Assessing the Carcinogenicity of Vadadustat, an Oral Hypoxia-Inducible Factor Prolyl-4-Hydroxylase Inhibitor, in Rodents

Abstract

Vadadustat is an investigational oral hypoxia-inducible factor (HIF) prolyl-4-hydroxylase inhibitor to treat anemia due to chronic kidney disease (CKD). Some studies suggest that HIF activation promotes tumorigenesis by activating angiogenesis downstream of vascular endothelial growth factor, while other studies suggest that elevated HIF activity may produce an antitumor phenotype. To evaluate the potential carcinogenicity of vadadustat in mice and rats, we dosed CByB6F1/Tg.rasH2 hemizygous (transgenic) mice orally by gavage with 5 to 50 mg/kg/d of vadadustat for 6 months and dosed Sprague-Dawley rats orally by gavage with 2 to 20 mg/kg/d for approximately 85 weeks. Doses were selected based on the maximally tolerated dose established for each species in previous studies. The tumors that were identified in the studies were not considered to be treatment-related for statistical reasons or within the historical control range. There was no carcinogenic effect attributed to vadadustat in mice or rats.

Anemia is a common complication seen in more than 50% of patients with late-stage chronic kidney disease (CKD) in the United States.^1,2 To treat anemia, current therapeutics include renal erythropoietin (EPO) and iron supplementation, but these may increase hemoglobin (Hb) levels, risking cardiovascular events in an already susceptible population.^2-4

Vadadustat is an investigational oral hypoxia-inducible factor prolyl-4-hydroxylase inhibitor (HIF-PHI) in development for the treatment of anemia due to CKD in adult patients who are dialysis-dependent (DD) or non–dialysis-dependent (NDD).^5,6 The HIF-PHIs stimulate production of endogenous EPO through the stabilization of HIF, a dimeric transcription factor that initiates pathways involved in regulating oxygen homeostasis. ⁷

The safety and efficacy of vadadustat have been established in clinical trials, but its carcinogenicity potential is under investigation. Some evidence suggests that HIF activation promotes tumorigenesis by activating angiogenesis downstream of vascular endothelial growth factor (VEGF). ⁸ While vadadustat stabilizes HIF to induce EPO secretion, there are no demonstrable effects on VEGF production. ⁹ Studies also suggest that elevated HIF activity may produce an antitumor phenotype. ¹⁰ Experimental evidence shows that vadadustat is not genotoxic (data not shown). At least 2 other PHIs display negative carcinogenicity potential in chronic rodent studies.^11,12 In one study in humans, the risk of neoplasia among >3600 patients treated with vadadustat was comparable with that in patients treated with darbepoetin alfa. ¹³

This program examined the carcinogenic potential of vadadustat in mice and rats, with dosing up to 2 years. Both studies were conducted with a daily dosing regimen similar to that used in the global phase 3 clinical trials for vadadustat.

Both the mouse and rat carcinogenicity study designs are detailed in Supplementary Table 1. Details pertaining to dosage justification, animal housing, and statistical analysis are elaborated in the Supplementary Methods.

Across species and sex, there were no statistically significant differences in survival for any of the vadadustat-treated groups compared with the vehicle group (Figure 1). In the mouse study, one male (cause of death was squamous cell carcinoma of the nonglandular region of the stomach) and one female (cause of death undetermined) in the 5 mg/kg vadadustat group were found dead on days 58 and 185, respectively, and one vehicle control female was found dead on day 69 (hemangiosarcoma with no clinical signs prior to death). There was no mortality in animals administered the water control, vehicle control (males), or 15 mg/kg/d or 50 mg/kg/d dose levels. As expected, there was a high incidence of mortality in the positive control (N-nitroso-N-methylurea [NMU]) group. In the rat study, dosing was intended to continue for up to 2 years. Specifically, predefined criteria indicated that dosing for either sex was to continue until the number of surviving control animals (water or vehicle controls) decreased to 20 or when the number of surviving animals in any drug-treated group decreased to 15. The rat study dosing duration was approximately 85 weeks. Females were terminated early in weeks 87 to 90, and males were terminated early in weeks 94 to 95. Summaries of survival/mortality (Supplementary Table 2) and cause of death (Supplementary Table 3) among rats are presented. The most frequent neoplastic causes of death/early euthanasia included tumors commonly observed in this strain of rat and are within the historical control limits of the testing facility. The most frequent non-neoplastic causes of death/early euthanasia included skin lesion and urinary infection in males, neither of which was associated with vadadustat administration.

Figure 1.

Vadadustat did not impact Tg.rasH2 mouse or Sprague-Dawley rat survival. Survival plots of male and female mice and rats taken over the 6-month and 85-week carcinogenicity studies, respectively. Test article concentrations from 2 to 50 mg/kg/d.

