Chemico-Biological Interactions 224 (2014) 1–12

Contents lists available at ScienceDirect

Chemico-Biological Interactions journal homepage: www.elsevier.com/locate/chembioint

Carcinogenicity in rats of the SGLT2 inhibitor canagliflozin Sandra De Jonghe a, Jim Proctor b, Petra Vinken a, Bianca Feyen a, Inneke Wynant a, Dirk Marien a, Helena Geys a, Rao N.V.S. Mamidi b, Mark D. Johnson b,⇑ a b

Janssen Research & Development, A Division of Janssen Pharmaceutica NV, Turnhoutseweg 30, B-2340 Beerse, Belgium Janssen Research & Development, LLC, 1000 Route 202 South, Raritan, NJ 08869, United States

a r t i c l e

i n f o

Article history: Received 30 May 2014 Received in revised form 29 August 2014 Accepted 23 September 2014 Available online 5 October 2014 Keywords: Canagliflozin SGLT2 inhibitor Rat carcinogenicity Kidney tumors Pheochromocytomas Leydig cell tumors

a b s t r a c t The carcinogenicity potential of canagliflozin, an inhibitor of SGLT2, was evaluated in a 2-year rat study (10, 30, and 100 mg/kg). Rats showed an increase in pheochromocytomas, renal tubular tumors, and testicular Leydig cell tumors. Systemic exposure multiples at the highest dose relative to the maximum clinical dose were 12- to 21-fold. Pheochromocytomas and renal tubular tumors were noted in both sexes at 100 mg/kg. Leydig cell tumors were observed in males in all dose groups and were associated with increased luteinizing hormone levels. Hyperplasia was increased in the adrenal medulla at 100 mg/kg, but only a limited increase in simple tubular hyperplasia was observed in the kidney of males at 100 mg/kg. Hyperostosis occurred and was accompanied by substantial effects on calcium metabolism, including increased urinary calcium excretion and decreased levels of calcium regulating hormones (1,25-dihydroxyvitamin D and parathyroid hormone). A separate study with radiolabeled calcium confirmed that increased urinary calcium excretion was mediated via increased calcium absorption from the gastrointestinal tract. It was hypothesized that, at high doses, canagliflozin might have inhibited glucose absorption in the intestine via SGLT1 inhibition that resulted in glucose malabsorption, which increased calcium absorption by stimulating colonic glucose fermentation and reducing intestinal pH. Pheochromocytomas and adrenal medullary hyperplasia were attributed to altered calcium homeostasis, which have a known relationship in the rat. In conclusion, Leydig cell tumors were associated with increased luteinizing hormone levels and pheochromocytomas were most likely related to glucose malabsorption and altered calcium homeostasis. Renal tubular tumors may also have been linked to glucose malabsorption. Ó 2014 Elsevier Ireland Ltd. All rights reserved.

1. Introduction SGLT2, a high-capacity, low-affinity glucose transporter, is expressed primarily in the proximal renal tubule and is responsible for most glucose reabsorption in the kidneys [1]. The transport of Abbreviations: 45Ca, calcium 45; 45CaCl2, calcium chloride labeled with calcium 45; ANOVA, analysis of variance; AUC, area under the curve; AV, amphophilicvacuolar; Cmax, peak plasma concentration; CPN, chronic progressive nephropathy; ELISA, enzyme-linked immunosorbent assay; GLP, Good Laboratory Practice; H&E, hematoxylin–eosin; IACUC, Institutional Animal Care and Use Committee; IC50, concentration causing 50% inhibition; K2EDTA, dipotassium ethylenediaminetetraacetic acid; LC–MS/MS, liquid chromatography coupled to tandem mass spectrometry; LCTs, Leydig cell tumors; LH, luteinizing hormone; NOEL, no observed effect level; PTH, parathyroid hormone; RIA, radioimmunoassay; RTTs, renal tubular tumors; SGLT1, sodium glucose cotransporter 1; SGLT2, sodium glucose cotransporter 2; T2DM, type 2 diabetes mellitus; Tmax, time to reach Cmax; v/v, volume to volume; w/v, weight to volume. ⇑ Corresponding author. Tel.: +1 (908) 704 4178; fax: +1 (908) 927 7172. E-mail address: [email protected] (M.D. Johnson). http://dx.doi.org/10.1016/j.cbi.2014.09.018 0009-2797/Ó 2014 Elsevier Ireland Ltd. All rights reserved.

glucose is against its concentration gradient, coupled to the active transport of sodium. Inhibition of SGLT2 decreases glucose reabsorption in the renal tubule, increases urinary glucose excretion, and thereby reduces plasma glucose [1]. Phlorizin, which is a nonselective inhibitor of SGLT1 (another isoform which is predominantly present in the gastrointestinal tract) and SGLT2, reduces plasma glucose in preclinical models of type 2 diabetes mellitus (T2DM) [2]. Due to its nonselective nature, with activity on SGLT1, SGLT2, and glucose transporters and its poor pharmaceutical properties, phlorizin is unsuitable for clinical development [3]. A number of selective, metabolically stable, SGLT2 inhibitors have been discovered and are in clinical development to treat T2DM [4]. Canagliflozin was the first SGLT2 inhibitor approved in the United States for the treatment of T2DM. Canagliflozin is a potent inhibitor of SGLT2 activity and has low potency SGLT1 inhibitory activity. Canagliflozin has been shown to increase urinary glucose excretion and decrease plasma glucose in animal models and in humans with T2DM [5,6]. In repeated-dose

2

S. De Jonghe et al. / Chemico-Biological Interactions 224 (2014) 1–12

toxicity studies with canagliflozin in mice up to 3 months, rats up to 6 months, and dogs up to 12 months, the compound was well tolerated and preneoplastic lesions were not observed [7,8]. In these studies, there were substantial effects on urinalysis parameters that were considered to be related to the mode of action. These effects included increases in urinary volume, glucose, and electrolytes. Rats appeared to be the most sensitive species, with additional effects on calcium homeostasis and bone that were not observed in mice or dogs. This report describes the preclinical assessment of the carcinogenicity potential of canagliflozin that was conducted as part of the development program to support world-wide registration. Where applicable, the results of the 3-month toxicity study in rats will be used for interpretation of findings in the carcinogenicity study. To assist interpretation of marked changes in calcium homeostasis (i.e., increased urinary excretion of calcium and increased trabecular bone formation) in rats treated with canagliflozin, a radiolabeled calcium absorption study was conducted in rats and the results are discussed in this article. In addition, possible mechanisms for induction of tumors in rats by canagliflozin and its clinical relevance are discussed in this article. Additional investigative studies that support the proposed mechanism of tumor formation are reported in a companion paper [9]. 2. Material and methods 2.1. Test substance Canagliflozin (CAS# 842133-18-0) with assay purity of 98.0% was formulated as an aqueous suspension containing 0.5% w/v hydroxypropyl methylcellulose in water. Formulations were stored refrigerated (2–8 °C) and in the dark. The concentrations, homogeneity, and stability of the formulations were checked during the studies using liquid chromatography and were confirmed to be within acceptance criteria. 2.2. Animals Studies were performed with Sprague-Dawley rats purchased from Charles River (Kingston, NY and Sulzfeld, Germany). The animals were acclimatized for approximately 2 weeks before the start of the experiments and were housed in stainless steel, grid-floor cages. Rats were individually housed for the 3-month study and group housed for the carcinogenicity, 45Ca absorption, and hormone evaluation studies in polysulfone cages with a wire-mesh roof, except for individual housing in stainless steel metabolism cages on the first and last study days for the 45Ca absorption study. The study rooms were maintained under standard conditions of temperature, relative humidity, ventilation, and illumination. Animals were fed ad libitum with a pelleted maintenance diet (Certified Rodent Diet 5002 from PMI Nutrition International, Inc., Richmond, IN), except as noted below, and allowed free and continuous access to drinking water. In the 45Ca absorption study, the calcium content was 1% in feed and 75 mg/l in water. In all instances, the animals were humanely handled in accordance to IACUC guidelines. 2.3. Experimental design 2.3.1. 2-Year rat carcinogenicity study In a GLP study conducted at Janssen Research & Development, a division of Janssen Pharmaceutica NV (Beerse, Belgium), SpragueDawley rats were administered canagliflozin at 0 (vehicle control; 0.5% w/v hydroxypropyl methylcellulose in water), 10, 30, or 100 mg/kg by oral gavage daily for 2 years (24 months). The dose

