Fenebrutinib

Bruton’s Tyrosine Kinase Small Molecule Inhibitors Induce a Distinct Pancreatic Toxicity in Rats s

Rebecca I. Erickson, Leah K. Schutt, Jacqueline M. Tarrant, Michelle McDowell, Lichuan Liu, Adam R. Johnson, Sock-Cheng Lewin-Koh, Maj Hedehus, Jed Ross, Richard A. D. Carano, Karin Staflin, Fiona Zhong, James J. Crawford, Shelly Zhong, Karin Reif, Arna Katewa, Harvey Wong, Wendy B. Young, Donna M. Dambach, and Dinah L. Misner
Genentech, Inc., South San Francisco, California (R.I.E., L.K.S., J.M.T., M.M., L.L., A.R.J., S.-C.L.-K., M.H., J.R., R.A.D.C., K.S., F.Z., J.J.C., S.Z., K.R., A.K., W.B.Y., D.M.D., D.L.M.); and University of British Columbia, Vancouver, British Columbia (H.W.); Primary Laboratory of Origin: Genentech, Inc., 1 DNA Way, MS59, South San Francisco, CA 94080
Received July 18, 2016; accepted October 31, 2016

ABSTRACT

Bruton’s tyrosine kinase (BTK) is a member of the Tec family of cytoplasmic tyrosine kinases involved in B-cell and myeloid cell signaling. Small molecule inhibitors of BTK are being investigated for treatment of several hematologic cancers and autoimmune diseases. GDC-0853 ((S)-2-(39-(hydroxymethyl)-1-methyl-5-((5-(2- methyl-4-(oxetan-3-yl)piperazin-1-yl)pyridin-2-yl)amino)-6-oxo- 1,6-dihydro-[3,49-bipyridin]-29-yl)-7,7-dimethyl-3,4,7,8-tetrahydro- 2H-cyclopenta[4,5]pyrrolo[1,2-a]pyrazin-1(6H)-one) is a selective and reversible oral small-molecule BTK inhibitor in development for the treatment of rheumatoid arthritis and systemic lupus erythe- matosus. In Sprague-Dawley (SD) rats, administration of GDC- 0853 and other structurally diverse BTK inhibitors for 7 days or longer caused pancreatic lesions consisting of multifocal islet- centered hemorrhage, inflammation, fibrosis, and pigment-laden macrophages with adjacent lobular exocrine acinar cell atro- phy, degeneration, and inflammation. Similar findings were not
observed in mice or dogs at much higher exposures. Hemorrhage in the peri-islet vasculature emerged between four and seven daily doses of GDC-0853 and was histologically similar to spontane- ously occurring changes in aging SD rats. This suggests that GDC-0853 could exacerbate a background finding in younger animals. Glucose homeostasis was dysregulated following a glucose challenge; however, this occurred only after 28 days of administration and was not directly associated with onset or severity of pancreatic lesions. There were no changes in other common serum biomarkers assessing endocrine and exocrine pancreatic function. Additionally, these lesions were not readily detectable via Doppler ultrasound, computed tomog- raphy, or magnetic resonance imaging. Our results indicate that pancreatic lesions in rats are likely a class effect of BTK inhibitors, which may exacerbate an islet-centered pathology that is unlikely to be relevant to humans.

Introduction
Bruton’s tyrosine kinase (BTK) is a cytoplasmic tyrosine kinase in the Tec kinase family (Bmx, Btk, Itk, Rlk, Tec) ex- pressed primarily in hematopoetic cell lineages that has essen- tial functions in B cells and myeloid cells. Antagonism of BTK enzymatic activity in B cells leads to inhibition of B-cell receptor (BCR)–dependent signaling. This results in the disruption of chemotactic signals important for B-cell survival, migration, and proliferation, and a reduction in inflammatory cytokine pro- duction from myeloid cells by preventing signaling through the

FCgRIII receptor. Small-molecule BTK inhibitors are cur- rently being developed for treatment of several hematologic cancers and autoimmune diseases.
Ibrutinib (Imbruvica; Pharmacyclics, Janssen Biotech, Inc., Sunnyvale, CA) is a highly effective BTK inhibitor for the treatment of B-cell malignancies that is generally well tolerated and largely devoid of leukopenia and hypogammaglobulinemia. Lymphocytosis is observed clinically with ibrutinib treatment and is thought to be related to inhibition of BCR signaling and efflux of cells from lymphoid tissues into the systemic cir- culation (Herman et al., 2014). Common adverse events for ibrutinib in patients with chronic lymphocytic leukemia include

This work was funded by Genentech, Inc. dx.doi.org/10.1124/jpet.116.236224.
s This article has supplemental material available at jpet.aspetjournals.org.
nausea, vomiting, diarrhea, constipation, petechiae, contusions, atrial fibrillation, and hypertension (Lipsky et al., 2015; Molica,

ABBREVIATIONS: AUC0-24, area under the concentration-time curve from time 0 to 24 hours; BCR, B-cell receptor; BTK, Bruton’s tyrosine kinase; CRL, Charles River Laboratories; CT, computed tomography; DAB, diaminobenzidine; DMSO, dimethylsulfoxide; ELISA, enzyme-linked immunosorbent assay; F-344, Fischer-344; GDC-0853, (S)-2-(39-(hydroxymethyl)-1-methyl-5-((5-(2-methyl-4-(oxetan-3-yl)piperazin-1-yl)pyridin-2-yl)amino)-6-oxo-1,6- dihydro-[3,49-bipyridin]-29-yl)-7,7-dimethyl-3,4,7,8-tetrahydro-2H-cyclopenta[4,5]pyrrolo[1,2-a]pyrazin-1(6H)-one; GNE-309, 2-(39-(hydroxymethyl)-5-((5- (2-methoxyethyl)-4,5,6,7-tetrahydropyrazolo[1,5-a]pyrazin-2-yl)amino)-1-methyl-6-oxo-1,6-dihydro-[3,49-bipyridin]-29-yl)-7,7-dimethyl-3,4,7,8-tetrahydro- 2H-cyclopenta[4,5]pyrrolo[1,2-a]pyrazin-1(6H)-one; HRP, horseradish peroxidase; IVGTT, intravenous glucose tolerance test; KO, knockout; MRI, magnetic resonance imaging; OGTT, oral glucose tolerance test; pBTK, phospho-BTK; PE, phycoerythrin; SD, Sprague-Dawley; TE, echo time; US, ultrasound; WH, Wistar-Han; XLA, X-linked agammaglobulinemia.

226

2015). It is unclear whether any of these findings are directly related to inhibition of BTK activity. Ibrutinib inhibits BTK covalently and irreversibly by targeting a cysteine residue (Cys481) conserved in 10 other kinases within the kinome. Consequently, ibrutinib potently inhibits several “off-target” kinases at therapeutic doses, which may also contribute to the adverse events observed (Pan et al., 2007; Evans et al., 2013). For example, EGFR, TEC, and PI3K-AKT pathway inhibition have been associated with diarrhea, bleeding, and atrial fibrillation, respectively (McMullen et al., 2014; Stephens and Spurgeon, 2015; Byrd et al., 2016).
Loss of BTK function through genetic mutation in humans or knockout (KO) in mice suggests that selective BTK inhibitors should cause suppression of antibody-mediated immune re- sponses while being generally well tolerated in patients. In humans, loss-of-function mutations in the gene encoding BTK result in X-linked agammaglobulinemia (XLA). Due to a block in B-cell development between the pro- and pre-B-cell stage, patients with XLA have a profound reduction in serum Ig concentration of all classes, and fail to mount effective humoral immune responses (Conley et al., 2000; Bao et al., 2012). Patients with XLA are susceptible to recurrent bacterial and enteroviral infections; however, with the advent of Ig infusion therapy, most of them live well into adulthood (Howard et al., 2006). Btk KO and X-linked immunodeficient mice have a less severe immunologic phenotype than humans due to only a partial block in B-cell development and some compensation by other BCR signaling components, including TEC kinase (Rawlings et al., 1993; Kerner et al., 1995; Ellmeier et al., 2000; Lindvall et al., 2005). Serum IgG and IgM concentrations in KO mice are reduced but not absent (Satterthwaite et al., 1997). Conversely, serum IgE concentration is increased, which has been attributed to an increase in class-switched B cells in the spleen (Iyer et al., 2011). Moreover, to the extent that BTK- deficient humans and mice have been evaluated, few clinical or histopathological abnormalities external to the hematopoetic system have been reported.
GDC-0853 ((S)-2-(39-(hydroxymethyl)-1-methyl-5-((5-(2-methyl- 4-(oxetan-3-yl)piperazin-1-yl)pyridin-2-yl)amino)-6-oxo-1,6- dihydro-[3,49-bipyridin]-29-yl)-7,7-dimethyl-3,4,7,8-tetrahydro- 2H-cyclopenta[4,5]pyrrolo[1,2-a]pyrazin-1(6H)-one) is a potent, selective, reversible, small-molecule BTK inhibitor in develop- ment for the treatment of autoimmune diseases (Young and Crawford, 2016). In preclinical assessments to characterize its toxicity profile, GDC-0853 was well tolerated when orally ad- ministered daily to Sprague-Dawley (SD) rats for up to 4 weeks. Findings that were considered related to pharmacologic activ- ity of GDC-0853 included mild increases in blood lymphocyte count, minimal B-cell depletion in lymphoid organs, and changes in serum Ig concentrations. Unexpectedly, the major finding was pancreatic toxicity, characterized microscopically by multifocal islet-centered hemorrhage, inflammation, fibrosis, and pigment- laden macrophages with adjacent lobular exocrine acinar cell atrophy, degeneration, and inflammation. We conducted a com- prehensive characterization of the distinct pancreatic lesions associated with GDC-0853 administration in SD rats. Our results suggest that inhibition of BTK enzymatic activity is involved in the pathogenesis of these lesions. The most sensi- tive strain was the SD rat; lesions in Fischer-344 (F-344) and Wistar-Han (WH) rats were of a lesser severity and/or required a longer treatment duration to develop. The observed lesions were histologically similar to spontaneously occurring changes

in aging SD rats. Pancreatic lesions were subclinical, with few changes in standard clinical pathology parameters in exocrine or endocrine pancreatic function, and were not detectable by imaging techniques. Our results indicate that pancreatic lesions in rats are likely a class effect of BTK inhibitors, which may exacerbate an islet-centered pathology that is unlikely to be relevant to humans.

