Activity of the novel BCR kinase inhibitor IQS019 in preclinical models of B-cell non-Hodgkin lymphoma

To assess the selectivity of the kinase inhibitor IQS019 (Fig. 1a and ref [13]), we first evaluated its inhibitory property against a panel of 400 kinases, including 70 disease-relevant protein kinase mutants and 13 lipid kinases, covering about 60% of the human kinome. We found the compound to be preferentially active against tyrosine kinase (TK) and tyrosine kinase-like (TKL) families, reaching a mean residual kinase activity of 28% at a 10 μM dose, while this activity remained above 70% in all the other kinase subgroups (Fig. 1b, *** p < 0.001 and Additional file 1: Figure S1). Of special interest, in a set of 17 TK/TKLs, the compound was able to inhibit at least 20% of the kinase activity at the lowest dose (0.1 μM) and to achieve an almost complete kinase inactivation at the 10 μM concentration. These kinases corresponded to leucocyte-, BCR-, or T-cell receptor (TCR)-related kinases (Lyn, Blk, Lck, Src, Frk, Csk, Hck, Fyn, Btk, Syk), the member of the Tec family of non-receptor tyrosine kinases, Bmx, and other receptor tyrosine kinases with lower relevance in B-NHL, such as Ddr2, Egfr, EphA, Erbb, Fgr and Braf (Additional file 1: Table S1). Among these potential targets, a radiometric kinase activity study further showed that IQS019 had an IC₅₀ in the low micromolar range for the BCR kinases Lyn (0.15 μM), Syk (1.6 μM) and Btk (2.1 μM), corresponding to those kinases able to bind the compound in their active site [13]. Accordingly, ectopic expression of each individual kinase in B lymphoid cells rendered them dependent of these kinases for their survival and increased cell sensitivity to IQS019. Indeed, while the calculated IC₅₀ of the compound was 5.4 μM in parental Ba/F3 cells, this value decreased to 2.2, 1.4, and 2.2 μM in Btk-, Lyn-, or Syk-overexpressing cells, respectively (Fig. 1c). Thus, these results confirm that IQS019 is a potent tyrosine kinase inhibitor, with the unique ability to bind and to simultaneously inhibit the three BCR kinases Lyn, Syk, and Btk.

We further assessed the activity of the compound in vitro using a panel of 21 B-NHL cell lines representative of the CLL, MCL, FL and DLBCL subtypes (Table 1). We show that a 5 μM dose of the compound decreased cell proliferation in all the cell lines (range: 29–100%), being MCL and FL cells significantly more sensitive to the compound (mean cytotoxicity at 48 h: 67.2 ± 15%) than CLL cells and DLBCL cells of either activated B-cell (ABC) or germinal centre B-cell (GCB) subtype (mean cytotoxicity at 48 h: 42 ± 9%) (p = 0.0002). Based on these results, a set of 13 CLL primary cultures were exposed for 24 h to increasing doses of IQS019 and cell viability was measured by MTT assay. Although a high variability was observed among cases, the viability decreased in a dose-dependent manner in all the samples treated with the compound (Fig. 1d). The calculated IC₅₀ was 6.1 μM in this set of samples, corresponding to the upper range of the values found in the cell lines. Similar responses were observed in FL and MCL primary cultures (data not shown). Of note, no association could be established between sensitivity to IQS019 and common cytogenetic alterations, TP53 mutation and/or deletion, or IGHV mutational status (Table 1, Additional file 1: Table S2 and Figure S2a). Of interest, a 24 h treatment with a 5 μM dose of the compound induced about 35% apoptosis increase in the representative cell lines UPN-1 and DOHH-2 (Additional file 1: Figure S2b). In CLL and primary cultures (n = 6) the average cell death induction reached 26% (range: 9.5–51.5%), as shown in the representative cases, CLL n.2 and CLL n.10 (Additional file 1: Figure S2b and data not shown). This phenomenon was completely abrogated in the presence of the pan-caspase inhibitor Q-VD-OPh. In parallel, the analysis of phospho-histone H3 levels as a surrogate of mitotic progression indicated a notable decrease of this marker in five out of six primary CLL cases treated with the compound (Additional file 1: Figure S2c). Thus, altogether these results demonstrate that IQS019 antitumor activity in B lymphoid cells involved both a blockade in cell proliferation and the induction of a caspase-dependent cell death.