volume was 5 ml/kg per day for all dose groups. Rats were approximately 6 weeks old at the initiation of dosing. The high dose was selected primarily based on decreased body weight gain in males (0.77) and females (0.83) relative to controls within 3 months of treatment at this dose in a 6-month study. Mortality checks and clinical observations were performed daily. Clinical observations included weekly mass palpation, with discovered masses checked and scored daily. An ophthalmologic examination was performed predose, at 12 months, and at the end of dosing on the first 20 surviving animals of each sex in the control and high dose groups. Body weight was recorded predose, on Day 0, and then weekly. Food consumption was measured weekly. Clinical pathology was performed for the first 20 surviving rats per sex per group after 12 months of dosing (prior to daily dosing) and in all surviving animals after 24 months of dosing (prior to terminal kill or towards the end of the dosing period). Blood samples were obtained from the sublingual veins of fasted rats into tubes containing K2EDTA for hematology, sodium citrate for coagulation, and no additive for serum chemistry determinations. Hematology evaluations included hemoglobin, hematocrit, red blood cell count and indices, white blood cell count, differential white blood cell count, reticulocytes, thrombocyte count, and normoblasts, as well as microscopic evaluation of blood smears. Coagulation evaluations included prothrombin time, activated partial thromboplastin time, and fibrinogen. Clinical chemistry evaluations included alanine aminotransferase, aspartate aminotransferase, alkaline phosphatase, gamma glutamyl transferase, total bilirubin, urea nitrogen, creatinine, sodium, potassium, chloride, calcium, phosphorus, total protein, albumin, glucose, cholesterol, and triglycerides. Urine samples were collected from fasted rats by housing the rats overnight (approximately 20 h) in diuresis cages. Urine was analyzed for volume, specific gravity, color, clarity, pH, protein, glucose, ketone, and blood. Full necropsy was performed on all animals, and all macroscopic changes were recorded. Formalin-fixed, paraffin-embedded tissues were processed into slides by sectioning and staining with H&E, for microscopic evaluation. All tissues and gross lesions were examined histologically in all animals at all dose levels. Satellite groups of rats were administered vehicle (3/sex) or canagliflozin at doses of 10, 30, or 100 mg/kg (4/sex/group) for 6 months for assessment of toxicokinetics. Blood samples (0.3 ml) were collected by puncture of the tail vein on Days 0, 90, and 180 (2, 7, and 24 h postdose for the vehicle group; 0.5, 1, 2, 4, 7, and 24 h postdose for the canagliflozin-treated groups) into tubes containing K2EDTA and placed on wet ice. Within 2 h of collection, blood samples were centrifuged (1500g and 5 °C for 10 min). Plasma was separated and stored at 20 °C until further analysis. Plasma concentrations of canagliflozin were quantified using a validated LC–MS/MS method with a lower limit of quantification of 50 ng/ml. Mean peak plasma concentration (Cmax) and corresponding time to reach Cmax (Tmax) were calculated. The mean area under the plasma concentration-time curve (AUC) was calculated using the linear up/log down trapezoidal rule [10]. 2.3.1.1. Data analysis. Statistical analysis for differences between each dosage group and the vehicle control group was assessed as follows. Mortality was assessed using two different tests for heterogeneity, the Logrank test and the generalized test according to Wilcoxon-Gehan. Survival analysis also included Kaplan-Meier Curves to show the survival distribution over time. Clinical observations and ophthalmology were analyzed by the Fisher exact test. Body weight, weight gain, food consumption, hematology, coagulation, clinical chemistry, urinalysis, and endocrinology were analyzed by the two-sided Mann–Whitney U test. Incidences of non-neoplastic and neoplastic lesions were analysed by the Fisher

S. De Jonghe et al. / Chemico-Biological Interactions 224 (2014) 1–12

Exact test with Bonferroni–Holm correction [11]. The presence of a positive dose-related trend across dose levels was tested using Peto’s prevalence, death rate, and onset-rate analysis for incidental, fatal, or mortality independent tumors [12]. In addition to the global decision on trend in tumor rates, simultaneous confidence intervals with an add-0.5-success-and failure Dunnett-type approach suggested by Hothorn and Hasler [13] were calculated for medullary hyperplasia and neoplasia. Survival intervals allow quantification of differences in tumor rates for the particular dose groups. 2.3.2. 3-Month rat study In a GLP study conducted at Janssen Research & Development, LLC (Raritan, NJ), Sprague-Dawley rats were administered canagliflozin at daily doses of 0 (vehicle control; 0.5% w/v hydroxypropyl methylcellulose in water), 4, 20, or 100 mg/kg by oral gavage for 3 months (13 weeks) followed by an 8-week recovery period for the control and high-dose groups (10/sex/group plus 5/sex for recovery groups). The dose volume was 10 ml/kg per day for all dose groups. Rats were approximately 8 weeks old at initiation of dosing, while rats in the carcinogenicity study were 6 weeks old at initiation of dosing as noted above. Organ systems are developed and functional in rats by 6 weeks of age [14]. In the kidney, which is the pharmacological site of action for canagliflozin, nephrogenesis is complete with minimal further development during weeks 6–8 [15]. Mortality checks were performed daily. Clinical observations were performed predose, twice weekly for the first 4 weeks, and then weekly. Body weight and food consumption were recorded predose and weekly. An ophthalmologic examination was performed predose and at Week 12. Blood samples were collected from the retro-orbital sinus of fasted rats immediately prior to necropsy into tubes containing K2EDTA for hematology, sodium citrate for coagulation, and no additive for serum chemistry determinations. Samples were analyzed for the same hematology, coagulation, and clinical chemistry parameters as those in the carcinogenicity study. Urine samples collected from fasted rats by housing the rats overnight (approximately 20 h) in diuresis cages were analyzed for the same urinalysis parameters as those in the carcinogenicity study and additional evaluations for calcium, phosphorus, chloride, potassium, sodium, magnesium, creatinine, urobilinogen, bilirubin, N-acetyl glucosaminidase, and gamma glutamyl transferase. Full necropsy was conducted for all animals, and organs weights and macroscopic changes were recorded. Formalin-fixed, paraffinembedded tissues were processed into slides, by sectioning and staining with H&E, for microscopic evaluation. All tissues (full tissue list) were examined for control, high-dose, and early death (regardless of group) animals. Evaluation at low and intermediate dose levels was limited to target tissue (bone) and tissues with gross lesions. Additional evaluations were bone biomarker analyses and quantitative histomorphometric analysis of one stifle joint with the proximal tibia per rat from animals at the terminal sacrifice. For the bone biomarker analysis, blood (4 ml) was collected at necropsy from the abdominal aorta/vena cava into tubes containing no additive, and the serum was separated. Serum samples were frozen and stored at 60 °C to80 °C for approximately 3 months until thawed at room temperature for analysis of biomarkers, including the calcium homeostasis biomarkers 1,25-dihydroxyvitamin D and intact parathyroid hormone (PTH). Radioimmunoassay (RIA) was used to determine 1,25-dihydroxyvitamin D in serum (1,25-dihydroxyvitamin D RIA kit from Immunodiagnostic Systems, Fountain Hills, AZ). An ELISA was used to determine PTH in serum (Rat Intact PTH ELISA kit from Immunotopics Inc., San Clemente, CA). For histomorphometry, the proximal tibiae

3

were processed, undecalcified, for histology preparation in plastic medium (methyl methacrylate). Thin frontal sections (approximately 4–6 lm) were obtained near the middle of the tibia and stained with Goldner’s trichome stain. Bone and tissue areas and osteoid tissue area were measured using an OsteoMeasure image analysis software program (OsteoMetrics, Inc., Atlanta, GA, version 4.1) interfaced with a Nikon Eclipse F400 light/epifluorescent microscope and video subsystem. Derived parameters (trabecular bone volume, mineralized bone volume, trabecular thickness, number and separation, unmineralized [osteoid] volume, thickness, and surface) were calculated. 2.3.2.1. Data analysis. The statistical significance for differences between each dosage group and the vehicle control group was assessed as follows. Body weight, weight gain, and food consumption were analyzed by the Dunn’s test [16]. Clinical pathology was analyzed by Dunnett’s test or Dunn’s test. Organ weight was analyzed by the Dunn’s test. Histopathology was analyzed by the Dunn’s test when there was more than one treated group for comparison and by the Wilcoxon–Mann–Whitney U test when there was one treated group for comparison. For bone biomarkers and histomorphometry, statistical analyses were performed by a oneway ANOVA procedure using SAS software (SAS Institute, Cary, NC, version 8), with individual group comparisons conducted using Dunnett’s procedure where each compound-treated group was compared to the vehicle-treated group. 2.3.3. Calcium absorption in rats In a pharmacokinetic study conducted at Janssen Research & Development, a division of Janssen Pharmaceutica NV (Beerse, Belgium) to determine the mechanism for increased urinary calcium observed in rat toxicity studies, male Sprague-Dawley rats (10/group plus 5/group for toxicokinetics; 8–9 weeks old at time of dosing) were administered 0 (vehicle control; 0.5% w/v hydroxypropyl methylcellulose in water) or 100 mg/kg canagliflozin by oral gavage once daily for 2 weeks (14 days). The dose volume was 1 ml/100 g per day for all dose groups. Rats were either fed or fasted for 24 h starting shortly before dosing. On Day 1 and Day 14, rats were administered an oral dose of 45Ca (25 mg/rat 45 CaCl2 (20 lCi/rat) at the same time as canagliflozin. Over a 0 to 24 h period postdose, labeled calcium was measured in urine (collected 0–6 h and 6–24 h postdose) and serum (0.5 ml blood collections without additives via tail vein at 1, 3, 6 and 24 h postdose, followed by centrifugation at 1700g during 10 min) using liquid scintillation counting using a Perkin Elmer counter on Day 1 and 14. AUC values were calculated for serum 45Ca concentrations on Days 1 and 14 (WinNonlinÒ, Pharsight). Cumulative urinary excretion of 45Ca was evaluated over a 0–24 h interval on Days 1 and 14. 2.3.4. Hormone evaluation in rats A GLP study was conducted at Janssen Research & Development, a division of Janssen Pharmaceutica NV (Beerse, Belgium) to assess the effects of canagliflozin on circulating luteinizing hormone (LH) and testosterone levels in rats of the same strain and purchased from the same vendor as those used in the rat carcinogenicity study. The study was conducted under the same conditions and at the same study site as the rat carcinogenicity study. In this study, male Sprague-Dawley rats (30/group plus 3/group for toxicokinetics) were administered 0 (vehicle control; 0.5% w/v hydroxypropyl methylcellulose in water) or 100 mg/kg canagliflozin by oral gavage once daily for 7 months. The dose volume was 10 ml/kg per day from Day 0 to 69 and then adjusted to 5 ml/kg per day from Day 70 onwards for all dose groups. Rats were approximately 8 weeks old at initiation of dosing. Blood samples collected monthly and at the same time each day (±30 min) to minimize effects associated with diurnal variation in hormone concentrations were analyzed

