Characterization of ibrutinib-sensitive and -resistant mantle lymphoma cells

Authors


Summary

Ibrutinib inhibits Bruton tyrosine kinase (BTK), a key component of early B-cell receptor (BCR) signalling pathways. A multicentre phase 2 trial of ibrutinib in patients with relapsed/refractory mantle cell lymphoma (MCL) demonstrated a remarkable response rate. However, approximately one-third of patients have primary resistance to the drug while other patients appear to lose response and develop secondary resistance. Understanding the molecular mechanisms underlying ibrutinib sensitivity is of paramount importance. In this study, we investigated cell lines and primary MCL cells that display differential sensitivity to ibrutinib. We found that the primary cells display a higher BTK activity than normal B cells and MCL cells show differential sensitivity to BTK inhibition. Genetic knockdown of BTK inhibits the growth, survival and proliferation of ibrutinib-sensitive but not resistant MCL cell lines, suggesting that ibrutinib acts through BTK to produce its anti-tumour activities. Interestingly, inhibition of ERK1/2 and AKT, but not BTK phosphorylation per se, correlates well with cellular response to BTK inhibition in cell lines as well as in primary tumours. Our study suggests that, to prevent primary resistance or to overcome secondary resistance to BTK inhibition, a combinatory strategy that targets multiple components or multiple pathways may represent the most effective approach.

Mantle cell lymphoma (MCL) is an incurable malignancy for which there are no generally accepted therapeutic standards, thus novel approaches are urgently needed. The B-cell receptor (BCR) signalling pathway plays an important role in the pathogenesis of several histological types of non-Hodgkin lymphomas, including MCL (Chiorazzi et al, 2005; Kuppers, 2005; Irish et al, 2006; Jares et al, 2007; Parekh et al, 2011; Perez-Galan et al, 2011; Deng et al, 2012; Young & Staudt, 2013). In normal B cells, antigen engagement of the BCR brings immunoglobulin Igα and Igβ into proximity to activate their immunoreceptor tyrosine-based activation motifs (ITAMs), which in turn recruits SRC tyrosine kinase LYN and spleen tyrosine kinase (SYK) from the cytoplasm to the plasma membrane (Dal Porto et al, 2004; Kurosaki et al, 2010; Buggy & Elias, 2012; Macias-Perez & Flinn, 2012). LYN and SYK subsequently phosphorylate Bruton tyrosine kinase (BTK), which in turn phosphorylates phospholipase C gamma 2 (PLCG2). The lipase then cleaves membrane phosphatidyl-inositol 4,5 bisphosphate into inositol triphosphate (IP3) and diacylglycerol (DAG), which subsequently mobilizes calcium and activates protein kinase C beta (PKCβ). Several downstream proteins, including AKT, NFκB and MAPKs, are activated in turn (Fluckiger et al, 1998; Jiang et al, 1998; Humphries et al, 2004). These signalling cascades eventually lead to increased cell survival, proliferation, differentiation and migration (Dal Porto et al, 2004; de Gorter et al, 2007; Harwood & Batista, 2010; Kurosaki et al, 2010).

Several lines of evidence suggest that BCR signalling is important in the pathogenesis of MCL. Previously, activated PKCβ was found in a panel of primary MCL cells (Boyd et al, 2009). In addition, SYK DNA amplification and RNA/protein overexpression was present in the Jeko-1 cell line as well as a subset of primary MCL cells as revealed by gene expression profiling (Rinaldi et al, 2006). More recently, the BCR signalosome proteins, BTK, LYN and SYK, were identified as the most abundant tyrosine-phosphorylated proteins in MCL cell lines and BCR signalling was found to be active in both cell lines and primary tumour tissues (Pighi et al, 2011). Moreover, sequencing of the immunoglobulin heavy chain variable region (IGHV) genes of more than 800 primary MCL cases revealed a biased repertoire and stereotyped BCR usage in nearly 50% of cases, implying MCL may be an antigen-driven process (Hadzidimitriou et al, 2011). Thus, blocking BCR signalling represents a promising approach for the treatment of MCL patients who do not respond well to conventional therapies.

