The interplay between microenvironmental signaling and novel targeted drugs in CLL - Chapter 4: Targeting antigen-independent proliferation in chronic lymphocytic leukemia through differential kinase inhibition

(1)

UvA-DARE is a service provided by the library of the University of Amsterdam (https://dare.uva.nl)

UvA-DARE (Digital Academic Repository)

The interplay between microenvironmental signaling and novel targeted drugs in

CLL

Thijssen, R.

Publication date

2016

Document Version

Final published version

Link to publication

Citation for published version (APA):

Thijssen, R. (2016). The interplay between microenvironmental signaling and novel targeted

drugs in CLL.

General rights

It is not permitted to download or to forward/distribute the text or part of it without the consent of the author(s)

and/or copyright holder(s), other than for strictly personal, individual use, unless the work is under an open

content license (like Creative Commons).

Disclaimer/Complaints regulations

If you believe that digital publication of certain material infringes any of your rights or (privacy) interests, please

let the Library know, stating your reasons. In case of a legitimate complaint, the Library will make the material

inaccessible and/or remove it from the website. Please Ask the Library: https://uba.uva.nl/en/contact, or a letter

to: Library of the University of Amsterdam, Secretariat, Singel 425, 1012 WP Amsterdam, The Netherlands. You

will be contacted as soon as possible.

(2)

TARGETING ANTIGEN-INDEPENDENT

PROLIFERATION IN CHRONIC LYMPHOCYTIC

LEUKEMIA THROUGH DIFFERENTIAL

KINASE INHIBITION

(3)

TARGETING ANTIGEN-INDEPENDENT PROLIFERA

TION IN CLL

4

Erik Slinger1,2_{, Rachel Thijssen}1,2_{, Arnon P. Kater}2,3*_{, and Eric Eldering}1,3*

Departments of Experimental Immunology1_{and Hematology}2_{, Academic Medical Center, Amsterdam,}

the Netherlands; 3_{Lymphoma and Myeloma Center Amsterdam (LYMMCARE), the Netherlands}

*Shared senior authorship

Submitted for publication

ABSTRACT

The clinical success of B cell receptor (BCR) signaling pathway inhibitors in chronic lymphocytic leukemia (CLL) is predominantly attributed to inhibition of adhesion in and migration towards the lymph node. Proliferation of CLL cells is also restricted to this protective niche, but the underlying mechanism(s) is/are not fully known. In contrast to normal B cells, triggering via TLR-, TNF- or cytokine- receptors can induce proliferation of CLL, but BCR stimulation does not. We here investigated the effects of various inhibitors that target B cell receptor (BCR) signaling, in the context of proliferation triggered either via CD40L/IL-21 or after CpG stimulation.

CD40L/IL-21-induced proliferation could be inhibited by idelalisib and ibrutinib. We demonstrate this was due to blockade of CD40L-induced ERK-signaling, whereas NF-kB signaling was not affected. Targeting JAKs, but not SYK, fully blocked CD40L/IL-21-induced proliferation. In contrast, PI3K, BTK as well as SYK inhibition fully prevented CpG-induced proliferation. Knock-down experiments showed that CD40L/IL-21 did not co-opt upstream BCR components such as CD79A, and was, in contrast to CpG-induced proliferation, not dependent on auto/paracrine BCR engagement. Our data indicate that currently applied BTK/PI3K inhibitors also target antigen-independent proliferation in CLL, and suggest that additional targeting of JAK and/or SYK might be clinically useful.

(4)

4

TION IN CLL

INTRODUCTION

A hallmark of chronic lymphocytic leukemia (CLL) is a dichotomy in survival and proliferation between the leukemic cells in peripheral blood (PB) and the lymph node (LN). PB CLL cells are in cell cycle arrest and sensitive to pro-apoptotic signals, whereas CLL cells in the LN proliferate and upregulate proteins that mediate resistance to apoptosis1,2_{. While normal} B cells that receive B cell receptor (BCR) stimulation proliferate vigourously3_{, proliferation} of CLL cells has been mostly described in antigen-independent context using combined stimulations with CD40L/CpG1,4_{, CD40L/IL-21}5_{, or IL-15/CpG}6_{. In fact, although} BCR-triggering can induce Ca2+_{-signals in CLL}7_{, it does not induce proliferation}5_{. This distinction} in proliferative signaling between non-malignant B cells and CLL cells is noteworthy, considering the clinical success of the BTK inhibitor ibrutinib and the PI3K

δ

inhibitor idelalisib that inhibit critical components of BCR signaling. Their immediate clinical effect is lymphocytosis combined with rapid reductions in LN sizes8,9_{. In vitro experiments have} confirmed that the effects of ibrutinib and idelalisib involve disruption of BCR-mediated adhesion in, and migration to, the LN10,11_{. Apart from this clear role in blocking adhesion} and homing, ibrutinib has been reported to interfere with proliferation, both in vitro4_and

clinically12_{. Furthermore, experiments with novel inhibitors targeting PI3K}

δ

_{have shown} that it may be involved in BCR-independent proliferation13_.

We have previously reported that a combination of CD40L and IL-21 signaling can induce antigen-independent proliferation in CLL cells5_{. However, the mechanism underlying this} mode of proliferation, and to which extent it engages JAK/STAT and/or BCR components remains unknown. In this context, the term antigen-independent proliferation applies to modes of CLL stimulation that do not involve specific antigen or direct triggering of the BCR. In addition, a recent report established an intriguing signaling route from the Toll-like receptor 9 (TLR9), to the BCR via secreted IgM, involving SYK and BTK in CLL cells harboring an unmutated BCR14_{. Whether this pathway is also involved in CpG-} and/or IL-21-mediated proliferation is as yet unknown. We investigated these aspects of intracellular signaling in antigen-independent proliferation and report that inhibitors of SYK, PI3K, BTK and JAKs differentially block various routes. In this study we compared and contrasted CD40L/IL-21-induced proliferation with CpG-induced proliferation and found that CpG-mediated proliferation is due to auto/paracrine BCR triggering via IgM secretion which can be blocked using SYK inhibitors. In contrast, CD40L/IL-21-mediated proliferation does not engage proximal BCR components and is fully blocked by JAK inhibitors.