4

S. De Jonghe et al. / Chemico-Biological Interactions 224 (2014) 1–12

for serum LH and testosterone levels (except at month 6 when levels may have been affected by dehydration from insufficient water supply prior to blood sampling). Validated radioimmunoassay methods were used to determine serum concentrations of LH (as described by Niswender et al. [17]) and testosterone (Coat-a-CountÒ Testosterone radioimmunoassay kit; Siemens Healthcare Diagnostics, Tarrytown, NY). Potential interference in the radioimmunoassay by serum binding proteins was eliminated by extracting testosterone from the serum using a mixture of benzene-hexane (1:2 v/v) prior to assaying. 2.3.4.1. Data analysis. The significance of differences in LH and testosterone levels between the control and 100 mg/kg group was assessed by the two-sided Mann–Whitney U test for pairwise comparison. 3. Results 3.1. 2-Year rat carcinogenicity study 3.1.1. Body weight, food consumption, and clinical observations Dose-related decreases were observed in body weight and body weight gain throughout the study. Decreases were especially pronounced in the high-dose group, with terminal body weights and body weight gains of 0.79 and 0.70 for males and 0.77 and 0.67 for females, respectively, compared to vehicle control rats that received 0.5% w/v hydroxypropyl methylcellulose without canagliflozin. Food consumption was increased in a generally dose-related manner during the entire study. In males/females at 10, 30, and 100 mg/kg, weekly increases were up to 1.34/1.49, 1.34/1.34 and 1.48/1.59, and maximum effects at study termination were seen at the high dose (1.42/1.39) compared to controls. The only clinical observations were wet bedding (both sexes) and waste of food (females) at all doses, and soft feces in males at 100 mg/kg, which were noted throughout the study. The decreases in body weight (gain) and increased food consumption were considered secondary to increased urinary glucose excretion, and the findings are consistent with the mode of action of the compound. Table 1 presents body weight and food consumption findings at the end of the study. 3.1.2. Survival Survival was numerically increased in all canagliflozin-dosed groups when compared to the survival rate of the vehicle controls (Fig. 1). This increase was more evident in males than in females. Survival in males at 30 mg/kg was statistically significantly greater than in vehicle controls (p < 0.001), with 49/65 male rats at 30 mg/ kg surviving to sacrifice vs 29/65 control males surviving to sacrifice. Additionally, mean duration of survival was 696 days for the 30 mg/kg males vs 586 days for the control males (p < 0.001). The survival in the male control rats was consistent with historical control survival of male rats at the study site.

3.1.3. Clinical pathology Decreases in serum glucose (both sexes) were observed at all canagliflozin doses, with maximum reductions at 100 mg/kg of approximately 0.5 to 0.6 relative to vehicle control values; the effect was observed under the conditions of this study in which treated animals were fasted overnight before sample collection (Table 2). Additional clinical pathology changes noted at all canagliflozin doses included slightly decreased white blood cell count (up to 0.79 at 1 year and 0.72 at 2 years; caused by decreases in lymphocytes and eosinophils), slight decreases in calcium (up to 0.94 at 1 year and 0.96 at 2 years), and slight to moderate increases in triglycerides (up to 3.6 at 1 year and 4.1 at 2 years) and blood urea nitrogen (up to 2.1 at 1 year and 2.4 at 2 years) in both sexes and in alkaline phosphatase (up to 2.1 at 1 year and 2.0 at 2 years) and alanine aminotransferase (up to 1.8 at 1 year and 1.3 at 2 years) in males. Additional findings at 30 and 100 mg/kg were decreases in albumin in females (up to 0.92 at 1 year and 0.94 at 2 years), and decreased bilirubin in males (0.56 at 1 year only) and females (0.7 at 1 year for both doses and 0.6 at 2 years for the high dose). Increases in inorganic phosphorus in males (1.1 at 1 and 2 years) and chloride in females (1.04 at 2 years only) were noted at 100 mg/kg. 3.1.4. Urinalysis Urinalysis findings at all canagliflozin doses were increases in urine specific gravity (1.02 at 1 and 2 years) and volume (up to 2.9 at 1 year and 3.4 at 2 years) in relation to increased glucose excretion. The urinary pH was decreased (up to 0.9 at 1 and 2 years) with an increased incidence of urinary ketones (up to 3 at 1 and 2 years). The urine also appeared less colored. 3.1.5. Toxicokinetics After 6 months of dosing (Day 180), systemic exposure (AUC0–24 h) to canagliflozin was dose related, slightly higher in females than in males, and comparable to exposure after 3 months (Day 90). Prolonged absorption phase was observed for canagliflozin following oral administration to rats and the mean Tmax values were in the range of 4–7 h after single dose and 1.75–4.75 h on Day 180. Exposure values at 6 months compared to those after a single dose were similar in males and slightly higher (3 at 10 and 30 mg/ kg, 1.4 at 100 mg/kg) in females. The Day 180 exposure values for males and female were 38 and 62 lg h/ml at the 10 mg/kg dose, 118 and 188 lg h/ml at the 30 mg/kg dose, and 316 lg h/ml and 559 lg h/ml at the 100 mg/kg dose, respectively. 3.1.6. Pathology Treatment-related macroscopic and/or microscopic findings were observed in the adrenal glands, kidneys, testes, and bone, and are described below. 3.1.6.1. Gross findings. Table 3 presents macroscopic findings in the main target tissues. Macroscopic changes in adrenal glands (enlargement) and testes (enlargement and discolored foci-areas,

Table 1 Body weight and food consumption findings in the 2-year rat carcinogenicity study with canagliflozin. Daily dose (mg/kg)

0

No. of animals

M:65

F:65

M:65

10 F:65

M:65

30 F:65

M:65

F:65

Body weighta,b (g) Body weight gaina,b (g) Food consumptiona,b (g)

663 469 189

463 311 141

0.885* 0.840* 1.238

0.875* 0.817* 1.248

0.852* 0.789* 1.307

0.788* 0.688* 1.255

0.787* 0.704* 1.418

0.767* 0.666* 1.390

F = females; M = males. a At end of dosing period. b For vehicle controls, group means are shown. For treated groups, multiples of control/baseline are shown. * statistically significant (⁄p < 0.01 or less) based on actual data (not on the multiples of control/baseline).

100

5

S. De Jonghe et al. / Chemico-Biological Interactions 224 (2014) 1–12

Fig. 1. (Color reproduction on the Web and in print). Kaplan-Meier survival plot for females and males in the 2-year rat carcinogenicity study with daily oral doses of canagliflozin. Control rats received 0.5% w/v hydroxypropyl methylcellulose in water.

