https://orcid.org/0000-0002-9168-2291
TO THE EDITOR:
Small molecule inhibitors targeting Bruton’s tyrosine kinase (BTK) have revolutionized treatment of chronic lymphocytic leukemia (CLL).1 BTK inhibitors (BTKi) used at present in routine practice bind covalently to cysteine 481.2-5 Resistance to BTKi is predominantly mediated by mutations in BTK at C481, affecting the covalent binding of BTKi and by activating mutations in the BTK downstream target, PLCG26,. A range of noncovalent BTKi are being developed to overcome resistance mediated by BTK mutation at C481 and are in early phase clinical assessment for treatment of CLL.7 Despite the clinical success of these inhibitors, a subset of patients develop refractory disease.8,9 The mechanisms mediating resistance to the different noncovalent BTKi are not completely characterized.
Recently, Wang et al., identified novel acquired mutations in BTK that conferred resistance in 7 of 55 patients treated with the noncovalent BTKi, pirtobrutinib.10 In contrast to ibrutinib, a single hotspot mutation was not identified with pirtobrutinib but different mutations were found to cluster in the protein kinase domain of BTK. However, it is unclear, if these kinase domain BTK mutations are acquired only in the context of pirtobrutinib resistance or they represent a general phenomenon upon resistance to noncovalent BTKi. In addition, it is unknown if there are additional, so far undescribed BTK mutations that are associated with noncovalent BTKi resistance.
Here, we used a long-term in vitro dose escalation method to generate resistance to ibrutinib and the noncovalent BTKi, pirtobrutinib (LOXO-305), vecabrutinib (SNS-062), nemtabrutinib (ARQ-531), fenebrutinib (GDC-0853), and RN-486, using the REC-1 mantle cell lymphoma (MCL) cell line. REC-1 cells were selected because of their high sensitivity to BTKi based on a cell viability screen of 6 CLL and MCL cell lines (supplemental Figure 1). Three to 6 independent lines of REC-1 cells were treated in parallel, with increasing concentrations of the inhibitors. The cell lines were designated as resistant to a given drug when the 50% inhibitory concentration (IC50) increased >10-fold compared with that of previously untreated REC-1 cells.
The different lines of REC-1 cells with acquired resistance to ibrutinib and the noncovalent inhibitors were analyzed using targeted next generation sequencing to identify acquired mutations in BTK and PLCG2 (Figure 1A; supplemental Figure 2). Although all the 6 lines of REC-1 cells treated long term with ibrutinib gained the C481F BTK mutation, in cells resistant to the noncovalent BTKi we identified 6 different mutations in BTK (variant allele frequency [VAF], 22% to 99%) and 3 mutations in PLCG2 (VAF, 10%, 30%, and 53%). Four of 6 vecabrutinib resistant lines acquired BTK G409R mutation whereas 1 acquired a L845F PLCG2 mutation. Three different BTK mutations, L528S, G480R, and D539H were found in REC-1 cells resistant to fenebrutinib. RN-486–resistant REC-1 cells acquired G480R and V416L BTK mutations, whereas pirtobrutinib-resistant lines acquired a BTK A428D mutation, and R727L and S1079R PLCG2 mutations in independent lines. Interestingly, mutations in BTK residues A428, L528, and V416 reported by Wang et al.,10 in pirtobrutinib resistant CLL were also identified in the REC-1 cell lines resistant to pirtobrutinib, fenebrutinib, and RN-486, respectively. Notably, nemtabrutinib-resistant REC-1 cells, despite the increased IC50, did not acquire a BTK or PLCG2 mutation, indicating the role of other mechanisms in the generation of resistance.