MATERIALS AND METHODS

Patient material

After written informed consent, patient blood was obtained during diagnostic or follow-up procedures at the Departments of Hematology and Pathology of the Academic Medical Center Amsterdam. This study was approved by the AMC Ethical Review Board and conducted in agreement with the Declaration of Helsinki. Blood mononuclear cells of patients with CLL, obtained after Ficoll density gradient centrifugation (Pharmacia Biotech, Roosendaal, The Netherlands) were frozen and stored as previously described

(5)

TION IN CLL

4

(13). Expression of CD5 and CD19 (both Beckton Dickinson (BD) Biosciences, San Jose, CA, USA) on leukemic cells was assessed by flow cytometry (FACScanto; BD Biosciences). CLL samples included in this study contained 81-99% CD5+_/CD19+_{cells. Patients are listed} in Supplemental Table 1.

Reagents

Both ibrutinib and biotinylated ibrutinib were obtained from Pharmacyclics (Sunnyvale, CA, USA). The MEK inhibitor U0126 was from BPS bioscience (San Diego, CA, USA). All other inhibitors were obtained from Selleckchem (Houston, TX, USA).

Cell culture and monitoring of proliferation

Cells were labeled with 0.25 μM CFSE (Life technologies) for 8 minutes in PBS at a concentration of 20 million cells per ml. Subsequently, the cells were co-cultured at a density of 1 million cells per ml on NIH-3T3 fibroblasts or NIH-3T3 cells expressing human CD40L for another 5 days, with the addition of 25 ng/ml IL-21 (R&D Systems, Minneapolis, MN, USA), 15 ng/ml IL-15 (Peprotech, Rocky Hill, NJ, USA) or 1.5 ug/ml CpG ODN2006 (Invivogen, San Diego, CA, USA). The cells were analyzed for CFSE dilution by flow cytometry and the division index was calculated. Relative division index was calculated over distinct experiments by dividing the index from experimental conditions by that of vehicle treated cells.

Western-blot analysis

Cell lysates were prepared by lysing in RIPA buffer (150 mM NaCl, 1 mM EDTA, 50 mM Tris-HCl pH 7.4, 0.1% SDS, and 1 % NP-40) and subsequently subjected to 10 seconds of sonication in a Branson sonificator (Danbury, CT, USA). Lysates were analyzed by SDS/ PAGE and subsequent Western-blotting. The following antibodies were used: mouse anti-BTK (BD), rabbit anti-phospho-STAT3 (Cell Signaling ), rabbit anti-cyclin D2, rabbit anti-phospho-p44/p42 , and anti-Actin (Santa Cruz).

Coupling of ibrutinib-biotin to avidin-agarose and pulldown

Ibrutinib-biotin was incubated for 30 minutes at room temperature at a concentration of 100 uM with avidin-agarose suspension in PBS on a rotator. The resulting ibrutinib-biotin agarose slurry was blocked with 10 mM biotin for an additional 10 minutes, and washed 3 times with PBS. Cytoplasmic lysates were prepared by lysing the cells using NP-40 lysis buffer as described above. Subsequently, the lysates were divided over the different pulldown conditions and incubated overnight at 4°C on a rotator. The ibrutinib-agarose was washed 3 times with NP-40 lysis buffer after which Laemmli buffer (60 mM Tris-HCl ph 6.8, 10% glycerol, 2% SDS, 100 mM DTT) was added and the samples were heated to 95°C for 10 minutes. The resulting eluates were then analyzed by Western-blot.

(6)

4

TION IN CLL

NF-

κ

B DNA binding and IgM ELISA

Nuclear lysates were prepared by lysing cells with the NucBuster kit (Merck Millipore, Darmstadt, Germany). The resulting nuclear lysates were normalized for protein content and 5 μg was used as input for the DNA-binding ELISA. The ELISA was performed with the TransAM NF-

κ

B kit (ActiveMotif, Carlsbad, CA, USA). Secreted IgM was measured in conditioned medium obtained after 72 hours of co-culture with NIH-3T3 or NIH-3T3 expressing CD40L, with 1.5 μg/ml CpG or 25 ng/ml IL-21 added to the medium. IgM levels were determined by sandwich ELISA, using a 96-well plate coated with anti-IgM (Dako, Glostrup, Denmark). The plate was blocked with 1% BSA and incubated for 2 hour with sample. Following this, the plate was washed and incubated with secondary antibody conjugated to HRP and quantified using TMB (Merck, Darmstadt, Germany).

Knockdown of BTK and CD79A in primary CLL cells

PBMCs were transfected with siRNA targeting BTK or CD79A (Ambion, Thermo Fisher Scientific, Waltham, MA, USA) by electroporation with an Amaxa machine (Lonza, Basel, Switzerland), siControl (Ambion, Thermo Fisher Scientific) was used as a negative control. BTX Express electroporation buffer (Harvard Apparatus, Holliston, MA, USA) was used to make a suspension of 60 million cells/ml which were then subjected to electroporation using program X-001. After which the cells were cultured for an additional 72 hours. Knock-down efficiency was monitored by intracellular FACS analysis of BTK and CD79A levels, using mouse anti-BTK-Alexafluor 647 and mouse anti-CD79A-PE.

Data analysis and statistics

Flow cytometry data was analyzed using FlowJo (Treestar, Ashland, OR, USA). All statistical analyses were performed using Graphpad Prism. Student’s t-test, and ANOVA analyses were used when appropriate, *=p<0.05, **=p<0.01, and ***=p<0.001 indicating level of significance. Error bars indicate standard error of the means.

RESULTS

Kinases involved in BCR signaling are required for antigen-independent

proliferation.

In order to identify kinases involved in CD40L/IL21 mediated proliferation, stimulated CLL cells were treated with different kinase inhibitors that target both non-BCR and BCR signaling pathways known to be active in CLL cells (Figure 1A). Inhibition of JAKs with INCB018424 (INCB) resulted in inhibition of proliferation, consistent with the requirement of IL-21 for proliferation. As a further control cells were treated with rapamycin, which completely abolishes proliferation, consistent with the block of cell metabolism associated with inhibition of mTOR. In contrast to the observed effects of c-Abl inhibitors in CD40L mediated pro-survival signaling15,16_{, Abl-specific inhibition by imatinib treatment did} not inhibit proliferation, whereas dasatinib which targets Abl and BTK did. Ibrutinib or idelalisib inhibited proliferation at 1 µM as well as at 100 nM (Supplemental Figure 1). These effects were consistent and independent of the mutational status of the BCR (Figure

(7)

