Antitumor Activity of the Novel BTK Inhibitor TG-1701 Is Associated with Disruption of Ikaros Signaling in Patients with B-cell Non–Hodgkin Lymphoma

B-cell non–Hodgkin lymphomas (B-NHL) account for up to 4% of globally diagnosed cancers (1). They are divided into low and high grades, typically corresponding to indolent (slow-growing) lymphomas, like chronic lymphocytic leukemia (CLL), and aggressive lymphomas, like diffuse large B-cell lymphoma (DLBCL), respectively (2). Targeting of the B-cell receptor (BCR) pathway through inhibition of Bruton tyrosine kinase (BTK) with the first-in-class irreversible inhibitor ibrutinib has demonstrated exceptional clinical activity as a monotherapy for various subtypes of B-NHL, most notably CLL (3) and mantle cell lymphoma (MCL; ref. 4), but also Waldenström macroglobulinemia (WM; ref. 5) and activated B-cell-like DLBCL (ABC-DLBCL; ref. 6). Nonetheless, resistance to ibrutinib has been observed due in part to the acquisition of mutations that either affect the irreversible binding of BTK inhibitors (BTKi) or activate the phospholipase C gamma 2 (PLCγ2) enzyme, a downstream enzyme in the BTK pathway (7). In the case of BTK, a cysteine-to-serine mutation (BTK^C481S) abrogates the covalent binding of ibrutinib to BTK and has been detected in up to 86% of relapsing patients with CLL, but only anecdotally in MCL relapsing patients (7–9). Additional intrinsic mechanisms of resistance involve activation of the noncanonical Nuclear factor κB (NF-κB)-inducible kinase (NIK)-NF-κB signaling in MCL (10, 11).

The use of distinct next-generation sequencing (NGS)-based genomic techniques including whole exome and targeted deep sequencing have been instrumental in identifying BTK^C481S mutation as a genetic cause of BTKi resistance (12, 13). More recently, global drug profiling using MS-based phosphoproteomics has been successfully employed to characterize drug mechanisms of action in single-agent therapy or multidrug combinations in solid cancers (14). In addition, a recent publication also showed that phosphoproteomic profiling can help to understand the role of BTKi in patients with CLL. In this study, CLL cells from patients with an unmutated immunoglobulin heavy chain gene (IGVH) status showed higher basal phosphorylation than patients with IGVH-mutated status (15).

Here, we present TG-1701, a novel, orally available, irreversible BTKi that exhibits improved selectivity when compared with ibrutinib (16), and shows activity in various models of B-NHL. A MS-based phosphoproteomic platform used to interrogate the effects of TG-1701 on patients with CLL enrolled in the phase I dose-escalation study (NCT03671590; refs. 17–19) pointed to the transcription factor Ikaros as both a potential biomarker of clinical activity and an important node downstream of BTK in the BCR pathway.

Blood samples were obtained from 6 patients with CLL treated on the NCT03671590 clinical trial (Table 1). Briefly, patients with relapsed or refractory histologically confirmed B-cell lymphoma or CLL were eligible and treated with TG-1701 orally once daily until disease progression or the occurrence of intolerable side effects. Patients received either 200, 300, or 400 mg TG-1701 daily (Table 1). On the first day of treatment, white blood cells were counted, the absolute lymphocyte count (ALC) calculated, and the normal lowest and normal highest ALC established (Supplementary Table S1). The percentage of CLL cells in each sample was the ratio of ALC subtracted from highest normal ALC and divided by total ALC. All patients provided written informed consent. Institutional Review Board approvals for the study protocol, amendments, and informed consent documents were obtained prior to study initiation; study procedures were conducted in accordance with the Declaration of Helsinki.

Seven MCL (REC-1^GFP+LUC+, JEKO-1, UPN-1, UPN-IbruR, GRANTA-519, Z-138, and MINO), two ABC-DLBCL (OCI-LY3 and HBL-1), and two follicular lymphoma (DoHH2 and RL) cell lines were cultured as described elsewhere (20, 21). JEKO-1, MINO, and REC-1 parental cells were obtained from ATCC cell bank (LGC Standards). DoHH-2, GRANTA-519, and RL cell lines were purchased at DSMZ. UPN-1 and Z-138 cells were provided by Dr. B. Sola (University of Caen). HBL-1 was provided by Dr. E. Valls (IDIBAPS). OCI-Ly3 was kindly provided by Dr. A. Staiger (Dr. Margarete Fischer-Bosch Institute of Clinical Pharmacology). Cell-line authentication was performed upon reception by short tandem repeat (STR) profiling, using AmpFlSTR Identifier Kit (Thermo Fisher Scientific), and based on available STR profiles. This analysis was then repeated every 6 months and up to 4 months prior to the submission of the present manuscript. Mycoplasm infection was routinely tested by PCR.

