L-428, L-540, HDLM-2, and KM-H2 HL cell lines were kindly provided from Prof. Carmelo Carlo-Stella from Humanitas Hospital (Milan, Italy). The expression of the immuno-phenotypic markers of the HL lines is reported in the Supplementary Table S1 . Testing for Mycoplasma infection was carried at a monthly basis. As positive controls, we used Kasumi-1 cells, an M2 acute myeloid leukemia cell line (Liebniz Institute Germane Collection of Microorganism and Cell Cultures DSMZ, Germania). Healthy donor B cells, serving as the normal control, were isolated from buffy-coat using EasySep™ kits (STEMCELL Technologies, USA). Two×10⁶ cells for each HL cell line were resuspended in 1ml of appropriate culture medium and plated in 12/24-well plates. Cells were incubated at different time points (24, 48, and 72 hours) with medium only or with CX-4945/Silmitasertib 0, 5, 10, and 15μM (provided from Selleck Chemicals, Munich - Germany) or with MMAE 5nM (20) (monomethyl auristatin E; Selleck Chemicals). For specific experiments HL cell lines were also treated (2x10⁶/ml) with the proteasome inhibitor Bortezomib (BTZ) 10nM up to 36 hours (Selleck Chemicals).

Aliquots of 100×10⁵ HL cell lines were harvested, washed in PBS 1x, and incubated for 10 min in the dark at room temperature and co-stained with the following antibodies: CD20 (APC-H7 conjugated), CD30 (PE conjugated) (Becton Dickinson; Franklin Lakes, NJ, USA), CD274/PD-L1 (PE-Cy7 conjugated, Fisher Scientific; Hampton, NH, USA). After incubation with antibody, cells were washed with PBS and analyzed by the flow cytometry. For each sample, 20,000 events were acquired and analyzed using the FACSCanto II™ A cytometer and data were processed using the DIVA Software (Becton Dickinson). For each antibody the mean fluorescence intensity was reported as compared to the untreated condition.

Apoptosis of different cell samples (pathological cells from different HL cell lines, normal B lymphocytes) were assessed using the Annexin V/Propidium Iodide (PI) staining (Apoptosis Detection Kit, Valter Occhiena, Turin, Italy), and by detection of PARP cleavage in western blotting (21).

L-428, L-540, HDLM-2, and KM-H2 HL cell lines were plated into 48 well plates (2×10⁵ cells/ml) in appropriate culture medium. CX-4945, MMAE were added at different concentrations: CX-4945 ranging from 1μM to 25μM and of MMAE ranging from 0.05nM to 10nM for 72h alone or in combination. Cell viability was measured through Trypan blue exclusion dye assay. The concentration of the single drug able to kill the 50% of cells (EC50) was determined by fitting the dose-response curve utilizing the Prism 7 software (GraphPad Software Inc. La Jolla, CA, USA). To calculate the Combination Index, cells were treated with a combination of CX-4945 and MMAE using the method of constant ratio drug combination proposed and described by Chou and Talalay (22).