Materials and Methods
Test Articles. All test articles, including GDC-0853, GNE-309 (2-(39-(hydroxymethyl)-5-((5-(2-methoxyethyl)-4,5,6,7-tetrahydropyr- azolo[1,5-a]pyrazin-2-yl)amino)-1-methyl-6-oxo-1,6-dihydro-[3,49-bipyr- idin]-29-yl)-7,7-dimethyl-3,4,7,8-tetrahydro-2H-cyclopenta[4,5]pyrrolo[1,2- a]pyrazin-1(6H)-one), ibrutinib, and spebrutinib, were synthesized at Genentech, Inc. (South San Francisco, CA) and formulated for oral administration in the same vehicle: 1.0% (w/v) 4000 cps hydroxypro- pylmethylcellulose (Fagron, Rotterdam, Netherlands), 0.2% (v/v) Tween 80 (Avantor, Center Valley, PA), and 100 mM citrate buffer (pH 3.0 6 0.1; Spectrum Laboratory Products, Inc., New Brunswick, NJ), prepared in reverse osmosis water. The same vehicle served as the control article in the in vivo studies.
Animal Care and Use. All in vivo experiments were performed in strict accordance with the guidelines in the National Institutes of Health Guide for the Care and Use of Laboratory Animals. All protocols were approved by the Institutional Animal Care and Use Committees at Genentech or Covance, Inc., and appropriate efforts were made to reduce animal suffering. Three strains of rats were used in the in vivo studies: Sprague-Dawley [Crl:SD(SD)], Wistar-Han [CRL:Crl:WI(Han)], and Fischer-344 (F344 Fischer). Sprague-Dawley rats were obtained from Charles River Laboratories (CRL; Hollister, CA or Portage, MI). Wistar-Han rats were obtained from CRL (Hollister, CA or Margate, UK). Fischer-344 rats were obtained from CRL (Kingston, NY). At the start of the studies, animals were 6–12 weeks old. Dose formulations (described earlier) were adminis- tered once daily by oral gavage at a dose volume of 5 or 10 ml/kg. Clinical observations (daily) and body weights (at least twice weekly) were collected for all animals.
Following the initial identification of pancreatic pathologic changes in rats administered GDC-0853 in 7- and 28-day toxicity studies, and the absence of similar findings in dogs administered GDC-0853 in a 28-day toxicity study, a series of experiments were conducted in rats to better characterize the pancreas lesions, to determine any strain differ- ences in susceptibility, and to determine whether the developing lesion could be monitored by glucose tolerance testing and/or imaging methods.
Toxicokinetics. Concentrations of test articles in plasma were quantified using the liquid chromatography–tandem mass spectrom- etry method. Toxicokinetic parameters [Cmax, time to maximal con- centration (Tmax), and area under the concentration-time curve from time 0 to 24 hours (AUC0-24)] were determined by noncompartmental methods using the extravascular input model (Phoenix WinNonlin, version 6.3.0; Certara USA Inc., Princeton, NJ). Concentrations of test articles were also evaluated in pancreas and liver tissue for compar- ison with concentrations in plasma at the same time point.
Clinical Pathology. Rats were fasted overnight prior to collection of blood for all clinical pathology measurements. Blood samples for hematology analyses were collected from the retro-orbital plexus under isoflurane-induced anesthesia into EDTA-containing tubes and analyzed on a Sysmex XT 2000iV (Sysmex America, Inc., Mundelein, IL). For measurement of IgE and IgG concentrations, blood was collected into serum separator tubes from the abdominal aorta under isoflurane-induced anesthesia and ketamine. Concentra- tion of serum IgE was measured by enzyme-linked immunosorbent assay (ELISA) using mouse anti-rat IgE for capture and detection (Clone B41-1 and Clone B41-3, respectively; BD Biosciences, San Jose, CA) and streptavidin–horseradish peroxidase (HRP; GE Life Sciences, Marlborough, MA). Purified rat IgE isotype control (BD Biosciences)

was used as a standard. Concentration of serum IgG was measured by ELISA with goat anti-rat IgG-Fc and HRP-conjugated goat anti-rat IgG-Fc as capture and detection antibodies, respectively (Bethyl Laboratories, Montgomery, TX). Rat reference serum (14 mg/ml of rat IgG; Bethyl Laboratories) was used as a standard. Serum amylase, lipase, glucose, and fructosamine were analyzed in blood collected from the jugular vein in conscious rats into serum separator tubes and measured on a Roche Modular Analytics (Roche Diagnostics Corpo- ration, Indianapolis, IN) or on a Beckman Coulter AU680 (Beckman Coulter Inc., West Sacramento, CA) for fructosamine using the manufacturer’s applications. Serum insulin was measured by ELISA (Mesoscale Discovery, Gaithersburg, MD).
Histopathology. In all in vivo studies, a necropsy was performed the day following administration of the last dose of test articles. The exceptions were studies that included a 4-week recovery-phase cohort. Pancreas was collected and processed for histopathological evalua- tion. The tissue was preserved in formalin, embedded in paraffin, and slides were prepared from 5-mm-thick sections that were subsequently stained with hematoxylin and eosin. Any macroscopic observations were recorded.
Histochemistry and Immunohistochemistry. Formalin-fixed, paraffin-embedded sections of pancreas from a subset of control, unaffected, and affected animals administered 30 mg/kg/day GDC- 0853 were cut at 4 mm for staining with Masson’s trichrome stain or immunolabeling with antibodies against CD68, clone ED1 (Serotec, Kidlington, UK); cytokeratin 19 (Lifespan Biosciences, Seattle, WA); or smooth muscle actin, clone E184 (Epitomics, Burlingame, CA).
For CD68, sections were deparaffinized and pretreated for antigen retrieval using Target Retrieval Solution (Dako, Carpinteria, CA). The slides were subsequently blocked for endogenous peroxidase activity using 3% H2O2 and for avidin/biotin using an avidin/biotin blocking kit (Vector Laboratories, Burlingame, CA). Sections were preblocked for nonspecific binding sites with blocking buffer. They were then in- cubated in mouse anti-CD68 at 10 mg/ml, followed by biotinylated horse anti-mouse IgG or biotinylated goat anti-rabbit IgG, respec- tively (Vector Laboratories). The sections were incubated in Vectas- tain ABC Elite reagent (Vector Laboratories), followed by Metal Enhanced Diaminobenzidine (DAB; Thermo Scientific, Rockford, IL).
For cytokeratin 19, after deparaffinization, the sections underwent antigen retrieval in Lab Vision EDTA (pH 8.0; Thermo Fisher Scientific, Waltham, MA), blocking, and then incubation in rabbit anti–cytokeratin 19 at 5 mg/ml. The primary antibody was detected with PowerVision Poly anti-Rabbit HRP (Leica, Newcastle, UK), followed by DAB.
For smooth muscle actin, all steps were performed on the Ventana Discovery XT Platform (Ventana Medical Systems, Tucson, AZ). Sections were deparaffinized using EZ Prep (Ventana Medical Systems, Tuscon, AZ), and pretreatment was accomplished with Cell Condition- ing 1 (Ventana Medical Systems, Tucson, AZ) using standard incubation time. Sections were then incubated with rabbit anti–smooth muscle actin at 0.075 mg/ml. The sections were subsequently incubated with anti-rabbit OmniMap-HRP reagent (Ventana Medical Systems), fol- lowed by DAB. All slides were counterstained with hematoxylin, dehydrated, cleared in xylene, and coverslipped.
Glucose Tolerance Testing. Animals were fasted overnight and administered vehicle or GDC-0853 2 hours prior to oral glucose toler- ance test (OGTT) or intravenous glucose tolerance test (IVGTT). For the OGTT studies (days 4, 14, and 32), animals were administered a 50% dextrose solution via oral gavage at a volume of 5 ml/kg (total dose of