TG-1701 was provided by TG Therapeutics, Inc. Ibrutinib was purchased from Selleckchem.

The MINO MCL xenograft model was generated by inoculating subcutaneously 6- to 7-week-old nude mice (Shanghai Ling Chang Experimental Animal Co., Ltd.) with 5 × 10⁶ MINO cells. When tumor volume reached 100 to 200 mm³, animals were randomized into five groups of 8 to10 mice each, and were dosed orally with TG-1701 (25, 50, or 100 mg/kg, orally, twice a day), ibrutinib (100 mg/kg, orally, twice a day), or vehicle for 21 days. The tumor volume (V) was calculated as: V = 1/2 × a × b², where a and b represent the length and width, respectively. In REC-1 and UPN-IbruR xenografts, CB17-SCID mice (Janvier labs) were inoculated subcutaneously with 10⁷ REC-1^GFP+Luc+ cells or UPN-IbruR cells and monitored for tumor growth, bioluminescence signal, and vital parameters as described previously (21), in compliance with the Animal Ethics Committee of the autonomous University of Barcelona (registry number 38/18). Tumor-bearing mice received either TG-1701 (25 mg/kg, every day) or ibrutinib (25 mg/kg, every day) for 17 days. Tumor samples were snap-frozen in OCT medium (Sakura Tissue Tek) or formalin-fixed and paraffin-embedded prior to immunohistochemical (IHC) staining with primary antibodies against Ikaros (Cell Signaling Technology) and CD20 (Beckman Coulter). Preparations were evaluated using an Olympus microscope and MicroManager software.

Total and nuclear protein extracts were obtained from MCL cell lines and tumor specimens using RIPA (Sigma-Aldrich), the Nuclear/Cytosol Fractionation Kit (BioVision), and T-PER (Thermo Fisher Scientific) buffers, respectively, and subjected to SDS-PAGE, as described previously (22, 23). Membrane-transferred proteins were revealed by incubating with primary and secondary antibodies (Supplementary Materials and Methods) followed by chemiluminescence detection using the ECL system (Pierce) and a Fusion FX imaging system (Vilber Lourmat). Band intensity was quantified using Image J software and normalized to housekeeping protein (GAPDH or TUBULIN). Values were referred to the indicated control (i.e., predose sample or DMSO-treated cells) and added below the corresponding band. If not otherwise specified, representative data from n = 2 experiments are shown.

Peripheral blood mononuclear cells (PBMC) were obtained from a total of 6 patients with CLL before and after 4-hour treatment with TG-1701 (Table 1), using standard Ficoll-Hypaque density gradient. Proteins were extracted using a Urea-based buffer (6 M Urea, 100 mmol/L Tris-HCl pH 7.5) on a Bioruptor sonicator, followed by standard Lys-C and trypsin-mediated digestion. Proteomic quantitative analyses were performed using 13 of the 16 channels of a 16plex - Tandem Mass Tag (TMT) system, according to manufacturer's instructions (Thermo Fisher Scientific). A small part of the sample (260 μg) was used for total proteome analysis, and remaining proteins were subjected to phosphoproteome enrichment using the High-Select TiO2 Phosphopeptide Enrichment Kit (Thermo Fisher Scientific). Total proteome and phosphopeptide fractions were fractionated in 24 and 9 fractions using a Zorbax Extent-C18 and the High pH Reversed-Phase Peptide Fractionation Kit (Thermo Fisher Scientific), respectively, with prior analysis on an Orbitrap Fusion Lumos Tribrid mass spectrometer. REC-1 and REC-BTK^C481S protein extracts were labeled in triplicates with two TMT10plex, using a single channel (131N) to label a pool of all the samples. TMT fractionation and MS analysis were performed as above. Raw data were processed in MaxQuant (RRID:SCR_014485) and analyzed as described in Supplementary Materials and Methods. Data have been deposited to the ProteomeXchange Consortium (http://proteomecentral.proteomexchange.org) via the PRIDE partner repository (PXD023231; ref. 24).