![Mutations in BTK and PLCG2 appear under selection pressure by BTK inhibitors in REC-1 cells. (A) Circos plot summarizing the distribution of BTK and PLCG2 mutations in REC-1 cells resistant to the different covalent and noncovalent BTK inhibitors. The innermost circle represents the number of independent lines of REC-1 cells that were used for generating resistance to each inhibitor and the number of REC-1 lines that acquired the specific mutations. (B) Comparison of intracellular calcium flux between WT and mutant REC-1 cells upon stimulation with 10 μg/mL of anti-IgM. Normalization was performed for baseline [Ca2+]i before anti-IgM stimulation and to maximum [Ca2+]i obtained for each cell line upon treatment with ionomycin. (C) Western blotting analysis for activation of phospho-BTK, phospho-PLCγ2, phospho-AKT, and phospho-ERK upon stimulation with 10 μg/mL anti-IgM for 15 minutes. In addition, total BTK, PLCγ2, AKT, and ERK1/2 levels were analyzed with glyceraldehyde-3-phosphate dehydrogenase (GAPDH) as the loading control. The arrow indicates the band for GAPDH. The data shown in B and C are representatives of 3 independent measurements.](https://ash.silverchair-cdn.com/ash/content_public/journal/bloodadvances/7/19/10.1182_bloodadvances.2022008955/2/m_blooda_adv-2022-008955-gr1.jpeg?Expires=1724165433&Signature=KcmZ2UpaImX0-EmXu2uGD9lEQjiW~1nhPaGygwRd1woBEDxZGTeChqoOyl9h3jpVo1wBRAeDQPbjY5XJ2gVXdOw9bRF~dWphq9JaiY2j1~bubKEKJ~aKoEjqPHWbod8YNHsF~O4JrRivoCC4Qj6AGMbTNR3F6LjlbdO~xDcBwynua4LSV1EhtYc~tPA3NKnI6foLwPH-02vk8f~btgZAxp6iJW3kM2nosriYHLcgTDN8CqB~fOjnskQatDI3nNrO2G5TTEp-EYi-UA~~~hbYY1SxI6XRLHdBO~1JFMfhqJQtSaScgIpG9AqxDhK1SIDNigFFIRipke3~hzk2Ywrt6A__&Key-Pair-Id=APKAIE5G5CRDK6RD3PGA)
Mutations in BTK and PLCG2 appear under selection pressure by BTK inhibitors in REC-1 cells. (A) Circos plot summarizing the distribution of BTK and PLCG2 mutations in REC-1 cells resistant to the different covalent and noncovalent BTK inhibitors. The innermost circle represents the number of independent lines of REC-1 cells that were used for generating resistance to each inhibitor and the number of REC-1 lines that acquired the specific mutations. (B) Comparison of intracellular calcium flux between WT and mutant REC-1 cells upon stimulation with 10 μg/mL of anti-IgM. Normalization was performed for baseline [Ca2+]i before anti-IgM stimulation and to maximum [Ca2+]i obtained for each cell line upon treatment with ionomycin. (C) Western blotting analysis for activation of phospho-BTK, phospho-PLCγ2, phospho-AKT, and phospho-ERK upon stimulation with 10 μg/mL anti-IgM for 15 minutes. In addition, total BTK, PLCγ2, AKT, and ERK1/2 levels were analyzed with glyceraldehyde-3-phosphate dehydrogenase (GAPDH) as the loading control. The arrow indicates the band for GAPDH. The data shown in B and C are representatives of 3 independent measurements.
Mutations in BTK and PLCG2 appear under selection pressure by BTK inhibitors in REC-1 cells. (A) Circos plot summarizing the distribution of BTK and PLCG2 mutations in REC-1 cells resistant to the different covalent and noncovalent BTK inhibitors. The innermost circle represents the number of independent lines of REC-1 cells that were used for generating resistance to each inhibitor and the number of REC-1 lines that acquired the specific mutations. (B) Comparison of intracellular calcium flux between WT and mutant REC-1 cells upon stimulation with 10 μg/mL of anti-IgM. Normalization was performed for baseline [Ca2+]i before anti-IgM stimulation and to maximum [Ca2+]i obtained for each cell line upon treatment with ionomycin. (C) Western blotting analysis for activation of phospho-BTK, phospho-PLCγ2, phospho-AKT, and phospho-ERK upon stimulation with 10 μg/mL anti-IgM for 15 minutes. In addition, total BTK, PLCγ2, AKT, and ERK1/2 levels were analyzed with glyceraldehyde-3-phosphate dehydrogenase (GAPDH) as the loading control. The arrow indicates the band for GAPDH. The data shown in B and C are representatives of 3 independent measurements.
Functional analysis of B-cell receptor (BCR) signaling capabilities of the different BTK mutant REC-1 cells was performed by measurement of intracellular Ca2+ ([Ca2+]i) upon stimulation with anti-IgM. Except for A428D, all BTK mutant REC-1 cell lines showed a similar or enhanced [Ca2+]i, but all mutants showed a more sustained increase in [Ca2+]i compared to wild-type (WT) cells (Figure 1B). Western blot analyses showed decreased anti-IgM–mediated activation of BTK in all the different BTK mutant cell lines compared to that in WT. A time course experiment additionally confirmed the lower BTK activation by anti-IgM in the mutant cells than in WT (supplemental Figures 3A and B). However, the downstream activation of phospho-AKT was enhanced in all the mutant lines, similar to earlier reports (Figure 1C; supplemental Figure 3c),11 whereas phospho-ERK levels changed marginally in the mutant cells compared to WT.