administered a 50% dextrose solution via tail vein at a volume of 5 ml/kg (total dose of 2500 mg/kg dextrose). Whole blood (0.1 ml) was collected prior to and at 2, 5, 10, 15, 30, and 60 minutes after administration of dextrose. Blood samples were collected via the tail vein onto a Nova StatStrip glucometer for analysis of blood glucose level. The remaining blood was processed to plasma for analysis of insulin levels.
Ultrasound Imaging Study. Micro-ultrasound imaging was performed with the Vevo 2100 microimaging system (VisualSonics, Toronto, Ontario, Canada) after 7 and 14 days of administration of vehicle, GNE-309, or GDC-0853. Pancreatic images were obtained with an MS400 256-element array transducer (30-MHz center fre- quency, VisualSonics, Toronto, ON, Canada), 12-mm field-of-view, axial resolution of 0.04 mm, and lateral resolution of 0.1 mm. Three- dimensional pancreatic images were acquired with a motorized drive mechanism that traverses the z-axis while acquiring x-y axis images at regular spatial intervals of 0.05 mm. Both b-mode and power Doppler images were acquired. Average b-mode intensity (in arbitrary units) for the pancreas was obtained from a single imaging slice that contains the largest transverse region of pancreas. Power Doppler images for the same slice were used to estimate percentage vascularity [(power Doppler vascular area/total pancreatic area) ti 100].
Computed Tomography Imaging Study. An eXplore Ultra small-animal computed tomography (CT) scanner (GE Healthcare, Little Chalfont, Buckinghamshire, United Kingdom) was used to scan the pancreas at baseline (day 21) and after 7 and 14 days of administration of vehicle or GDC-0853. To attain dynamic contrast enhancement, a bolus of 0.6 ml of iohexol (Omnipaque-300, GE Healthcare, Little Chalfont, Buckinghamshire, United Kingdom) was administered via lateral tail vein at a rate of 7.2 ml/s, with a delay of 5 seconds before beginning the scan. Regions of interest were generated for the head, body, and tail of pancreas as well as for the abdominal aorta, liver, and portal vein to calculate the following parameters:
•Time-attenuation curve
•Pancreatic enhanced and unenhanced phases
•Hepatic enhanced and unenhanced phases
•Functional perfusion data: blood volume (ml), blood flow (ml/min), mean transit time (minutes), and permeability (ml/100 g of tissue/min).
The ratio of the late to early enhancement obtained from the hepatic and pancreatic phases (L:E) was quantified. For a healthy human with no pancreatic disease, the L:E ratio is #1 (Hashimoto et al., 2011).
Magnetic Resonance Imaging Study. Magnetic resonance im- aging was conducted after 7 days (4 males from each group on day 8, and 4 males from each group on day 9) and 14 days (day 15 or 16) of administration of vehicle or GNE-309. Experiments were performed on a 4.7T horizontal imaging system (Agilent Technologies, Santa Clara, CA) using a 63-mm quadrature volume coil. The transverse relaxation time (T2) was measured using a two-dimensional multislice fast spin echo experiment with field of view of 60 ti 60 mm2; 16 coronal slices of 1.0-mm thickness; matrix size of 128 ti 128 (zero-filled to 256 ti 256); echo train length of 16; four echo times (TEs) of 7, 16, 26, and 36 ms; and four averages. Data acquisition was synchronized with respiration by triggering with the respiration signal produced by the SA Instruments (Stony Brook, NY) equipment, which resulted in a repetition time of approximately 2000 ms. For each image pixel, T2 and S0 were determined by fitting to the equation:

2500 mg/kg dextrose). Blood (approximately 0.25 ml) was collected prior to administration of GDC-0853 and dextrose and at 15, 30, 60, and 120 minutes after administration of dextrose. Blood samples were collected via the tail vein (days 4 and 14) or jugular vein (day 32) onto a glucometer (days 4 and 14: Nova StatStrip; Nova Biomedical, Waltham, MA; day 32: AlphaTRAK2; VWR, Radnor, PA) for analysis of blood

2

TE T2

2
TE
2T2 TE
T2

glucose level. The remaining blood was processed to plasma for anal- ysis of insulin levels. For the IVGTT study (days 4, 14, and 28), rats were cannulated in the femoral artery. Animals were intravenously
T2, or transverse relaxation time, is the decay constant for the component of magnetization that is perpendicular to the static magnetic field of the magnet, S0. The pancreas was manually outlined

on the S0 image. The volume and average T2 of the pancreas was calculated. Pancreatic lesions could lead to an increase in volume or T2 due to neoplasm, inflammation, or edema, or a decrease in T2 due to hyperdense tissue (hyperplasia).
Kinase Panel Selectivity Testing. GDC-0853, GNE-309, and ibrutinib were tested at a concentration of 1 mM against a panel of 221 recombinant human kinase biochemical assays, including cyto- plasmic and receptor tyrosine kinases, serine/threonine kinases, and lipid kinases (Invitrogen SelectScreen Kinase Profiling Services, Thermo Fisher Scientific). For kinase activity assays, the ATP concentrations used were within 2-fold of the apparent Michaelis constant (Km) for each kinase, whereas for ATP-site competition binding assays, the active-site competitive probe concentration used was approximately equal to its dissociation constant (Kd).
Human Ex Vivo Whole-Blood CD69 Assay. Heparinized whole blood (100 ml) from healthy human volunteers in 96-well square top/tapered V-bottom deep-well plates (Analytical Sales, Pompton Plains, NJ) was incubated for 1 hour at 37°C with a titration of BTK inhibitor or vehicle in duplicate [0.38% dimethylsulfoxide (DMSO)]. Blood was then stimulated with 50 mg/ml goat anti-human IgM F(ab9)2 (Southern Biotech, Birmingham, AL). After an 18-hour incubation, cells in the blood were stained with 45 ml per well of a cocktail containing anti-mouse CD19-PerCP-Cy5.5, clone SJ25C1 (1.1 mg/ml; BD Biosciences); anti-mouse CD27-fluorescein isothiocyanate, clone L128 (0.18 mg/ml; BD Biosciences), and anti-mouse CD69-PE (CD69- phycoerythrin), clone FN50 (7.5 ml per 45 ml of cocktail; BD Pharmin- gen, San Jose, CA) or PE mouse IgG1, k isotype control (10 ml per 45 ml cocktail; BD Pharmingen) for 30 minutes at room temperature. This was followed by lysis of red blood cells using 1ti BD Pharm lysis buffer (BD Biosciences). Cells were washed with fluorescence-activated cell sorting buffer and fixed with 2% paraformaldehyde. Samples were acquired and analyzed on a BD LSRII flow cytometer (BD Biosciences, San Jose, CA). Activation of B cells (CD191CD27–) was assessed based on CD69 PE mean fluorescence intensity using FACSDiva software (BD Biosciences). The CD69 PE mean fluorescence intensity was plotted against the log10 of the BTK inhibitor concentration and fit by nonlinear regression using Prism version 5.0 (GraphPad Software, La Jolla, CA) to the variable-slope four-parameter sigmoidal inhibition equation to determine the half-maximal inhibitory concentration (IC50). Inhibitors were tested against blood from at least three different donors, and the mean IC50 (6SD) was calculated.
Human Ex Vivo Whole-Blood Phospho-BTK Assay. Heparinized whole blood (304 ml) from healthy human volunteers in 96-well square top/tapered V-bottom deep-well plates was incubated for 6 hours at 37°C with a titration of BTK inhibitor or vehicle in duplicate (8% DMSO) in a water reservoir within a tissue culture incubator (0.4% DMSO). After the incubation, the plate was put on ice, and the blood was diluted with 320 ml of 2ti lysis buffer (Cell Signaling Technology, Danvers, MA) containing complete protease inhibitors (Roche Diagnostics), 2ti phosphatase inhibitor cocktails 2 and
3(Sigma-Aldrich, St. Louis, MO), and 200 mM sodium orthovanadate (New England Biolabs, Ipswich, MA). The diluted blood was thor- oughly mixed, and the plate was sealed and stored at –80°C. For phospho-BTK (pBTK)(Y223) detection, High Bind 96-well plates (Mesoscale Discovery, Gaithersburg, MD) were spotted with 10 ml per well of 10 mg/ml mouse anti-BTK monoclonal antibody (BD Biosciences) and then allowed to dry before use. Phospho-BTK from 250 ml of thawed lysate was detected using rabbit anti-human pBTK (Y223) monoclonal antibody (13.2 mg/ml; Epitomics) and SULFO-TAG conjugated goat anti-rabbit IgG antibody (1 mg/ml; Mesoscale Discov- ery). The electrochemiluminescence signal was recorded with a SECTOR Imager 6000 MSD reader (Mesoscale Discovery). The pBTK electrochemiluminescence signal was plotted against the log10 of the inhibitor concentration and fit by nonlinear regression using Prism version 5.0 (GraphPad Software) to the variable slope four-parameter sigmoidal inhibition equation to determine the IC50. Inhibitors were tested against blood from at least three different donors, and the mean IC50 (6SD) was calculated.