Total RNA was extracted using TRizol (Thermo Fisher Scientific) following manufacturer's instructions and poly-A-tailed enriched mRNA selected. Paired-end stranded RNA libraries with 51 read-inward facing paired mates were prepared, following sequencing with Illumina's NovaSeq6000 at the Centro Nacional de Analisis Genomico (CNAG). Data analysis was performed as described in Supplementary Materials and Methods.

The reverse transcription reaction was performed using a High-Capacity cDNA Reverse Transcription Kit (Applied Biosystems). The mRNA expression was analyzed in triplicate by quantitative real-time PCR. Amplification was performed using SYBR Green-based detection (GoTaqPCR Master Mix; Promega). The relative expression of each gene was quantified by the comparative cycle threshold method (ΔΔCt). β-Actin, GAPDH, and B2M were used as endogenous controls.

BTKi-treated REC-1 and JEKO-1 cells (2–3 × 10⁵) were seeded on poly-L-lysine-coated glass coverslips and stained as previously (23) with anti-Ikaros antibody (Cell Signaling Technology). Fluorescence signal was acquired on a Leica microscope and quantified using the LAS X (Leica) and Image J (RRID:SCR_003070) softwares.

CRISPR-Cas9 gene editing tool were used to generate REC-BTK^C481S and REC-BTK^KO cells. Briefly, REC-1^GFP+LUC+ cells were electroporated with SpCas9 nuclease and BTK-specific gRNA using a Neon Transfection System (Thermo Fisher Scientific). After 30 days, single clones obtained by limit dilution were checked for BTK protein levels and analyzed by Sanger sequencing. Complete gene edition protocol can be found in Supplementary Data.

Presented data are the means ± Standard Deviation or SEM of three independent experiments. All statistical analyses were done by using GraphPad Prism 4.0 software (RRID:SCR_002798). Comparison between two groups of samples was evaluated by nonparametric Mann–Whitney test to determine how response is affected by two factors. Pearson test was used to assess statistical significance of correlation. Results were considered statistically significant when P value < 0.05.

In a binding assay on a panel of 441 human kinases, TG-1701 was more selective than ibrutinib, with a comparable BTK Kd (3 nmol/L vs. 1.5 nmol/L, respectively) and a lower binding to EGFR, ITK, TXK, and JAK3 (Kd 135-, >48-, 68-, and >94-fold higher than those of ibrutinib, respectively; Fig. 1A and Supplementary Table S2). A BTK kinase activity assay revealed a TG-1701 EC₅₀ of 6.70 nmol/L, slightly higher than ibrutinib IC₅₀ (1.65 nmol/L; Fig. 1B and Supplementary Table S3). Accordingly, in an in vitro BTK occupancy assay, TG-1701 and ibrutinib showed a similar dose-dependent capacity to displace a BTK-specific fluorescent probe in the BTKi-sensitive follicular lymphoma cell line DoHH-2 (20), with complete BTK occupancy at 30 and 10 nmol/L, respectively (Fig. 1C). Consistently, BCR downstream signaling was impaired in a concentration-dependent manner in IgM-stimulated cells, with maximal effects observed at 100 nmol/L for TG-1701 and 10 nmol/L for ibrutinib (Fig. 1D). These effects were associated with a mean TG-1701 GI₅₀ of 6.4 μmol/L at 72 hours, in a set of 10 parental B-NHL cell lines, which was slightly inferior to mean ibrutinib GI₅₀ (14.1 μmol/L), especially in MCL cell lines (4.3 μmol/L vs. 10.8 μmol/L, respectively; Supplementary Table S3). Given that the activity of both BTK inhibitors in MINO cells was similar to these mean values, the tumor growth inhibition (TGI) of TG-1701 was evaluated in vivo in the corresponding xenograft model. A 16-day treatment of mice bearing MINO-derived tumors with 25, 50, and 100 mg/kg TG-1701 achieved a 56%, 72%, and 78% TGI, respectively, comparable with the 70% TGI observed in the ibrutinib (100 mg/kg) arm (Fig. 1E). A single oral gavage with 50 mg/kg TG-1701 further confirmed a rapid dephosphorylation of BTK and AKT as early as 2 and 4 hours, respectively, which was maintained for at least 24 hours due to the irreversible nature of TG-1701–mediated BTK inhibition (Fig. 1F).