Whole cell extracts (WCE) were obtained by RIPA Lysis and extraction buffer (Tris-HCL 20mmol/l, NaCl 150mmol/l, EDTA 5.0mmol/l, Niaproof 1.5%, Na3VO4 1.0mmol/l, SDS 0.1%, Thermo Fisher Scientific, Waltham, MA, USA), added with protease inhibitors (Halt Protease Inhibitor Cocktail, Thermo Fisher Scientific), phosphatase inhibitors (Phosphatase Inhibitor Cocktail, Thermo Fisher Scientific), and EDTA (Thermo Fisher Scientific) on ice. Nuclear and cytoplasmic Subcellular Fractionation were prepared using a commercial kit (Thermo Scientific, Rockford, IL, USA). The supernatant was quantified by BCA protein quantification assay (Thermo Fisher Scientific). Equal amounts of protein sample were added to 3× Red loading sample buffer (Cell Signaling Technology, Danvers, USA) and boiled for 5 min. Western blotting was conducted according to standard protocols (23). Briefly, 15 to 25 μg of WCE or nuclear and cytoplasmic fractions were subjected to SDS-PAGE, transferred to Nitrocellulose/PVDF membranes and immunoblotted with the following primary antibodies: anti-Akt-Ser129or anti-Akt-Ser473 (Cell Signaling), anti-Akt, anti-STAT3-Ser727, anti-STAT3 (Abcam, Cambridge, UK), anti-CK2β and anti-CK2α (kindly provided by Prof. Maria Ruzzene, University of Padova), anti-NF-κB-Ser529 (p65) (Cell Signaling), anti-NF-κB-p65 (Abcam), anti-PARP (Cell Signaling), mono- and polyubiquitinylated protein conjugates (Enzo Life Science Ltd, Exeter, UK), anti-β-actin (Sigma-Aldrich, St. Louis, MO), anti-α-tubulin (Sigma Aldrich), anti-GAPDH (Cell Signaling). Detection was performed using chemiluminescence reaction (ECL, Euroclone, Milan, Italy). Images were acquired using the Amersham Imager 600 (GE Healthcare; Chicago, IL, USA), the protein bands were scanned and quantified by densitometry, using the Image J program (Github, San Francisco, USA).

HL cell lines (L-540, L-428, HDLM-2, KM-H2) were lysed in Blast R™ Lysis Buffer by Signal-Seeker™ Kit (Cytoskeleton-Thermofisher Scientific) and the enriched total ubiquitinated proteins from HL cell line lysates were pulled out by an affinity beads system. Ubiquitinated protein fractions were obtained from 1mg total protein lysates. After extensive washes in IP buffer the immunocomplex was resuspended in Laemmli buffer with β-mercaptoethanol and processed for WB analysis.

Aliquots of the different cell samples (pathological cells from different HL cell lines and normal B lymphocytes from healthy subjects) were collected, washed and plated on poly-L-lysine coated slides for 15 min at RT. Cells were then fixed in 4% paraformaldehyde for 10min, washed twice with PBS 1x and permeabilized with 0,1% Triton X-100 (Sigma-Aldrich) for 4 min (24). Before staining, non-specific protein binding were blocked by incubating the slides for at least 30 min in 2% BSA. Cells were then stained with antibodies against CK2α and CK2β (the same used for WB) and DAPI for nuclear staining, washed three times with PBS and incubated with anti-mouse-Alexa488 secondary antibody for 30 minutes in the dark. Slides were mounted with cover slips and fluorescence was detected using the UltraView LCI confocal system (Perkin Elmer; Waltham, MA, USA) equipped with a fluorescence filter set for excitation at 488 and 360nm.

Tissue microarrays were prepared from cases with adequate diagnostic material, as previously reported (25). In details, tumor areas enriched in HRS cells were selected and 3 tissue cores (diameter=1 mm) were obtained from each donor block. Appropriate positive and negative controls were also included. The tissue microarray was prepared by using the Galileo TMA CK3500 arrayer (Integrated System Engineering; Milan – Italy. Immunohistochemical staining for CK2α (EP1963Y, Epitomics, CA, USA) and CK2β (PA5-27416, ThermoFisher, Massachusetts, USA) were performed in duplicate. Antigen retrieval was performed with heat/ethylenediamine-tetra-acetic acid (EDTA) in an automated immunostainer (University Hospital of Padova). All cases with discordant immunohistochemical results were assessed in joint sessions at the microscope by two haemato-pathologists. The positivity for CK2α and CK2β were graded as: 0 = negative; 1 = positive <30% of HRS; 2 = positive 30-60% HRS or week-moderate intensity; 3 = positive >60% HRS or strong diffuse intensity. The clinical features, including age, sex and therapies of the 25 analyzed patients are summarized in Supplementary Table S2 . Inclusion criteria were diagnosis of cHL before 2000 and having signed the informed consent. This part of the study was conducted according to the declaration of Helsinki, approved by the ethic committee of the Padova University Hospital (protocol # 4,089/AO/17) and informed consents were collected.