Expression of Btk in Pancreas. Laser capture microdissec- tion was conducted on exocrine and endocrine tissue from SD rat pancreatic tissue (three samples) using Zeiss Palm (Zeiss, Thornwood, NY). RNA was isolated using the Arcturus PicoPure kit (Applied Biosystems, Carlsbad, CA) and DNase treated on-column before bioanalyzer quantification and qualification (RNA 6000 Pico Kit; Agilent, Santa Clara, CA). Reverse transcription to cDNA (Super- Script VILO Master Mix; Thermo Fisher Scientific) was subsequently performed on the laser capture microdissection samples: human islet RNA (four anonymous samples, sex unknown; Prodo Laboratories, Irvine, CA), universal RNA (rat/human; Clontech, Mountain View, CA), and total RNA from pooled isolated rat islet preparations (AllCells, Alameda, CA). This was followed by preamplification utilizing PreAmp Master Mix (Fluidigm, South San Francisco, CA). Subsequent polymerase chain reaction for relative gene expression used Taqman Fast Universal PCR Master Mix (Applied Biosystems) and the 96.96 Dynamic Array Integrated Fluidic Circuit (Fluidigm, South San Francisco, CA) on the Biomark HD System (Fluidigm). All target and housekeeping gene assays used for preamplification and relative gene expression were species-specific Taqman Assays pur- chased from Life Technologies. Nonblank values in all samples were all detectable with a good dynamic range, and there were no false positives in any assay blanks. Standard curves from diluted islet samples were within the linear range for robust assay detection. Finally, gene expression was normalized against the ribosomal endogenous control gene, 18S. Data are represented as delta CTs, relative to the highly expressed 18S, where negative numbers are lower than 18S and positive numbers are greater than 18S.
Statistical Analyses. Data are expressed as the mean 6 standard deviation. Repeated-measures models were fitted to the longitudinal blood glucose and insulin data with treatment, time points, and their interaction as fixed effects and animal as a random effect. Effect of treatment on change from baseline was evaluated at each time point using contrast t test with Bonferroni correction for multiple compar- isons. Log transformation was applied prior to analysis to address heteroskedasticity where appropriate.
Levene’s test was used to test for heteroskedasticity in b-mode intensity and percentage vascularity for ultrasound (US) and mag- netic resonance imaging (MRI) T2 data. If Levene’s test was signifi- cant, treatment effect was evaluated using Welch’s t test; otherwise, a two-sample t test assuming equal variances was conducted.
For all comparisons, statistical analyses were evaluated at a 5% significance level, and statistical analysis of data were performed using SAS 9.2 (SAS Institute, Cary, NC).
For the CT study, differences between animals administered GDC- 0853 or vehicle were tested using a t test with P , 0.05 in the statistical package.
For the MRI study, differences in volume or overall pancreatic T2 between animals administered GNE-309 or vehicle were tested using a t test with P , 0.05 in the statistical package JMP (SAS Institute).

Results
BTK Inhibitors Induce Distinct Pancreatic Lesions in Sprague-Dawley Rats. GDC-0853, a potent, selective, reversible, small-molecule BTK inhibitor, was orally admin- istered to SD rats for 1–28 consecutive days at doses up to 100 mg/kg/day (Table 1). Exposure to GDC-0853 (assessed by Cmax and AUC0-24) increased over the dose range tested. Generally, increase in exposure was dose-proportional up to 30 mg/kg/day and greater than dose-proportional between 30 and 100 mg/kg/day. No toxicologically significant accumu- lation or sex-dependent differences in exposure were observed (Supplemental Fig. 1; Table 2). GDC-0853 was generally well tolerated. Occasional hypoactivity, reduced mean body weight gain, and reduced food consumption (approximately 80% of

TABLE 1
In vivo study design

Purpose

Duration (days) Test Article

Rat Strain

Number (N) of Animals

Dose Levels

Endpoints Evaluated

mg/kg
Pilot toxicity 7 GDC-0853 SD N = 4/sex/group 0, 5, 15, 30 TKt, CP, PHP

General toxicity
28–32 GDC-0853
SD
N = 15/sex/group, with N = 5/sex/group undergoing a 28-day recovery period
0, 0.5, 10, 30, 100 TKs, CP, OGTT,
PHP

Characterization of pancreas
pathology
21
GDC-0853
SD
N = 6 males/group
0, 0.5, 30
TKt, CP, PHP,
IHC

Oral glucose metabolism 14 GDC-0853 SD N = 6 males/group 0, 10, 100 TKt, OGTT,
PHP

Intravenous glucose
metabolism
28 GDC-0853 SD N = 10 males/group 0, 30 TKt, IVGTT,
PHP

Strain sensitivity 14 GDC-0853 SD, WH, F-344 N = 6 males/group 0, 30 TKt, PHP
Strain sensitivity 14 GNE-309 SD, WH, F-344 N = 6 males/group 0, 100 TKt, PHP
Doppler ultrasound imaging 14 GNE-309 SD N = 8 males/group 0, 100 TKt, US, PHP
Doppler ultrasound imaging 14 GDC-0853 SD N = 16 males/group 0, 30 TKt, US, PHP

Computed tomography
imaging
14 GDC-0853 SD N = 16 males/group 0, 30 TKt, CT, PHP

Magnetic resonance imaging 14 GNE-309 SD N = 8 males/group 0, 100 TKt, MRI, PHP
CP, standard clinical pathology; IHC, immunohistochemistry of pancreatic tissue; PHP, standard pancreatic histopathology (hematoxylin and eosin); TKs, toxicokinetics in satellite animals; TKt, toxicokinetics in toxicity animals.

control means) were observed only after 28 days at the highest dose tested (100 mg/kg/day).
GDC-0853–related changes in clinical pathology parame- ters included mildly increased blood lymphocyte counts (up to 50% over control mean) and elevated serum IgE concentration (up to 5-fold over control mean) with no clear change in IgG concentration (Supplemental Fig. 2). These changes were consistent with reported effects of BTK inhibition following administration of ibrutinib or through loss of function in KO mice (Satterthwaite et al., 1997; Iyer et al., 2011; Herman et al., 2014).
GDC-0853–related microscopic findings in the pancreas were observed at doses $5 mg/kg/day after dosing durations of 7–28 days. Multifocal pancreatic lesions, often islet- centered but sparing islet cells, were characterized by varying
degrees (minimal to moderate) of islet/peri-islet hemorrhage, fibrosis, mixed cell infiltrates, and pigment-laden macro- phages. Additionally, significant involvement of adjacent exocrine acinar tissue was noted in some animals, consisting of lobular acinar cell atrophy with acinar cell degeneration/
apoptosis, interstitial mixed-cell (predominantly macro- phages with fewer lymphocytes and neutrophils) infiltrates, and/or interlobular hemorrhage and edema. Despite substan- tial involvement of islet tissue, there was little microscopic evidence of islet cell degeneration. Pancreatic lesions were not observed in rats administered 30 mg/kg/day GDC-0853 for 1 or 3 days, suggesting the onset of the lesion was between 4 and 7 days of dosing. Representative photomicrographs depicting the variety of pancreatic findings present are shown in Fig. 1. The most subtle microscopic change was minimal hemorrhage

TABLE 2
Comparison of BTK inhibitors and results in Sprague-Dawley rats: BTK potency and selectivity, toxicokinetics, and incidence of pancreatic lesions

BTK Inhibitor
Whole-Blood Potencya
pBTK IC50 CD69 IC50

b

Study Duration

Dose

c AUC0-24c

Incidence of Pancreatic
d

nM nM days mg/kg/day mM hour×mM
GDC-0853 11 6 4 8 6 6 BTK (.80%), SRC, FGR, BMX (.50%) 7 5 1.52 7.55 1/4 M, 0/4 F
10 4.29 23.9 2/4 M, 0/4 F
30 15.1 149 2/4 M, 0/4 F
GNE-309 15 6 2 13 6 3 BTK (.80%), SRC, FGR (.50%) 7 30 4.17 17.4 0/4 M, 1/4 F
100 17.7 183 2/4 M, 2/4 F
300 54.8 566 2/4 M, 3/4 F

Ibrutinib 7 6 2 12 6 1 BTK, SRC, FGR, BMX, BLK, BRK, TEC, YES, RIPK2, HCK, CSK, LCK, LYN, TXK, ERBB4, ERBB2, EGFR, FRK, RET, SRM, TNK2, FLT3, ITK, JAK3 (.80%), FLT4, KDR, B-RAF, TIE2, CSF1R, FGF1R, PDGFRa (.50%)
14 5
25
0.349 0.608
2.20 5.56
0/6 M, 1/6 F 1/6 M, 1/6 F

Spebrutinib 186 6 214 392 6 146 BTK (.80%), SRC (.50%), BMX, STK16, TXK, TEC, JAK3 (.80%), ERBB4, ARK5, JNK1, RET, FLT3, AuroraA, JNK3, AuroraB, DRAK1, WEE1, JAK2, FLT4, IRAK1 (.50%)
14
30
100
300
1.00 2.66
2.26 10.5
6.64 36.3
3/6 M 2/6 M 1/6 M

F = female; M = male.
baThe mean IC50 6 standard deviation from three to six donors is presented.
c BTK inhibitors were tested against a panel of 221 active recombinant human kinases.
dToxicokinetics performed on the last day of dosing: group mean Cmax and AUC0-24 (male and female combined where relevant due to no appreciable sex differences). Not observed in any control animal treated with vehicle.