To identify potential biomarkers of TG-1701 activity in B-NHL, PBMCs from 6 patients with CLL enrolled in the TG-1701–101 phase I trial were isolated at different time points for omics analysis. In all patient samples but one, the percentage of circulating cancer cells was comprised between 60% and 98% (Table 1; Supplementary Table S1). All the patient samples harbored a wild-type BTK gene, as confirmed by Sanger sequencing of BTK exon 11 (data not shown).

Near-complete BTK occupancy was obtained in all samples tested (Fig. 2A) and linear kinetics was observed with approximately dose-proportional increases in C_max and AUC_{0–8 hours} from 100 to 300 mg (Supplementary Fig. S2; ref. 18), with a positive correlation between the daily dose and TG-1701 C_max for the 6 patients studied (Fig. 2B). After three cycles of treatment, the best change in tumor burden as assessed by CT scan ranged from 38% to 87%, with 1 tumor-free patient with a lymphocytosis that defined his best response as a stable disease (Fig. 2C).

In PBMC protein extracts from the six predose and the six 4-hour postdose samples, a total of 5,585 proteins and 2,438 phosphosites were identified. An initial principal component analysis (PCA) did not discriminate the pretreatment (PRE) from the posttreatment (POST) profiles. However, a specific unsupervised clustering of the data into two subgroups of 3 patients allowed a clear differentiation between the PRE and POST samples (Fig. 2D). These two subgroups, blindly selected to uncover changes due to TG-1701 treatment, clearly fitted the clinical outcome of the patients (Fig. 2C), separating de facto a group encompassing the 3 best responses (called the early responders) and a group gathering the 3 other patients with lower responses (called the late responders). Supporting this partition of the patients, the early responders exhibited a stronger IL10 de-repression and a stronger, although not significant, decrease in IL2 and IL6 expression (Supplementary Fig. S3A), suggesting that they developed a more potent early anti-inflammatory response, a common feature of BTKi activity (25). In addition, CCL3 and CCL4 chemokine genes, two bona fide biomarkers for BCR pathway activation (26), were more downregulated in early-responder patients (Supplementary Fig. S3A). A set of 118 phosphopeptides were differentially phosphorylated (7 down- and 111 upregulated) after TG-1701 treatment in the early-responder subgroup only (adjusted P value < 0.1; Fig. 2E). These sites corresponded to the putative modulation of 14 protein kinases (Supplementary Tables S4 and S5). Importantly, these quantitative phosphoproteomic changes showed a strong homogeneity in the 3 early-responder patients and enabled a clear distinction between PRE and POST samples (Fig. 2F). Only one single phosphosite (UBA1-pSer4) was found significantly upregulated in the late-responder patients after TG-1701 treatment.