Total cellular RNA was extracted from the four HL cell lines, Kasumi, and purified healthy B lymphocytes, using the RNeasy Mini Kit (Qiagen, Germany). cDNA was generated by Reverse Transcription System (Promega, USA). SYBR Green (FastStart Universal SYBR Green Master, ROX) real-time polymerase chain reaction (RT-PCR) was carried out in a Sequence Detection System 7000 (Applied Biosystem) with the ABI PRISM 7000 software (AppliedBiosystems, Foster City, CA). The primers used are the following: CSNK2A1 (CK2α): Forward 5′-3′ GTTTGGGTTGTATGCTGGCA and Reverse 5′-3′ TTTCGAGAGTGTCTGCCCAA; CSNK2A2 (CK2α’): Forward 5′-3′ CGACCATCAACAGAGACTGACTG and Reverse 5′-3′ GTGAGACCACTGGAAAGCACAG; CSNK2B (CK2β): Forward 5′-3′ TCCTTACCAACCGTGGCATC and Reverse 5′-3′ CATGCACTTGGGGCAGTAGA; ACTIN (β-actin): Forward 5′-3′ CCAGCTCACCATGGATGATG and Reverse 5′-3′ ATGCCGGAGCCGTTGTC.

Statistical analyses were performed using Prism 7 (GraphPad Software Inc. La Jolla, CA, USA) for paired Student’s t test, Kruskall-Wallis, and Mann-Whitney Test, Wilcoxon matched-pairs signed rank test. Data are reported as mean ± standard deviation (SD). p<0.05 or less were considered statistically significant.

We assessed expression of CK2 subunits in cHL cell lines. By WB, we found that all the four HL-derived cell lines L-428, L-540, KM-H2 and HDLM-2 expressed higher levels of the catalytic subunit CK2α than normal B lymphocytes ( Figure 1A ). As positive control, we stained protein extracts from Kasumi-1 cell line, which notably shows high expression levels of both the CK2 subunits (26). The median densitometry of CK2α/Tubulin for each cell lines was 0.46 ± 0.05 (p=0.0250), 0.51 ± 0.10 (p=0.0285), 0.86 ± 0.05 (p=0.0005), 0.89 ± 0.05 (p=0.0044) and 0.26 ± 0.09 for L-428, HDLM-2, L-540, KM-H2 and normal B-cell from heathy donors ( Figures 1A, B ). The median CK2α/Tubulin ratio in HL cell lines was 2.7-fold higher than healthy B lymphocytes (p=0.0044) ( Figure 1B ). Conversely, the non-catalytic subunit CK2β was found to be expressed at lower levels as compared to normal B lymphocytes (p<0.05) ( Figure 1A ). The median densitometry of CK2β/Tubulin was 0.07 ± 0.03 (p<0.0001), 0.20 ± 0.02 (p<0.0001), 0.31 ± 0.10 (p=0.0009), 0.47 ± 0.10 (p=0.0025) and 0.98 ± 0.09 for L-428, HDLM-2, L-540, KM-H2 and B-cell from heathy donors ( Figures 1A, C ). The median CK2β/Tubulin ratio in HL cell lines was 3.8-fold fewer than B lymphocytes (p=0.0040) ( Figure 1C ).