TION IN CLL

4

Figure 1. CD40L/IL-21-induced proliferation can be inhibited by various kinase inhibitors. (A) Left panel: relative division index calculated from CFSE dilution assays from thawed CLL cells, stimulated with CD40L and IL-21 for 5 days, and treated with 1 µM of indicated kinase inhibitors (n=8). Right panel; representative example of primary data. (B) Relative division index calculated from CFSE dilution assays plotted for CLL cells with a mutated BCR (M-CLL) compared with CLL cells with unmutated BCR (U-CLL) (n=4). (C) Relative division index calculated from CLL cells stimulated with CD40L and IL-21 and treated with 1 µM of ibrutinib, idelalisib or both (n=4). (D) CFSE dilution assay for CLL cells stimulated with CD40L and IL-21. The cells were treated with 1 µM of ibrutinib for the indicated times, after which the medium was replaced (n=7). (E) CFSE dilution assay performed on CLL cells from patients before start of treatment (white bars) and after 7 days of treatment with ibrutinib (black bars). The cells were stimulated with the indicated stimuli and treated with 1 µM ibrutinib (n=3). Significance was calculated compared to vehicle treated cells by a one-way ANOVA with a Bonferroni post-test (*p < 0.05, **p < 0.01, ***p < 0.001, ns not significant).

1B). A combination of ibrutinib and idelalisib did not result in a significant further reduction of proliferation compared to either compound alone, suggesting that PI3K

δ

and BTK are acting in the same pathway (Figure 1C). Inhibition of the more proximal kinase SYK with R406, did not affect proliferation. To assess whether the effect of ibrutinib on proliferation was specific, we applied wash-out experiments making use of ibrutinib’s covalent binding to BTK to differentiate between potential non-covalent interactions. CD40L/IL-21 stimulated CLL cells were treated for 1 hour, 24 hours, or 5 days with ibrutinib. A significant effect on proliferation was already apparent after 1 hour treatment with ibrutinib and maximum inhibition was achieved after 24 hours (Figure 1D), indicating that the effect is on-target via BTK.

A

Figure 1

CD40L + IL-21 CD40L + IL-21 + Ibrutinib CD40L CD40 L -Ibruti nib Idelalisib 0.0 0.5 1.0 1.5 _M-CLL U-CLL CD40L + IL-21 R el at ive di vi sion ind ex Events CFSE

B

C

CD40 L -Ibruti nib Idelalisib Ibruti nib + Idelal isib 0.0 0.5 1.0 1.5 R el at ive di vi sion ind ex ns ns CD40 L - _1h 24h 5d 0.0 0.5 1.0 1.5 CD40L + IL-21 + Ibrutinib R el at ive di vi sion ind ex * *** ***

D

CD40 L -Ibruti nib 0.0 0.5 1.0 1.5 Baseline 7d ibrutinib R el at ive di vi sion ind ex *

E

CD40 L -Imatin ib Dasa tinib Ibruti nib Idelal isib_R406 Rapa mycinINCB 0.0 0.5 1.0 1.5 CD40L + IL-21 *** ** ** *** *** R el at iv e di vi si on in de x CD40L + IL-21 CD40L + IL-21

(8)

4

TION IN CLL

To support these in vitro findings, paired CLL samples obtained at baseline and seven days after start of ibrutinib treatment were compared (Figure 1E). The cells obtained after seven days of ibrutinib treatment were unable to proliferate upon stimulation with CD40L and IL-21, demonstrating that in vivo inhibition of BTK also blocks non-BCR-mediated proliferation in vitro.

CpG signaling was recently linked with BCR triggering in unmutated CLL14_{, and} ibrutinib has been described to affect CpG-induced proliferation17_{. Therefore the effects of} the same panel of kinase inhibitors used in the experiment described above were assessed in two different models of CpG-induced proliferation. Both CD40L/CpG (Figure 2A) and IL-15/CpG (Figure 2B) induced proliferation showed a more pronounced response to ibrutinib and idelalisib, compared to CD40L/IL-21 mediated proliferation. A major difference between CpG mediated proliferation and CD40L/IL-21-mediated proliferation was the sensitivity to the specific SYK inhibitor R406 for CpG stimulated cells, which fits with the recently reported involvement of SYK specifically in CpG signaling14_.

Ibrutinib interferes with CD40L-mediated ERK-signaling but not with JAK/

STAT and NF-kB signaling.

The clinical efficacy of ibrutinib in CLL is undisputed, but there is uncertainty whether all its effects result from displacement of cells from the LN by interfering with adhesion10,11_or

A

B

Figure 2

CD40 L -Imatin ib Dasa tinib Ibruti nib Idelal isib_R406 Rapa mycinINCB 0.0 0.5 1.0 1.5 CD40L + CpG *** *** *** _*** _*** R el at iv e di vi si on in de x -IL-15CpG -Imatin ib Dasa tinib Ibruti nib Idelal isib_R406 Rapa mycinINCB 0.0 0.5 1.0 1.5 IL-15 + CpG *** *** *** *** *** * R el at iv e di vi si on in de x

Figure 2. CpG-induced proliferation is impaired by inhibition of BCR-associated kinases.Relative division index calculated for CLL cells, stimulated with CD40L and CpG (A) or CpG and IL-15 (B) for 5 days, and treated with 1 µM of indicated kinase inhibitor (n=8). Significance was determined by a one-way ANOVA with a Bonferroni post-test as in Figure 1.

(9)

TION IN CLL

4

whether inhibition of BTK can also have more direct effects on proliferation. Therefore, we further explored the underlying mechanism of BTK involvement in CD40L/IL-21 mediated proliferation, in comparison with PI3K

δ

inhibition where relevant. First, considering that one of the effects of CD40L stimulation of CLL cells is an increase in IL-21R membrane expression18_{, the effect of ibrutinib on IL-21R or TLR9 expression was assessed. Flow} cytometry analysis showed that ibrutinib did not affect IL-21R (Supplemental Figure 2A) or TLR9 expression (Supplemental Figure 2B).

