PBMC, MCL cell lines Jeko-1, Rec-1, Granta-519 and primary PBMC from patients were isolated and cultured as previously described (25). Malignant and healthy B cells were isolated with EasySep™ kits (STEMCELL Technologies, USA), after achieving informed consent according to the declaration of Helsinki, with approval of the internal Institutional Board (protocol # 4089/AO/17). The clinical features of the patients analyzed are described in Table 1 .

IPTG was from Sigma-Aldrich, (Italy); Ibrutinib, Duvelisib, Bortezomib and Z-VAD-FMK was from Selleck chemicals (USA); D4476 (4-[4-(2,3-Dihydro-1,4-benzodioxin-6-yl)-5-(2-pyridinyl)-1H-imidazol-2-yl]benzamide) was from abcam (UK). Working dilution of D4476 was prepared using Fugene VI reagent (Promega, Italy).

Cell viability was measured through Trypan blue exclusion dye assay. Apoptosis was assessed by Annexin V/Propidium Iodide (PI) staining (IMMUNOSTEP, Spain) and FACS analysis as in (17).

It was performed as described in (17).

Jeko-1 and Granta-519 cells were plated into 96 well plates in 100 μl media. D4476, Ibrutinib, and Duvelisib were added at different concentrations for 72 h alone or in combination. Cell viability was analyzed with 3-(4,5-dimethylthiazol-2-yl)- 2,5-diphenyltetrazolium bromide (MTT) and the CI was calculated as in (26).

It was performed as in (26) with the following antibodies: anti-CK1α (abcam, UK), secondary Alexa-Fluor 594-conjugated goat anti-rabbit (Life technologies, Italy), MALT1 and CARD11, (Santa Cruz Biotechnology, Inc (USA) using Alexa-Fluor 488 conjugated anti-mouse. Specimens were mounted in Vectashield medium with DAPI (Vector Laboratories, USA) and analyzed using Zeiss LSM 700 E90 confocal microscope, oil objective 63x (Italy).

WB was performed as described (27). Antibodies used were the following: CK1α, PARP, Mcl1, total β-catenin, Ser 473 AKT, total AKT, Ser 176/180, Ser 177/181 IKKα/α, total IKKα, IKKβ, Ser 536 NF-κB p65, Ser 652 CARD11, Tyr 223 BTK, total BTK, Ser 32 Ikbα, BCL10 (Cell signaling Technology, MA, USA); GAPDH (Ambion, USA), β-actin (Sigma-Aldrich, Italy); p21 (Becton Dickinson, Italy); Caspase 3 (Enzo Life Science, UK); total p65 (abcam, UK), DEPTOR (Millipore, Itlay); CARD11 and BTK for immunoprecipitation (Santa Cruz Biotechnology, Inc; Italy). Images were acquired using the Image Quant LAS 500 chemiluminescence detection system (GE Healthcare, USA).

Cells were lysed in immunoprecipitation (IP) buffer containing Tris 50mM, NaCl 150 mM, Triton 1%, NP-40 1%, EDTA 2mM, phosphatases and proteases inhibitors (Life technologies and SIGMA Aldrich). 500-800 µg of total protein lysates were precleared for 30 min at 4°C with anti-mouse or anti-goat agarose beads (e-bioscience, Italy), according to the antibody used for the IP. After centrifugation, the supernatant was incubated overnight 4°C with the dedicated antibody (BTK, CARD11 or CK1α) or with the corresponding pre-immune serum. Agarose beads were added to the protein suspension for 2h at 4°C. After extensive washes in IP buffer the immunocomplex was resuspended in Laemmli buffer with β-mercaptoethanol and processed for WB analysis.

RNA interference (RNAi) was performed with nucleofection of double strand (ds) siRNA and by the generation of inducible shRNA MCL cell clones. Nucleofection was performed by electroporation of Jeko-1 using the Amaxa system (Lonza, Rockland, Inc.) with 100 pmol of scrambled siRNAs or CK1α targeting siRNAs ON-TARGET plus SMARTpool siRNA (Thermo Scientific, USA), as described (17) through the nucleofector solution R with the program A-023. Cells were harvested at 48h after nucleofection. Cells were simultaneously electroporated with a fluorescent oligonucleotide (siGLO green). The transfection rate, evaluated by FACS analysis, has always been around 95%. To perform shRNA lentiviral transduction, Jeko-1 and Granta-519 cells were transduced with the IPTG inducible lentiviral particles carrying CSNK1A1-specific shRNA (pLKO_IPTG_3XLacO, Sigma-Aldrich, Italy) with the sequence TRCN0000006044. 2x10⁴ cells were infected with a multiplicity of infection of 12, using the spinfection method, in the presence of 8µg/ml polybrene (Sigma-Aldrich, Italy). Puromycin selection (1µg/ml) was initiated two days after transduction. A titration curve for puromycin resistance was performed for Jeko-1 and Granta-519 cells by an Antibiotic Kill Curve Assay. Once a cellular clone was established, to induce CK1α silencing, cells were incubated with 500/1000µM IPTG every two/three day for a total of one (in the case of Granta-519) or two (in the case of Jeko-1) weeks, time lapse in which the best knockdown efficacy was obtained.