Immunofluorescence experiments ( Figure 1D ) revealed a strong and diffuse signal for CK2α in HL cells, while CK2β exhibited a lower intensity signal, particularly in the KM-H2 and HDLM-2 cell lines. CK2α seems to be localized both in the nucleus and cytosol of HRS cells, including the formation of discrete foci, as observed in the L-428 and KM-H2 cell lines. Instead, CK2β exhibits a diffuse and weak signal with a less distinct localization which hampers the quality of picture. In order to determine the subcellular distribution of α and β subunits more precisely in HL cell lines, we analyzed subcellular protein fractions, unveiling the presence of CK2α in both the nuclear and cytosolic compartments of HL cell lines. In contrast, CK2β predominantly localizes to the cytosol but exhibits a faint representation in the nuclear compartment of the L-540 and L-428 cell lines ( Figure 1E ).

To assess whether the dysregulation among CK2 subunits was linked to alterations in CK2 mRNA levels, we performed RT-PCR analysis on the mRNA of CK2 catalytic (α - CSNK2A1 and α^I - CSNK2A2) and regulatory (β - CSNK2B) subunits. The mRNA levels in HL lines were then compared to those observed in control B lymphocytes. We observed statistically significant differences only for the CSNK2A1 gene, specifically in the L-428 (p<0.01) and L-540 (p<0.0001) cell lines (Supplementary Results, Supplementary Figure S2C ).

We performed a tissue microarray of neoplastic lymph nodes derived from 25 patients with cHL ( Supplementary Table S3 ) and 5 with reactive adenopathies to evaluate the expression and localization of CK2 subunits in primary HL specimens. As shown in Figure 1F, G and Supplementary Figures S1A–C , we observed that 71% of patients strongly expressed CK2α in HRS cells (i.e., grade 3), while 29% were at grade 2 and 4% at grade 1 but no one was grade 0. Conversely, no patient expressed CK2β at grade 3, 67% of HL patients expressed CK2β at grade 2, 21% at grade 1 and 17% at grade zero ( Figures 1F, G ). Moreover CK2α was expressed both in the nucleus and the cytosol of all HRS, while CKβ was found in the cytosol in 96% of cases but in the nucleus only in 1 patient (p<0.0001) ( Figure 1H , Supplementary Figures S1A–C ) confirming our above-mentioned findings ( Figures 1D, E ). We also compared the clinical characteristics of our HL patients with the grading of CK2 subunits expression, but we did not find any correlation between age, gender, histological subtypes, interim 18-FDG PET-CT results, or relapsed cases. However, higher CK2β levels (i.e., grade 2) were associated with shorter progression free survival (PFS) than patient displaying lower levels (i.e., grade 0-1) (p=0.0421, Supplementary Figure S2A ).

To elucidate the mechanisms accounting for the downregulation of CK2β, we investigated whether CK2β was degraded by the proteasome. For this purpose, we treated HL cell lines in vitro with 10nM Bortezomib (BTZ) for 36 h. Inhibition of the proteasome resulted in an increase of the CK2β subunit protein levels compared to the untreated condition in the HL cell lines ( Figures 2A, B ). To further investigate this issue, all polyubiquitinated proteins from both the untreated and treated conditions were subjected to purification using specific affinity beads. The enriched protein samples were analyzed by WB, revealing a pronounced presence of polyubiquitinated CK2β in the BTZ-treated condition, strongly indicating proteasome-dependent degradation of β subunits in HL cell lines ( Figure 2B ). To verify if the proteasome was effectively inhibited, we assessed the nonspecific increase in polyubiquitin residues through WB, which demonstrated a marked increase of ubiquitinated proteins across a broad range of molecular weights, as expected. ( Figure 2B ).