Ibrutinib is an orally available small molecule that binds covalently to the active site of BTK at cysteine 481 and inhibits BTK's enzymatic and downstream signalling activities (Pan et al, 2007; Honigberg et al, 2010). Very recently, an international multicentre phase 2 trial of ibrutinib in patients with relapsed or refractory MCL demonstrated a response rate of 68% with a median response duration close to 18 months (Wang et al, 2013). The drug was approved by the US Food and Drug Administration in November 2013 for the treatment of patients with MCL who have received at least one prior therapy.

Despite the remarkable activity of ibrutinib, approximately 30% of patients with MCL have primary resistance to the drug while some patients appear to develop secondary resistance. Currently, it is largely unknown why some MCL patients respond and others resist ibrutinib treatment. Understanding the underlying molecular mechanism is of paramount importance. In this study, we investigate the molecular mechanisms underlying cellular sensitivity and resistance of MCL cells to BTK inhibition with ibrutinib.

Materials and methods

Cell lines and reagents

Human MCL cell lines Jeko-1 and Mino were purchased from the American Type Culture Collection (Manassas, VA, USA) and authenticated in April 2014 by the same supplier. Granta-519 cells were obtained from the Deutsche Sammlung von Mikroorganismen und Zellkulturen GmbH (Braunschweig, Germany). Jeko-1 and Granta-519 were cultured in RPMI 1640 medium supplemented with 10% Fetal Bovine Serum (FBS) and 100 μg/ml penicillin/streptomycin. Mino cells were cultured in RPMI 1640 medium supplemented with 20% FBS and 100 μg/ml penicillin/streptomycin. Cells were maintained in a humidified 37°C/5% CO2 incubator. Ibrutinib was purchased from Selleckchem (Houston, TX, USA). Ten millimolar stock solution was divided into aliquots, and stored at −20°C for later use.

Preparation of primary MCL cells

Frozen primary MCL cells were obtained from the tumour bank of the Department of Pathology at Weill Cornell Medical College after Institutional Review Board review and approval. All primary cells were frozen in RPMI 1640 medium supplemented with 20% FBS and 5% dimethyl sulfoxide, and stored in liquid nitrogen. Only samples with >85% lymphoma cells were selected for the study. Aliquots of cells were thawed at 37°C for 1–2 min followed by the addition of 100 μl DNase and then incubated at 37°C for 90 s. Cell pellets were obtained by centrifugation at 500 g for 5 min at room temperature, and washed twice in ice-cold RPMI 1640 medium added dropwise. Cells were then re-suspended in 5 ml of room temperature RPMI 1640 medium, and viable cells were counted by Muse (Millipore, Billerica, MA, USA). Cells were then cultured in RPMI 1640 medium supplemented with 20% FBS at 37°C for 1 h in the presence or absence of ibrutinib followed by BCR stimulation using 5 μg/ml anti-IgM antibody at 37°C for 10 min. The clinical characteristics of these patients are listed in Table SI.

Isolation of normal human B cell

Normal B cells were isolated from the buffy coat of the healthy donors obtained from the New York Blood Bank using the EasySep Human B cell Enrichment kit (StemCell Technologies Inc. Vancouver, Canada). The purity of the enriched normal B-cell preparations was >96% by CD19 staining.

Immunoblotting assays

Whole cell extracts were prepared by lysing the cells in radioimmunoprecipitation assay buffer [50 mmol/l Tris-HCl, pH 7·5, 150 mmol/l NaCl, 1% Triton X-100, 0·5% Sodium deoxycholate, 0·5% sodium dodecyl sulfate, 2 mmol/l EDTA, 1 × protease inhibitor cocktail II and III (Calbiochem, San Diego, CA, USA), and 1 × phosphotease inhibitor (Sigma-Aldrich, St Louis, MO, USA)]. Proteins were separated on 4–12% NuPAGE gels and transferred onto Immobolin-FL polyvinylidene difluoride membranes (Millipore). The membranes were then probed with antibodies for BTK (BD Biosciences, San Jose, CA, USA), phospho-BTK (Y223), p-RB (Cell Signaling, Danvers, MA, USA), MYC, CCND1 (Santa Cruz, Dallas, TX, USA) and β-actin (Sigma-Aldrich) at 4°C overnight, incubated with IRDye 800CW goat anti-mouse IgG and IRDye 680 goat anti-rabbit IgG (Li-Cor, Lincoln, NE) at room temperature for 1 h, and were scanned with Odyssey imager (Li-Cor). Levels of protein or phosphor-protein expression were normalized to β-actin.