Based on the above results, we analyzed the effect of IQS019 on the phosphorylation status of Syk, Lyn and Btk in four cell lines representative of MCL (UPN-1), FL (DOHH-2), CLL (JVM-13), and DLBCL (OCI-LY10) subtypes. Cells were incubated for 6 h with increasing concentrations of IQS019 and phosphorylation levels of Syk and Lyn at their respective Tyr352 and Tyr396 residues, were evaluated by Western blot, while Btk phosphorylation at Tyr223 residue was analyzed by flow cytometry. As observed in Fig. 2a, IQS019 treatment led to a dose-dependent dephosphorylation of Syk and Lyn in the four cell lines tested. Consistent with the cytotoxicity of the compound (Table 1), a complete dephosphorylation of the two kinases was observed in UPN-1 and DOHH-2, while a slight, persistent phosphorylation of both Syk and Lyn was detected in OCI-LY10 and JVM-13 cells (Fig. 2a). Regarding Btk phosphorylation, flow cytometry analysis showed a 30% (UPN-1 and OCI-LY10) and a 60% (DOHH-2 and JVM-13) decrease in the relative mean fluorescence intensity ratio (r) of phospho-Btk signal in cells exposed to a 5 μM dose of the compound (Fig. 2b).

In a second step, the four previous cell lines and two representative primary CLL cases were BCR-stimulated with their corresponding anti-Ig in the presence of increasing concentrations of IQS019, and phospho-Syk, phospho-Lyn and phospho-Btk levels were analyzed as above. As shown in Fig. 3a, BCR ligation induced an increase in the phosphorylation levels of Syk and Lyn in all the samples tested, that was hampered by IQS019 in a dose-dependent manner. Remarkably, a dose of IQS019 as low as 1 μM was sufficient to completely counteract the anti-IgM-mediated activation of Syk and Lyn in the highly IgM-responsive (unmutated IGHV) CLL sample showing the greatest efficacy of the stimulation (CLL#10, Fig. 3a). Similarly, IQS019 efficiently counteracted Ig-induced Btk phosphorylation in cell lines, as shown by a 30 to 70% reduction in relative phospho-Btk levels (Fig. 3b). In CLL primary cells, for all but 1 cases out of the 6 examined, IQS019 achieved a 30% reduction in phospho-Btk levels (p = 0.0005), as shown in the representative CLL no.10 (Fig. 3b and data not shown). Of note, in the representative cell line UPN-1, Syk-dependent phosphorylation of Btk at Tyr551 was negligible upon BCR triggering and remained unaffected in the presence of IQS019 (data not shown), suggesting that IQS019-mediated inhibition of Btk requires a direct interaction of the compound with the kinase, rather than an indirect, Syk-mediated signal transduction. Altogether, these results suggest than IQS019 counteracts both constitutive and antigen-induced BCR signaling in B lymphoid cell lines and primary cells.

Migration of neoplastic B cells has been shown to be heavily affected upon exposure to drugs targeting the BCR-associated kinases, as these latest tightly regulate the re-organization of the cytoskeleton required for cell chemotaxis [15]. Thus, we evaluated the effect of IQS019 on the migratory capacity of malignant B cells, using a CXCL12-dependent chemotaxis assay with 3 cell lines harboring detectable levels of CXCR4 (Additional file 1: Figure S3) and in a set of seven CLL primary samples, either untreated or pre-treated with IQS019 or with the standard CXCR4 antagonist AMD3100. The migration induced by recombinant CXCL12 in MCL, FL and DLBCL cell lines was significantly inhibited by the compound at all the doses tested (Fig. 4a). The statistical significance of this effect was higher at the 2.5 μM than at the 1 μM dose in DOHH-2 and OCI-LY10 cells, while an almost complete inhibition of cell migration was achieved in UPN-1 cells at the lowest dose. In the case of CLL primary cells, since the stimulation of BCR has been shown to facilitate CXCL12-mediated migration [15], we evaluated the activity of IQS019 after BCR crosslinking. As shown in Fig. 4b, c, IQS019 significantly overcame IgM-activated, CXCL12-dependent chemotaxis in all the primary samples tested, either at the 1 μM dose (mean inhibition: 51.5%; range: 23.9–85.5%; p = 0.0013) or at the 2.5 μM dose (mean inhibition: 82.9%; range: 63.4–97.6%; p < 0.0001), when compared to untreated control cells. Accordingly, the mean fraction of cells with detectable F-actin polymerization shifted from 13.8% in control cells to 75.1% after CXCL12 stimulation, and was lowered down to 25.9% in the presence of IQS019 (Fig. 4d, *** p = 0.0003). Of special interest, when comparing with AMD3100, IQS019 showed similar, or even superior anti-migratory activity (Fig. 4a, b and c). These results indicate that IQS019-mediated inhibition of BCR upstream kinases may interfere with B cell chemotaxis and tumor cell dissemination.