Fig. 1. Representative photomicrographs of pancreatic histopathology observed in Sprague-Dawley rats following daily oral administration of GDC-0853 for 21 or 28 days are presented. (A) Minimal peri-islet hemorrhage, day 29 (bar = 50 mm). (B) Islet/peri-islet hemorrhage and fibrin with mixed-cell infiltrates and exocrine degeneration/atrophy, day 29 (bar = 100 mm). (C) Islet/peri-islet fibroplasia/fibrosis with adjacent exocrine atrophy, day 22 (bar = 100 mm). (D) Multifocal lesions at different stages present concurrently, day 29 (bar = 500 mm). (E) Masson’s trichrome stain highlighting collagen (blue) in peri-islet fibrosis, day 22 (bar = 100 mm). (F) Immunohistochemical labeling of smooth muscle actin, day 22 (bar = 100 mm). (G) Immunohistochemical labeling of CD68, day 22 (bar = 100 mm). (H) Immunohistochemical labeling of cytokeratin 19, day 22 (bar = 100 mm).

in the small blood vessels of the peri-islet vasculature at the interface of the islet and exocrine tissue (Fig. 1A). In more severely affected islets, the hemorrhage was more extensive, and islet/peri-islet lesions had the appearance of “exploding islets,” with clusters or individual islet cells floating in lakes of hemorrhage, fibrin and secondary degeneration of adjacent exocrine acinar cells, and varying amounts of mixed-cell inflammatory infiltrates and pigment-laden macrophages (Fig. 1B). Other islets were surrounded and/or dissected by large areas of dense fibroplasia and fibrosis (Fig. 1C). Atrophy of adjacent lobules of exocrine pancreas was suggestive of impaired blood flow through the affected peri-islet vascula- ture. Multifocal lesions at different stages of development were often present in the same animal, suggesting an ongoing insult (Fig. 1D). Pancreatic lesions were not observed in any animal treated with vehicle or 0.5 mg/kg/day GDC-0853 for up to 28 days. Thus, 0.5 mg/kg GDC-0853 was considered to be the no-observed-adverse-effect level for this finding. There was considerable resolution of the pancreatic pathology at the end of the 28-day recovery period, with only small amounts of mature fibrous connective tissue dissecting islet/peri-islet areas and pigment-laden macrophages in the interstitial and interlobular connective tissue remaining.
Immunohistochemical and histochemical evaluation of the pancreatic lesion was performed to further characterize the pathologic changes observed microscopically by hematoxylin and eosin staining. Histochemical staining of the affected pancreatic sections with Masson’s trichrome stain demon- strated an increase in blue-staining collagen fibers consistent with fibrosis within affected islet/peri-islet and interstitial areas (Fig. 1E). Increased smooth muscle actin labeling within

areas of pancreatic islet/peri-islet fibrosis could originate from several different cell types, including blood vessel–associated smooth muscle pericytes or a localized activation of pancreatic stellate cells (Fig. 1F). Increased CD68 labeling confirmed a macrophage infiltrate (Fig. 1G). Increased numbers of weakly cytokeratin 19–positive tubular structures were present in surrounding areas of islet/peri-islet fibrosis and within atro- phic exocrine lobules. This was consistent with pancreatic ductular proliferation and/or formation of duct-like tubular complexes by atrophic acinar cells (Fig. 1H). These findings confirmed the presence of fibrosis and a CD68-positive mac- rophage infiltrate and suggested that blood vessel–associated pericytes or activated pancreatic stellate cells may be involved in the production of the fibroplastic reaction.
To help understand whether inhibition of BTK enzyme activity was driving the SD rat pancreatic pathologic changes, we evaluated three additional BTK inhibitors in SD rats: GNE-309, ibrutinib, and spebrutinib (Fig. 2). GNE-309 is a potent and highly selective BTK inhibitor that is structurally related to GDC-0853. Ibrutinib and spebrutinib are structur- ally distinct, potent, and less-selective BTK inhibitors. In contrast to GDC-0853, which reversibly inhibits BTK at the ATP binding site, ibrutinib and spebrutinib inhibit BTK covalently and irreversibly by targeting a cysteine residue (Cys481) near the ATP binding site that is conserved in 10 other kinases within the kinome (Pan et al., 2007; Evans et al., 2013). As a consequence, covalent-binding BTK inhib- itors potently inhibit many of these “off-target” kinases.
The on-target potency of the four molecules was compared using two human ex vivo whole-blood assays to evaluate BTK- pathway inhibition. Additionally, the selectivity was assessed

Fig. 2. Molecular structures of BTK inhibitors GDC-0853, GNE-309, ibrutinib, and spebrutinib.

using a panel of 221 in vitro kinase assays, including cytoplasmic and receptor tyrosine kinases, serine/threonine kinases, and lipid kinases. This single point testing at 1 mM showed that, relative to ibrutinib and spebrutinib, GDC-0853 and GNE-309 were much more selective for BTK over other kinases (Table 2). GDC-0853 and GNE-309 were also evalu- ated in a broad panel of biochemical radioligand-binding and enzyme assays, including 42 targets of major classes of biogenic amine receptors, neuropeptide receptors, ion channel binding, and neurotransmitter transporter. The results of these assays did not reveal any significant off-target binding (data not shown).
Doses of GNE-309, ibrutinib, and spebrutinib were selected at levels that would result in clinically relevant plasma exposures (Pharmacyclics, Inc., 2015; Brown et al., 2016). When administered to SD rats for 7–14 consecutive days, exposure to each of the molecules (Cmax and AUC0-24) in- creased with the dose over the dose range tested (Table 2). All molecules were well tolerated with no clinical signs or body weight changes attributed to treatment. Microscopic pancre- atic lesions similar to those observed in SD rats administered GDC-0853 were present with GNE-309, ibrutinib, and spe- brutinib at all dose levels examined, but not in animals treated with vehicle (Table 2).
Taken together with the highly selective properties of GDC- 0853, where no off-target kinase activity is expected at the low-observed-adverse-effect level for pancreatic lesions in SD rats (7 days at 5 mg/kg/day), these data indicate that in- hibition of BTK enzymatic activity in SD rats is likely involved in the development of the pancreatic lesions (Fig. 3).
Sprague-Dawley Rat Was the Most Sensitive Strain Tested. During the histopathologic characterization of the GDC-0853–related pancreatic lesions, it was noted that the more mildly affected pancreatic islets in young (10- to 12-week-old) SD rats treated with GDC-0853 closely resem- bled the microscopic changes of the age-related spontaneous pancreatic islet hemorrhage and fibrosis in SD rats (Reaven and Reaven, 1981; Dillberger, 1994; Imaoka et al., 2007). This similarity suggested that BTK inhibitors might exacerbate a background pancreatic change in younger SD rats, and that the SD strain may be highly sensitive to the pancreatic effects

of GDC-0853. To test this hypothesis, we investigated the effects of BTK inhibitors on the pancreas of two other commonly used laboratory rat strains, F-344 and WH. Fol- lowing administration of GDC-0853 (30 mg/kg/day) or GNE- 309 (100 mg/kg/day) for 14 consecutive days, exposure was confirmed in all treated animals. Plasma concentrations of both GDC-0853 and GNE-309 were similar across strains at 3 hours postdose, near the anticipated Tmax, on day 14 (Fig. 4). With both molecules, the same incidence and severity of BTK inhibitor–related pancreatic lesions was observed across the strains: in 5 of 6 SD (moderate), 1 of 6 F-344 (minimal), and 0 of 6 WH rats. Thus, SD rats were particularly sensitive to the effects of BTK inhibitors on the pancreas (Fig. 4). The reduced sensitivity of WH rats was further confirmed in longer-term studies with GDC-0853. No GDC-0853–related pancreatic lesions were observed in WH rats administered GDC-0853 at doses up to 30 mg/kg/day for 28 days. However, following 6 months of daily dosing up to 30 mg/kg/day in WH rats, minimal-to-mild pancreatic islet/peri-islet fibrosis and pigment-laden macrophages were observed in males administered $2 mg/kg/day GDC-0853 (data not shown). These pancreatic changes were similar in extent and severity to the islet changes occurring spontaneously in aging SD rats, but did not occur in any vehicle-treated animals, and therefore were considered related to GDC-0853 administration.
BTK Is Expressed at Low Levels in the Rat and Human Pancreas. BTK is predominantly expressed in most hematopoietic cells and tissues harboring these cell types. To date, there is no reported expression or function of BTK in the endocrine or exocrine pancreas. To better understand whether BTK enzyme inhibition in the pancreas itself could be causing the lesions in rats, we evaluated local drug concentrations of GDC-0853 and Btk expression in pancreatic tissue. Following 21 days of dosing at 30 mg/kg in SD rats, GDC-0853 concen- tration near Tmax (3 hours postdose) in plasma (8.12 mM) was greater than in pancreas (0.319 mM) or liver (2.44 mM). This confirmed local exposure but demonstrated no preferential drug accumulation in pancreas and liver. Transcriptional profiling of Btk was performed on laser-captured microdis- sected endocrine and exocrine tissue from SD rats and purchased human islet samples. Compared with lymph node

Fig. 3. Exposure to GDC-0853 in Sprague- Dawley rats relative to BTK (on-target) and off- target (BMX, FGR, SRC) kinase half-maximal inhibitory concentrations (IC50) is presented. Pharmacokinetic profiles were plotted for the no-observed-adverse-effect level (NOAEL; 0.5 mg/kg/day) and low-observed-adverse-effect level (LOAEL; 5 mg/kg/day) doses identified in 28- and 7-day studies, respectively. Kinase IC50 values were determined in in vitro biochemical activity assays at Genentech (for BTK) or Invitrogen (for BMX, FGR, SRC) using ATP concentrations equal to the apparent Michaelis constant (Km) for each kinase.