Cell metabolic activity, cell growth and viability determination

MCL cell lines were treated with various concentrations of ibrutinib. The metabolic activities of cells were measured following 72 h ibrutinib treatment with MTT assay per manufacturer's instruction (Roche Applied Science, Indianapolis, IN, USA). IC50 was calculated using the Sigma Plot generated with Prism 5 (GraphPad, La Jolla, CA, USA). Cells were collected every 24 h and re-suspended in 300 μl of florescence-activated cell sorting (FACS) buffer (PBS containing 0·5% FBS) with 2 μg/ml propidium iodide (PI). Cell viability and cell numbers were determined by flow cytometry. Cell growth and viability graphs were generated with Prism 5 (GraphPad).

Cell cycle analysis

MCL cells were treated with various concentrations of ibrutinib for 48 h. Cells were incubated with 10 μmol/l 5-bromo-2-deoxyuridine (BrdU; BD Biosciences) at 37°C for 4 h, and stained with phycoerythrin-conjugated anti-BrdU antibody (BD Biosciences) according to the supplier's manual. The percentage of cell cycle distribution was measured by flowJo (Tree Star Inc. Ashland, OR, USA).

siRNA transfection

The siRNA against human BTK was customized and purchased from Thermo Scientific (Waltham, MA, USA). The sense strand sequence of BTK siRNA was: 5′-GUAUGAGUAUGACUUUGAAUU-3′, antisense sequence was: 5′-UUCAAAGUCAUACUCAUAGUU-3′. A non-targeting siRNA was included as a negative control. Jeko-1 cells were electroporated using Amaxa® cell line Nucleofactor® kit V with X-001 Nucleofactor program. Mino and Granta-519 cells were eletroprorated using Nucleofactor kit T with program T-001 and O-017 prior to further analyses.

Intracellular phosphospecific flow cytometry assays

About 5 × 105 MCL cells were treated with ibrutinib at 37°C for 1 h. Cells were then re-suspended in 500 μl of RPMI 1640 medium with 10% FBS, and stimulated with 5 μg/ml of goat F (ab') 2 anti-human IgM antibody (Southern Biotech, Birmingham, AL, USA) at 37°C for 15 min. Cells were fixed in 4% formaldehyde for 10 min at room temperature and permeabilized with 100% methanol on ice for 20 min. Cells were then incubated with p-AKT (S473) (193H12) (Alexa®647) and p-ERK (T202/Y204) (D13, 14, 4E) (Alexa®488) XP Rabbit monoclonal antibodies (Cell Signaling) at room temperature for 1 h, and subjected to flow cytometry analysis.

Statistical analyses

Analyses were performed and graphs were generated using GraphPad Prism software (GraphPad). An unpaired student T test was performed to compare the phosphorylated or total BTK levels between MCL patient and normal control samples, or the MFI fold changes between untreated and treated MCL cell lines/primary cells. The relationships between the percentage of cell death response and inhibition of ERK1/2 or AKT phosphorylation were analysed by Spearman correlation.

Results

Primary MCL cells display a higher BTK activity than normal B cells

Expression levels and activity of BTK has not yet been thoroughly examined in primary MCL cells. To demonstrate the rationale of targeting BTK in MCL, we compared MCL primary tumour cells (n = 11) with B cells from healthy donors (n = 5) for their levels of total BTK and phosphorylated BTK. Upon BCR stimulation, tyrosine 223 (Y223) of BTK is auto-phosphorylated, which subsequently activates the enzymatic activity of BTK (Honigberg et al, 2010). p-BTK at Y223 is thus used as a surrogate marker of BTK activity. Figure 1 shows that total BTK was similarly expressed in primary MCL cells and resting B cells from the peripheral blood of control subjects (Fig 1A, B). However, p-BTK (Y223) was expressed at significantly higher levels in primary MCL cells compared to resting B cells (P = 0·0001; Fig 1A, C). Among three MCL cell lines with similar total BTK expression, Jeko-1 had the highest level of p-BTK (Y223), whereas Mino and Granta-519 cells displayed lower levels of p-BTK. Overall, these results indicated that BTK is constitutively active in MCL. These observations support the rationale for therapeutically targeting the BCR pathway in MCL.