Table 2 Clinical pathologya findings in the 2-year rat carcinogenicity study with canagliflozin. Daily dose (mg/kg)

0

No. of animals

M:65

Hematology and coagulationb Week 52 Red blood cells (106/ll) Hemoglobin (g/dl) Hematocrit (%) White blood cells (103/ll) Lymphocytes (103/ll) Eosinophils (103/ll) Week 105–106 White blood cells (103/ll) Lymphocytes (103/ll) Eosinophils (103/ll) Serum chemistryb Week 52 Calcium (mg/dl) Inorg. phosphorus (mg/dl) Albumin (g/dl) Glucose (mg/dl) Triglycerides (mg/dl) Urea nitrogen (mg/dl) Total bilirubin (mg/dl) Alkaline phosphatase (U/l) Alanine aminotransferase (U/l) Week 105–106 Chloride (mmol/l) Calcium (mg/dl) Inorg. phosphorus (mg/dl) Albumin (g/dl) Glucose (mg/dl)

10 F:65

30

100

M:65

F:65

M:65

F:65

M:65

F:65

– – – 9.9 7.24 0.14

7.72 14.5 42.7 6.5 4.76 0.14

– – – 0.929* 0.862* 0.857*

– – – – – –

– – – 0.768* 0.620* 1.357*

0.964* 0.959* 0.963* – – –

– – – 0.778* 0.626* 0.429*

0.913* 0.924* 0.927* 0.785* 0.679* 0.500*

9.6 5.50 0.11

7.8 4.17 0.12

0.823* 0.807* 0.72*

0.872 0.844* 0.583*

0.990* 0.736* 0.636*

0.718* 0.705* 0.583*

0.990* 0.642* 0.636*

0.718* 0.600* 0.500*

9.9 5.9 – 110 63 14.0 0.09 53 37

10.3 – 5.1 109 47 15.8 0.15 – –

0.960* – – 0.636* 1.238* 1.636* – 1.491* 1.541*

0.951* – – 0.706* 1.532* 1.405* – – –

0.960* – – 0.527* 1.603* 2.014* 0.778* 1.642* 1.649*

0.951* – 0.941* 0.596* 1.851* 1.804* 0.733* – –

0.960* 1.102* – 0.518* 1.540* 2.143* 0.556* 2.057* 1.784*

0.942* – 0.922* 0.532* 3.596* 1.930* 0.733* – –

– 9.8 5.5 – 89

101 10.4 – 4.6 94

– 0.990* – – 0.640*

– 0.962* – – 0.723*

– 0.990* – – 0.584*

– 0.942* – 0.935* 0.649*

– 0.990* 1.055* – 0.483*

1.040* 0.962* – 0.935* 0.564*

– No noteworthy findings. F = females; M = males. a At end of dosing period. b For vehicle controls, group means are shown. For treated groups, multiples of control/baseline are shown. Statistical significance is based on actual data (not on the multiples of control/baseline). * Statistically significant (p < 0.05 or less).

6

S. De Jonghe et al. / Chemico-Biological Interactions 224 (2014) 1–12

Table 3 Macroscopic findings in main target tissues in the 2-year rat carcinogenicity study with canagliflozin. N = 65/group

Males

Daily dose (mg/kg)

0

10

Adrenal glands Enlarged

6

5

Kidneys Calculus/gravel in pelvis Discoloration: pale Focus/area depressed Focus/area discolored Irregular surface Mass Pelvic dilatation Swollen

– 5 – 3 – – 4 4

Testes Focus/area discolored Mass Swollen Accessory sex glands Small

Females 30

100

0

10

30

100

7

24*

13

21*

26*

34*

5* 9* 3* 4 1* – 6* 5

5* 14* 4* 1 1* 2 10* 6

5* 24* – 10* 9* 8* 8* 19*

3 1 – 1 – – 1 –

11* 7* – 1 – – 4* 1

14* 9* – 3 1* 1 6* 2

21* 34* 4* 4 12* 3* 4* 11*

1 – –

6* – 2*

14* 1* 10*

24* 1* 12*

2

3

11*

20*

Table 5 Survival-adjusted pheochromocytoma tumor rates in males in the 2-year rat carcinogenicity study with canagliflozin. Daily dose (mg/kg)

0

10

30

100

Number of tumors (y) Number of animals (n) Poly-3 adjusted n Raw rate (y/n) Poly-3 adjusted (+ADD1 correction) rate

4 65 43.3 6.15% 10.16%

4 65 52.0 6.15% 8.48%

7 65 58.8 10.77% 12.54%

28 65 49.5 43.08% 56.49%

Table 6 Survival-adjusted adrenal hyperplasia rates in males in the 2-year rat carcinogenicity study with canagliflozin.

– No noteworthy findings. N = number of animals. * Indicates a treatment-related effect.

mass) usually correlated to histological findings (hyperplasia or neoplasia). For the testes, this was associated with small accessory sex glands. Macroscopic changes in the kidneys were related to renal tubular tumors (RTTs) (masses), chronic progressive nephropathy (CPN) (pale discoloration, depressed focus/areas, irregular surface, enlargement), chronic hypercalciuria/mineralization (calculi or gravel), or increased urine volume (dilated pelvis). 3.1.6.2. Histopathology: proliferative findings. 3.1.6.2.1. Adrenal glands. An increased incidence of pheochromocytomas (benign and malignant) accompanied by medullary hyperplasia was noted in the adrenal glands of both sexes at 100 mg/kg (Table 4). As discussed below, the higher incidence of pheochromocytomas in males at 30 mg/kg was considered to be not treatment related based on survival adjusted analysis; therefore, the no observed effect level (NOEL) was 30 mg/kg. Survival-adjusted analysis of pheochromocytoma (benign and malignant) tumor rates were 10%, 8%, 13%, and 57% in males at 0, 10, 30, and 100 mg/kg, respectively, confirming the absence of a treatment-related effect at 30 mg/kg (Table 5). Survival-adjusted analysis of hyperplasia showed the absence of a statistically significant (p > 0.05) treatment-related effect at 30 mg/kg (Table 6).

Daily dose (mg/kg)

0

10

30

100

Number of hyperplasia (y) Number of animals (n) Poly-3 adjusted n Raw rate (y/n) Poly-3 adjusted hyperplasia rate

15 65 53 23% 34%

17 65 53 26% 32%

30 65 61 46% 49%

29 65 44 45% 55%

3.1.6.2.2. Kidneys. Basophilic RTT (adenomas and carcinomas) were noted in both sexes at 100 mg/kg (Table 7); these were localized in the cortex and/or outer stripe of the outer medulla. The tumors occurred late during the study and were associated with a marginal increase in simple tubular hyperplasia in the high-dose males. The 2 RTTs (1 benign, 1 malignant) in males at 30 mg/kg had a distinctly different appearance from all but 1 of the RTTs observed at 100 mg/kg. The 2 RTTs in males at 30 mg/kg and 1 RTT in males at 100 mg/kg were of the amphophilic-vacuolar (AV) phenotype characteristic of spontaneously occurring tumors that have been reported to occur in rats [18,19]. Representative histologic micrographs show the markedly different appearance of the AV RTTs and treatment-related basophilic RTTs (Fig. 2). Based upon categorization of the kidney tumors at 30 mg/kg as AV tumors, the RTTs observed in males at this dose were considered to be incidental and unrelated to canagliflozin treatment. Therefore, the NOEL for RTT was 30 mg/kg.

3.1.6.2.3. Testes. In the testes, interstitial cell (Leydig cell) adenoma and interstitial cell hyperplasia were noted at all canagliflozin dose levels and increased in a dose-related fashion (Table 8). The Leydig cell tumors (LCTs) appeared late during the study, mainly in terminal males and after 600 days of dosing (except for 1 rat). In many rats, LCTs were associated with atrophy of the secondary sex glands.

Table 4 Pheochromocytomas and medullary hyperplasia incidence in the 2-year rat carcinogenicity study with canagliflozin. Daily dose (mg/kg)

N Pheochromocytoma Benign

Malignant

Hyperplasia medullary

Male

Female

0

10

30

100

HCDa

0

10

30

100

HCDa

65

64

64

65

540

65

63

62

64

540

4

4

7

26*

2

1

3

7





1

2*

54 Mean: 10% Range: 3.1–20% 6 Mean: 2.59% Range: 1.54–4.61%









14 Mean: 1.11% Range: 0–3.1% 6 Mean: 1.11% Range: 0–4.6%

15

17

30*

29*

8

6

8

27

– No noteworthy findings. HCD = historical control data (historical rate of incidence observed in vehicle control animals at the study site); N = number of animals. a Historical control data at the study site. * Indicates a treatment-related effect following survival-adjusted analysis.

7

S. De Jonghe et al. / Chemico-Biological Interactions 224 (2014) 1–12 Table 7 Kidney tumors and hyperplasia in the 2-year rat carcinogenicity study with canagliflozin. Daily dose (mg/kg)

Male

Female

N

0 65

10 65

30 64

Renal tubular cell tumors, basophilic Adenoma renal tubule, basophilic Carcinoma renal tubule, basophilic Renal tubular cell tumors, spontaneous Adenoma renal tubule, AV

– – – – –

– – – – –

– – 2 1

Carcinoma renal tubule, AV Hyperplasia renal tubule simple Grade 1 Grade 2

– 3 3 –

– 1 1 –

1 3 3 –

a

100 65

HCD 540

12*,b 8* 4* 1

0 0

1 6* 5 1

1d Mean: 0.18% Range: 0–1.4% 0

0 65

10 64

30 65

100 65

– – – – –

– – – – –

– – – – –

8*,c 7* 2* – –

– 3 3 –

– 2 2 –

– – – –

– 2 1 1

HCDa 540 1 0 0

0

AV = amphophilic-vacuolar (a spontaneous tumor type); HCD = historical control data (historical rate of incidence observed in vehicle control animals at the study site); N = number of animals. a Incidence at the study site. b Animal No. 333 had an adenoma and a carcinoma. c Animal No. 740 had an adenoma and a carcinoma. d The renal tubular tumor was considered AV (spontaneous). * Indicates a treatment-related effect.