In order to validate the activity of IQS019 in vivo, we first synthesized the salt form of the compound, thereafter labeled as IQS019-2MeSO3H, and evaluated its single-dose toxicity over 14 days after intravenous administration in healthy immunodeficient (SCID) mice (details in Additional file 1 Methods). As the maximum tolerated dose was not reached, 2 and 10 mg/kg doses were selected for further in vivo experiments. We then developed two different, complementary xenotransplant animal models of the two entities showing increased sensitivity to the compound in vitro, i.e., MCL and FL. Heterotopic MCL tumors were generated in SCID mice subcutaneously inoculated with UPN-1 cells, while a systemic (i.e., characterized by homing of tumor B cells from peripheral blood to spleen) FL tumor model was obtained by intravenous injection of DOHH-2 cells in SCID mice. As shown in Fig. 5a, after two weeks of treatment, mice bearing MCL tumors and dosed with IQS019-2MeSO₃H showed a 63% reduction in tumor volume, when compared to the vehicle group (*p < 0.05). There was not subsequent improvement of the anti-tumor activity of the compound between the 2 mg/kg and the 10 mg/kg dosing, suggesting that optimal activity was reached at the lowest dose. Consistently, tumor metabolism was similarly decreased in both treatment groups, as glucose uptake fell to 50–52% in tumors from all IQS019-2MeSO₃H-exposed animals, irrespective of the dose (Fig. 5b). This effect was closely related to the inhibition of the three BCR-related kinases Syk, Lyn, and Btk, as shown by a complete reduction of their phosphorylated forms in the drug-treated specimens, when compared to the control group (Fig. 5c). Immunohistochemical analysis of representative tumor sections further revealed that IQS019 therapy efficiently reduced the mitotic index and induced apoptosis in UPN-1-derived tumors, as shown by a decreased labeling of phospho-histone H3 and an intracellular increase in the activated form of caspase-3 (Fig. 5c).

In the systemic DOHH-2 mouse model, mice dosing was initiated at day 7 post-inoculation, with a 2 mg/kg IQS019-2MeSO₃H regimen, daily, for 15 days. Once inoculated, FL cells rapidly migrate to the spleen [16]. Therefore, at the end of the procedure, entire spleens were processed, and the presence of malignant B cells was evaluated by labeling with anti-human CD45 antibody and tumor cell recounting on a flow cytometer. IQS019-2MeSO₃H treatment induced a 52% reduction in tumor cell infiltration into the spleen, when compared to vehicle group (Fig. 5d, * p = 0.01). Accordingly, the r fluorescence values of phospho-Syk, phospho-Btk, and phospho-Lyn, decreased by 83, 57, and 33% in tumors B cells purified from IQS019-treated animals (Fig. 5e). Altogether, these results demonstrate that IQS019 is safe and exhibits in vivo efficacy against MCL and FL tumor burden, involving the inhibition of BCR signaling and the blockade of tumor cell homing to lymphoid compartment.