Performed as in (25) using the QuantStudio 5 detection system (Applied Biosystem, CA, USA) with the QuantStudio™ Design and Analysis Software v.1.4.3. The primers used are the following: BTK Forward 5’-3’ GGGGTTTGCTCAGACTGTCC and Reverse 5’-3’ AATCACTGCGGCCATAGCTT; RELA Forward 5’-3’ CCCCACGAGCTTGTAGGAAAG and Reverse 5’-3’ CCAGGTTCTGGAAACTGTGGAT; BIRC3 Forward 5’-3’ GACAGGAGTTCATCCGTCAAG and Reverse 5’-3’ TTCCACGGCAGCATTAAT; GAPDH Forward 5’-3’ AATGGAAATCCCATCACCATCT and Reverse 5’-3’ CGCCCCACTTGATTTTGG.

Data were examined for their statistical significance with the two-tail unpaired Student’s t test or ANOVA analysis of variance with post-hoc corrections. Values were considered statistically significant at p values below 0.05.

We analyzed CK1α mRNA expression in different subtypes of B cell derived cancers (through the publicly available database Oncomine), finding that CSNK1A1 mRNA is highly expressed in MCL human samples as compared to normal B lymphocytes (www.oncomine.org, “Basso Lymphoma” data set). Thus, we evaluated CK1α protein level in purified B lymphocytes from healthy controls, MCL patients and the MCL cell lines Jeko-1, Granta-519 and Rec-1.

Immunoblot and densitometric analysis revealed that all the MCL cell lines and most of MCL patients overexpressed CK1α when compared to normal B lymphocytes, even if with some variability between different MCL samples ( Figure 1A ). Among a total of fifteen MCL patients analyzed, eleven showed CK1α overexpression in malignant B cells at variance from what seen in B cells from healthy controls. CK1α was overexpressed in 100% of patients in Figure 1A left, in 80% of patients in Figure 1A middle, and in 50% of patients in Figure 1A right. On average, 77% of MCL patients and 100% of the MCL cell lines tested showed overexpressed CK1α at the protein level.

To study the localization of the kinase inside B cells, we performed a confocal immunofluorescence analysis in purified B lymphocytes from a healthy control, a MCL patient and in Jeko-1 cells. CK1α localized both in the cytoplasm and in the nucleus of Jeko-1 and MCL patient cells while it was restricted in cytoplasm in healthy B lymphocytes ( Figure 1B ). These data, which are similar to what we observed in another B cell malignancy like MM (17), suggest an abnormal localization of CK1α in MCL cells.

To examine if CK1α is essential for MCL cell growth, Jeko-1, Granta-519, Rec-1 and B cells isolated from 7 MCL patients, were treated with increasing concentrations of the CK1 chemical inhibitor D4476 or with the vehicle DMSO. Apoptosis was investigated through Annexin V (AV) and PI labelling and FACS analysis. Treatment with DMSO alone did not produce any significant changes in cell viability compared to untreated cells. All the MCL cell lines were sensitive to D4476, starting from the 20 µM concentration for Granta-519 and Rec-1 cells and 30 µM for Jeko-1 ( Figures 2A , S1 ). Even if with some variability, malignant B cells from most of MCL patients proved to be sensitive to D4476 (6 out of 7) used at 40 µM. Of note, B cells from 3 patients were responsive also to the 20 µM D4476 concentration ( Figure 2B ). We previously showed that D4476 is not toxic for healthy B lymphocytes at any concentration used (17).