3.1.6.3. Histopathology: non-proliferative findings. Table 9 presents non-proliferative findings. The kidneys showed a dose-related CPN, tubular and pelvic dilatation (all doses; related to the increased urinary volume), increased mineralization (related to hypercalciuria), and tubular hypertrophy (males at 30 mg/kg; both sexes at 100 mg/kg). Atrophy of male accessory sex glands was noted from 30 mg/kg onwards. A dose-related increase in hyperostosis (an increase in trabecular bone volume in the metaphysis beneath the growth plate) of the sternum and/or stifle (distal femur and proximal tibia) was apparent in both sexes at 30 and 100 mg/kg. 3.2. 3-Month rat study Key findings from the 3-month rat study are presented in Table 10. Canagliflozin caused increases in urine volume and marked increases in urinary glucose at all doses that were related to the pharmacology of the compound to the same extent as was seen in the rat carcinogenicity study. These effects were accompanied by increases in urinary calcium excretion, especially at the high dose (up to 11-fold), compared to the vehicle controls. In addition, there was a modest increase (2) in urinary excretion of other electrolytes (eg, phosphorus, sodium, potassium, and magnesium) that are attributed to glucose-related osmotic diuresis related to the mode of action. Decreases were also observed in the calcium regulating hormones (1,25-dihydroxyvitamin D and PTH). 3.3. Calcium absorption in rats Increased urinary calcium excretion in rats is attributed to increased calcium absorption, which was shown by measuring uptake of orally administered radiolabeled calcium (Table 11). Treatment with canagliflozin increased calcium absorption from the gastrointestinal tract by 40% (based upon AUC0–24 h for labeled calcium) on the first day of treatment, and this increase was maintained throughout the 2-week treatment period. 3.4. Hormone evaluation in rats Sustained elevations in LH can lead to the formation of LCTs in rats. LH and testosterone measurements in canagliflozin-treated and vehicle control male rats in the 7-month repeated-dose study are presented in Table 12. Relative to vehicle treated males, LH

levels increased in rats treated with canagliflozin. Increases in LH were seen at the 1-month time point and were generally maintained throughout the treatment period. Testosterone levels were unchanged at all intervals compared to vehicle control.

4. Discussion The carcinogenicity profile of canagliflozin showed treatmentrelated tumors in three target organs in rats. These consisted of RTTs, pheochromocytomas, and LCTs. Canagliflozin is not genotoxic based on a battery of in vitro (Ames, mouse lymphoma) and in vivo (micronucleus, Comet) tests [7,8]. Therefore, the tumors are considered to occur by a non-genetic mechanism. Basophilic RTTs (adenomas and carcinomas) were noted in both sexes at 100 mg/kg. The 2 RTTs (1 benign, 1 malignant) in males at 30 mg/kg and 1 RTT observed at 100 mg/kg were of the AV phenotype characteristic of spontaneously occurring tumors that have been reported to occur in rats [18]. Based on a survey of National Toxicology Program carcinogenicity studies, it was concluded that the AV tumors occur randomly across all groups and are not related to test article treatment [19]. Spontaneous AV tumors are more commonly seen in recent years, with the majority being reported in studies commencing since 2009. When they occur, spontaneous AV tumors are often observed in more than 1 rat within the same study (up to 7) and were documented to occur in rats as young as 7–10 weeks of age [18]. This literature reaffirms that AV tumors are spontaneous, non-treatment-related lesions. Further, it was recommended that AV tumors be recorded separately from conventional treatment-related RTTs in order to clearly distinguish these two renal tubule neoplasms from one another and allow for appropriate interpretation of a compound’s potential carcinogenic effect in the kidney [18]. Therefore, in the canagliflozin rat carcinogenicity study, only the basophilic RTTs at the high dose (100 mg/kg) were considered treatment related. In addition, the kidneys showed findings related to chronic hypercalciuria, increased urinary volume (tubular and pelvic dilatation), and increased CPN, with only a marginal increase in simple tubular hyperplasia in the high-dose males. There was a slight increase in incidence and severity (predominantly scores of Grade 2 or 3) of CPN in treated groups compared to the control groups. However, it was not sufficient to explain the formation of renal tubule tumors, especially at the observed tumor incidence. Generally, renal tubule tumors are associated with severe CPN having a score

8

S. De Jonghe et al. / Chemico-Biological Interactions 224 (2014) 1–12

Fig. 2. (Color reproduction on the Web and in print). Representative histology micrographs of AV tumor, basophilic renal tubular tumor, and unaffected kidney in the 2-year rat carcinogenicity study. 20 magnification for the main photos and 40 for the inserts. (A) A spontaneous AV tumor in a male dosed at 30 mg/kg. The tumor cells form a solid mass of cells. In the insert, the cytoplasm of tumor cells is eosinophilic to amphophilic, finely granular, and contains numerous vacuoles. The nuclei of these cells are large with prominent nucleoli. (B) A treatment-related basophilic renal tubular tumor in a male dosed at 100 mg/kg. The tumor cells (photo and insert) contain hyperchromatic nuclei that are surrounded by a small amount of basophilic cytoplasm. (C) Contralateral unaffected kidney of the male with basophilic renal tubular tumor dosed at 100 mg/kg shown in B. A glomerulus is evident at the top and proximal tubules are evident around the glomerulus. In the insert, a brush border is evident on the luminal surface of the proximal tubules.

of 4 or 5 (end-stage nephropathy) [20]. Neither tubular hyperplasia nor CPN was observed in the 6-month rat repeated-dose toxicity study at the same dose level [7]. Pheochromocytomas (benign and malignant) accompanied with medullary hyperplasia were observed in the adrenal glands

of males at 30 mg/kg and both sexes at 100 mg/kg. The slightly higher incidence of pheochromocytomas in males at 30 mg/kg was considered to be not toxicologically relevant based upon two observations. First, the frequency of pheochromocytomas observed in the mid-dose males (and females) was within the range of the

9

S. De Jonghe et al. / Chemico-Biological Interactions 224 (2014) 1–12 Table 8 Testicular tumors and hyperplasia in the 2-year rat carcinogenicity study with canagliflozin. Daily dose (mg/kg)

N

Male

Adenoma interstitial cell (Leydig cell) Hyperplasia interstitial cell Grade 1 Grade 2 Grade 3 Grade 4

0 65

10 65

30 64

100 65

HCDa 540

1 12 9 2 1 –

8* 24* 14 8 – 2

20* 35* 13 20 2 –

24* 32* 10 17 5 –

4

HCD = historical control data; N = number of animals. a Incidence at the study site. * – Indicates a treatment-related effect.

Table 9 Non-proliferative histopathology findings in main target tissues in the 2-year rat carcinogenicity study with canagliflozin. N = 65/group

Males

Daily dose (mg/kg)

0

10

30

100

0

10

30

100

44

60* 10* 47* 14 5 4* 2* 15* "*

56* "* 18* 53* 27* 24* 44* 19* 27* "*

23

4 13 13 3 – – 12

59* "* 17* 49* 20* 7 9* 6* 22* "*

37* "* 7* 40* 5 4 11* 6* 38 "*

47* "* 14* 46* 6 4 22* 3* 39 "*

50* "* 15* 50* 14* 29* 32* 30* 56 "*

1

2*

11*

53*

2*

4*

11*

56*

10

5

11 "*

24* "*

Kidneys Chronic progressive nephropathy Increased severity Dilatation pelvic Dilatation tubule(s) Hypertrophy tubule(s) Mineralization cortex Mineralization medulla Mineralization papilla Mineralization pelvis Increased severity Bone Hyperostosis, trabecular Accessory sex glands Atrophy Increased severity *

Females

3 14 7 4 7 1 56

Indicates a treatment-related effect.

Table 10 Bone and renal findings in the 3-month rat study with canagliflozin. Daily dose (mg/kg)

0 (Control)

4

M:15 10.87 1.0371

F:14 8.50 1.0289

M:10 1.43 1.75*

F:10 1.41* 1.93*

M:10 2.31* 1.48*

F:10 2.12* 1.90*

M:15 3.39* –

F:14 2.85* 1.61*

Urine chemistry (week 13) Glucose/creatinine (mg/mg) Calcium/creatinine (mg/mg) Magnesium/creatinine (mg/mg) Phosphorus/creatinine (mg/mg) NAG/creatinine (U/mg) NAG-EXCRETED (U/16 h) Protein/creatinine (mg/mg) Protein-EXCRETED (mg/16 h)

0.128 0.052 0.190 1.396 0.013 0.141 0.607 6.813

0.129 0.171 0.356 2.074 0.014 0.097 0.138 0.895

1141* 2.24* – – – – – –

1539* 2.29* – – – – – –

1733* 3.38* 2.49* 2.90* 2.26* 2.00* – –

2039* 2.88* – 2.10* – – 2.92* 2.77*

2023* 10.54* 3.76* 3.57* 2.59* 2.58* 2.12* 2.14*

2660* 7.36* 2.28* 2.49* 2.35* 2.22* 3.25* 3.15*

Hormone analysis 1,25-Dihydroxyvitamin D (pg/ml) Parathyroid hormone (pg/ml)

60.27 1849

16.83 1496

– –

– –

0.18* –

– –

0.03* 0.35*

0.09* 0.39*

M:10 – – 1 – –

F:10 1 – – – –

M:10 – – 4 – –

F:10 2 – 1 – –

M:10 2 4 5 – –

F:10 5 – 1 2 –

M:10 6 8 7 – –

F:10 9 4 2 4 –

Urinalysisa (week 13) Urine volume (ml) Specific gravity

Histopathology Bone, stifle: hyperostosis, trabecular, metaphyseal (minimal to moderate) Bone, sternum: hyperostosis, trabecular (minimal to mild) Kidneys: mineralization, interstitial (minimal) Kidneys: mineralization, pelvis (minimal) Kidneys: tubular dilatation

N=

N=

20

For controls, group means are shown. For treated groups, multiples of controls are shown. – No noteworthy findings. F = female; GGT = gamma glutamyl transferase; M = male; N = number of animals; NAG = N-acetyl-b-D-glucosaminidase. a At end of dosing period. * indicates statistical significance (p < 0.05 or less) based on actual data (not on the multiples of control).