Figure 1.

BTK is more active in MCL primary tumour and cell lines than in resting B cells. (A) Western blotting analysis of phosphorylated and total BTK expression in MCL patient samples (N = 11), normal B cells from healthy donors (N = 5) and MCL cell lines (N = 3). β-actin was included as a loading control. An unpaired student T test was performed to compare the relative expression of total BTK (B) or phosphorylated-BTK (C) in different groups. t-BTK, total BTK; p-BTK, phosphorylated BTK; MCL, mantle cell lymphoma.

MCL cell lines show differential sensitivity to BTK inhibition

To determine the sensitivity of MCL to BTK inhibition and the consequences of such inhibition, we first evaluated the activity of ibrutinib in MCL cell lines. A cell-based MTT (3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide) assay, which reflects cellular metabolic activity, was performed following 72 h of ibrutinib treatment at varying doses (Fig 2A). Among three cell lines, Jeko-1 showed the highest sensitivity to ibrutinib with a 50% inhibitory concentration (IC50) of 0·115 μmol/l. Mino had an intermediate sensitivity with IC50 of 0·491 μmol/l and Granta-519 was the least sensitive cell line with no detectable inhibition until the concentration of ibrutinib reached 16 μmol/l (Fig 2A). Given that the pharmacokinetic and pharmacodynamic studies derived from a human trial suggested that the in vivo maximal achievable concentration is 0·408 μmol/l with a once daily dose of 560 mg (Advani et al, 2013), we classify Jeko-1 as a sensitive cell line, and Mino, Granta-519 as resistant cell lines. These results suggest that standard, clinically achievable doses of ibrutinib elicit anti-lymphoma activity in some, but not in all MCL cells.

Figure 2.

Ibrutinib inhibits metabolic activity and cell growth of sensitive but not resistant MCL cells. (A) Mantle cell lymphoma (MCL) cell lines were treated with indicated doses of ibrutinib for 72 h followed by MTT assay. 50% inhibitory concentration (IC50) values were then calculated using GraphPad Prism 5 software (GraphPad). (B) MCL cells were treated with indicated doses of ibrutinib for up to 96 h. Cell numbers were determined at the indicated time points and normalized to vehicle treated control. Error bars represent the standard error of the mean (SEM) from technical replicates.

We next investigated the effects of ibrutinib on MCL cell growth by counting cell numbers over 96 h of ibrutinib treatment. The growth of Jeko-1 cells was drastically reduced by ibrutinib in dose- and time-dependent manners (Fig 2B). In contrast, some inhibition was observed in Mino cells, whereas little inhibition was observed in Granta-519 cells. These results are in agreement with the MTT data demonstrating differential sensitivity of MCL cells to BTK inhibition.

MCL cell lines show differential survival and cell cycling response to BTK inhibition

The effects of ibrutinib on cell numbers or cellular metabolism can result from either cell death or cell cycle inhibition or both. We first analysed the effects of ibrutinib on MCL cell viability. Using PI staining, we showed that the three MCL cell lines displayed different survival response to ibrutinib (Fig 3A). In Jeko-1 cells, significant cell death was induced by ibrutinib in a dose- and time-dependent manner. Meanwhile, cell death was seen in Mino cells only upon treatment with 20 μmol/l of ibrutinib. In contrast, Granta-519 cell survival was minimally affected by ibrutinib treatment.

Figure 3.