Secondly, ibrutinib has been described to bind JAK319_{, which possesses a cysteine} residue at the homologous site of BTK. Therefore, as read-out for JAK3 activity, STAT3-phosphorylation after CD40L/IL-21 stimulation was investigated for sensitivity to ibrutinib (Figure 3A). Stimulation with CD40L alone resulted in modest STAT3 phosphorylation, which fits with an earlier report on direct interaction of CD40 with JAK320_{. Additional} stimulation with IL-21 resulted in pronounced phosphorylation of STAT3, which was completely unaffected by ibrutinib. As a control, the JAK-inhibitor INCB reduced phospho-STAT3 levels. CD40L/IL-21 induced phospho-STAT3 phosphorylation is not completely abolished after treatment with INCB, because INCB is 130-fold more selective for JAK1/2 than for JAK321_{. Thus, upstream IL-21 signaling was apparently unperturbed by BTK inhibition.}

Thirdly, CD40 signaling was investigated, NF-

κ

B signaling was analyzed by examining nuclear lysates from CD40L stimulated cells for p65 (canonical) and p52 (non-canonical) activity using DNA binding ELISA (Figure 3B). This showed that CLL cells respond to CD40L stimulation by signaling through both canonical and non-canonical NF-

κ

B pathways. Ibrutinib did not affect CD40L-induced NF-

κ

B signaling. Also, idelalisib did not show any effect on the NF-

κ

B signaling cascade. As a control, cells treated with the IKK

β

inhibitor BMS-345541 were also analyzed and as expected showed an impaired NF-

κ

B response to CD40L stimulation. These results were verified by Western-blot analysis of p100/p52 cleavage of samples stimulated in the aforementioned manner (Supplemental Figure 3).

Fourthly, CD40 signaling also activates MAPK, which results in PKC-independent phosphorylation of ERK22_{. Indeed, stimulation with CD40L resulted in ERK1/2} phosphorylation, which was not further increased by IL-21. Treatment of the cells with ibrutinib strongly inhibited CD40L-induced ERK1/2 phosphorylation, suggesting that this is the relevant signaling node controlled by BTK (Figure 3C,D). To probe for a direct association between BTK and ERK1/2, we investigated ibrutinib binding to kinases, using a pulldown strategy utilizing a biotinylated form of ibrutinib immobilized onto avidin-agarose. Evidently, biotinylated ibrutinib binds to BTK, which can be prevented by pretreating the lysates with free ibrutinib, but we could not detect association to ERK1/2 (Figure 3E). A comprehensive pharmacoproteomics approach has shown that in primary CLL cells ibrutinib only functionally binds to BTK (manuscript in preparation), arguing that the observed effect on ERK1/2 is on-target.

Concerning the anti-proliferative aspect, we reasoned that since both STAT323_and MAPK24_{signaling control cyclin D2 levels, interfering with CD40L-induced MAPK signaling} with ibrutinib or idelalisib might decrease cyclin D2. When CLL cells were stimulated with CD40L there was a minor increase in cyclin D2 levels, which was strongly enhanced by IL-21, reflecting that stimulation of both CD40 and IL-21 pathways is required to induce proliferation. Confirming the expectation above, treatment with ibrutinib and also idelalisib

(10)

4

TION IN CLL

Figure 3

A

B

P-STAT3 β-actin CD40L IL-21 Ibrutinib INCB -+ -+ + -+ + + -+ + -+ β-actin P-ERK1/2 CD40L IL-21 Ibrutinib INCB + + -+ + + -+ + -+ + -+ + -+ -+

-C

D

E

BTK ERK1/2 + -Biot.-ag. Ibr.-biot.-ag. Ibr -+ -+ + Input 0.0 0.5 1.0 1.5 CanonicalNon-canonical -CD40L Ibrutinib Idelalisib BMS-345541 + -+ + -+ -+ -+ -+ R el at iv e A4 50 Mock CD40L Ibrutinib 0.0 0.5 1.0 1.5 ** ** R el at iv e ER K p ho sp ho ry la tio n

Figure 3. Analysis of downstream effects of ibrutinib on combined CD40L/IL-21 stimulation demonstrates CD40L-induced ERK1/2 phosphorylation is prevented. (A) Western-blot analysis showing STAT3 phosphorylation in CLL cells subjected to the indicated conditions. Beta-actin was used as a loading control. Representative result shown from 3 independent experiments. (B) DNA-binding ELISA for canonical NF-

κ

B signaling (measuring p65 binding) and non-canonical NF-

κ

B signaling (measuring p52 binding) (n=4). (C) Western-blot analysis showing ERK1/2 phosphorylation in primary CLL cells subjected to the indicated conditions. Actin was used as a loading control. Representative result of 4 independent experiments. (D) Quantitative analysis of ERK1/2 phosphorylation as determined by Western-blot analysis comparing CD40L stimulated cells with or without treatment with 1 µM ibrutinib (n=4). (E) Pulldown performed with ibrutinib-biotin coupled to avidin agarose (Ibr-biot.-ag.). Pulldowns with free biotin-coupled avidin-agarose (Biot.-ag.) or lysates that were pre-treated with free ibrutinib (Ibr) as indicated were used as controls. The eluates were analyzed by Western-blot analysis for BTK and ERK. Representative result of 3 independent experiments is shown.

indeed results in reduced CD40L/IL-21-induced cyclin D2 expression (Figure 4A), and similarly for CD40L/CpG induced proliferation (Figure 4B). Consistent with the observed partial inhibition of proliferation, cyclin D2 levels are not fully reduced in the case of CD40L/IL-21 stimulated cells as opposed to the more pronounced reduction in CD40L/ CpG stimulated cells. In agreement with these results, inhibition of upstream MEK by U0126 or the downstream target of cyclin D2, CDK4 by PD0332991, resulted in impaired proliferation (Figure 4C and 4D).

Collectively, these data indicate that both BTK and PI3K

δ

are involved in CD40 signaling through the MAPK axis, and impact BCR-independent proliferation through converging JAK-STAT and MAPK signals to control cyclin D2 levels.

