Comparative Assessment of Cell Line Models for Optimal Representation of Chronic Lymphocytic Leukemia

Background

Cell line models have proven to be helpful tools for the study of many types of hematologic malignancies. However, chronic lymphocytic leukemia (CLL) has historically been challenging to immortalize into a cell line without introducing a more aggressive biology that calls into question the fidelity of the model to the more typical indolent CLL biology. Recently, several cell lines have been purported to more accurately represent CLL biology. We utilized a variety of laboratory techniques including BH3 profiling, a functional assay that assesses the propensity of the cells to undergo apoptosis (“apoptotic priming”) and cellular dependence on anti-apoptotic BCL-2 family proteins, to perform a comparative analysis of 5 such cell lines to identify the one(s) that most closely model CLL biology in the laboratory setting.

Methods

Five different cell lines (HG3, MEC-1, OSU-CLL, OCI-LY1 and OCI-LY1-R) were evaluated and compared with CLL primary samples that were derived from the peripheral blood of patients with previously untreated CLL and included diverse CLL genetic markers. Baseline BH3 profiling to measure cytochrome C (cytoC) release was performed on each cell line as previously described (Ryan et al., 2016). Each cell line was then co-cultured with ibrutinib, pirtobrutinib or venetoclax at increasing concentrations in a series of CellTiter-Glo (CTG) viability assays in the presence and absence of conditioned media derived from the HS5 stromal cell line to assess cellular viability in response to each drug. In addition, dynamic BH3 profiling (DBP) was performed (Montero et al., 2015) whereby each cell line was treated with equimolar concentrations of each drug for 24 hours and the change in priming (delta-priming) comparing drug-treatment with DMSO control was calculated.

Results

Consistent with our prior work, baseline BH3 profiling of 20 primary CLL patient samples showed relatively uniform dependence on BCL-2 (median 60.23% cytoC release), with CLL cells with mutated IGHV having higher BCL-2 dependence but unmutated IGHV having higher levels of overall apoptotic priming. In contrast, profiling of all 5 cell lines revealed significant heterogeneity in anti-apoptotic protein dependence. The MEC-1 cell line showed no clear dependence and low levels of overall priming, while both the HG3 and OSU-CLL cell lines showed low dependence on BCL-2 (17.47% and 10.35%, respectively) and higher dependence on MCL-1 at 48%. The OCI-LY1-R cell line also had high MCL-1 dependence at 58.47% and is inherently resistant to BCL-2 inhibition. The OCI-LY1 cell line demonstrated the highest level of BCL-2 dependence at 98%.

In viability experiments with co-culture with the BTKis ibrutinib and pirtobrutinib in the presence of conditioned media, all 5 cell lines performed similarly to primary CLL samples, with about 60% viability after 24 hours of drug treatment at a concentration of 10 μM. However, consistent with the BH3 profiling results, when exposed to venetoclax, only OCI-LY1 behaved similarly to the primary CLL samples, with decreased viability from 50% at 1 nM to nearly 0% at 1 μm, with similar results in both mutated and unmutated IGHV patient samples.

DBP of primary CLL cells demonstrated an increase in BCL-2 dependence (media delta priming of +11.6) after ibrutinib exposure, a decrease in dependence across all peptides after pirtobrutinib exposure (-16.68), a decrease in dependence on BCL-2 (delta of -14.7), and an increase in MCL-1 (+40.0) after venetoclax exposure. Similar patterns were seen in the OCI-LY1 cell line, which showed an increase in BCL-2 dependence after ibrutinib (+26), decrease after pirtobrutinib exposure (-17.62), a decrease in dependence on BCL-2 (-9), and an increase in dependence on MCL-1 (+23) after venetoclax. The other 4 cell lines displayed distinctly different changes in priming patterns by DBP after 24h.

Conclusion

In this comparative assessment of primary CLL cells and 5 cell lines commonly used in laboratory investigations to represent CLL, the OCI-LY1 cell line most accurately mimics CLL biology with respect to baseline BCL-2 dependence, viability after treatment with the BCL-2 inhibitor venetoclax, and delta-priming in response to BTKi and venetoclax. Further studies are ongoing to compare these cell lines to primary CLL cells, including additional functional assays and in vivo models.

Davids:Eli Lilly: Consultancy; AstraZeneca: Consultancy, Research Funding; BeiGene: Consultancy; Ascentage Pharma: Consultancy, Research Funding; Merck: Consultancy; Janssen: Consultancy; AbbVie: Consultancy, Research Funding; BMS: Consultancy; Surface Technology: Research Funding; Genentech: Consultancy, Research Funding; TG Therapeutics: Consultancy, Research Funding; Genmab: Consultancy; Adaptive Biosciences: Consultancy; Novartis: Research Funding; MEI Pharma: Research Funding.