The Bruton’s Tyrosine Kinase (BTK) Inhibitor Ibrutinib Abrogates Bispecific Antibody-Mediated T-cell Cytotoxicity

LETTER TO THE EDITOR

Bispecific antibodies (BiAbs) have emerged as promising therapeutics for hematologic malignancies, with the CD19/CD3 Bispecific T-cell Engager blinatumomab approved for patient application and many more investigational agents in clinical testing (Viardot and Bargou 2018). Still, many patients fail T-cell-directed BiAb therapy despite target antigen expression on their cancer cells. Better understanding of these failures is crucial to improve BiAb efficacy through optimized construct design and rational combination therapy.

Bruton’s Tyrosine Kinase inhibitors (BTKi) might be such combination partners. By interfering with signalling downstream of the B-cell receptor, BTKi such as ibrutinib and acalabrutinib are highly active against chronic lymphocytic leukaemia (CLL) and other B-cell malignancies (Byrd, et al 2016, Deeks 2017). Beyond their direct anti-cancer effects, BTKi also modulate T-cell function, at least partly through inhibition of the IL-2-inducible T-cell kinase (ITK) (Dubovsky, et al 2013). Several studies suggested ibrutinib may increase the efficacy of chimeric antigen receptor (CAR)-modified T-cells and decrease their toxicity (Fraietta, et al 2016, Ruella, et al 2016, Ruella, et al 2017). Two studies, including one recently reported in this journal, indicated T-cells from ibrutinib-treated patients have greater in vitro anti-tumour efficacy than T-cells from ibrutinib-naïve patients when combined with BiAbs (Gohil, et al 2019, Robinson, et al 2018). On the other hand, also recently reported in this journal, broad-spectrum tyrosine kinase inhibitors can acutely abrogate cytotoxicity of T-cell-directed BiAbs and CAR T-cells in vitro (Fasslrinner, et al 2019). These findings prompted our interest in evaluating how acute exposure of leukaemia cells to ibrutinib and related compounds modulates the anti-tumour efficacy of T-cell-directed BiAbs.

For our studies, we generated CD19/CD3 and CD33/CD3 BiAbs from published sequences and exposed a panel of CD19+ human acute lymphoblastic leukaemia (ALL) cell lines, CD33+ human acute myeloid leukaemia (AML) cell lines, and AML patient specimens to BiAbs and various concentrations of ibrutinib, other BTKi (AVL-292, acalabrutinib, and GDC-0853 (Crawford, et al 2018)), the ITK inhibitor BMS-509744 (Kutach, et al 2010), and the Src family kinase inhibitors PP2 and SU6656 (Blake, et al 2000). All inhibitors were used at non-toxic concentrations (<10% cell death as single agent). In vitro cytotoxicity was determined in 48-hour assays as done previously using T-cells purified from unstimulated healthy donor peripheral blood mononuclear cells (Laszlo, et al 2014) (see Online Supplement for detailed methods).

As shown in Figure 1A, ibrutinib and AVL-292 markedly reduced CD33/CD3 and CD19/CD3 BiAb cytotoxicity. In contrast, the more specific BTKi acalabrutinib and GDC-0853 at similar concentrations did not significantly impair BiAb killing in vitro, with only some cell lines showing decreased cytotoxicity with GDC-0853 (Figure 1A). Qualitatively similar results were obtained with a second healthy T-cell donor (data not shown). Ibrutinib did not inhibit cytotoxicity induced by gemtuzumab or inotuzumab ozogamicin (antibody-drug conjugates targeting CD33 and CD22) or cytarabine, suggesting an inhibitory effect specific to T-cell mediated cytotoxicity (Supplemental Figure 1).

Figure 1 -. Effect of ibrutinib and related compounds on BiAb cytotoxicity.

A. AML cell lines (ML-1, NB4) and ALL cell lines (REH, RS4;11) were treated with 200 pg/mL of CD33/CD3 BiAb or CD19/CD3 BiAb, respectively, and indicated inhibitors for 48 hours in the presence of CellVue-membrane labelled healthy donor T-cells at an effector:target (E:T) ratio of 1:1. Dead target cells were then identified as being negative for CellVue dye and positive for 4’,6-diamidino-2-phenylindole (DAPI) staining via flow cytometry. Results are presented as change in dead cells compared to the same inhibitor condition without BiAb. Note: separate experiments with similar baseline cytotoxicity (i.e. no inhibitor) using the inhibitors GDC-0853, PP2 and SU6656 have been combined on the same axis for clarity. Results are presented as mean±SEM of at least three separate experiments. *, p<0.05; **, p<0.01, ***, p<0.001; ****, p<0.0001 compared to no inhibitor control. B. AML patient specimens were treated with CD33/CD3 BiAb at a dose of 500 pg/mL as well as the indicated inhibitor in the presence of healthy donor T-cells at an E:T of 3:1, and dead cells analysed after 48 hours as in A. Mean±SEM of the 4 samples is also shown (grey bars). Results are presented as change in dead cells as in A. Abbreviations: BTKi, Bruton’s tyrosine kinase inhibitor; ITKi, IL-2-inducible T-cell kinase inhibitor; SRCi, Src family kinase inhibitor.