To validate the results obtained with the chemical inhibitor of CK1 and to investigate the specific role of the α isoform, we used RNAi to knockdown CK1α protein in Jeko-1 cells. CK1α knockdown was obtained through the generation of IPTG-inducible CSNK1A1 specific shRNAs or with electroporation of double-strand (ds) CSNK1A1 directed siRNAs ( Figures 2C , S1 ). The number of apoptotic cells was assessed by FACS analysis of AV/PI positive cells, which showed increased apoptotic rate upon silencing. To exclude off-target effects induced by IPTG per se, wt Jeko-1 cells were treated with the same concentration of IPTG and this treatment did not increase apoptosis. The proapoptotic effect of CK1α inhibition or silencing was confirmed by a reduction of uncleaved PARP, pro-caspase 3 and Mcl1 ( Figure 2D ). An effective decrease of CK1α protein levels upon silencing (around 40-50% with both the techniques used) was confirmed ( Figure 2E ). As expected, IPTG treatment did not affect the expression level of CK1α and of pro- and anti-apoptotic proteins in wt Jeko-1 cells, confirming the efficacy/specificity of the silencing strategy. To evaluate the effect of CK1α silencing on MCL cell proliferation, we performed cell cycle analysis of Jeko-1 wt and CK1α–silenced clones treated with IPTG ( Figure 2F ). The pro-apoptotic effect of CK1α knockdown was confirmed by the observation of a substantial increase in the fraction of CK1α-directed inducible shRNA-bearing Jeko-1 cells in the sub-G₁ (apoptotic) phase of the cell cycle upon IPTG treatment. Alongside with the prevalent accumulation of sub-G1 (likely apoptotic) cells, CK1α silencing in the same cellular clone determined a minimal accumulation of cells in G₀/G₁ phase and a more evident reduction of cells in the S phase, suggesting an impairment also of proliferation upon CK1α knockdown. As expected, IPTG treatment of wt cells did not affect cell cycle. Moreover, CK1α silencing, through IPTG treatment for 6 days and for 13 days caused a reduction in the total count of live cells as judged by Trypan blue exclusion assay analysis, indicating a proliferation dysregulation ( Figure 2G ).

CK1α is known to be involved in the regulation of several survival signaling pathways associated with BCR to which MCL cells are addicted for their growth. Bidere et al. (21) defined CK1α as a “conditionally essential malignancy” gene demonstrating that CK1α is required for constitutive NF-κB pathway in DLBCL. To understand if this holds true also for MCL cells, we treated three different MCL cell lines with D4476 40μM or DMSO for 48h and analyzed the NF-κB signaling pathway, in particular the activating phosphorylation of CARD11 on Ser 652, of NF-κB p65 on Ser 536, of the upstream IKKα/β, and of IkBα on Ser 32. Remarkably, the results showed a significant reduction of the activation of the NF-κB pathway upon CK1 chemical inhibition ( Figure 3A ). Most importantly, the specific effects of the inhibition on NF-κB were confirmed also by RNAi, both with IPTG-inducible shRNA transduced Jeko-1 and Granta-519 cells, and with electroporation of ds CSNK1A1-directed siRNA in Jeko-1 cells ( Figures 3B, C ). Surprisingly, CK1α inactivation not only caused a decrease in the activating phosphorylations of CARD11 on Ser 652, NF-κB-p65 on Ser 536, IKKα/β on Ser 176/180, Ser 177/181, but also in their total protein levels. Consequently, CK1α loss-of-function determined a reduced mRNA expression of the NF-κB-p65 target BIRC3 ( Figure 3D ). BIRC3 encodes for c-IAP-2 (cellular inhibitor of apoptosis), which confers resistance to apoptosis.

It has previously been demonstrated that CK1α controls canonical NF-κB activation in stimulated T and in DLBCL cells, participating to the activation of the upstream CBM1 complex (CARD11, BCL10, MALT1), through the phosphorylation of CARD11 and MALT1 (21, 22). We therefore evaluated a possible interaction between CK1α and the CBM1 complex in MCL cells. We stimulated Jeko-1, Granta-519 and MCL patients B cells with anti-IgM to fully activate the BCR and NF-κB signaling, and we performed immunoprecipitation and immunofluorescence experiments using CARD11 and CK1α antibodies. We found that CK1α associates with the CBM1 complex at basal and anti-IgM stimulated conditions ( Figures 4A–C ), as evidenced by coimmunoprecipitation of CK1α and CARD11 and as suggested by cellular colocalization studies of CK1α with CARD11 and MALT1 ( Figure S2 ). Of note, CK1α silencing in Jeko-1 determined a reduction of CARD11 and BCL10 protein levels ( Figure 4D ) accompanied by a decrease in basal and anti-IgM-induced NF-κB pathway activation ( Figure 4E ). Of note, the perturbation of NF-κB signaling is achieved also within 72h (shorter) incubation time with IPTG treatment (silencing), indicating an acute/direct effect of CK1α silencing.