Fig. 4. Strain sensitivity to BTK inhibitor–induced pan- creatic lesions is presented. GDC-0853 (30 mg/kg/day) and GNE-309 (100 mg/kg/day) were administered orally for 14 consecutive days to SD, F-344, and WH rats (n = 6 males per group per strain). With each test article, pancreatic lesions were observed in 5, 1, and 0 SD, F-344, and WH rats, respectively. All lesions in SD rats were moderate in severity, whereas those in F-344 rats were minimal in severity. Plasma concentrations (mM) of GDC-0853 and GNE-309 at 3 hours postdose are presented as the mean 6 standard deviation.

or universal RNA samples, Btk was expressed at significantly lower levels across islet/endocrine and exocrine tissue from both rats and humans (Fig. 5). It is not known whether these Btk transcripts would translate to protein expression.
BTK Inhibition Has Mild Effects on Glucose Metabolism. After administration of GDC-0853 to SD rats at doses up to 100 mg/kg/day for 28 consecutive days, despite the presence of pancreatic lesions, there were no changes in serum biomarkers of exocrine and endocrine pancreatic func- tion, including amylase, lipase, insulin, or fructosamine (Fig. 6). There was a mild increase in blood glucose (10–30%, P , 0.05) in fasted rats administered 100 mg/kg/day GDC-0853 after 12 and 28 days of dosing (Fig. 7A); however, given that pancreatic lesions of a similar incidence and severity were observed in animals administered $10 mg/kg/day in this study, the relationship between fasted blood glucose levels and pancreatic lesions was uncertain. An OGTT was per- formed as a more sensitive test of subclinical effects on glucose homeostasis. At 28 days, there was an increased and pro- longed peak in blood glucose (without corresponding changes in insulin) relative to animals administered vehicle at $10 and $30 mg/kg/day GDC-0853, respectively (Fig. 7, B and C). To evaluate the effectiveness of the OGTT as a premonitory endpoint in predicting the onset of pancreatic changes, a subsequent study evaluated the effects of 10 and 100 mg/kg/
day GDC-0853 after 4 and 14 days of administration. At these
earlier time points, there were no differences in the OGTT response of blood glucose or serum insulin between animals treated with vehicle and GDC-0853, despite the presence of pancreatic lesions in the majority of animals administered GDC-0853 (9 of 12) at 14 days. An IVGTT in SD rats admin- istered 30 mg/kg/day GDC-0853 evaluated at 4, 14, and 28 days of dosing showed minor changes only after 28 days (Supplemental Fig. 3). Collectively, these findings show that GDC-0853 administration results in mild glucose dysregula- tion after prolonged treatment in rats. This effect is not clearly related to the pancreatic lesions; therefore, fasted glucose or glucose challenge tests are not biomarkers of BTK-related pancreatic toxicity in SD rats.
Doppler Ultrasound, Computed Tomography, and Magnetic Resonance Imaging Were Not Adequately Sensitive to Identify BTK Inhibitor–Related Pancreatic Pathologic Changes In Vivo. Histopathologic characteriza- tion of the BTK inhibitor–related pancreatic changes in SD rats identified hemorrhage, edema, and fibrosis as the major structural alterations. We used several clinically relevant imaging techniques to determine whether the pancreatic changes in SD rats could be visualized in vivo to evaluate their potential application as a clinical monitoring tool.
In studies evaluating US, GDC-0853 (30 mg/kg/day) or GNE-309 (100 mg/kg/day) was orally administered to male SD rats for 14 days, and imaging was performed on days 7 and

Fig. 5. Relative Btk transcript expression (dCT; delta Cycle Threshold) in pancreatic tissue from humans and Sprague-Dawley rats is presented. BTK is expressed at low levels in endocrine and exocrine tissue from rat and human pancreas. Laser capture microscopy was performed on exocrine and endocrine pancreas from rats. Transcriptional expression was analyzed by Fluidigm in human islets (four samples), rat exocrine and islet tissues (three samples each), and rat lymph nodes (three samples). The results are normalized to 18S/baseline. Values are presented as the mean 6 standard deviation.

Fig. 6. Amylase, lipase, insulin, and fructosamine levels in Sprague-Dawley rats (n # 15 per sex per group) following administration of GDC-0853 are presented. GDC-0853 was administered orally at doses of 0.5–100 mg/kg/day for 28 days. Values are presented as the mean 6 standard deviation.

14. Histopathologic evaluations at the day 15 necropsy con- firmed pancreatic lesions in seven of eight and eight of eight animals administered GDC-0853 and GNE-309, respectively. There were no significant differences detected by b-mode intensity or percentage vascularity parameters after 7 days of treatment with either molecule. After 14 days of treatment, b-mode intensity and percentage vascularity were signifi- cantly elevated over the respective controls for animals treat- ed with GNE-309, but not GDC-0853 (Fig. 8). Together with the lack of association between individual severity of pancre- atic findings and these parameters, a consistent relationship with the histologic changes could not be supported.
CT and MRI evaluations were conducted following admin- istration of GDC-0853 or GNE-309 for 14 or 17 days, re- spectively. Pancreatic lesions were confirmed in all treated animals. The CT imaging demonstrated no clear GDC- 0853–related effects detected by time-attenuation curve and L:E ratio, pancreatic enhanced and unenhanced phases, or functional perfusion (Supplemental Fig. 4). Similarly, there were no clear treatment-related changes in MRI T2 maps (Supplemental Fig. 5).

Discussion
In these experiments, we showed that administration of a selective BTK small molecule inhibitor, GDC-0853, to SD rats results in distinct microscopic pancreatic lesions character- ized by multifocal islet/peri-islet hemorrhage, inflammation, fibrosis, and pigmented macrophages with adjacent lobular exocrine acinar cell atrophy, degeneration, and inflammation. These lesions demonstrated significant reversibility following
a 4-week recovery period, with only mature islet/peri-islet fibrosis and pigmented macrophages remaining. Additional BTK inhibitors—two structurally distinct, irreversible inhib- itors (ibrutinib and spebrutinib) and a second highly selective, reversible inhibitor (GNE-309)—caused the same pancreatic lesions when administered to SD rats under the same conditions. For the selective BTK inhibitors, the pancreatic lesions were observed at doses where no significant off-target activity was expected. Collectively, these data strongly sug- gest that inhibition of BTK enzymatic activity is involved in the pathogenesis of these lesions, which may be considered a class effect of BTK inhibitors in rats.
As evidence of species specificity, no similar pancreatic findings were observed in CD-1 mice or beagle dogs adminis- tered GDC-0853 or other potent BTK inhibitors for up to 9 months, despite achieving exposures up to 24 times the established low-observed-adverse-effect level in SD rats (Sup- plemental Fig. 6). In addition, there are no reports of pancre- atic changes in mice with BTK mutations (knockout or X-linked immunodeficient) or pancreatic disease or dysfunc- tion in male patients who lack functional BTK enzyme (XLA) in at least six published clinical series and/or registries in- cluding more than 400 patients (Lederman and Winkelstein, 1985; Hermaszewski and Webster, 1993; Plebani et al., 2002; Moin et al., 2004; Aghamohammadi et al., 2006; Winkelstein et al., 2006). Also of note, ibrutinib has been administered to thousands of patients and is not associated with pancreatic toxicity. Although SD rats were the most sensitive strain tested in the current studies, F-344 and WH rats developed BTK inhibitor–related pancreatic changes of a lesser severity after 2 weeks or 6 months, respectively. Taken together, these

Fig. 7. Glucose and insulin levels in Sprague-Dawley rats following administration of GDC-0853 are presented. GDC-0853 was administered orally at doses of 0.5–100 mg/kg/day for up to 32 days. (A) Fasted serum glucose in rats (n # 15 per sex per group) was measured on days 13 and 28. An oral glucose tolerance test in rats (n = 5 per sex per group) on day 32 included blood glucose (B) and serum insulin (C) measurements. Values are presented as the mean 6 standard deviation. Significance was defined at P , 0.05. Veh, vehicle.

results support the hypothesis that biologic differences pre- dispose rats to BTK inhibitor–induced pancreatic changes, with the SD rat being an exceptionally sensitive strain. Furthermore, the likelihood of clinical translatability is low.
Spontaneous pancreatic islet hemorrhage and fibrosis in SD rats is well described (Reaven and Reaven, 1981; Dillberger, 1994; Imaoka et al., 2007). The changes are characterized by an initial extravasation of red blood cells from the peri-islet capillaries and hemosiderin (pigment) deposition in the center or periphery of the islet. The earliest observations are recorded in 12-week-old animals, but the incidence and severity of the changes increase considerably by 26 weeks of age (Imaoka et al., 2007). These spontaneous islet changes are more commonly observed in males, and a similar sex pre- dilection for males was observed in the GDC-0853–related pancreatic findings. The relative sex- and strain-susceptibility differences may be related to body weight gain and overall metabolic status. Male SD rats have a significantly faster growth rate compared with male WH rats, owing to higher daily food consumption. At 12 weeks of age, male SD rats are already 30–70% larger than male WH rats (Hayakawa et al., 2013; Charles River Laboratories International, Inc.). Re- ducing body weight by exercise and/or caloric restriction (from 800 to 500 g at 12 months of age) significantly reduced the incidence of spontaneous pancreatic islet pathologic changes when compared with sedentary SD rats fed ad libitum (Reaven and Reaven, 1981). SD rats have higher rates of hyperglycemia and a higher predisposition to streptozotocin- induced diabetes. These findings may be associated with strain differences in the developing islet (Ojiro et al., 1993). Pancreatic islet pathology observed in rat strains with a
genetic predisposition to obesity/diabetes suggests that rats are sensitive to developing islet hemorrhage in response to increased metabolic demands. Spontaneous pancreatic islet