Ibrutinib reduces survival and proliferation of sensitive but not resistant MCL cells. (A) Viability of mantle cell lymphoma (MCL) cells treated with indicated doses of ibrutinib for up to 96 h. Cells were collected at indicated time points, and stained with propidium iodide. Cell viability was measured using flow cytometry, and shown as the percentage relative to vehicle-treated control. (B) Cell cycle analysis of MCL cells treated with indicated doses of ibrutinib for 48 h. Cells were labelled with 10 μmol/l 5-bromo-2-deoxyuridine (BrdU) for 4 h, followed by intracellular staining with BrdU and 7-aminoactinomycin D and flow cytometric analysis. Percentages of events at different stages in the cell cycle are shown. (C) Western blotting analyses for cell cycle regulatory proteins. MCL cells were treated with 0·2 or 1 μmol/l Ibrutinib for 48 h. Western blotting was performed using antibodies for MYC, p-RB, CCND1 and β-actin. One-way anova tests were used to analyse the differences in band intensity between treated and untreated cells. Bar graph was generated using GraphPad Prism 5 (GraphPad). Error bars represent the SEM from three independent experiments. *, P ≤ 0·05; **, P ≤ 0·01; ns, not significant.

Previously, we found that ibrutinib primarily targets proliferation in CLL (Cheng et al, 2014). We thus investigated the effect of ibrutinib on the cell cycle of the MCL tumour cells. In Jeko-1 cells, the percentage of S phase fraction was reduced from 27% (untreated control cells) to 16·6% following treatment with 0·2 μmol/l of ibrutinib for 48 h (Fig 3B). This was accompanied by a concomitant increase in G0/G1 phase. In comparison, the distribution of cells among cell cycle stage remained unchanged in Mino as well as in Granta-519 cells (Fig 3B).

To corroborate this finding, we tested whether CCND1 along with other cell cycle regulatory proteins were affected by ibrutinib. The expression of phosphorylated RB (p-RB), CCND1 and MYC proteins were examined by immunoblotting assays. Phosphorylated RB and CCND1 protein levels in Jeko-1 cells were down-regulated upon treatment with increasing doses of ibrutinib (Fig 3C, top and bottom panels, P < 0·05 and P < 0·01). Meanwhile, MYC level was also significantly reduced by ibrutinib (Fig 3C, top and bottom panels, P < 0·05). In contrast, no significant reduction in p-RB, CCND1 or MYC protein levels was observed in Mino and Granta-519, the two resistant cell lines. Together, the data demonstrates that MCL cell lines show a differential cell cycling response to BTK inhibition.

Collectively, these data suggest MCL cells have different sensitivity to ibrutinib in terms of both cell viability and cell cycle that largely correlates with cellular metabolic activities (MTT assay) and cell growth in these cell lines.

Genetic knockdown of BTK inhibits the growth, survival and proliferation of sensitive but not resistant MCL cell lines

To demonstrate that the effects of ibrutinib were mediated through BTK inhibition rather than a result of non-specific targeting, we introduced BTK-specific siRNA into the cells to determine whether the effects of ibrutinib can be reproduced with RNA interference. BTK specific or control siRNA was transiently transfected into the three MCL cell lines and Western blotting assays were performed to determine the BTK knockdown efficiency. Significant reduction in both total and p-BTK was achieved in siBTK-transfected cells compared to the controls (Fig 4A). BTK knockdown inhibited cell growth in Jeko-1, but not in Mino or Granta-519 cells (Fig 4B). Knockdown of BTK also led to reduced viability in Jeko-1 (Fig 4C, P < 0·05) but not in Mino or Granta-519 cells. Further analysis of the cell cycle demonstrated that BTK reduction suppressed the G1/S transition (S phase fraction reduced from 51·5% to 37·3%) in Jeko-1 but not in Mino or Granta-519 cells (Fig 4D). Together, these results provide strong evidence that BTK is required for lymphoma cell viability, growth and proliferation. The data further suggests that ibrutinib acts through BTK to produce its anti-lymphoma effects.

Figure 4.

BTK siRNA inhibits cell growth, viability and proliferation in sensitive but not resistant MCL cells. (A) Mantle cell lymphoma (MCL) cells were transfected with either control siRNA or BTK siRNA. Immunoblotting for total BTK (t-BTK) and phosphorylated BTK (p-BTK) protein was carried out at 72 h post-transfection. β-actin was included as a loading control. (B) Cell growth of transfected cells at indicated time course following transfection. (C) Viability of transfected cells at 72 h post-transfection. (D) Analysis of cell cycle of transfected cells at 72 h post-transfection. Error bars represent the SEM from three independent experiments. An unpaired student T test was performed to compare the viability and percentage of cell cycle phase distribution in siBTK and vehicle treated cells. *, P ≤ 0·05; ns, not significant; BrdU, 5-bromo-2-deoxyuridine; 7-AAD, 7-aminoactinomycin D.