As previously mentioned, PI3K/AKT is an essential cascade constitutively active in a subset of MCL, including all the aggressive blastoid variants and in MCL cell lines (28). It is known that CK1α may impinge on the PI3K/AKT pathway (17). Therefore, we sought to investigate whether CK1α could regulate this cascade also in MCL. CK1 chemical inhibition with D4476 in Jeko-1, Granta-519, Rec-1 MCL cells ( Figure 5A ) and CK1α silencing through the IPTG inducible method in Jeko-1 and Granta-519 led to a reduction of the phosphorylation of AKT on Ser 473 and of total AKT ( Figures 5B, C ). Wild-type Jeko-1 or Granta-519 cells treated with IPTG did not show any changes in the phosphorylated/total AKT protein content. It is known that CK1α phosphorylates the mTOR inhibitor DEPTOR, leading to its proteasomal degradation. As expected, as a consequence of CK1α silencing, DEPTOR level increased ( Figure 5B ). The reduction in AKT activating cascade is present also in CSNK1A1 directed siRNA transiently transfected Jeko-1 cells ( Figure 5B ).

Since CK1α may have a broader regulative role on the BCR-transduced signal, we next investigated whether CK1α inhibition could affect other BCR-triggered events, such as BTK activation. Figure 5 shows that both the BTK activating phosphorylation on Tyr 223 and the total amount of the BTK protein are downregulated when CK1α is chemically inhibited or silenced, indicating that CK1α could sustain BTK activity in MCL cells.

The observed CK1α modulation of some BCR dependent signaling cascades was present also at early time points of D4476 treatment, such as 6h (activating phosphorylation of AKT and CARD11) and 24h (activating phosphorylation of AKT and BTK) ( Figure S3 ). To exclude a global rewiring of survival signaling events downstream the BCR as a direct effect of CK1α inhibition dependent cytotoxicity, and to selectively ascribe the effects observed on NF-κB, AKT and BTK cascades directly to the activity of CK1α, we also evaluated the effects of CK1α inactivation on ERK/MAP kinase signaling. To note, neither the chemical inhibition nor silencing of CK1α modified ERK 1,2 activating phosphorylation on Thr 202, Tyr 204, ( Figures 5 and S3 ), pointing to a role of CK1α in sustaining specifically NF-κB, AKT and BTK BCR dependent cascades. Altogether, the observation of early signaling changes upon CK1α inactivation indicates that they are very unlikely due to the ongoing apoptotic process. To investigate the mechanism of CK1α dependent regulation of BTK and NF-κB p65 stability, we checked whether the effects observed upon CK1α inactivation were due to transcriptional or post-translational mechanisms. Interestingly, we found that the mRNA expression of BTK and REL A (p65) did not change upon CK1α chemical inhibition or silencing, suggesting a post-translational mechanism of regulation of these two proteins ( Figures S4A, B ). We next asked whether this mechanism could be proteasome- or caspase-dependent. To this aim, the CK1α-directed shRNA transduced Jeko-1 clone was treated with Bortezomib (7.5nM), a clinically used proteasome inhibitor or with Z-VAD-FMK (2μM), a cell permeant pan-caspase inhibitor, along with IPTG to silence CK1α. Remarkably, p65 and BTK protein levels remained downregulated with the association of CK1α and bortezomib, while the CK1α silencing dependent reduction was rescued by Z-VAD-FMK treatment, suggesting that BTK and p65 degradation might be due to a caspase- rather than to a proteasome-dependent mechanism ( Figure S4C ).

Given the novel, unanticipated potential role of CK1α in regulating BTK function, we analyzed if these two kinases could physically interact. To this aim, we treated MCL cells with anti-IgM for 5 min and BTK was subsequently immunoprecipitated. Strikingly, CK1α was present in the immunocomplex ( Figure 5D ). To test specificity and BCR activation, phosphorylated AKT on Ser 473, indicative of fully activated BCR, was evaluated. Interestingly, CK1α physically interacts with BTK also at basal condition in patient isolated B cells from a particularly aggressive form of MCL. Thus, CK1α physically interacts with BTK downstream from the BCR in MCL cells.