100

10

S. De Jonghe et al. / Chemico-Biological Interactions 224 (2014) 1–12

Table 11 Effect of canagliflozin on 45Ca absorption in rats following oral administration of canagliflozin (100 mg/kg) and 45Ca (20 lCi/rat) daily for 2 weeks. Time (h)

Plasma Concentration (45Ca lg/g) Day 1 Vehicle

0 1 3 6 24 AUC0–24 h (lg h/ml)

Urinary excretion (0–24 h)

Day 14 Canagliflozin

0.0 0.0 17.9 13.6 10.8 17.2 6.61 12.3 2.42 3.06 137 200 Urinary excretion of dose) 0.41 1.67

Vehicle

Canagliflozin

0.0 0.0 18.8 11.4 9.45 16.1 5.96 10.8 2.51 2.94 131 182 45 Ca (% of administered 0.24

5.07

Table 12 Luteinizing hormone and testosterone levels in male rats orally administered canagliflozin for 7 months and fed a standard diet. Daily dose (mg/kg)

0

100

Luteinizing hormone (ng/ml)

1 month 2 months 3 months 4 months 5 months 7 monthsa

0.64 0.50 0.54 0.45 0.44 0.33

1.75* 1.50* 1.46* 1.56* 1.84* 1.30

Testosterone (ng/ml)

1 month 2 months 3 months 4 months 5 months 7 monthsa

1.50 2.11 1.53 1.57 2.08 1.39

1.49 0.97 1.26 1.24 1.00 1.15

30 males/group.  = multiple compared to control. a Due to a collection error at the 6-month interim, the sample was taken at 7 months. * statistically significant change (p < 0.05 or less; two-sided Mann–Whitney U test) compared to standard diet control.

laboratory’s historical control data. Second, survival in the treated male rats was statistically significantly greater in mid-dose males relative to vehicle controls (p < 0.001). Adrenal medullary tumor formation generally occurs late in the life of rodents and spontaneous pheochromocytomas are relatively common in male rats [21]. The probability of observing spontaneously occurring tumors is increased by the substantially greater survival in mid-dose males relative to vehicle controls (with nearly twice the number of animals at terminal sacrifice). Survival-adjusted analysis of hyperplasia and pheochromocytoma showed the absence of a statistically significant (p > 0.05) treatment-related effect at 30 mg/kg (Table 6). The NOEL for pheochromocytomas and RTTs is considered to be 30 mg/kg, and safety margins are 5–7 based on the ratio of systemic exposures (AUC) at the NOEL dose and the highest clinical dose of 300 mg once daily. In rats, pharmacologically and structurally diverse groups of non-genotoxic agents can induce pheochromocytomas [22] through a variety of pharmacodynamic mechanisms including dysregulation in calcium homeostasis, hypoxia, hypoglycemia, neurogenic or hormonal signals (growth hormone and prolactin), overfeeding and stress [23,24]. While different non-genotoxic mechanisms (hormonal perturbations, overfeeding, calcium homeostasis, etc.) either alone or in combination may play a role in induction of pheochromocytomas, based on pronounced increase (up to 11-fold) in urinary calcium excretion associated with hyperostosis and changes in 1,25 vitamin D and PTH observed in the

3-month repeated-dose toxicity study in rats (Table 10) at a dose corresponding to the high dose used in the rat carcinogenicity study, dysregulation of calcium homeostasis was considered as a predominant pathway for induction of pheochromocytomas by canagliflozin in rats. Further, in the 45Ca radiolabeled study (Table 11), increased plasma concentrations of radiolabeled calcium accompanied by increased urinary calcium excretion were observed and these results indicate that canagliflozin causes increased absorption of dietary calcium. Disturbed calcium homeostasis, through several different modes of action, has been associated with an increased incidence of adrenal medullary hyperplasia and pheochromocytomas in rats. Treatments that cause carbohydrate malabsorption in rats, such as lactose, sorbitol, xylitol, mannitol, and lactitol, are associated with abnormalities in calcium handling (including marked increases in calcium absorption and urinary calcium excretion) and treatment-related adrenal medullary hyperplasia and pheochromocytomas [24–28]. Male rats, regardless of the strain, show a higher incidence of polyol- and lactose-induced pheochromocytoma than do female rats [29]. While inducing pheochromocytomas in rats, lactose and xylitol do not induce pheochromocytomas in mice [29]. Agents causing carbohydrate malabsorption in humans have not been linked to an increased risk of pheochromocytomas [25]. Based on canagliflozin’s mechanism of action and possible inhibition of glucose absorption from the gastrointestinal tract (see discussion below), it is likely that canagliflozin at high doses can induce a carbohydrate malabsorption syndrome in rats similar to that observed with poorly digestible sugars. Carbohydrate malabsorption, increased urinary calcium excretion, and hyperostosis have also been described in rats with other SGLT1/2 inhibitors at high doses [30,31]. A series of studies was conducted that provided further evidence of carbohydrate malabsorption with canagliflozin in rats [9]. As noted above, canagliflozin induces marked increases in calcium absorption and urinary calcium excretion in rats, thus, providing a similar setting for the induction of pheochromocytomas as observed with poorly digestible sugars and sugar alcohols (lactose, sorbitol, xylitol, mannitol, and lactitol). As observed in sugar and sugar alcohol studies and in the 2-year rat carcinogenicity study with canagliflozin, a higher incidence of pheochromocytomas and medullary hyperplasia was observed in males when compared to females. Similar to sugar and sugar alcohol studies, adrenal medullary tumors or medullary hyperplasia were not observed in the 2-year mouse carcinogenicity study with canagliflozin [7,8]. Based on the similarities between canagliflozin and poorly digestible sugars and sugar alcohols, it is hypothesized that carbohydrate malabsorption mediated calcium dysregulation is likely the mechanism for formation of pheochromocytomas by canagliflozin in rats. In humans, there is no known association between lactose consumption and the incidence of pheochromocytomas [32]. Based on species differences, 5–7 safety exposure multiples relative to the clinical dose, and experience with use of sugar and sugar alcohols, pheochromocytomas observed in the rat carcinogenicity study with canagliflozin are not expected to have safety implications in humans. LCTs accompanied by interstitial cell hyperplasia were noted at all canagliflozin dose levels and increased in a dose-related fashion. The LCTs appeared late during the study, mainly in terminal males and after 600 days of dosing (except for 1 rat). In the mid- and high-dose groups, LCT’s were associated with atrophy of the accessory sex glands (seminal vesicles, prostate, and coagulating gland). Based upon the observed incidence of tumors, a NOEL could not be established for LCTs. It is important to note that both spontaneous and drug-induced LCTs are commonly observed in rats [33]. A wide range of medications causing LCTs in rats has not been reported to cause LCTs in humans [34,35]. Although the mechanism for LCT is not known for all drugs associated with this tumor, disruptions in