(11)

TION IN CLL

4 Dichotomy between CD40L/IL-21- versus CpG-induced proliferation

depends on involvement of the BCR complex

In the experiments described above we demonstrated that CD40L/IL-21-mediated proliferation can be inhibited by interfering with CD40L-induced signaling through ERK. In stark contrast with CD40L/IL-21-induced proliferation, CpG-induced proliferation was effectively blocked by inhibition of SYK. We next focused on the distinctions between these pathways. This distinction was reflected in the capacity to produce secretory IgM (sIgM); CLL cells stimulated with CpG produced more sIgM than those stimulated with CD40L/IL-21. CpG-induced sIgM production was quite variable among four unmutated CLL samples tested (100-5000 pg/ml after 72 hours of culture), yet was completely dependent upon SYK, PI3K

δ

, and BTK activity, whereas the modest CD40L/IL-21-induced sIgM production was unperturbed upon inhibition of these kinases. This distinction was most noticeable in CLL samples with high IgM secretion upon CpG stimulation as shown in Figure 5A. In agreement with data shown in Figure 3C, ibrutinib and idelalisib inhibited ERK phosphorylation in cells stimulated with either CD40L/IL-21 or CpG. In contrast, the SYK inhibitor R406 showed a marked dichotomy between these stimuli, inhibiting CpG-induced ERK phosphorylation but not CD40L-induced ERK phosphorylation

Figure 4

Cyclin D2 -CD40L IL-21 Ibrutinib Idelalisib + -+ + -+ + + -+ + -+ β-actin

A

B

Cyclin D2 -CD40L CpG Ibrutinib Idelalisib + -+ + -+ + + -+ + -+ β-actin Mock_CD40L -Ibruti nib Idelal isib 0.0 0.5 1.0 1.5 CD40L + IL-21 ** ** R el at iv e C yc lin D 2 le ve l Mock_CD40L -Ibruti nib Idelal isib 0.0 0.5 1.0 1.5 CD40L + CpG * _** R el at iv e C yc lin D 2 le ve l

C

CD40 L -U012 6 PD03 3299 1 0.0 0.5 1.0 1.5 CD40L + IL-21 *** *** R el at iv e di vi si on in de x CD40 L Vehic le U012 6 PD03 3299 1 0.0 0.5 1.0 1.5 CD40L + CpG *** *** R el at iv e di vi si on in de x

D

Figure 4. Downstream cyclin D2 expression is prevented by BTK or PI3K

δ

inhibition. (A) Western-blot analysis showing cyclin D2 expression under the indicated conditions, after 72 hours of culturing in the presence of IL-21 and or CD40L. Actin was used as a loading control. Representative result of 3 independent experiments is shown. (B) Western-blot analysis showing cyclin D2 expression under the indicated conditions, after 72 hours of culturing in the presence of CpG and or CD40L. Actin was used as a loading control. Representative result of 2 independent experiments is shown. (C) and (D) CFSE dilution assay showing response of CLL cells stimulated with CD40L/IL-21 or CpG to treatment with 1 µM of the MEK inhibitor U0126 or the CDK4 inhibitor PD0332991 (n=4).

(12)

4

TION IN CLL

(Figure 5B). To further substantiate this distinction we examined involvement of CD79A as an upstream component of the BCR which is critical for BCR signaling25_{, and compared} siRNA knock-down of BTK or CD79A. Intracellular flow-cytometry showed approximately 50% knockdown of either BTK or CD79A protein compared to cells transfected with control siRNA after 72 hours (Figure 5C). Proliferative capacity were generally reduced after the nucleofection procedure, so functional effects after knock-down were analyzed using phospho-ERK1/2 as read-out. Using CD40 stimulation, phospho-ERK1/2 levels showed a significant decrease with siRNA targeting BTK (siBTK), but not with siRNA targeting CD79A (siCD79A) (Figure 5D). This indicates that CD40L-induced ERK1/2 phosphorylation involves recruitment of BTK independent of the upstream BCR component CD79A. In summary, these two modes of antigen-independent proliferative stimulation differ in involvement of proximal BCR components.

Figure 5

A

0 2 4 6 -+ -+ -+ -+ -+ -+ -+ + -+ -+ -+ -+ + -+ -+ -+ -+ -+ -+ CD40L CpG IL-21 Ibrutinib Idelalisib R406 Ig M (u g/ m l)

B

CD40L IL-21 CpG Ibrutinib Idelalisib R406 + + -+ -+ -+ + -+ -+ -+ -+ + -+ + -+ + -+ -+ + -+ -P-ERK1/2 β-actin Btk _CD79A siControl siBtk/CD79A Isotype Ev en ts

C

siControl siBTK siCD79A BTK β-actin + -+ -+

D

siBtk siCD79A 0.0 0.5 1.0 1.5 * R el at iv e ER K p ho sp ho ry la tio n P-ERK1/2

Figure 5. Differential involvement of BTK and the BCR in CD40L/IL-21 versus CpG induced stimulation. (A) Analysis of secreted IgM levels in the medium of CLL cells after 72 hours of culture under indicated conditions. (B) Western-blot analysis of cells stimulated as indicated with CpG or CD40L/IL-21, in the presence or absence of indicated kinase inhibitors. (C) FACS analysis showing decreased BTK and CD79A expression in CLL cells transfected with siRNA against BTK and CD79A, respectively. Representative result of 4 independent experiments is shown. (D) Western-blot analysis of cells transfected with siRNA targeting BTK or CD79A and which were subsequently stimulated with CD40L for 72 hours. Effects on downstream signaling were assessed by probing for BTK and phospho-ERK1/2. Actin was used as a loading control. Quantification of 4 separate knock-down experiments is shown in the right panel. Significance was determined with a Student’s t-test (*p < 0.05).

(13)

TION IN CLL

4 DISCUSSION

In the present study, the molecular mechanisms underlying antigen-independent proliferation in CLL were investigated with the aid of currently approved drugs, as well as other kinase inhibitors under clinical development. Combined, the data uncovered two essentially distinct routes to proliferation; CD40L/IL-21 engages JAKs, BTK and PI3K independent of proximal BCR components, while CpG indirectly triggers the BCR via sIgM secretion and thus involves SYK, PI3K and BTK. This difference is underlined by the ability of JAK inhibitors to block CD40L/IL-21-mediated proliferation, versus the capacity of SYK inhibitors to prevent proliferation induced by CpG.