We previously reported that IQS019 presented an increased anti-proliferative activity in vitro when compared to ibrutinib, in a single MCL cell line and at a single time point [13]. To confirm this preliminary experiment, we compared by MTT assay the anti-proliferative effect of IQS019 and ibrutinib at 24, 48, and 72 h, using doses ranging from 0.5 to 10 μM, in a panel of eight cell lines that included the ibrutinib-sensitive MINO, REC-1, UPN-1, DOHH-2, and WSU-NHL and the ibrutinib-resistant Z-138, GRANTA-519, and JVM-2 cells. Figure 6a shows that the mean IC₅₀ of ibrutinib remained significantly high (i.e., > 10 μM) in this set of cell lines, even after a 72-h drug exposure, mainly due to the high values observed in the resistant cell lines (152.4 μM for Z-138, 22.1 μM for GRANTA-519 and 77.4 μM for JVM-2). In contrast, IQS09-mediated proliferation blockade was almost completely reached in all the cell lines after only 24 hours, with a mean IC₅₀ of 6.7 μM (range: 2.7–8.7 μM), which decreased down to 4.1 μM (range: 2.2–11.7 μM) and 3.3 μM (range: 2.2-5.1 μM) at 48 h and 72 h, respectively. Of particular interest, after 72 h these IC₅₀ values were much lower in Z-138 (4.5 μM), GRANTA-519 (5.1 μM), and JVM-2 cells (4.1 μM) than observed after ibrutinib treatment. Accordingly, while a short exposure to ibrutinib only marginally affected CXCL12-dependent cell migration in the UPN-1 cell line, this process was blocked up to 39% in cells cultured with IQS019 (Fig. 6b). To validate these observations in in vivo settings, MCL tumor-bearing mice were treated with a standard 25 mg/kg dose of ibrutinib [17], the equivalent dose of IQS019-2MeSO₃H, or vehicle. While ibrutinib allowed to a significant 25.1% reduction in tumor growth after 2 weeks of treatment (* p = 0.049), IQS019-2MeSO₃H showed superior activity, as it could inhibit the tumor outgrowth up to 42% when compared to vehicle group (Fig. 6c, ** p = 0.006, * p = 0.048). At the pharmacokinetic level, while both compounds presented similar half-life and C_max values in mice after a single oral administration, the total drug exposure over time was considerably improved in the case of IQS019-2MeSO₃H, as shown by a 12 fold increase in the AUC value. Consequently, the global bioavailability dropped from 2 to 4% in the case of ibrutinib, to about 70% in the case of IQS019-2MeSO₃H (Additional file 1: Figure S4b and Table S3), thus suggesting that a better PK profile may account for the improved activity of IQS019 vs ibrutinib in vivo.

To unravel at the molecular level the mechanisms underlying this superior activity of IQS019 over the Btk inhibitor, we established an ibrutinib-resistant cell line designated UPN-IbruR, derived from the parental UPN-1 by repeated drug selection (Additional file 1 Methods). When compared with the parental cell line, UPN-IbruR presented approximately a 10-fold increase in the ibrutinib IC₅₀ after 72 h of treatment (24.6 vs 2.4 μM for parental cells), with negligible difference in IQS019 IC₅₀ (5.6 vs 2.3 μM for parental cells) (Additional file 1: Figure S5a).The ibrutinib resistance phenotype of UPN-IbruR cells was not associated to mutations in BTK or PLCG2 genes, which both harbored a wild type sequence (Additional file 1: Figure S5b), but may rather be associated to the activation of non-canonical NF-κB pathway, as suggested by the overexpression of p52 (Additional file 1: Figure S5c). While a similar dose-dependent decrease in phospho-Lyn and phospho-Btk levels was found in IQS019- and in ibrutinib-treated UPN-1 cells, the expression of phospho-Syk was almost completely lost only in cells exposed to 1 μM IQS019 (Fig. 6d,e). In sharp contrast, in UPN-IbruR cells, the Btk inhibitor failed to modulate the phosphorylation of the three kinases, while IQS019 showed significant inhibitory activity of phospho-Syk and phospho-Lyn at a dose as low as 1 μM (Fig. 6d). However, the compound was unable to downregulate phospho-Btk (Fig. 6e), suggesting that in ibrutinib-resistant cells, the capacity of IQS019 to inhibit Syk and Lyn may allow the compound to maintain a significant antitumoral activity independent of the expression of a non-druggable form of Btk. Altogether, these results point out a significant superior antitumoral activity of pleiotropic BCR kinase targeting by IQS019 over the sole inhibition of Btk, in in vitro and in vivo models of B-NHL.