S. De Jonghe et al. / Chemico-Biological Interactions 224 (2014) 1–12

the hypothalamic-pituitary-gonadal axis resulting in increased LH, a Leydig cell mitogen, is an established mechanism shared by a number of non-genotoxic agents causing LCTs in rats [33]. Increases in LH leading to LCT in rats can be caused by agents inhibiting testosterone synthesis (ie, primary hypogonadism with a compensatory increase in LH) or inhibiting testosterone feedback on the hypothalamicpituitary-gonadal axis [33]. As shown in Table 12, sustained increase in plasma LH (1.46–1.84 fold) with no changes in testosterone levels was observed in rats treated with canagliflozin. The increased LH was observed within 1 month of treatment and the effects persisted throughout the course of the 7-month repeated-dose study. Further, rats are considered to be genetically predisposed to develop LCTs, in contrast to man where this tumor-type occurs at a very low incidence [33]. This predisposition may occur because rat Leydig cells possess LH-releasing hormone receptors and a 14 higher number of LH receptors when compared to humans [34,35]. In general, drug-induced LCTs in rats are not predictive for humans and their occurrence in rats is considered to be non-relevant to humans. Further, there were no changes observed in LH or testosterone levels in subjects with T2DM treated with canagliflozin for 12 weeks at the clinical doses of 100 mg and 300 mg once daily [7,8]. Therefore LCT formation seen in rats treated with canagliflozin was not deemed to be of clinical relevance. Hyperostosis, an increase in trabecular bone volume in the metaphysis beneath the growth plate, was seen in the carcinogenicity study and in repeated-dose toxicity studies in rats with 3 months of daily oral administration. Reversibility was shown in the 3-month rat study after an 8-week recovery period. Hyperostosis was not observed in the 3-month toxicity or 2-year carcinogenicity studies in mice [7,8]. In the 2-year rat carcinogenicity study, hyperostosis of the sternum and/or stifle (distal femur and proximal tibia) was more pronounced in both sexes at 100 mg/kg than at lower doses of 10 and 30 mg/kg. Hyperostosis was attributed to hyperabsorption of calcium that was shown in the labeled calcium absorption study, and this was accompanied by substantial effects on calcium homeostasis, including increased urinary calcium excretion and a decrease in the levels of calcium regulating hormones (1,25-dihydroxyvitamin D and PTH) as shown in the 3month toxicity study. Canagliflozin is a potent SGLT2 inhibitor, but also possesses less potent SGLT1 inhibitory activity. The IC50s are 3.7 nM for rat SGLT2 and 571 nM for rat SGLT1 [5]. As SGLT1 is expressed on the luminal surface of intestinal enterocytes and is responsible for glucose and galactose absorption, prolonged exposure of canagliflozin due to slow absorption phase (prolonged Tmax) in the rat gut lumen after oral administration of high doses would lead to inhibition of SGLT1 resulting in glucose malabsorption [36]. The unabsorbed glucose provides a substrate for bacterial fermentation in the distal gastrointestinal tract and thereby lowers the luminal pH which increases calcium solubility and enhances the 1,25-dihydroxyvitamin D independent absorption of calcium. The study with radiolabeled calcium confirmed increased calcium absorption in rats treated with canagliflozin following a single dose and the effect was maintained at the same level through 14 days of repeated dosing. Hence, the observed increase in urinary calcium excretion is attributed to increased calcium absorption. Increased urinary calcium excretion was dose related with maximal effect greater than 10-fold at the high dose of 100 mg/kg in the 3-month study. In contrast, near maximal urinary glucose excretion (pharmacological effect) was observed at the low dose of 4 mg/kg. These results indicate that increased urinary calcium is due to SGLT1 inhibition and is not related to SGLT2 inhibition. Due to the slow oral absorption phase of canagliflozin in rats, SGLT1 inhibition effect and associated calcium changes are unavoidable at the high doses needed to meet the requirements of dose selection as per ICH guidelines [37]. The high dose (100 mg/kg,  12 and 21 higher exposure multiples

11

in male and female rats, respectively relative to exposure at the highest clinical dose of 300 mg once daily) selected in the rat carcinogenicity study was based on the 10% body weight reduction relative to vehicle controls observed at 100 mg/kg dose at the 3 month interim of the 6-month repeated dose toxicity in rats. At the 3month interim, there were no effects on the body weights at low (4 mg/kg) and mid (20 mg/kg) doses. Due to lack of body weight changes and lower exposure multiples at lower doses, 100 mg/kg was considered to be the maximum tolerated dose and therefore appropriate as a high dose in the carcinogenicity study. Testing a sufficiently high dose with calcium effects was unavoidable and as the high dose did not negatively impact the overall survival of rats in the carcinogenicity study, the 100 mg/kg high dose was considered appropriate. Worldwide health authorities including the U.S. Food and Drug Administration and the European Medicines Agency concurred with the dose selection and considered the study sufficient to evaluate the carcinogenic potential of canagliflozin [7,8]. In mice, the lack of findings related to SGLT1 inhibition and calcium hyperabsorption (eg, bone hyperostosis) can be explained by the greater oral bioavailability and much faster absorption of canagliflozin in mice compared to rats. Similar to findings with canagliflozin, carbohydrate malabsorption increased urinary calcium excretion, and hyperostosis have also been described in rats with poorly absorbable SGLT1 inhibitors and other SGLT2 inhibitors at doses in excess of the exposure providing maximal SGLT2 inhibition [30,31]. Similar tumor profiles have been reported under conditions of carbohydrate malabsorption, induced by an alpha glucosidase inhibitor (acarbose) and by diets of poorly absorbable carbohydrates (eg, lactose). Specifically, in rats, acarbose is associated with RTT and LCT, while lactose and other poorly absorbable sugars are associated with LCT and pheochromocytoma [23,28]. Acarbose has not been associated with an increase of these tumors in humans [38,39], nor has higher dietary intake of poorly absorbable carbohydrates been associated with such tumors in humans [40,41]. While induction of tumors via carbohydrate malabsorption is a plausible mechanism, additional investigative studies were conducted and are reported in the companion paper [9] that supports this mechanism of tumor formation with canagliflozin in rats. Based on several years of safe clinical use of drugs like acarbose, together with absence of canagliflozin treatment related tumor findings in mice and dogs, and the proposed mechanisms for tumor formation in rats, the tumor findings in the rat carcinogenicity study were not deemed to have clinical relevance. 5. Conclusion In conclusion, the carcinogenicity potential of canagliflozin was investigated in a 2-year study in rats. Three types of treatmentrelated tumors (LCTs, pheochromocytomas, and RTTs) were observed. LCTs were attributed to increased LH, an established mechanism in rats. Pheochromocytomas were attributed primarily to altered calcium homeostasis following increased calcium absorption secondary to carbohydrate malabsorption. RTTs may also have been linked to carbohydrate malabsorption, and additional studies that were reported in the companion paper further supported this mechanism. Based on lack of tumor findings in the mouse and dog, evidence that the mechanisms that underlie the occurrence of these tumors are rat-specific, safety exposure multiples and experience with drugs like acarbose, tumor findings in the rat carcinogenicity study with canagliflozin were not deemed to be of clinical relevance. Funding This work was supported by Janssen Research & Development, LLC and Janssen Research & Development, a Division of Janssen

12

S. De Jonghe et al. / Chemico-Biological Interactions 224 (2014) 1–12

Pharmaceutica NV. The authors are all employees and, as such, designed the studies, participated in data collection and analysis, made the decision to publish, and prepared the manuscript. Conflict of Interest The authors declare that there are no conflicts of interest. Transparency Document The Transparency document associated with this article can be found in the online version.

Acknowledgements The authors are grateful to Dr. Richard Sharpe (University of Edinburgh) for advice on the mechanistic studies related to LCTs and Dr. Terry Nett (Colorado State University) for assistance with the mechanistic studies and for the conduct of assays for testosterone and LH. The authors are grateful to Dr. Gordon Hard and Dr. Jerry Hardisty (Experimental Pathology Laboratories) for assistance with microscopic evaluation of the kidneys. The authors thank Verna Hillsamer (Janssen Research & Development, LLC) for writing assistance in the development of this manuscript. References [1] E.M. Wright, B.A. Hirayam, D.F. Loo, Active sugar transport in health and disease, J. Intern. Med. 261 (2007) 32–43. [2] J.R.L. Ehrenkranz, N.G. Lewis, C.R. Kahn, J. Roth, Phlorizin: a review, Diabetes Metab. Res. Rev. 21 (2005) 31–38. [3] A. Mather, C. Pollock, Renal glucose transporters: novel targets for hyperglycemia management, Nat. Rev. Nephrol. 6 (2010) 307–311. [4] E.C. Chao, R.R. Henry, SGLT2 inhibition – a novel strategy for diabetes treatment, Nat. Rev. Drug Discov. 9 (2010) 551–559. [5] Y. Liang, K. Arakawa, K. Ueta, Y. Matsushita, C. Kuriyama, T. Martin, F. Du, Y. Liu, J. Xu, B. Conway, J. Conway, D. Polidori, K. Ways, K. Demarest, Effect of canagliflozin on renal threshold for glucose, glycemia, and body weight in normal and diabetic animal models, PLoS ONE 7 (2) (2012) e30555, http:// dx.doi.org/10.1371/journal.pone.0030555 (Advance Access published February 15, 2012). [6] J. Rosenstock, N. Aggarwal, D. Polidori, Y. Zhao, D. Arbit, K. Usiskin, G. Capuano, W. Canovatchel, W. Canagliflozin DIA 2001 Study Group, Dose-ranging effects of canagliflozin, a sodium-glucose cotransporter 2 inhibitor, as add-on to metformin in subjects with type 2 diabetes, Diabetes Care 35 (6) (2012) 1232– 1238. [7] U.S. Food and Drug Administration, Center for Drug Evaluation and Research. Pharmacology/Toxicology Review(s) of Canagliflozin NDA 204042 for the treatment of type 2 diabetes submitted by Janssen Pharmaceuticals Inc. on 31 May 2014, Pharmacology/Toxicology NDA Review and Evaluation. 264 pgs. . Accessed July 2, 2014. [8] European Medicines Agency, Committee for Medicinal Products for Human Use. Marketing Authorisation Application Assessment Report: Canagliflozin, Procedure No. EMEA/H/C/002649/0000. EMA/374133/2013, 19 September 2013. . Accessed July 2, 2014. [9] R.N.V.S. Mamidi, J. Proctor, S. De Jonghe, B. Feyen, E. Moesen, P. Vinken, J.Y. Ma, S. Bryant, S. Snook, C. Louden, G. Lammens, K. Ways, M.F. Kelley, M.D. Johnson, Carbohydrate malabsorption mechanism for tumor formation in rats treated with the SGLT2 inhibitor canagliflozin, Chem. Biol. Interact. 221 (2014) 109–118. [10] J. Gabrielsson, D. Weiner, Chapter 3.7.2: non-compartmental analysis, in: J. Gabrielsson, D. Weiner (Eds.), Pharmacokinetic and Pharmacodynamic Data Analysis. Concepts & Applications, third ed., Swedish Pharmaceutical Press, Stockholm, 2000, pp. 141–146. [11] S. Siegel, N.J. Castellan Jr., Nonparametric Statistics for the Behavioral Sciences (International Edition), second ed., McGraw-Hill, Singapore, 1988. [12] R. Peto, M.C. Pike, N.E. Day, R.G. Gray, P.N. Lee, S. Parish, J. Peto, S. Richards, J. Wahrendorg, Guidelines for simple, sensitive significance tests for carcinogenic effects in long-term animal experiments, IARC Monogr. Eval. Carcinog. Risk Chem. Hum. Suppl. 2 (1980) 365–367. [13] L.A. Hothorn, M. Hasler, Proof of hazard and proof of safety in toxicological studies using simultaneous confidence intervals for differences and ratios to control, J. Biopharm. Stat. 18 (2008) 915–933.