Inhibition of ERK1/2 and AKT phosphorylation, but not BTK phosphorylation, correlates with cellular response to BTK inhibition in MCL cell lines

To understand the mechanism for ibrutinib sensitivity/resistance, we first checked all cell lines for a previously reported ibrutinib-resistant BTKC481S mutation using an established Taqman genotyping assay as well as the PLCG2R665W mutation using Sanger sequencing (Chang et al, 2013a; Furman et al, 2013, 2014). The mutations were not detected in any of the three cell lines (data not shown). We then determined whether the phosphorylation of BTK responds to ibrutinib. We performed immunoblotting assays in MCL cells treated with or without ibrutinib. Phosphorylated BTK (Y223) expression was reduced by ibrutinib not only in the sensitive cell line Jeko-1 (Fig 5A, left panel) but also in resistant cell lines Mino and Granta-519 (right and bottom panels). As a control, total BTK remained unchanged by the treatment (Fig 5A). These results suggest that ibrutinib indeed directly inhibits the activity of BTK in MCL cells. However, this inhibition does not translate to cellular responses in terms of cell growth, viability and proliferation. Thus, the determinant of sensitivity versus resistance does not reside at the BTK level.

Figure 5.

Phosphorylation of ERK1/2 and AKT, but not BTK, was inhibited in sensitive but not resistant MCL cells. (A) Western blotting of BTK phosphorylation. Mantle cell lymphoma (MCL) cells were treated with indicated doses of ibrutinib, cell lysates were collected at 1 or 4 h, and subjected to Western blotting analysis using phosphorylated BTK (p-BTK) (Y223) and total BTK (t-BTK) antibodies. (B) Phospho-specific flow cytometric assays. Three MCL cell lines were treated with 0·2 μmol/l ibrutinib for 1 h followed by BCR stimulation using 5 μg/ml anti-IgM antibody: Left column, p-ERK1/2 (T202/Y204) and right column, p-AKT (S473). (C) Quantification of ERK and AKT phosphorylation. MFI, mean fluorescence intensity of 10 000 events. Error bars represent the SEM from three independent experiments. *, P ≤ 0·05; ns, not significant.

To see whether other signalling events better correlate with cellular responses of cells to the drug, we measured the phosphorylation levels of ERK and AKT in ibrutinib-treated MCL cells using phospho-flow assays that are sensitive and quantitative (Yang et al, 2008; Lu et al, 2009; Song et al, 2010; Cheng et al, 2011). Phosphorylation of ERK in BCR-stimulated Jeko-1 cells was inhibited by 0·2 μmol/l ibrutinib compared to the vehicle-treated cells (Fig 5B, top left panel). p-AKT inhibition was similarly observed in Jeko-1 cells (Fig 5B, top right panel). In contrast, inhibition of p-ERK or p-AKT was less pronounced in Mino or Granta-519 cells (Fig 5B, middle and bottom panels). Quantitative analysis of the phosphorylation data revealed a significant reduction of either p-ERK1/2 (Fig 5C, left panel) or p-AKT (Fig 5C, right panel) in Jeko-1 cells, but not in Mino or Granta-519 cells. Overall, these results demonstrate that the activity of ERK and AKT, but not BTK, correlate with the cellular responses of MCL cells to ibrutinib, suggesting events downstream from BTK may be important for the regulation of cell proliferation and apoptosis.

Inhibition of ERK1/2 and AKT activities correlate with cell death response in primary MCL cells

To determine whether data obtained with the cell lines are reproducible with primary tumour cells, we tested 10 human MCL samples (Table SI) for their apoptotic response to 1 μmol/l ibrutinib. These 10 primary MCL samples displayed varying degree of apoptotic response to ibrutinib treatment that ranged from 23–72% with a median response at 41% (Fig 6A).

Figure 6.