By means of WB analysis, we confirmed the overexpression of AKT, STAT3, and NF-kB (p65) as compared to their normal B cell counterparts (see Figures 2C–F ). As depicted in the representative WB images, we found higher levels of both the whole protein and of the phosphorylated counterpart at the phosphorylation target sites of CK2α. Specifically, CK2 can phosphorylate AKT on Serine 473 (AKT S473, Figure 2C ) by recruiting the mTOR2 complex (27) and directly phosphorylates AKT on Serine 129 (AKT S129, Figure 2D ), NF-kB p65 on Serine 529 (NF-kB S529, Figure 2E ), and STAT3 on Serine 727 (STAT3 S727, Figure 2F ). These molecules exhibited phosphorylation at CK2-related residues under basal conditions, as compared to normal B lymphocytes. Protein phosphorylation levels diverged across different cell lines, reflecting the clinical heterogeneity observed in patients with HL. In addition, AKT, STAT3, and NF-kB (p65) were significantly more expressed in HL cell lines compared to normal B cells (p<0.05 unpaired Student’s t test, Figures 2C–F ). Densitometric values have been summarized in Supplementary Table S4 .

CD30 and PD-L1/CD274 are known to be expressed on the cell membrane of HRS cells, while CD20 is usually absent. According to data coming from the literature, in HRS the locus of the gene CD274, that is mapped on chromosome 9 is usually amplified (5). Moreover, STAT3, NF-kB, and AKT molecular pathways can regulate the expression PD-L1 gene (5, 28, 29). Considering the role of CK2 in activating all these proteins within cHL cell lines, we examined whether CK2 could potentially affect PD-L1 expression. To delve into this question, we treated HRS cell lines with CK2α-specific inhibitor CX-4945/silmitasertib at a concentration of 10μM for 24h and 48h and we assessed expression of CD20, CD30 and PD-L1. We found a decrease in the mean fluorescence intensity (MFI) of PD-L1 after treatment with CX-4945 in HL cell lines, while no significant changes were observed for CD20 and CD30 ( Figure 3A and Supplementary Figures S2B ). Specifically, after 24 hours of treatment, the decrease of CD274/PD-L1 MFI was 46.0% (p<0.05), 51.4% (p<0.01), 29.0% (p<0.05), and 29.6% (p<0.05), for HDLM-2, L-428, L-540, and KM-H2, respectively, as compared to untreated conditions. Similarly, after 48 hours of treatment, the CD274/PD-L1 MFI decreased of 31.3% (p<0.05) for HDLM-2, 50.0% (p<0.01) for L-428, 55.7% (p<0.01) for L-540, 42.6% (p<0.01) for KM-H2, compared to untreated conditions (Mann-Whitney test, Figure 3A ). The downregulation of PD-L1 was also confirmed by WB analysis ( Figures 3B, C ). The PD-L1 expression values, determined through protein densitometry and normalized to β-actin, align with flow cytometry results, confirming a substantial reduction in protein levels following treatment with the CK2 inhibitor. Specifically, after 24 hours of culture, the PD-L1/β-actin ratios shifted from 1.17 ± 0.14, 0.85 ± 0.07, 1.74 ± 0.5, and 0.82 ± 0.06 under untreated conditions to 0.53 ± 0.08, 0.74 ± 0.02, 0.4 ± 0.11, and 0.60 ± 0.03 after treatment with CX-5945 at 10 μM, respectively, for L-428, L-540, KM-H2, and HDLM-2 cell lines. Specifically, HL cell lines L-428 and HDLM-2 displayed a significant difference in the means of the experimental triplicates, with p<0.05 and p<0.01, respectively (unpaired t test). The association between CK2 and PD-L1 became more evident after 48 hours of treatment. The PD-L1/β-actin ratios shifted from fold change values of 0.83 ± 0.08, 1.0 ± 0.1, 0.90 ± 0.25, and 0.90 ± 0.06 to values of 0.36 ± 0.01, 0.4 ± 0.18, 0.30 ± 0.12, and 0.25 ± 0.01, respectively, following treatment with CX-4945/silmitasertib at 10 μM for the L-428, L-540, KM-H2, and HDLM-2 cell lines. Once again, the L-428 and HDLM-2 cell lines exhibited a treatment-related difference that was statistically significant (p<0.01 and p<0.001, respectively, unpaired t test).