Furthermore, we tested, if each of these mutant cell lines showed differential sensitivity to other covalent and noncovalent BTKi (Figure 2A). Among all the BTKi analyzed, nemtabrutinib showed the highest efficacy across the different BTK mutant REC-1 cells. This effect of nemtabrutinib may be attributable to its unique protein kinase inhibitory profile.12 The C481F mutant REC-1 cells showed significantly higher resistance to the covalent BTKi than to noncovalent BTKi. The G409R and D539H mutant cells were resistant only to the noncovalent BTKi, whereas the cells harboring the A428D, L528S, G480R, or V416L mutations were resistant to both covalent and noncovalent BTKi (supplemental Figure 4 and 5). In line with the drug efficacy data, [Ca2+]i measurements after drug treatment of BTK WT cells showed a complete inhibition of anti-IgM–induced calcium mobilization whereas the BTK mutants showed differential sensitivity to the noncovalent and covalent BTKi (Figure 2A and supplemental Figure 6).
![Response of BTK mutant REC-1 cells to the different covalent and noncovalent BTK inhibitors. (A) VAF of the different BTK mutations acquired in the REC-1 cells are indicated. Each of the cell lines (WT or BTK mutant) were treated in triplicates with increasing concentrations of the indicated covalent and noncovalent BTK inhibitors. Cell viability was measured after 96 hours using CellTiter Glo (Promega). Heatmap on the left shows the fold change in the IC50s of the respective cell lines compared to the WT cells for each of the BTK inhibitors. The cell viability heatmap is a summary of 3 independent experiments. Intracellular calcium flux was determined after exposing the different REC-1 cell lines (WT or BTK mutant) to 1 μM of the covalent or noncovalent BTK inhibitors for 1 hour, followed by stimulation with 10μg/mL of anti-IgM. The heatmap on the right shows percentage inhibition of the maximum calcium flux obtained upon BCR stimulation. The values for calculating the percentage inhibition were taken from representative experiments shown in supplemetal Figure 7A. Calcium flux measurement for each independent REC-1 line treated with 8 different BTK inhibitors and dimethyl sulfoxide control was performed at least twice. (B) COS-7 cells were transfected as indicated with 100 ng/well of vector encoding PLCγ2 and 400 ng/well of empty vector (∅) or increasing amounts (100, 200, and 400 ng/well) of vector encoding WT, C481S, C481F, G409R, L528W, L528S, G480R, D539H or K430R mutant BTK. Twenty-four hours after transfection, the cells were incubated for 18 hours with myo-[2-3H] inositol, and inositol phosphate formation was then determined. Results are one representative from 3 independent experiments and illustrated as mean ± SD of 3 technical replicates (C) Representative western blot showing IGF1R expression in BTK mutant REC-1 cells in comparison to WT. (D) Expression of IGF1R in the BTK mutant REC-1 cells normalized to that of WT. IGF1R expression was quantified using ImageJ from 2 repetitions of the western blots and normalized to the corresponding GAPDH controls. (E) The BTK mutant and WT REC-1 cells were treated with increasing concentrations of the IGF1R inhibitor linsitinib for 96 hours and cell viability was measured using CellTiter Glo (Promega). The data shown is a summary of 3 independent experiments performed in triplicates.](https://ash.silverchair-cdn.com/ash/content_public/journal/bloodadvances/7/19/10.1182_bloodadvances.2022008955/2/m_blooda_adv-2022-008955-gr2.jpeg?Expires=1724165433&Signature=c5F9qKWZs~Bj~UjPDOcF3fVtjOluGcXuoYsRW4dlTpBLo0CmZs0gvAyg2FZn1hfu-RrdjKM2pb0tgCtYAGPkL0z6qLNn2jAQgcJLcuoXqJ4Ohmi5RZqLY1ebppkQy4oIlP2USocxKVliGRWAJTZcNPd1ci2~cgO5DKzO4QNFjvaYLFPPLmtOaNDbLnbrNRRwEe~bIAa9bsBRDcSoaL-i2LLthdwoylZyrJjM5LLM6Xo3rLug155enAsZeZrSrC~1VnKUIvBe8cRz4GxkdcU89IvnlRCKwOhGlllRX59HbDNwipFqY3K6tD1uGRiAqlvuq1lqN86vs0X155xUasGNAg__&Key-Pair-Id=APKAIE5G5CRDK6RD3PGA)