Given the aforementioned role of CK1α on BCR related signaling cascades, we next asked whether CK1α could interfere with BCR inhibitors (such as Ibrutinib or Duvelisib) -induced MCL cell apoptosis. Since we showed an unprecedented role of CK1α in sustaining BTK levels and activity and given the known role of CK1α on the BTK downstream CBM1 complex, we reasoned that its inhibition or silencing combined with direct BTK targeting would produce cooperative/synergic cytotoxic effects. Therefore, we treated three MCL cell lines, primary MCL B cells and healthy B lymphocytes with different doses of D4476 and Ibrutinib, alone or in combination. Annexin V staining and FACS analysis showed a cooperative effect of D4476 and Ibrutinib in inducing cell death in MCL cells ( Figures 6A, B and S5A ). This did not hold true for normal B cells, which were spared by the cytotoxic activity of D4476 ( Figure 6C ). The therapeutic potential of the association of Ibrutinib with CK1α inactivation was confirmed also in the models of CK1α silencing. Treatment with IPTG of CSNK1A1-directed shRNA transduced Jeko-1 clones significantly empowered Ibrutinib induced apoptosis ( Figures 6D and S5B ). Of note IPTG-treated wt Jeko-1 cells did not show increase in apoptosis compared to Ibrutinib-only treated cells, confirming the efficacy and the specificity of the silencing method. These results were confirmed also in experiments of nucleofection of CSNK1A1-directed siRNAs in Jeko-1 cells ( Figure 6D ). The rise in apoptosis in the combination treatments was confirmed by immunoblot analysis of PARP cleavage ( Figure S5C ). Moreover, MTT viability assays in the presence of increasing concentrations of Ibrutinib, D4476 or the combination of the two drugs for 72 h showed a strong synergic effect of the two compounds, as judged by the calculated Combination Index (CI) below 1 ( Figure 6E ). Of note, the synergy was present also in Granta-519 cells, a line known to be less sensitive to Ibrutinib (IC50 in Granta-519 28 µM versus 11 µM for Jeko-1), suggesting that CK1α inhibition could partly overcome Ibrutinib resistance.

Analysis of intracellular signaling revealed that Ibrutinib cooperated with CK1 inactivation (both chemical inhibition with D4476 and gene silencing) in modulating NF-ĸB, AKT and BTK-dependent cascades by further reducing the levels of phosphorylated NF-ĸB-p65 on Ser 536, AKT on Ser 473 and BTK on Tyr 223, without affecting ERK1/2 MAP kinase activity ( Figures 7A, B ). To note, we noticed that in Ibrutinib-treated Jeko-1 cells CK1α expression was decreased. We therefore treated all the MCL cell lines available with different concentrations of Ibrutinib for 24h. Strikingly, Ibrutinib effectively caused a reduction of CK1α protein expression in Jeko-1, Rec-1 and - even if to a lower extent - in Granta-519 cells ( Figure 7C ). The effectiveness of Ibrutinib was confirmed by the abolition of phosphorylated BTK on Tyr 223.

We next tested the interaction of CK1α inhibition with Duvelisib, a PI3Kγ/δ inhibitor approved for the treatment of Chronic Lymphocytic Leukemia (CLL) and Follicular Lymphoma (FL). However, its use in MCL is still under scrutiny (29, 30).

We first verified the efficacy of Duvelisib on a panel of MCL cell lines and purified B cells from six MCL patients. Cells were treated with different doses of Duvelisib for 24 and 48 h. Duvelisib caused MCL cell apoptosis in a dose and time dependent manner. The efficacy of the drug was monitored by assessing the levels of phosphorylated AKT on Ser 473, which were always reduced in all the cells treated ( Figure S6 ). Next, the effects on apoptosis of the combination of CK1α inhibition together with Duvelisib was analyzed through Annexin V staining and FACS analysis. Jeko-1, Granta- 519, Rec-1 MCL cells were treated with D4467 and Duvelisib alone or in combination ( Figures 8A and S7 ). Also, Jeko-1 CK1α silenced cells were treated with Duvelisib ( Figure 8B ). The combination of Duvelisib and CK1α inactivation cooperatively increased the cytotoxicity induced by the single agents. Of note, this cooperation was absent in healthy normal B cells, which were spared by D4476-induced cytotoxicity ( Figure 8C ). The higher rate of apoptosis obtained in the combination experiments was confirmed also by immunoblot analysis of PARP, pro-caspase 3 and Mcl1 protein cleavage ( Figure 8D ). The effect was synergic since the calculated combination index was lower than 1 ( Figure 8E ). Of note, CK1α inactivation caused a significant reduction of the IC50 of Duvelisib in the more resistant Granta-519 cells ( Figure 8E ). Interestingly, as observed with Ibrutinib, treatment with Duvelisib at different concentrations caused a reduction of CK1α protein expression in MCL cells ( Figure S8 ).