Since CK2 is known to sustain pro-survival signals in cancer cells, we examined the effect of its chemical inhibition in HL cell line. Treatment with CX-4945/silmitasertib resulted in time- and dose-dependent apoptosis, as confirmed through Annexin V/Propidium Iodide flow cytometry testing ( Figures 4A, B , Kruskal-Wallis’s test, p<0.05). For all the four HL cell lines, the percentage of viable cells exhibited a proportional decrease, starting from a dose of 5µM, in comparison to DMSO-treated cells (p<0.05). Notably, in vitro treatment with silmitasertib at 10µM concentration for 48 hours reduced the number of viable HL cells by half ( Figure 4A ). Furthermore, to elucidate the mechanisms underlying silmitasertib-induced apoptosis, we exposed HL cell lines to increasing doses of silmitasertib for 24 hours. This treatment resulted in the cleavage of PARP, a recognized marker of apoptosis ( Figure 4B ). Furthermore, that after 24 hours of cell treatment with CX-4945 (at concentrations of 5μM and 10μM), there was a significant decrease in the phosphorylation levels of AKT-S473 ( Figure 4C ), AKT-S129 ( Figure 4D ), STAT3-S727 ( Figure 4E ), and NF-kB-S529 ( Figure 4F ).

Since CD30 levels were not affected after CK2 inhibition ( Supplementary Figure S2B ), we subsequently proceeded to investigate whether CX-4945/silmitasertib might enhance the effectiveness of this innovative therapy. Specifically, we investigated whether CK2 inhibition could enhance the cytotoxicity of MMAE, the microtubule inhibitor conjugated to the anti-CD30 monoclonal antibody brentuximab vedotin.

To this aim, HL cell lines were treated with 5µM CX-4945 and 5nM MMAE, according to the literature (20), or a combination of both drugs for either 24 or 48 hours duration. As shown in Figure 5A , combination of CX-4945+MMAE significantly decreased the proportion of viable cells in comparison to MMAE. After 24 hours of in-vitro treatment, the percentage of viable cells, i.e. AV- PI-, decreased from 90.1%, 87.3%, 87.9%, and 94.8% with MMAE alone to 82.7% (p=0.12), 70.6% (p<0.05), 63.3% (p<0.05), and 85.9% (p<0.05) when treated with CX-4945 + MMAE in L428, L540, HDLM-2, and KM-H2, respectively (determined by the Wilcoxon matched-pairs signed rank test, Figure 5A ). After 48 hours of treatment the rate of alive cells decreased from 79%, 62.2%, 66.5%, and 80.3% with MMAE alone to 64.3% (p=0.052), 51.8% (p=0.051), 48.4% (p<0.05), and 75.5% (p<0.05) with CX-4945+MMAE in L428, L540, HDLM-2, and KM-H2, respectively (Wilcoxon matched-pairs signed rank test, Figure 5A ). The cell lines tested responded differently to the combination treatment of CX-4945 and MMAE. Specifically, KM-H2 and L-428 cells exhibited the lowest sensitivity, whereas HDLM-2 cells demonstrated the highest sensitivity to the combination treatment. The increased apoptotic effect resulting from the combination of CX-4945 and MMAE was further supported by WB analysis, revealing a decreased full-length PARP band and an increased cleaved protein band compared to cells treated exclusively with MMAE.

The possible synergism between CX-4945 and MMAE in reducing cell viability was evaluated using the Chou-Talalay method through the evaluation of trypan blue exclusion. The EC50 for each cell lines of the single treatment (CX-4945 or MMAE) and the EC50 for the drugs used in combination are reported in Figure 5B . The combination effect analyses identified that the combination index values were all below 1 for each cell line, indicating a synergistic cytotoxic effect between CX-4945 and MMAE. In particular, the obtained combination indexes were 0.91, 0.78, 0.67, and 0.23 for L-428, L-540, KM-H2, and HDLM-2, respectively ( Figure 5C ).