Response of BTK mutant REC-1 cells to the different covalent and noncovalent BTK inhibitors. (A) VAF of the different BTK mutations acquired in the REC-1 cells are indicated. Each of the cell lines (WT or BTK mutant) were treated in triplicates with increasing concentrations of the indicated covalent and noncovalent BTK inhibitors. Cell viability was measured after 96 hours using CellTiter Glo (Promega). Heatmap on the left shows the fold change in the IC50s of the respective cell lines compared to the WT cells for each of the BTK inhibitors. The cell viability heatmap is a summary of 3 independent experiments. Intracellular calcium flux was determined after exposing the different REC-1 cell lines (WT or BTK mutant) to 1 μM of the covalent or noncovalent BTK inhibitors for 1 hour, followed by stimulation with 10μg/mL of anti-IgM. The heatmap on the right shows percentage inhibition of the maximum calcium flux obtained upon BCR stimulation. The values for calculating the percentage inhibition were taken from representative experiments shown in supplemetal Figure 7A. Calcium flux measurement for each independent REC-1 line treated with 8 different BTK inhibitors and dimethyl sulfoxide control was performed at least twice. (B) COS-7 cells were transfected as indicated with 100 ng/well of vector encoding PLCγ2 and 400 ng/well of empty vector (∅) or increasing amounts (100, 200, and 400 ng/well) of vector encoding WT, C481S, C481F, G409R, L528W, L528S, G480R, D539H or K430R mutant BTK. Twenty-four hours after transfection, the cells were incubated for 18 hours with myo-[2-3H] inositol, and inositol phosphate formation was then determined. Results are one representative from 3 independent experiments and illustrated as mean ± SD of 3 technical replicates (C) Representative western blot showing IGF1R expression in BTK mutant REC-1 cells in comparison to WT. (D) Expression of IGF1R in the BTK mutant REC-1 cells normalized to that of WT. IGF1R expression was quantified using ImageJ from 2 repetitions of the western blots and normalized to the corresponding GAPDH controls. (E) The BTK mutant and WT REC-1 cells were treated with increasing concentrations of the IGF1R inhibitor linsitinib for 96 hours and cell viability was measured using CellTiter Glo (Promega). The data shown is a summary of 3 independent experiments performed in triplicates.
Response of BTK mutant REC-1 cells to the different covalent and noncovalent BTK inhibitors. (A) VAF of the different BTK mutations acquired in the REC-1 cells are indicated. Each of the cell lines (WT or BTK mutant) were treated in triplicates with increasing concentrations of the indicated covalent and noncovalent BTK inhibitors. Cell viability was measured after 96 hours using CellTiter Glo (Promega). Heatmap on the left shows the fold change in the IC50s of the respective cell lines compared to the WT cells for each of the BTK inhibitors. The cell viability heatmap is a summary of 3 independent experiments. Intracellular calcium flux was determined after exposing the different REC-1 cell lines (WT or BTK mutant) to 1 μM of the covalent or noncovalent BTK inhibitors for 1 hour, followed by stimulation with 10μg/mL of anti-IgM. The heatmap on the right shows percentage inhibition of the maximum calcium flux obtained upon BCR stimulation. The values for calculating the percentage inhibition were taken from representative experiments shown in supplemetal Figure 7A. Calcium flux measurement for each independent REC-1 line treated with 8 different BTK inhibitors and dimethyl sulfoxide control was performed at least twice. (B) COS-7 cells were transfected as indicated with 100 ng/well of vector encoding PLCγ2 and 400 ng/well of empty vector (∅) or increasing amounts (100, 200, and 400 ng/well) of vector encoding WT, C481S, C481F, G409R, L528W, L528S, G480R, D539H or K430R mutant BTK. Twenty-four hours after transfection, the cells were incubated for 18 hours with myo-[2-3H] inositol, and inositol phosphate formation was then determined. Results are one representative from 3 independent experiments and illustrated as mean ± SD of 3 technical replicates (C) Representative western blot showing IGF1R expression in BTK mutant REC-1 cells in comparison to WT. (D) Expression of IGF1R in the BTK mutant REC-1 cells normalized to that of WT. IGF1R expression was quantified using ImageJ from 2 repetitions of the western blots and normalized to the corresponding GAPDH controls. (E) The BTK mutant and WT REC-1 cells were treated with increasing concentrations of the IGF1R inhibitor linsitinib for 96 hours and cell viability was measured using CellTiter Glo (Promega). The data shown is a summary of 3 independent experiments performed in triplicates.