We first focused on the involvement of BTK and PI3K

δ

in CD40L/IL-21 mediated proliferation. BTK is typically associated with BCR-signaling, but has also been reported to act in CD40-signaling26_{in lymphoma cell lines using agonistic anti-CD40. In human} microvascular endothelial cells PI3K

δ

activity has been linked to TRAF627_{, which is recruited} to the CD40 signalosome upon ligation28_{. Experiments performed in the aforementioned} study showed that CD40L-induced activity could be blocked by the use of PI3K-inhibitors. The inhibition of proliferation observed in the present study is most likely the result of inhibition of CD40L-induced ERK activation, while leaving IL-21-induced signaling intact as evidenced by our data which demonstrate that STAT3-phosphorylation is not affected by ibrutinib. In addition, we found that CD40L-induced NF-

κ

B signaling was not impaired by ibrutinib nor idelalisib. Earlier analyses of CLL cells from PB, LN and bone marrow from patients treated with ibrutinib indicated inhibition of BCR and NF-

κ

B signaling12_. The beneficial clinical effects of BTK and PI3K

δ

inhibitors has generally been interpreted as confirmation of a major contribution of BCR signaling29_{. In addition, inhibition of adhesion,} either via BCR- or chemokine signaling, at LN sites is also a major biological effect11,30_{, and} this would indirectly lead to loss of all microenvironmental stimulatory signals, such as from TLRs, CD40, BAFF and cytokines. Indeed, the BCR versus NF-

κ

B and ERK profiles under ibrutinib treatment do suggest such indirect effects12_{. In addition, we have found evidence} for IL-21 signaling in CLL LN5_{, and Wagner et al recently put forward arguments for TLR9} signaling14_{. In summary, to what extent direct BCR triggering versus antigen-independent} signals drive CLL proliferation in vivo is difficult to discriminate, but the available evidence is clearly compatible with a contribution from indirect signals.

In case of CD40L/IL-21-induced proliferation, several arguments exclude involvement of upstream BCR components. Experimentally, knock-down of CD79A did not affect CD40L-induced ERK-phosphorylation, and inhibition of SYK did not block proliferation. In addition, the fact that there is no difference in CD40L/IL-21 proliferation between unmutated CLL and mutated CLL indicates that, in contrast to CpG-induced proliferation4_, the BCR is not directly involved. A recent paper showed that stimulation with CpG oligonucleotides results in the production of autoreactive anti-IgM in ZAP-70 positive CLL14_{, resulting in pro-survival signals in a SYK-dependent manner. Our data using the SYK} inhibitor R406 corroborate and extend these findings. Both CD40L/IL-21 and CpG-induced proliferation involve BTK and PI3K

δ

, but engagement of the BCR signalosome is distinct among these pathways. While in CD40L/IL-21 induced proliferation the BCR itself is completely circumvented, in CpG-induced proliferation the BCR is required. This was shown by stimulating cells with CpG or CD40L/IL-21 in the presence of various

(14)

4

TION IN CLL

kinase inhibitors. Whereas ibrutinib and idelalisib were capable of inhibiting both CpG and CD40L induced proliferation and ERK1/2 phosphorylation, R406 was only able to inhibit CpG-induced effects. Accordingly, CpG-induced sIgM production was inhibited by ibrutinib, idelalisib, and R406, while these compounds failed to impact upon the more modest CD40L/IL-21-induced sIgM production.

In the context of CLL, cyclin D2 has been described to be overexpressed in proliferation centers of CLL31,32_{. Cyclin D2 expression has been described to be controlled by the signaling} pathways investigated in the present study23,33_{. The critical role of cyclin D2 and CDK4 in} antigen-independent proliferation is underlined by the complete dependence of CLL cells on CDK4 for proliferation. Our data support the following model (Figure 6): The CD40 and IL-21R signal transduction routes converge with the upregulation of cyclin D2, by signaling through the MAPK cascade and the JAK/STAT axis, respectively. PI3K

δ

and BTK are engaged by CD40 ligation resulting in phosphorylation of ERK. Alternatively, stimulation of CLL cells with CpG results in the production of autoreactive IgM which engages the BCR resulting in proliferative signaling. These events culminate in upregulation of cyclin D2, which acts as a master regulator of proliferation. Cyclin D2 engages into a complex with CDK4, promoting phosphorylation of retinoblastoma-associated protein by CDK4 which subsequently leads to activation of the proliferative program34_.

Our data demonstrate that apart from the known effects of ibrutinib and idelalisib on adhesion and migration, both compounds also act on antigen-independent proliferation. The difference in potency of ibrutinib and idelalisib in curtailing CD40L/IL-21 and CpG-induced proliferation is intriguing. IL-21 signaling provides a strong proliferative signal35_, which is counteracted due to the induction of BIM-dependent apoptosis36_{. By co-stimulating} the CLL cells via CD40, IL-21-induced apoptosis is prevented. The contribution of CD40 to proliferation in this setting is minor compared to IL-21 signaling as JAK inhibition completely abrogates proliferation. In contrast to IL-21 signaling, CpG stimulation has been reported to result in apoptosis only in ZAP70-low cells which were suggested to be unable to mount a sufficient production of autoreactive IgMs37_{. In cells that are capable} of producing adequate amounts of autoreactive IgMs, CpG stimulation by itself may be sufficient to drive proliferation. This distinction explains why ibrutinib and idelalisib can completely block CpG-induced proliferation.

Of note, Chiron et al. have explored the possibility of manipulating the cell cycle with the CDK4 inhibitor PD03332991 in order to overcome ibrutinib resistance in mantle cell lymphoma (MCL)38_{. Our data suggest that a similar strategy may be possible in CLL.} Alternatively, JAK inhibitors may be employed to block CD40L/IL-21 induced proliferation. Similarly, CpG-induced proliferation can be efficiently targeted with the use of SYK inhibitors. Understanding the molecular mechanisms underlying antigen-independent proliferation will be of use in developing new treatment strategies which will improve the efficacy of treatment of CLL with kinase inhibitors that target BCR signaling.

ACKNOWLEDGEMENTS

We would like to thank the CLL patients for their blood donations and collaboration in this study. This work was supported by a grant from the Dutch Cancer Foundation (number UVA 2011-5097) to APK.

(15)

TION IN CLL

4 REFERENCES

1. Tromp JM, Tonino SH, Elias JA, et al. Dichotomy in NF-kappaB signaling and chemoresistance in immunoglobulin variable heavy-chain-mutated versus unmutated CLL cells upon CD40/TLR9 triggering. Oncogene. 2010;29(36):5071-5082.

2. Herishanu Y, Perez-Galan P, Liu D, et al. The lymph node microenvironment promotes B-cell receptor signaling, NF-kappaB activation, and tumor proliferation in chronic lymphocytic leukemia. Blood. 2011;117(2):563-574.

3. De Silva NS, Klein U. Dynamics of B cells in germinal centres. Nat Rev Immunol. 2015;15(3):137-148. 4. Guo A, Lu P, Galanina N, et al. Heightened BTK-dependent cell proliferation in unmutated

chronic lymphocytic leukemia confers increased sensitivity to ibrutinib. Oncotarget. 2015. 5. Pascutti MF, Jak M, Tromp JM, et al. IL-21 and CD40L signals from autologous T cells can induce

antigen-independent proliferation of CLL cells. Blood. 2013;122(17):3010-3019.