In primary MCL tumour cells, phosphorylation of ERK1/2 and AKT predicts apoptotic response to ibrutinib. Primary mantle cell lymphoma (MCL) cells were treated with indicated doses of ibrutinib for 1 h prior to BCR stimulation. (A) Apoptotic response of 10 primary MCL cells. The median response was at 41%. (B) BTK immunoblotting assay in tumour cells with different cell death response. β-actin was included as a loading control. (C) Phospho-flow assays for p-ERK1/2 and p-AKT in two representative cases that have different cell death response. (D) Phospho-flow assays for p-ERK1/2 and p-AKT in all 10 cases treated with different doses of ibrutinib. Mean fluorescence intensity (MFI) was plotted. (E) Inhibition of ERK1/2 and AKT phosphorylation is linearly correlated with the cell death response following treatment with 1 μmol/l ibrutinib. Spearman correlation was applied to analyse the relationship between the percentage of cell death response and percentage of inhibition of phosphorylation using the formula: [(MFIuntreated−MFItreated)/MFIuntreated] × 100%.

We next determined whether the signalling activities correlate with this cellular response in ibrutinib-treated primary cells. Immunoblotting analysis showed that p-BTK was inhibited in all tested samples regardless of their degree of apoptotic response (Fig 6B). For example, MCL05, the sample with the least apoptotic response, had a similar degree of p-BTK inhibition as MCL10, the sample with the most apoptotic response (Fig 6A, B).

We subsequently determined the degree of p-ERK and p-AKT inhibition in response to ibrutinib treatment in these primary patient samples. Two representative cases are presented in Fig 6C. In response to increasing doses of ibrutinib, there was a marked and simultaneous down-regulation of both p-ERK and p-AKT in MCL09 (Fig 6C, top panels). Meanwhile, there was minimal inhibitory response in MCL05 (Fig 6C, bottom panels). Inhibition of p-ERK and p-AKT was highly variable across all 10 primary samples (Fig 6D). However, a linear correlation was found between the degree of p-ERK inhibition and the cell death response (Fig 6E, left, P = 0·0087). Similarly, a significant linear correlation was also found between the degree of p-AKT inhibition and the cell death response (Fig 6E, right, P = 0·0102). These data corroborate our findings with the cell lines: although p-BTK was inhibited in all MCL cells, only the degree of p-ERK and p-AKT inhibition correlated well with the final cellular outcomes. These data suggest that both ERK and AKT play essential roles for ibrutinib to achieve its therapeutic effects.

Discussion

Several other groups have investigated the effects of ibrutinib in MCL. Chang et al (2013b) characterized the MCL cells mobilized from the lymphoid tissue to the peripheral blood in ibrutinib-treated MCL patients who experienced peripheral lymphocytosis. Dasmahapatra et al (2013) showed that ibrutinib increased bortezomib-induced mitochondrial injury and apoptosis in MCL cell lines. However, ibrutinib by itself had insignificant effects in MCL cells, including Granta-519, at a concentration as high as 7·5 μmol/l. Meanwhile, another study showed that apoptosis was induced in both Jeko-1 and Mino cells with 10 or 20 μmol/l of ibrutinib (Cinar et al, 2013). We, however, approached the issue somewhat differently. Using 0·4 μmol/l as the cut-off, we defined the MCL cell lines as either sensitive (Jeko-1) or resistant (Mino and Granta-519). The concentration (0·4 μmol/l) was chosen because it is the highest concentration clinically achievable in patients treated with a daily dose of 560 mg ibrutinib (Advani et al, 2013). We then focused on the differences in the molecular and cellular behaviours between ibrutinib-sensitive and ibrutinib-resistant MCL cells that were not addressed previously. Given approximately one-third of the MCL patients have primary resistance to ibrutinib and a fraction of responding patients develop secondary resistance, understanding the molecular mechanisms underlying the cellular sensitivity is of particular importance.