To evaluate the ability of the novel BTK mutants to activate PLCγ2, we transiently transfected COS-7 cells with plasmids encoding PLCγ2 or BTK variants and determined their lipase activity in terms of inositol phosphate formation13 (Figure 2B). Although WT BTK and to a lesser extent the C481S mutant showed activation of PLCγ2, we observed diminished or no activation of the lipase with the novel BTK mutants G409R, L528W, L528S, G480R, and D539H, indicating impaired kinase function similar to the K430R kinase-dead BTK mutant.14 Notably, the functional effects of the BTK mutants observed in Figure 2B are not caused by differences in protein expression (supplemental Figure 7a).
Recently, dependencies of kinase-dead BTK on TLR9, UNC93B1, CNPY3, and scaffolding interactions with HCK and LYN were reported.15-17 Mechanisms by which kinase-dead BTK modulates signaling in the REC-1 model are yet to be ascertained but the differential response of C481F mutants to covalent vs noncovalent BTKi indicates that drug binding may affect the scaffolding function of kinase-dead BTK variants.18 Alternative survival signaling pathways by receptor tyrosine kinases are known to contribute to BCR signaling inhibitor resistance.19,20 Therefore we verified expression of insulin-like growth factor receptor (IGF1R) in BTKi resistant cell lines. All the BTK mutant REC-1 cells except for G409R showed an increase in IGF1R expression (Figure 2C and D) and a greater sensitivity to IGF1R inhibitor, linsitinib (Figure 2E). A similar increase in IGF1R expression was also identified in adoptively transferred Eμ-TCL1 tumors isolated from recipient mice that had stopped responding to ibrutinib (supplemental Figure 8), indicating that the observed increase in IGF1R is not an exclusive phenomenon of BTKi resistant REC-1 cells.
Overall, our findings indicate that resistance to noncovalent BTKi is associated with different mutations in the protein kinase domain of BTK, similar to the reports on pirtobrutinib.10 Subset of these mutations are shared phenomena associated with resistance to different noncovalent BTKi.10 Despite potential limitations of an in vitro system, the long-term dose escalation REC-1 model has proven to be efficient in both recapitulating BTK mutations that have been previously reported (3 of 5) in the context of in vivo pirtobrutinib resistance and in identifying novel mutations, namely L528S, G409R, G480R, and D539H that have not yet been implicated in BTKi resistance. The resistant cell line model also represents an ideal system that allows genetic manipulation to further explore additional mechanisms contributing to drug resistance, in contrast to resistant primary samples from patients with CLL. Our study also highlights the importance of sequencing either all exons or at least the kinase domain (Exons 13-19) of BTK to identify BTK mutations in the context of progression under noncovalent BTKi treatment.
In line with the report by Wang et al, the novel BTK mutations clustered in the protein kinase domain and negatively affected kinase activity of BTK.10 Notably, we have already shown that cells harboring kinase-dead BTK are still capable of activating BCR signaling, whereas the signaling is fully abrogated in BTK-deficient cells.18 Also, BTK-independent mechanisms, such as IGF1R activation, could additionally contribute to noncovalent BTKi resistance and may represent important alternative target for treating noncovalent BTKi resistance.
Acknowledgments: J.Q. was supported by funding from Changzhou Jinse Medical Information & Science Technology Co. Ltd., China. X.G. received funding from China Scholarship Council. The project was supported by funding from Deutsche Forschungsgemeinschaft SFB1074, projects B1, B2, and B6.
Contribution: J.Q., S.E., M.W., and B.M.C.J. designed and performed experiments, analyzed and interpreted data, and edited the manuscript; D.Y. and E.T. analyzed and interpreted sequencing data and edited the manuscript; R.P.D. and X.G. performed experiments and analyzed the data; A.M., C.S., D.M., and P.G. interpreted data and edited the manuscript; M.W., B.M.C.J., and S.S. designed the study, analyzed and interpreted data, and wrote the manuscript; and all authors reviewed the manuscript.
Conflict-of-interest disclosure: The authors declare no competing financial interests.
Correspondence: Stephan Stilgenbauer, Division of CLL, Department of Internal Medicine III, Ulm University, Albert-Einstein-Allee 23, 89081, Ulm, Germany; e-mail: stephan.stilgenbauer@uniklinik-ulm.de; and Billy Michael Chelliah Jebaraj, Division of CLL, Department of Internal Medicine III, Ulm University, Albert-Einstein-Allee 23, 89081, Ulm, Germany; e-mail: billy.jebaraj@uniklinik-ulm.de.