Beside the 118 phosphosites mentioned above and depicted in Fig. 2E, another set of 96 phosphopeptides were present in pretreatment samples and totally dephosphorylated upon treatment with TG-1701. The total absence of phosphorylation preempted the statistical analysis and incorporation of these 96 samples in the volcano plot, even though these sites were the most impacted by TG-1701 treatment. Of special interest, the corresponding list of phosphosites comprised the p-Ser442/445 residue of Ikaros, a zinc finger-containing DNA-binding protein that plays a pivotal role in B-cell homeostasis. Ikaros-Ser442/445 dephosphorylation was indeed the strongest event associated with TG-1701 activity (Supplementary Table S6). Because Ikaros nuclear localization and transcriptional activity both depend on BTK-mediated phosphorylation at Ser214/215 residues (27), we investigated whether, analogously, Ikaros function was differentially affected by TG-1701 in early-responder versus late-responder CLL patient samples. Using previously validated Ikaros-repressed and Ikaros-enhanced gene signatures (28), we identified 21 proteins of the repressed signature that were upregulated, whereas another set of 17 factors from the Ikaros-enhanced gene signature were depleted only in early responders, suggesting that Ikaros was functionally impaired after TG-1701 treatment (Fig. 3A). Although a comparative multidimensional (MDS) analysis of RNA-seq data obtained from the same subset of samples failed to show a clear difference between early-responsive and late-responsive patient clusters (Supplementary Fig. S3B), a clear trend in the upregulation of Ikaros-repressed genes and downregulation of Ikaros-enhanced genes was seen in TG-1701 early-responsive patients only (Fig. 3B). Accordingly, the IKZF1-repressed gene, YES1, was significantly upregulated, whereas the IKZF1-enhanced gene MYC was downregulated in early responders but not in late responders (Fig. 3C). Importantly, this effect was not due to Ikaros protein destabilization, as its level of expression did not vary upon TG-1701 treatment in early-responder patients, whereas the expression of p-BTK and MYC proteins underwent a 70% and 48% downregulation, respectively, and the levels of YES1 increased by almost 5-fold (Fig. 3D; Supplementary Fig. S3C). Such modifications were not seen in late-responder patients, although BTK was notably dephosphorylated (Fig. 3D; Supplementary Fig. S3C). These data were confirmed in vitro using the ibrutinib-sensitive MCL cell line REC-1 characterized by ibrutinib and TG-1701 GI₅₀ values of 5.82 and 3.83 μmol/L at 72 hours, respectively (Supplementary Table S3). In these cells, as observed in early-responder patients with CLL, TG-1701 treatment led to efficient BTK dephosphorylation, YES1 upregulation, and MYC downregulation both at mRNA and protein levels (Fig. 3E and F). Of note, among the IKZF1-target genes studied here, YES1 reactivation was significantly higher in TG-1701–treated than in ibrutinib-exposed cells (Fig. 3E). Finally, Fig. 3G shows that ibrutinib and TG-1701 elicited a 67% to 83% and a 80% to 85% reduction in nuclear Ikaros, consistent with a nearly 50% to 70% decrease of Ikaros in REC-1 and JEKO-1 nuclear protein fraction (Fig. 3H), suggesting that dual dephosphorylation of Ikaros at Ser442 and Ser445 was associated with the nuclear exclusion of this factor.

To further explore TG-1701 mechanisms of action and potential mechanisms of resistance, we CRISPR-engineered the REC-1^GFP+LUC+ cell line to express the BTK^C481S. This mutation was associated with a 10.3- and a 54.8-fold decrease in ibrutinib and TG-1701 inhibitory kinase activity, respectively (Supplementary Fig. S4A; Supplementary Table S3). The REC-1–BTK^C481S cell line was 4.2- and 2.8-fold less sensitive to ibrutinib and TG-1701 respectively, compared with parental REC-1 cells (Fig. 4A; Supplementary Table S3). A washout experiment further showed that irreversible BTK inhibition—illustrated by kinase phosphorylation over 24 hours after BTKi removal in REC-1 cells—was mostly lost in REC-1–BTK^C481S cells (Fig. 4B). Although total proteome composition was modified solely in the parental cells treated with either ibrutinib or TG-1701 as assessed by PCA analysis (Supplementary Fig. S4B), a set of 16 Ikaros-repressed proteins were upregulated and another set of 14 Ikaros-upregulated proteins were downregulated in REC-1, but not in REC-1–BTK^C481S cells exposed to TG-1701 (Fig. 4C). Despite the short drug exposure, transcriptional upregulation of YES1, and repression of MYC were observed in BTK^wt cells treated with both BTKis. In REC-1–BTK^C481S cells, a pronounced basal Ikaros-repressed gene signature was observed in the absence of any treatment (black bars, Fig. 4D), including YES1 expression, and this pattern was not modified upon BTKi exposure (Fig. 4D; Supplementary Fig. S4C).

To confirm the role of BTK as an upstream regulator of Ikaros signaling in BTKi-exposed cells, we generated a BTK knockout (KO) model derived from the REC-1 cell line, using a CRISPR-Cas9 method, as described above. The obtained REC-BTK^KO derivative, characterized by an almost complete depletion of BTK (Fig. 4E) was refractory to both ibrutinib and TG-1701 (Supplementary Table S3) and did not undergo significant modulation of YES1 and MYC expression after exposure to TG-1701 (Fig. 4E and F). To validate the relevance of the BTK-Ikaros signaling axis in the regulation of YES1 and MYC upon BTK inhibition, two CRISPR-Cas9–engineered cell lines either devoid of Ikaros (REC-IKZF1^KO) or with de novo expression of BTK after BTK gene knockout (REC-BTK^KO-OE) were generated (see Supplementary Materials and Methods). Supplementary Figure S5A and S5B shows that Ikaros depletion was accompanied by a 2- to 3-fold increase in YES1 expression in REC-1 cells, and that no further upregulation of this gene could be achieved after BTK inhibition. MYC expression did not undergo significant variation upon treatment with either TG-1701 or ibrutinib. Most interestingly, the re-introduction of an IgM-responsive form of BTK within REC-BTK^KO cells, restored the capacity of TG-1701, and in a lesser extent ibrutinib, to suppress the Ikaros-dependent repression of YES1 and upregulation of MYC (Supplementary Fig. S5C and S5D). Altogether, these data strongly suggest that TG-1701–dependent impairment of Ikaros signaling in MCL cells is a BTK-dependent process.

