), as an exploratory pharmacodynamic (PD) analysis. [abstract]. In: Proceedings of the 104th Annual Meeting of the American Association for Cancer Research; 2013 Apr 6-10; Washington, DC. Philadelphia (PA): AACR; Cancer Res 2013;73(8 Suppl):Abstract nr 3527. doi:10.1158/1538-7445.AM2013-3527

Chinese hamster ovary cells contain transcriptionally active full-length type C proviruses

We have isolated a genomic locus from Chinese hamster ovary (CHO) cells that contains a full-length provirus. Nucleotide sequence analysis indicates that it is a defective member of the rodent type C retrovirus family with an env region that is similar to those of mouse amphotropic retrovirus and subgroup B feline leukemia virus. We were able to demonstrate that this provirus is a member of a closely related family of full-length proviruses in CHO cells and Chinese hamster liver. Hybridization probes generated from this genomic clone were used to characterize type C retrovirus RNA expression in CHO cells. Full-length genomic RNA and subgenomic envelope mRNA were detected in CHO cell lines but not in the human-derived 293 cell line. Interestingly, we discovered that the site of retrovirus integration lies within a G repeat sequence belonging to the short interspersed element family of retroposons.