Since ibrutinib inhibits not only BTK but also ITK (Dubovsky, et al 2013), we evaluated whether ITK inhibition might account for the suppressive activity of ibrutinib using the ITK-specific inhibitor BMS-509744 (Kutach, et al 2010). Unlike ibrutinib, however, BMS-509744 did not significantly inhibit BiAb cytotoxicity (Figure 1A). Since ibrutinib also inhibits Src family kinases (Crawford, et al 2018), we tested the broad-spectrum Src inhibitor PP2 and found it to decrease BiAb-induced cytotoxicity in a manner similar to ibrutinib. On the other hand, the more specific Src inhibitor SU6656 had a less pronounced inhibitory effect (Figure 1A). Qualitatively similar results were obtained with these inhibitors when 4 AML patient specimens were treated with CD33/CD3 BiAb (Figure 1B). Together, these data suggest a profound inhibitory effect of ibrutinib on BiAb-induced cytotoxicity that is unlikely to be mediated through BTK inhibition.

BiAb-mediated T-cell cytotoxicity depends on both T-cell and tumour cell factors (Viardot and Bargou 2018). To determine whether ibrutinib abrogated BiAb cytotoxicity through a T-cell- or target cell-dependent mechanism, we pre-treated either healthy donor T-cells or leukaemia cells with ibrutinib for 24 hours, after which we removed the BTKi and then performed co-culture cytotoxicity assays with fresh target or T-cells, respectively. Pre-treatment of T-cells but not leukaemia cells significantly reduced BiAb-induced cytotoxicity (Figure 2A), an effect that was not due to a reduction of T-cell numbers during ibrutinib pre-treatment (data not shown). Consistent with this inhibitory effect on T-cells, ibrutinib abrogated BiAb-induced T-cell activation as measured by CD25 and CD69 cell surface display, whereas acalabrutinib had no effect (Figure 2B). These data suggest ibrutinib-mediated inhibition of BiAb-mediated cancer cell killing is due to an inhibitory effect on T-cell effector function rather than induction of leukaemia cell resistance.

Figure 2 -. Ibrutinib suppresses BiAb-mediated killing via inhibition of T-cell activation.

A. Leukaemia target cells (REH, NB4) or healthy donor T-cells were incubated for 24 hours with 10 μM ibrutinib before cells were washed and 48-hour co-culture assays performed in the presence of CD19/CD3 BiAb (200 pg/mL, REH cells) or CD33/CD3 BiAb (500 pg/mL, NB4 cells); pre-treated target cells were cultured with fresh healthy donor T-cells, and pre-treated T-cells were cultured with untreated target cells. T-cells from both conditions were stained with CellVue membrane dye and were incubated with target cells at an E:T of 1:1. Dead target cells were then identified as negative for CellVue dye and positive for 4’,6-diamidino-2-phenylindole (DAPI) staining via flow cytometry, and change in dead target cells (compared to same pre-treatment condition without BiAb) was then calculated. Results are presented as mean±SEM of at least three separate experiments. B. ML-1 and RS4;11 cells were co-cultured with CD33/CD3 or CD19/CD3 BiAb, respectively, each at a dose of 100 pg/mL, as well as healthy donor T-cells at an E:T of 1:1. T-cells were then stained with anti-CD25 or anti-CD69 and positive cells identified via flow cytometry compared to isotype control. Results are presented as mean±SEM of at least three separate experiments. *, p<0.05; **, p<0.01 compared to no inhibitor control.

Together, our studies indicate acute exposure to BTKi impairs T-cell activation and lysis of target cells upon treatment with CD3-directed BiAbs, at least partially due to effects that are independent of BTK inhibition. Further mechanistic studies will be necessary to identify which signalling pathways are involved in this inhibitory effect of BTKi. Of note, this acute effect may be compensated for in malignancies in which the BTKi exert direct toxicity to tumour cells as, for example, in CLL. However, particularly in malignancies where BTKi have no such direct anti-tumour effects, acute BTKi treatment may reduce the overall treatment efficacy but also side effects related to T-cell activation (e.g. cytokine release syndrome), a possibility clinicians should be aware of. As BiAbs and small molecular inhibitors beyond BTKi are increasingly combined with each other (e.g. fms-related tyrosine kinase 3 [FLT3] inhibitors and BiAbs for AML) (Fasslrinner, et al 2019, Reiter, et al 2018), such antagonistic effects should be considered and carefully examined preclinically and during early clinical testing.

On the other hand, prolonged BTKi exposure may enhance BiAb-induced cytolytic activity of T-cells (Gohil, et al 2019, Robinson, et al 2018). Based on our data indicating some BTKi have a dual effect with regard to modulation of T-cell-induced cytotoxicity, one may speculate – but future experiments will need to address – whether for patients with prolonged exposure to BTKi, BTKi discontinuation during BiAb or CAR T-cell therapy might lead to enhanced efficacy of the immunotherapeutic.