Given the low recurrence of BTK^C481S mutation in patients with MCL (9), the UPN-IbruR noncanonical NF-κB–driven ibrutinib resistance model was also studied (20). This subclone is characterized by the absence of mutations in the BTK and PLCG2 genes and the constitutive activation of p52-dependent signaling, driving a 2- to 3-fold increase in ibrutinib and TG-1701 GI₅₀ at 72 hours (Supplementary Fig. S4D; Supplementary Table S3; ref. 20). When compared with the BTKi-sensitive REC-1 xenograft model in which a 17-day dosing with TG-1701 achieved a 53% tumor TGI versus vehicle, UPN-IbruR tumors were almost insensitive to TG-1701 (Fig. 5A). Accordingly, p-BTK was efficiently downregulated in representative REC-1 tumor specimens, but not in UPN-IbruR xenografts treated with TG-1701 (Fig. 5B). In agreement with in vitro results, MYC was downregulated, and YES1 upregulated at protein and/or mRNA levels, in association with Ikaros nuclear exclusion and a decrease in CD20⁺ malignant B cells, in BTKi-sensitive, but not in BTK-insensitive MCL xenografts (Fig. 5B–D).

During the past decade, BTK inhibitors have increasingly replaced chemotherapy-based regimens in patients with CLL and MCL. TG-1701 is a novel second-generation BTKi currently under clinical development. TG-1701 is more selective than ibrutinib, with a comparable BTK Kd and similar in vitro and in vivo characteristics. TG-1701 is currently being tested in a phase I trial comprised of a single-agent arm and a combination arm with ublituximab (a novel CD20 antibody) and umbralisib (a dual PI3Kδ and CKIε inhibitor). With a median follow-up of 7 months in a 200 mg daily monotherapy expansion cohort, preliminary overall response rates (ORR) are 95% (19/20) in CLL, 50% (9/18) in MCL, and 95% (18/19) in WM. No complete responses (CR) are confirmed on TG-1701 monotherapy (17). In this study, we show that phosphoproteomic analysis of pre- and posttreatment samples clustered patients with CLL according to their early clinical responses (early responders vs. late responders) and helped decipher the mechanisms underlying drug responsiveness. According to the results, the response from these patients did not depend on the pharmacokinetic or pharmacodynamic properties of TG-1701, but rather on differences in BTK downstream signaling. In early-responder patients, Ikaros-p-Ser442/445 was the phosphopeptide most impacted (dephosphorylated) by TG-1701 treatment and several Ikaros-dependent factors were transcriptionally deregulated, suggesting that Ikaros may represent a biomarker for early response, and/or an important new node downstream BTK inhibition. Interestingly, a recent study in acalabrutinib-treated patients with CLL (15) has shown that the unmutated IGVH cells displayed a higher basal phosphorylation level compared with the mutated IGVH cells, showing that the phosphoproteomic profile of BTKi-treated patients can cluster specific subgroups of patients.