6. Mongini PK, Gupta R, Boyle E, et al. TLR-9 and IL-15 Synergy Promotes the In Vitro Clonal Expansion of Chronic Lymphocytic Leukemia B Cells. J Immunol. 2015;195(3):901-923.

7. Duhren-von Minden M, Ubelhart R, Schneider D, et al. Chronic lymphocytic leukaemia is driven by antigen-independent cell-autonomous signalling. Nature. 2012;489(7415):309-312.

8. Byrd JC, Furman RR, Coutre SE, et al. Targeting BTK with ibrutinib in relapsed chronic lymphocytic leukemia. N Engl J Med. 2013;369(1):32-42.

9. Brown JR, Byrd JC, Coutre SE, et al. Idelalisib, an inhibitor of phosphatidylinositol 3-kinase p110delta, for relapsed/refractory chronic lymphocytic leukemia. Blood. 2014;123(22):3390-3397. 10. de Rooij MF, Kuil A, Geest CR, et al. The clinically active BTK inhibitor PCI-32765 targets B-cell

receptor- and chemokine-controlled adhesion and migration in chronic lymphocytic leukemia.

Blood. 2012;119(11):2590-2594.

11. de Rooij MF, Kuil A, Kater AP, Kersten MJ, Pals ST, Spaargaren M. Ibrutinib and idelalisib synergistically target BCR-controlled adhesion in MCL and CLL: a rationale for combination therapy. Blood. 2015;125(14):2306-2309.

12. Herman SE, Mustafa RZ, Gyamfi JA, et al. Ibrutinib inhibits BCR and NF-kappaB signaling and reduces tumor proliferation in tissue-resident cells of patients with CLL. Blood. 2014;123(21):3286-3295. 13. Thijssen R, Ter Burg J, van Bochove GG, et al. The pan phosphoinositide 3-kinase/mammalian

target of rapamycin inhibitor SAR245409 (voxtalisib/XL765) blocks survival, adhesion and proliferation of primary chronic lymphocytic leukemia cells. Leukemia. 2016;30(2):337-345. 14. Wagner M, Oelsner M, Moore A, et al. Integration of innate into adaptive immune responses in

ZAP-70-positive chronic lymphocytic leukemia. Blood. 2016;127(4):436-448.

15. Hallaert DY, Jaspers A, van Noesel CJ, van Oers MH, Kater AP, Eldering E. c-Abl kinase inhibitors overcome CD40-mediated drug resistance in CLL: implications for therapeutic targeting of chemoresistant niches. Blood. 2008;112(13):5141-5149.

16. Dicker F, Kater AP, Prada CE, et al. CD154 induces p73 to overcome the resistance to apoptosis of chronic lymphocytic leukemia cells lacking functional p53. Blood. 2006;108(10):3450-3457. 17. Herman SE, Gordon AL, Hertlein E, et al. Bruton tyrosine kinase represents a promising

therapeutic target for treatment of chronic lymphocytic leukemia and is effectively targeted by PCI-32765. Blood. 2011;117(23):6287-6296.

18. Jahrsdorfer B, Blackwell SE, Wooldridge JE, et al. B-chronic lymphocytic leukemia cells and other B cells can produce granzyme B and gain cytotoxic potential after interleukin-21-based activation. Blood. 2006;108(8):2712-2719.

(16)

4

TION IN CLL

19. Honigberg LA, Smith AM, Sirisawad M, et al. The Bruton tyrosine kinase inhibitor PCI-32765 blocks B-cell activation and is efficacious in models of autoimmune disease and B-cell malignancy.

Proc Natl Acad Sci U S A. 2010;107(29):13075-13080.

20. Hanissian SH, Geha RS. Jak3 is associated with CD40 and is critical for CD40 induction of gene expression in B cells. Immunity. 1997;6(4):379-387.

21. Quintas-Cardama A, Vaddi K, Liu P, et al. Preclinical characterization of the selective JAK1/2 inhibitor INCB018424: therapeutic implications for the treatment of myeloproliferative neoplasms.

Blood. 2010;115(15):3109-3117.

22. Li YY, Baccam M, Waters SB, Pessin JE, Bishop GA, Koretzky GA. CD40 ligation results in protein kinase C-independent activation of ERK and JNK in resting murine splenic B cells. J

Immunol. 1996;157(4):1440-1447.

23. Scuto A, Krejci P, Popplewell L, et al. The novel JAK inhibitor AZD1480 blocks STAT3 and FGFR3 signaling, resulting in suppression of human myeloma cell growth and survival.

Leukemia. 2011;25(3):538-550.

24. Dey A, She H, Kim L, et al. Colony-stimulating factor-1 receptor utilizes multiple signaling pathways to induce cyclin D2 expression. Mol Biol Cell. 2000;11(11):3835-3848.

25. Sanchez M, Misulovin Z, Burkhardt AL, et al. Signal transduction by immunoglobulin is mediated through Ig alpha and Ig beta. J Exp Med. 1993;178(3):1049-1055.

26. Brunner C, Avots A, Kreth HW, Serfling E, Schuster V. Bruton’s tyrosine kinase is activated upon CD40 stimulation in human B lymphocytes. Immunobiology. 2002;206(4):432-440.

27. Yoon K, Jung EJ, Lee SR, Kim J, Choi Y, Lee SY. TRAF6 deficiency promotes TNF-induced cell death through inactivation of GSK3beta. Cell Death Differ. 2008;15(4):730-738.

28. Deregibus MC, Buttiglieri S, Russo S, Bussolati B, Camussi G. CD40-dependent activation of phosphatidylinositol 3-kinase/Akt pathway mediates endothelial cell survival and in vitro angiogenesis. J Biol Chem. 2003;278(20):18008-18014.

29. Wiestner A. BCR pathway inhibition as therapy for chronic lymphocytic leukemia and lymphoplasmacytic lymphoma. Hematology Am Soc Hematol Educ Program. 2014;2014(1):125-134.

30. Chang BY, Francesco M, De Rooij MF, et al. Egress of CD19(+)CD5(+) cells into peripheral blood following treatment with the Bruton tyrosine kinase inhibitor ibrutinib in mantle cell lymphoma patients. Blood. 2013;122(14):2412-2424.