[14] G.P. Bailey, T.P. Coogan, L.M. De Schaepdrijver, Preclinical development of a pharmaceutical product for children, in: A.M. Hoberman, E.M. Lewis (Eds.), Pediatric Nonclinical Drug Testing: Principles, Requirements, and Practices, John Wiley & Sons Inc., New Jersey, 2012, pp. 129–139. [15] T. Zoetis, M.E. Hurtt, Species comparison of anatomical and functional renal development, Birth Defects Res. B Dev. Reprod. Toxicol. 68 (2) (2003) 111–120. [16] O.J. Dunn, Multiple comparisons sing rank sums, Technometrics 6 (1964) 241– 252. [17] G.D. Niswender, A.R. Midgley Jr., S.E. Monroe, L.E. Reichert Jr., Radioimmunoassay for rat luteinizing hormone with antiovine LH serum and ovine LH-131I, Proc. Soc. Exp. Biol. Med. 128 (1968) 807–811. [18] T.A. Crabbs, S.R. Frame, V.A. Laast, J. Thomas, B. Zimmerman, J.F. Hardisty, Occurrence of spontaneous amphophilic-vacuolar renal tubule tumors in Sprague Dawley rats from sub-chronic toxicity studies, Toxicol. Pathol. 41 (6) (2013) 866–887. [19] G.C. Hard, J.C. Seely, G.E. Kissling, L.J. Betz, Spontaneous occurrence of a distinctive renal tubule tumor phenotype in rat carcinogenicity studies conducted by the National Toxicology Program, Toxicol. Pathol. 36 (2008) 388–396. [20] G.C. Hard, K.N. Khan, A contemporary overview of chronic progressive nephropathy in the laboratory rat, and its significance for human risk assessment, Toxicol. Pathol. 32 (2004) 171–180. [21] A.S. Tischler, R.E. Coupland, Changes in structure and function of the adrenal medulla, in: U. Mohr, D.L. Dungworth, C.C. Capen (Eds.), Pathobiology of the Aging Rat, vol. 2, ILSI Press, Washington, D.C, 1994, pp. 245–258. [22] A.S. Tischler, R.A. DeLellis, The rat adrenal medulla. 1. The normal adrenal, J. Am. Coll. Toxicol. 7 (1988) 1–21. [23] H. Greim, A. Hartwig, U. Reuter, H.B. Ricihter-Reichhelm, H.W. Thielmann, Chemically induced pheochromocytomas in rats: mechanisms and relevance for human risk assessment, Crit. Rev. Toxicol. 39 (2009) 695–718. [24] A.S. Tischler, J.F. Powers, J.C. Downing, J.C. Riseberg, M. Shahsavari, J. Ziar, R.M. McClain, Vitamin D3, lactose, and xylitol simulate chromaffin cell proliferation in the rat adrenal medulla, Toxicol. Appl. Pharmacol. 140 (1996) 115–123. [25] A. Baer, Sugars and adrenomedullary proliferative lesions: The effects of lactose and various polyalcohols, Int. J. Toxicol. 7 (1988) 71–81. [26] A.P. de Groot, B.A.R. Lina, A.J.M. Hagenaars, V.M. Hollanders, M. Andringa, V.J. Feron, Effects of a dietary load of acid or base on changes induced by lactose in rats, Food Chem. Toxicol. 33 (1995) 1–14. [27] F.J.C. Roe, A. Bär, Enzootic and epizootic adrenal medullary proliferative disease of rats: influence of dietary factors which affect calcium absorption, Hum. Toxicol. 4 (1985) 27–52. [28] E.J. Sinkeldam, R.A. Woutersen, V.M.H. Hollanders, H.P. Til, A. Van GarderenHoetmer, A. Bär, Subchronic and chronic toxicology/carcinogenicity feeding studies with lactitol in rats, J. Am. Coll. Toxicol. 11 (1992) 165–188. [29] B.S. Lynch, A.S. Tischler, C. Capen, I.C. Munor, L.M. McGirr, R.M. McClain, Low digestible carbohydrates (polyols and lactose): significance of adrenal medullary proliferative lesions in the rat, Regul. Toxicol. Pharmacol. 23 (1996) 256–297. [30] T. Kissner, N. Doyle, E. Krupp, M. Heinrichs, S. Haile, S.Y. Smith. Bone as a target tissue in the toxicological assessment of anti-diabetes drug class of SGLT1 inhibitors. Poster presented at: 49th Annual Meeting of the Society of Toxicology; March 7–11, 2010; Salt Lake City, Utah. [31] M. Tirmenstein, T.E. Dorr, E.B. Janovitz, D. Hagan, L.M. Abell, J.M. Onorato, J.M. Whaley, M.J. Graziano, T.P. Reilly, Nonclinical toxicology assessments support the chronic safety of dapagliflozin, a first-in-class sodium glucose cotransporter 2 inhibitor, Int. J. Toxicol. 32 (5) (2013) 336–350. [32] P. Greaves, Chapter 13. Endocrine glands, adrenal medulla, in: Histopathology of Preclinical Toxicity Studies, third ed., Academic Press, London, UK, 2007, pp. 812–819. [33] E.D. Clegg, J.C. Cook, R.E. Chapin, P.M. Foster, G.P. Daston, Leydig cell hyperplasia and adenoma formation: mechanisms and relevance to humans, Reprod. Toxicol. 11 (1997) 107–121. [34] S.M. Cohen, L.L. Arnold, Chemical carcinogenesis, Toxicol. Sci. 120 (Suppl. 1) (2011) 76–92. [35] J.C. Cook, G.R. Klinefelter, J.F. Hardisty, R.M. Sharpe, P.M. Foster, Rodent Leydig cell tumorigenesis: a review of the physiology, pathology, mechanisms, and relevance to humans, Crit. Rev. Toxicol. 29 (2) (1999) 169–261. [36] L. Drozdowski, A.B.R. Thomson, Intestinal sugar transport, World J. Gastroenterol. 12 (11) (2006) 1657–1670. [37] ICH S1C(R2). Dose Selection for Carcinogenicity of Pharmaceuticals. International Conference on Harmonization of Technical Requirements for Registration of Pharmaceuticals for Human Use. Harmonized Tripartite Guideline. March 2008. [38] F.J.C. Roe, Relevance for man of the effects of lactose, polyols and other carbohydrates on calcium metabolism seen in rats: a review, Hum. Toxicol. 8 (1989) 87–98. [39] P. Hollander, Safety profile of acarbose, an a-glucosidase inhibitor, Drugs 44 (1992) 47–53. [40] A. Bär, Significance of Leydig cell neoplasia in rats fed lactitol or lactose, J. Am. Coll. Toxicol. 11 (1992) 189–207. [41] G. Ursin, E. Bjelke, S.E. Vollset, Milk consumption and cancer incidence: a Norwegian prospective study, Br. J. Cancer 61 (1990) 456–459.

Carcinogenicity in rats of the SGLT2 inhibitor canagliflozin.

The carcinogenicity potential of canagliflozin, an inhibitor of SGLT2, was evaluated in a 2-year rat study (10, 30, and 100 mg/kg). Rats showed an inc...
1MB Sizes 0 Downloads 8 Views