Fig. 8. Ultrasound imaging of the pancreas in Sprague-Dawley rats following administration of GNE-309 (n = 8 males per group) or GDC-0853 (n = 16 males per group) is presented. GNE-309 and GDC-0853 were administered orally at doses of 100 and 30 mg/kg/day, respectively, for 14 days. B-mode intensity and percentage vascularity were measured on day 14. Values are presented as the mean 6 standard deviation. Significance was defined at P , 0.05. Veh, vehicle.

changes in the OLETF, WY/Kob, ZDF, and Goto-Kakizaki rats (Lacraz et al., 2009; Jones et al., 2010; Katsuda et al., 2014) appear to be related to increased insulin demand and hyper- insulinemia. They commonly include early islet hypertrophy/
hyperplasia and degranulation of beta cells, followed by hemorrhage, hemosiderin deposition, inflammation, and/or fibrosis. Evidence of vascular injury in response to increased metabolic demand, including hemorrhage and hemosiderin deposition, appears to be unique to rats. In mouse models of type 2 diabetes and in human patients with type 2 diabetes, islet changes have been observed that include beta-cell hyperplasia followed by decreased beta-cell mass, amyloid deposition, and inflammation; notably, however, hemorrhage has not been reported as a feature characterizing the pancre- atic pathology (Junger et al., 2002; Hull et al., 2005; Iizuka et al., 2005; Bonner-Weir and O’Brien, 2008; Donath et al., 2008; Talchai et al., 2009). Compared with mice and other mammals, including dogs and primates, rats have unique anatomic features within the microvasculature perfusing the endocrine and exocrine tissue (Greaves, 2012), which may be associated with increased susceptibility to injury.
The spontaneous pancreatic islet lesions observed in SD rats are thought to be preceded by alterations in glucose metabolism. In contrast, GDC-0853–treated SD rats do not have significantly altered glucose metabolism at the early stages of development of lesions, indicating that the BTK inhibitor–induced pancreatic changes likely involve a distinct pathogenic mechanism. In the acute stage, occurring within as few as 7 days of administration of GDC-0853, pancreatic pathology was characterized by islet/peri-islet hemorrhage, suggesting drug-induced microvascular injury to the thin- walled capillaries within this region. A similar hypothesis was suggested by Brenneman et al. (2014), who described a very similar drug-related pancreatic finding in SD rats for which the drug target was not disclosed. In their investigations, an increase in immunohistochemical markers of endothelial cytotoxicity in the islet microvasculature between 1 and 5 days of administration was identified. This was suggestive of test article–induced vascular injury, although it was acknowl- edged that a primary versus secondary drug effect could not be determined. In light of the histopathologic characteristics of the islet-centered changes seen with GDC-0853 involving compromised vascular integrity with resulting local acute hemorrhage and edema and chronic fibrosis, we attempted to image the rat pancreatic lesions using several modalities. By US, there were no significant differences between GDC- 0853–treated and control animals in either b-mode or per- centage vascularity after 14 days. Furthermore, there were no differences in MRI or CT imaging parameters, the latter having shown promise in detecting fibrotic changes in patients experiencing pancreatic anastomotic failure following pan- creatoduodenectomy (Hashimoto et al., 2011).
Despite the presence of significant pancreatic pathology, SD rats administered GDC-0853 demonstrated good overall tolerability with no related clinical signs. Additionally, no significant changes in standard clinical pathology parameters associated with exocrine or endocrine function were observed in studies up to 4 weeks in duration. There were no significant changes in serum amylase, lipase, insulin, or fructosamine (glycated serum albumin used to assess glycemic control during the preceding approximately 2 weeks). Fasted blood glucose was mildly increased in rats administered the highest

dose of GDC-0853. However, given that pancreatic lesions of a similar incidence and severity were observed in animals receiving lower doses in this study with no apparent changes in glucose levels, the relationship between the minor increase in glucose levels and pancreatic lesions is uncertain. Elevated relative glucose levels were observed following OGTT at all doses of GDC-0853, where pancreatic lesions were observed, but these effects were detected only after 28 days (not after 4 or 14 days). Thus, there appear to be GDC-0853–related changes in glucose homeostasis, but the relationship to pancreatic findings has not been definitively established and may be considered either secondary to the development of the lesion or even a separate effect of BTK inhibitors in rats. Moreover, the relative timing and mild magnitude of the changes in glucose regulation in rats suggest that similar tests would be of questionable clinical utility for monitoring the onset of any pancreatic changes in a heterogeneous patient population.
In summary, our results suggest that BTK inhibitors, as a class, cause pancreatic lesions in rats that may be due to exacerbation of a unique susceptibility of rats (especially in the SD strain) to develop peri-islet hemorrhage and sub- sequent inflammation and fibrosis. These pancreatic changes are subclinical, with few changes in standard clinical pathol- ogy parameters associated with exocrine or endocrine pancre- atic function, suggesting significant functional reserve. These lesions were not detected by three imaging methods, making monitoring of development of the lesions difficult. Nonethe- less, the absence of similar changes in other nonclinical species administered BTK inhibitors and in BTK mutant mouse models, and the lack of clinical reports of pancreatic dysfunction in patients with XLA and patients treated with ibrutinib argue that this type of injury is very unlikely to occur with BTK inhibitor therapeutics in humans.

Acknowledgments
The authors thank Laura de Forge, Arna Katewa, Joseph Lubach, Michael Sweeney, and Covance Laboratories, Inc., for assistance in the conduct of the in vivo and ex vivo experiments.

Authorship Contributions
Participated in research design: Erickson, Schutt, Tarrant, McDowell, Lewin-Koh, Hedehus, Ross, Carano, Staflin, Craw- ford, S. Zhong, Reif, Wong, Young, Dambach, Misner.
Conducted experiments: McDowell, Hedehus, Ross, Staflin, Katewa.
Contributed new reagents or analytic tools: McDowell, Hedehus, Ross, Carano, Crawford.
Performed data analysis: Erickson, Schutt, Tarrant, McDowell, Liu, Johnson, Lewin-Koh, Hedehus, Ross, Carano, Staflin, F. Zhong, Katewa, Wong, Misner.
Wrote or contributed to the writing of the manuscript: Erickson, Schutt, Tarrant, Johnson, Carano, Staflin, Wong, Young, Dambach, Misner.
References
Aghamohammadi A, Fiorini M, Moin M, Parvaneh N, Teimourian S, Yeganeh M, Goffi F, Kanegane H, Amirzargar AA, Pourpak Z, et al. (2006) Clinical, immuno- logical and molecular characteristics of 37 Iranian patients with X-linked agam- maglobulinemia. Int Arch Allergy Immunol 141:408–414.
Bao Y, Zheng J, Han C, Jin J, Han H, Liu Y, Lau Y-L, Tu W, and Cao X (2012) Tyrosine kinase Btk is required for NK cell activation. J Biol Chem 287: 23769–23778.
Bonner-Weir S and O’Brien TD (2008) Islets in type 2 diabetes: in honor of Dr. Robert C. Turner. Diabetes 57:2899–2904.
Brenneman KA, Ramaiah SK, Rohde CM, Messing DM, O’Neil SP, Gauthier LM, Stewart ZS, Mantena SR, Shevlin KM, Leonard CG, et al. (2014) Mechanistic investigations of test article-induced pancreatic toxicity at the endocrine-exocrine interface in the rat. Toxicol Pathol 42:229–242.