31. Paul JT, Henson ES, Mai S, et al. Cyclin D expression in chronic lymphocytic leukemia. Leuk

Lymphoma. 2005;46(9):1275-1285.

32. Igawa T, Sato Y, Takata K, et al. Cyclin D2 is overexpressed in proliferation centers of chronic lymphocytic leukemia/small lymphocytic lymphoma. Cancer Sci. 2011;102(11):2103-2107. 33. Piatelli MJ, Doughty C, Chiles TC. Requirement for a hsp90 chaperone-dependent MEK1/2-ERK

pathway for B cell antigen receptor-induced cyclin D2 expression in mature B lymphocytes. J Biol

Chem. 2002;277(14):12144-12150.

34. Kato JY, Matsuoka M, Strom DK, Sherr CJ. Regulation of cyclin D-dependent kinase 4 (cdk4) by cdk4-activating kinase. Mol Cell Biol. 1994;14(4):2713-2721.

35. Zeng R, Spolski R, Casas E, Zhu W, Levy DE, Leonard WJ. The molecular basis of IL-21-mediated proliferation. Blood. 2007;109(10):4135-4142.

36. Gowda A, Roda J, Hussain SR, et al. IL-21 mediates apoptosis through up-regulation of the BH3 family member BIM and enhances both direct and antibody-dependent cellular cytotoxicity in primary chronic lymphocytic leukemia cells in vitro. Blood. 2008;111(9):4723-4730.

(17)

TION IN CLL

4

37. Jahrsdorfer B, Muhlenhoff L, Blackwell SE, et al. B-cell lymphomas differ in their responsiveness to CpG oligodeoxynucleotides. Clin Cancer Res. 2005;11(4):1490-1499.

38. Chiron D, Di Liberto M, Martin P, et al. Cell-cycle reprogramming for PI3K inhibition overrides a relapse-specific C481S BTK mutation revealed by longitudinal functional genomics in mantle cell lymphoma. Cancer Discov. 2014;4(9):1022-1035.

(18)

4

TION IN CLL

99

SUPPLEMENTARY MATERIALS

Figure S1. CD40/IL-21-mediated proliferation is inhibited by ibrutinib and idelalisib at lower concentrations. Relative division index calculated from CFSE dilution assays from primary CLL cells, stimulated with CD40L and IL-21 for 5 days, and treated with 100 nM of indicated kinase inhibitor (n=7).

Figure S2. Non-canonical NF-kB signaling is unaffected by ibrutinib. Western-blot analysis of CLL lysates after 24 hours stimulation with CD40L. Lysates were analyzed for p100 and p52 presence.

Supplemental Figure 1

CD40 L -Ibruti nib Idelal isib 0.0 0.5 1.0 1.5 *** *** CD40L + IL-21 R el at iv e di vi si on in de x

Supplemental Figure 2

Isotype Mock CD40L 0 500 1000 1500 Vehicle 1 µM Ibrutinib IL -2 1R e xp re ss io n (G M I) IL-21R Ev en ts Isotype Mock CD40L CD40L + ibrutinib

A

Isotyp e Mock -Ibruti nib 0 1000 2000 3000 4000 CD40L TL R 9 ex pr es si on (G M I)

B

TLR9 Ev en ts Isotype Mock CD40L CD40L + ibrutinib

Supplemental Figure 3

p100 p52 Vehicle Ibrutinib BMS Vehicle Ibrutinib BMS

Mock CD40L β-Actin 100 kD 50 kD 50 kD 37 kD

Supplemental Figure 1

Supplemental Figure 2

A

B

Supplemental Figure 3

Mock CD40L

100 kD 50 kD 50 kD

(19)

TION IN CLL

4

Figure S3. CD40L-induced IL-21R expression is unaffected by ibrutinib. TLR9 expression is also not affected by ibrutinib. (A) IL-21R expression was assessed by FACS analysis after 72 hours of culturing on NIH-3T3 cells (Mock) or NIH-3T3 expressing CD40L (CD40L) in the presence or absence of 1 µM ibrutinib. IL-21R levels are displayed as geometric mean intensity (GMI). (B) TLR9 expression was asssessed by FACS analysis after 72 hours of culturing on Mock cells or CD40L expressing NIH-3T3 cells. TLR9 levels are displayed as geometric mean intensitiy (GMI).

Supplemental Figure 1

Supplemental Figure 2

A

B

Supplemental Figure 3

Mock CD40L β-Actin 100 kD 50 kD 50 kD 37 kD

(20)

4

TION IN CLL Ta b le S 1. P at ie nt c ha ra ct er is tic s Patient # Age Gender

IgVH mutation status

WBC (10^9/l) % lymphocytes % CD19+/CD5+ FISH Tr eatment befor e sampling 1 47 m unmutated 398 100 98.1 13q- + 11q-chlorambucil, FC, dasatinib 2 80 m mutated 149 92 98.2 13q- + 11q-chlorambucil 3 55 v unmutated 90.3 90 96.4 13q- + 11q-none 4 60 m mutated 56.5 90 95.2 unknown none 5 62 v mutated 27 87.3 91.9 13q-chlorambucil, pr ednison 6 72 v mutated 14.7 59 69.6 unknown none 7 65 v mutated 59 92 87.6 no abnor malities none 8 64 v mutated 71 91 97 no abnor malities FCR 9 64 v unmutated 64.8 82 88 unknown none 10 50 v mutated 80.2 96 91.4 no abnor malities unknown 11 71 v mutated 44.8 90.5 95.3 unknown none 12 71 v mutated 43.4 91 94.1 unknown none 13 59 v unmutated 52.6 79 87.3 tris 12 none 14 62 m unmutated 192 92.9 94.9 13q-none 15 73 m ND 5.7 ND 81.9 13q- + 17p-unknown 16 60 v ND 291 ND 96.6 unknown unknown 17 50 m unmutated 21.8 78.7 48.3 13q-FCR 18 50 m unmutated 47.4 ND 70.9 13q-FCR 19 60 v mutated 88.9 ND 97.5 13q-chlorambucil 20 77 m ND 66.9 97 97.3 unknown none 21 67 v unmutated 494 97 96.4 11q-none 22 72 m ND 104 ND 91.2 no abnor malities chlorambucil, R-CVP , FCR 23 62 m ND 12.5 ND 86.3 13q-Rituximab + Bentamustine