Several findings in the current study have not been described previously: (i) We demonstrated that there was a higher level of BTK phosphorylation in primary MCL cells and cell lines than in normal resting B cells. Higher p-BTK renders lymphoma cells more susceptible to BTK inhibition than normal B-cells, supporting the rationale of targeting BTK in MCL; (ii) MCL cell lines had differential responses to ibrutinib. In particular, Jeko-1 cells are sensitive while Mino and Granta-519 cells are resistant to ibrutinib treatment; (iii) The sensitive Jeko-1 cells respond to ibrutinib with apoptosis and cell cycle arrest at clinically achievable concentrations in a dose-dependent manner; (iv) Genetic knocking-down of BTK in MCL cells generated effects similar to ibrutinib on cell growth and cell cycle, suggesting that these cellular outcomes were indeed mediated by specific BTK inhibition; (v) Further dissection of the signalling pathway revealed that, although phosphorylation of BTK was universally inhibited by ibrutinib in all cell lines, only the degree of either p-ERK or p-AKT inhibition correlated well with cellular sensitivities; and (vi) The cell line observations were validated with primary patient samples, in which the p-BTK was inhibited in all cases tested, but only p-ERK or p-AKT inhibition exhibited a linear correlation with the cellular death response caused by ibrutinib.

These data reveal that ERK and AKT remain active in a subset of MCL tumours that enable cells continue to survive and grow even in the presence of BTK inhibition. The results imply that alternative upstream pathways or kinases other than BTK mediate the activation of these downstream events. Further studies are definitely required to understand the precise mechanisms that keep ERK or AKT active in such ibrutinib-resistant MCL tumours.

Similar situations are present in diffuse large B cell lymphoma (DLBCL). Using a genomic approach, both pre-clinical and clinical studies have shown that when comparing the germinal centre B (GCB) with the activate B-cell (ABC) subtype, the ABC subtype, which displays chronic active BCR signalling due to mutations in the components of the BCR pathway, respond better to BTK inhibition (Davis et al, 2010; Wilson et al, 2012; Yang et al, 2012). Further, in the ABC subtype, when cases carry upstream BCR mutations were compared with the ones carry downstream mutations, only those carry upstream mutations respond well to BTK inhibition (Davis et al, 2010). Similar genomic and genetic analyses may also be useful in MCL to unveil the mutations or other types of genetic abnormalities that contribute to ibrutinib sensitivity.

Through the current investigation, it has become apparent that simply inhibiting BTK is not sufficient to suppress the entire activity of the BCR signalling pathway in every MCL tumour. As a matter of fact, acquired mutations in BTK and PLCG2 have been identified in CLL patients receiving ibrutinib both by our group and others (Chang et al, 2013a; Furman et al, 2013, 2014). We further demonstrated that the BTKC481S mutation affects ibrutinib binding, reactivates BCR signalling downstream of BTK, and increases the expression of a panel of BCR-targeted genes (Furman et al, 2013, 2014).

Acquiring of mutations or re-wiring of the BCR pathway to retain the downstream signalling also appears to be a common theme for resistance to other BCR-targeting agents in other types of non-Hodgkin lymphoma as well. In earlier studies, we demonstrated that the activities of particular BCR downstream signalling events determine the response of CLL to dasatinib (primarily targets LYN) (Song et al, 2010), DLBCL to dasatinib (Yang et al, 2008) and DLBCL to PRT060318 (selectively targets SYK) (Cheng et al, 2011). In conclusion, these studies suggest that to prevent the primary resistance or to overcome the secondary resistance to molecularly targeted therapies, a strategy involving the combined targeting of both upstream and downstream events or targeting other pathways in addition to BCR would be the most rational approach.

Acknowledgements

The authors would like to thank Dr. Wayne Tam and Ms. Sharon S. Barouk-Fox for providing the primary MCL frozen samples. We would also like to thank Mr. Chaojie Zhen for manuscript editing.

Authorship

J.M. and P.L. developed the assays, designed and performed the experiments, solved technical problems, analysed the data, and wrote part of the manuscript. A.G., S.C. and H.Z. performed the experiments and analysed the data, P.M. and M.C. contributed useful discussions and suggestions. Y.L.W. formed the hypothesis, directed and coordinated the project, designed the experiments, analysed the data, and wrote the manuscript.

Conflict of interests

P.M. received honoraria from Janssen Pharmaceuticals. The other authors declare no conflicts of interests for this study.

Ancillary