Brown JR, Harb WA, Hill BT, Gabrilove J, Sharman JP, Schreeder MT, Barr PM, Foran JM, Miller TP, Burger JA, et al. (2016) Phase I study of single-agent CC-292, a highly selective Bruton’s tyrosine kinase inhibitor, in relapsed/refractory chronic lymphocytic leukemia. Haematologica 101:e295–e298.
Byrd JC, Harrington B, O’Brien S, Jones JA, Schuh A, Devereux S, Chaves J, Wierda WG, Awan FT, Brown JR, et al. (2016) Acalabrutinib (ACP-196) in Relapsed Chronic Lymphocytic Leukemia. N Engl J Med 374:323–332.
Conley ME, Rohrer J, and Minegishi Y (2000) X-linked agammaglobulinemia. Clin Rev Allergy Immunol 19:183–204.
Dillberger JE (1994) Age-related pancreatic islet changes in Sprague-Dawley rats. Toxicol Pathol 22:48–55.
Donath MY, Schumann DM, Faulenbach M, Ellingsgaard H, Perren A, and Ehses JA (2008) Islet inflammation in type 2 diabetes: from metabolic stress to therapy. Diabetes Care 31 (Suppl 2):S161–S164.
Ellmeier W, Jung S, Sunshine MJ, Hatam F, Xu Y, Baltimore D, Mano H, and Littman DR (2000) Severe B cell deficiency in mice lacking the tec kinase family members Tec and Btk. J Exp Med 192:1611–1624.
Evans EK, Tester R, Aslanian S, Karp R, Sheets M, Labenski MT, Witowski SR, Lounsbury H, Chaturvedi P, Mazdiyasni H, et al. (2013) Inhibition of Btk with CC-292 provides early pharmacodynamic assessment of activity in mice and hu- mans. J Pharmacol Exp Ther 346:219–228.
Greaves P (2012) Exocrine pancreas/endocrine pancreas, in Histopathology of Pre- clinical Toxicity Studies: Interpretation and Relevance in Drug Safety Evaluation, pp 489–510, Academic Press, Amsterdam, Netherlands.
Hashimoto Y, Sclabas GM, Takahashi N, Kirihara Y, Smyrk TC, Huebner M, and Farnell MB (2011) Dual-phase computed tomography for assessment of pan- creatic fibrosis and anastomotic failure risk following pancreatoduodenectomy. J Gastrointest Surg 15:2193–2204.
Hayakawa K, Mimura Y, Tachibana S, Furuya M, Kodama T, Aoki T, Hosokawa S, Fukui M, Shibata S, Yoshida M, et al. (2013) Study for collecting background data on Wistar Hannover [Crl:WI(Han)] rats in general toxicity studies–comparative data to Sprague Dawley rats. J Toxicol Sci 38:855–873.
Herman SEM, Mustafa RZ, Gyamfi JA, Pittaluga S, Chang S, Chang B, Farooqui M, and Wiestner A (2014) Ibrutinib inhibits BCR and NF-kB signaling and reduces tumor proliferation in tissue-resident cells of patients with CLL. Blood 123: 3286–3295.
Hermaszewski RA and Webster AD (1993) Primary hypogammaglobulinaemia: a survey of clinical manifestations and complications. Q J Med 86:31–42.
Howard V, Greene JM, Pahwa S, Winkelstein JA, Boyle JM, Kocak M, and Conley ME (2006) The health status and quality of life of adults with X-linked agamma- globulinemia. Clin Immunol 118:201–208.
Hull RL, Shen Z-P, Watts MR, Kodama K, Carr DB, Utzschneider KM, Zraika S, Wang F, and Kahn SE (2005) Long-term treatment with rosiglitazone and metformin reduces the extent of, but does not prevent, islet amyloid deposition in mice expressing the gene for human islet amyloid polypeptide. Diabetes 54: 2235–2244.
Iizuka S, Suzuki W, Tabuchi M, Nagata M, Imamura S, Kobayashi Y, Kanitani M, Yanagisawa T, Kase Y, Takeda S, et al. (2005) Diabetic complications in a new animal model (TSOD mouse) of spontaneous NIDDM with obesity. Exp Anim 54: 71–83.
Imaoka M, Satoh H, and Furuhama K (2007) Age- and sex-related differences in spontaneous hemorrhage and fibrosis of the pancreatic islets in Sprague-Dawley rats. Toxicol Pathol 35:388–394.
Iyer AS, Morales JL, Huang W, Ojo F, Ning G, Wills E, Baines JD, and August A (2011) Absence of Tec family kinases interleukin-2 inducible T cell kinase (Itk) and Bruton’s tyrosine kinase (Btk) severely impairs Fc epsilonRI-dependent mast cell responses. J Biol Chem 286:9503–9513.
Jones HB, Nugent D, and Jenkins R (2010) Variation in characteristics of islets of Langerhans in insulin-resistant, diabetic and non-diabetic-rat strains. Int J Exp Pathol 91:288–301.
Junger E, Herberg L, Jeruschke K, and Leiter EH (2002) The diabetes-prone NZO/Hl strain. II. Pancreatic immunopathology. Lab Invest 82:843–853.
Katsuda Y, Ohta T, Miyajima K, Kemmochi Y, Sasase T, Tong B, Shinohara M, and Yamada T (2014) Diabetic complications in obese type 2 diabetic rat models. Exp Anim 63:121–132.

Kerner JD, Appleby MW, Mohr RN, Chien S, Rawlings DJ, Maliszewski CR, Witte ON, and Perlmutter RM (1995) Impaired expansion of mouse B cell progenitors lacking Btk. Immunity 3:301–312.
Lacraz G, Giroix M-H, Kassis N, Coulaud J, Galinier A, Noll C, Cornut M, Schmidlin F, Paul J-L, Janel N, et al. (2009) Islet endothelial activation and oxidative stress gene expression is reduced by IL-1Ra treatment in the type 2 diabetic GK rat. PLoS One 4:e6963.
Lederman HM and Winkelstein JA (1985) X-linked agammaglobulinemia: an anal- ysis of 96 patients. Medicine (Baltimore) 64:145–156.
Lindvall JM, Blomberg KEM, Väliaho J, Vargas L, Heinonen JE, Berglöf A, Mohamed AJ, Nore BF, Vihinen M, and Smith CIE (2005) Bruton’s tyrosine kinase: cell biology, sequence conservation, mutation spectrum, siRNA modifications, and expression profiling. Immunol Rev 203:200–215.
Lipsky AH, Farooqui MZ, Tian X, Martyr S, Cullinane AM, Nghiem K, Sun C, Valdez J, Niemann CU, Herman SE, et al. (2015) Incidence and risk factors of bleeding- related adverse events in patients with chronic lymphocytic leukemia treated with ibrutinib. Haematologica 100:1571–1578.
McMullen JR, Boey EJH, Ooi JYY, Seymour JF, Keating MJ, and Tam CS (2014) Ibrutinib increases the risk of atrial fibrillation, potentially through inhibition of cardiac PI3K-Akt signaling. Blood 124:3829–3830.
Moin M, Aghamohammadi A, Farhoudi A, Pourpak Z, Rezaei N, Movahedi M, Gharagozlou M, Ghazi BMS, Zahed A, Abolmaali K, et al. (2004) X-linked agam- maglobulinemia: a survey of 33 Iranian patients. Immunol Invest 33:81–93.
Molica S (2015) The clinical safety of ibrutinib in chronic lymphocytic leukemia. Expert Opin Drug Saf 14:1621–1629.
Ojiro K, Kitamura H, Shimada T, and Nakamura M (1993) A morphometrical study of the postnatal development of rat pancreatic islets, with special regard to the differences between Wistar and Sprague-Dawley strains. Kaibogaku Zasshi 68: 190–203.
Pan Z, Scheerens H, Li SJ, Schultz BE, Sprengeler PA, Burrill LC, Mendonca RV, Sweeney MD, Scott KC, Grothaus PG, et al. (2007) Discovery of selective irre- versible inhibitors for Bruton’s tyrosine kinase. ChemMedChem 2:58–61.
Pharmacyclics, Inc. (2015) Imbruvica: Highlights of prescribing information. Retrieved from http://www.accessdata.fda.gov/drugsatfda_docs/label/2015/
205552s002lbl.pdf
Plebani A, Soresina A, Rondelli R, Amato GM, Azzari C, Cardinale F, Cazzola G, Consolini R, De Mattia D, Dell’Erba G, et al.; Italian Pediatric Group for XLA- AIEOP (2002) Clinical, immunological, and molecular analysis in a large cohort of patients with X-linked agammaglobulinemia: an Italian multicenter study. Clin Immunol 104:221–230.
Rawlings DJ, Saffran DC, Tsukada S, Largaespada DA, Grimaldi JC, Cohen L, Mohr RN, Bazan JF, Howard M, Copeland NG, et al. (1993) Mutation of unique region of Bruton’s tyrosine kinase in immunodeficient XID mice. Science 261:358–361.
Reaven EP and Reaven GM (1981) Structure and function changes in the endocrine pancreas of aging rats with reference to the modulating effects of exercise and caloric restriction. J Clin Invest 68:75–84.
Satterthwaite AB, Cheroutre H, Khan WN, Sideras P, and Witte ON (1997) Btk dosage determines sensitivity to B cell antigen receptor cross-linking. Proc Natl Acad Sci USA 94:13152–13157.
Stephens DM and Spurgeon SE (2015) Ibrutinib in mantle cell lymphoma patients: glass half full? Evidence and opinion. Ther Adv Hematol 6:242–252.
Talchai C, Lin HV, Kitamura T, and Accili D (2009) Genetic and biochemical path- ways of beta-cell failure in type 2 diabetes. Diabetes Obes Metab 11 (Suppl 4): 38–45.
Winkelstein JA, Marino MC, Lederman HM, Jones SM, Sullivan K, Burks AW, Conley ME, Cunningham-Rundles C, and Ochs HD (2006) X-linked agammaglob- ulinemia: report on a United States registry of 201 patients. Medicine (Baltimore) 85:193–202.
Young W and Crawford J (2016) Discovery of GDC-0853: A Highly Potent, Selective and Non-Covalent BTK Inhibitor, American Chemistry Society, San Diego, CA.Fenebrutinib

Address correspondence to: Leah K. Schutt, 1 DNA Way, MS59, South San Francisco, CA 94080. E-mail: [email protected]