Ikaros is a zinc finger protein involved in gene regulation and chromatin remodeling. Its nuclear localization, stability, and transcriptional activity depend on its phosphorylation status, which is regulated by BTK, casein kinase II (CKII), and protein phosphatase 1 (PP1) interplay (27, 29). Although the exact role of serine residues at position 442 and 445 is still unknown, their juxtaposition to the conserved PP1 binding motif in the C-terminal end of Ikaros protein (30) might confer them a role in the PP1-mediated regulation of Ikaros stability and pericentromeric localization (29). Interestingly, CKII was ranked #4 in the list of kinases with reduced activity in TG-1701 early-responder patients, whereas phosphorylation of the PP1 inhibitory subunit, PPP1R14A, and dephosphorylation of the PP1 inhibitor, PPP1R2, both associated with reduced PP1 activity (31), were among the top four modifications detected in TG-1701 early-responder patients (Supplementary Tables S4 and S5). Ikaros expression was affected by TG-1701 treatment, neither in CLL primary cells nor in BTKi-sensitive REC-1 models, indicating that the inhibition of Ikaros signature in early responders was more likely due to a nuclear exclusion of the transcription factor. Importantly, our results obtained in REC-1–BTK^C481S and REC-1–BTK^KO subclones demonstrated that both the cell proliferation blockade and the modulation of Ikaros activity upon TG-1701 treatment are tightly dependent on the presence of a wild-type BTK protein, supporting the idea that Ikaros phosphorylation at Ser442/445 is controlled by the kinase, and discarding a potential off-target effect of the compound. Downregulation of the Ikaros program in BTK^C481S cell line is not understood at that point. One possible explanation is that the C481S mutation does not only abrogate the irreversibility of ibrutinib or TG-1701 binding but also modifies the substrate specificity of the newly formed substrate pocket that would not phosphorylate Ikaros anymore. Further studies involving the production of a phospho-specific Ikaros-Ser442/445 antibody will be required to confirm this hypothesis.

The number of samples tested in this study is too small to draw definitive conclusions, and because responses to BTKis continue to deepen over time (as long as 2 years), the early (C3D1) timepoint may not fully represent the overall drug efficacy. However, the differences presented here between “early-responders” and “late-responders” are clear enough to warrant future studies, for example testing an enlarged pool of patient samples treated with BTKis or exploring other NHLs. In addition, other (non-Ikaros) potential biomarkers for which an existing phospho-antibody is commercially available could be evaluated. In the study presented here, the assessment of the clinical outcome available at the time of analysis was the first CT scan, after two cycles of treatment. Assessing the phosphorylation status of the biomarkers presented here in patients who progressed under, or become resistant to BTKi treatment, might also help understand the underlying mechanisms of activity or resistance.

H. Miskin reports personal fees from TG Therapeutics, Inc. during the conduct of the study. P. Menéndez reports personal fees from OneChain Immunotherapeutics outside the submitted work. E. Normant reports employment and ownership of stock with TG Therapeutics. G. Roué reports grants from TG Therapeutics and Instituto de Salud Carlos III during the conduct of the study. No disclosures were reported by the other authors.

M.L. Ribeiro: Resources, investigation, methodology, writing–original draft. D. Reyes-Garau: Investigation, methodology. M. Vinyoles: Resources, validation, investigation, writing–review and editing. N. Profitós Pelejà: Investigation, methodology. J.C. Santos: Investigation, methodology. M. Armengol: Investigation. M. Fernàndez-Serrano: Investigation. A. Sedó Mor: Investigation, methodology. J.J. Bech-Serra: Data curation, software, visualization. P. Blecua: Data curation, software, visualization. E. Musulen: Investigation, methodology. C. De La Torre: Formal analysis, investigation, methodology. H. Miskin: Validation, writing–review and editing. M. Esteller: Data curation, writing–review and editing. F. Bosch: Writing–review and editing. P. Menéndez: Validation, investigation, methodology, writing–review and editing. E. Normant: Conceptualization, resources, supervision, validation, writing–original draft, writing–review and editing. G. Roué: Conceptualization, resources, supervision, funding acquisition, validation, investigation, visualization, writing–original draft, project administration, writing–review and editing.

We are very grateful to Dr. Chan Cheah at Linear Clinical Research, Dr. Constantine Tam at St. Vincents private hospital in Melbourne, and Dr. Nick Wickham at Adelaide Cancer Center in Ashford, as well as the TG Therapeutics clinical team led by Dr. Alejandro Ricart for their precious help. The authors would like to thank the CNAG-CRG for assistance with RNA sequencing. This study was financially supported by Fondo de Investigación Sanitaria PI18/01383, European Regional Development Fund (ERDF) “Una manera de hacer Europa” (to G. Roué), and TG Therapeutics. J.C. Santos and M. Fernández-Serrano were recipients of a Sara Borrell research contract (CD19/00228) and a predoctoral fellowship (FI19/00338) from Instituto de Salud Carlos III, respectively. M. Armengol was a fellow of PROTEOblood (EFA360/19), a project co-financed by the ERDF through the Interreg V-A Spain-France-Andorra (POCTEFA) program. This work was carried out under the CERCA Program (Generalitat de Catalunya).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.