There were no apparent vadadustat-related effects on mean body weight and mean body weight gain during the study period for either mice or rats (Supplementary Figure 1). Occasionally, there were statistical differences in mean body weight or mean body weight gain during the study in the treated male groups in mice (relative to vehicle control); however, mean body weight on day 182 and overall mean body weight gain (days 1-182) were statistically similar and within biological variation for males and females. In the rat study, mean body weight (and occasionally mean body weight gain) was increased in females in one test group (20 mg/kg/d dose) over 2 time intervals (days 50-120: up to 7.2% from water control; 6.4% from vehicle control, and days 204-288: up to 9.4% from water control; 7.6% from vehicle control). However, this finding was not considered to be toxicologically relevant, as there were no other notable intervals during the study and no correlating changes in mean food consumption or overall changes in mean body weight gain. Experimental evidence suggests that the mechanism of action of vadadustat is through the stimulation of erythropoiesis, resulting in increased hematocrit. ⁹ As hematocrit increases beyond normal homeostatic levels, there is an associated increase in blood viscosity. Depending on the extent of hematocrit increase, microscopic changes may be mild in nature, such as increased cellularity in the spleen and bone marrow, or more significant in nature (e.g., hemorrhage, congestion, necrosis, and thrombosis/infarction formation in tissues such as the nonglandular stomach, heart, and kidney). Therefore, measurement of hematocrit is considered a good biomarker of exposure and likelihood of tissue damage. Vadadustat-related changes in hematology were observed in both studies (Figure 2). In mice, blood for clinical pathology was collected from the main study animals on the last day of the study. Vadadustat administration resulted in mildly higher circulating erythrocyte mass at the 50 mg/kg/d dose level (data not shown). Vadadustat-related minimally higher hematocrit, red blood cell count, and Hb concentration (data not shown) were observed in the 50 mg/kg/d group males. Females administered both 15 and 50 mg/kg/d had higher reticulocyte counts, but accompanying differences in erythrocyte indices were only observed in females at the highest dose. In the highest dose group, the large variability in hematocrit was driven mostly by one mouse, yet was still within normal biological variation. In rats, increases in hematocrit (Figure 2), red blood cell count, Hb, mean cell volume, and mean cell Hb (data not shown) at all time points were seen in males administered 20 mg/kg/d. Hematocrit in females administered 20 mg/kg/d was not statistically significantly different from controls due to high variability driven by 3 animals in the study (within normal biological variation). As seen in Figure 2, hematology parameters at weeks 26 and 52 are comparable with the results from the end of the study. Males demonstrated decreases in platelet count at weeks 26 and 52 in the 7 mg/kg/d group and in the 20 mg/kg/d group at all time points (data not shown). Decreased reticulocyte count was observed in males administered 20 mg/kg/d at week 26 and in males administered ≥7 mg/kg/d at week 52 (data not shown). These observations were not seen in females. Other differences in hematology parameters (including those for females), including those determined to be statistically significantly different, were due to biological variation and not related to administration of vadadustat.

Figure 2.

Tg.rasH2 mouse and Sprague-Dawley rat hematocrit levels. Percent hematocrit of male and female mice and rats following 6 months or 85 weeks of vadadustat exposure, respectively. Statistical significance is indicated by (*). Test article concentrations from 2 to 50 mg/kg/d. VC indicates vehicle control; WC, water control.

For toxicokinetic analyses, blood samples were collected at 0, 1, 2, 4, 8, and 24 hours from 3 males and 3 females at each time point (days 1 and 180 in mice; day 1 and weeks 26 and 52 in rats). Blood samples were processed in K3EDTA tubes and cold centrifugated (within 2 hours of collection). The resultant plasma was separated, transferred to chilled polypropylene tubes containing 40 μL of 0.48 g/mL anhydrous citric acid in deionized water per mL of plasma, and stored frozen at −70°C. Plasma samples were analyzed using a validated ultra-high performance liquid chromatography tandem mass spectrometry (UHPLC/-MS/MS) bioanalytical procedure (WIL protocol no. WIL-797038).

There were no quantifiable concentrations of vadadustat in toxicokinetic samples collected and analyzed from any groups of mice prior to dosing on days 1 and 180 (data not shown). All samples were below the limit of quantification (<60.0 ng/mL). A qualitative review of the toxicokinetic data revealed an apparent sex-related difference in systemic exposure (as assessed by maximum observed peak plasma concentration [C_max] and area under the plasma concentration–time curve [AUC]_0-24h) across dose groups (Table 1). In general, systemic exposure was higher (>2-fold) for females after doses of 5 and 15 mg/kg/d, but similar (approximately 2-fold) after doses of 50 mg/kg/d. Vadadustat was rapidly absorbed, with composite time to maximum plasma concentration (T_max) values of approximately 1.0 hour postdose across all dose groups and days of dosing for both sexes. The terminal half-life for vadadustat in mice ranged from 1.0 to 2.5 hours across all dose groups and days of dosing. In males and females, systemic exposure (as assessed by C_max and AUC_0-24h) increased with increasing dose of vadadustat. No accumulation or decrease of vadadustat (decrease not less than 0.5-fold and increase not greater than 3-fold) was observed upon repeat administration of vadadustat on day 180 for mice.

Table 1.

Summary of AUC_0-24h and C_max exposure levels for vadadustat in mouse plasma: days 1 and 180.

Dose (mg/kg/d)	Sex	Vadadustat
		Cmax (ng/mL)	AUC0-24h (ng·h/mL)
Day 1
5	Male	1160	2190
	Female	4080	9120
15	Male	6130	10,700
	Female	14,200	38,800
50	Male	44,300	85,500
	Female	44,500	157,000
Day 180
5	Male	699	2190
	Female	3830	8570
15	Male	4160	8240
	Female	12,600	33,700
50	Male	50,400	94,700
	Female	56,700	167,000

Abbreviations: AUC, area under the plasma concentration–time curve; C_max, maximum observed peak plasma concentration.

In rats, no relevant sex differences were observed for C_max or AUC_0-t values on the protocol-specified collection days of 1, 178, and 360, and the pharmacokinetic parameters presented below represent sex-combined values (Table 2). T_max was generally observed at 1.0 hour postdose on days 1, 178, and 360. In general, on days 1, 178, and 360, AUC_0-t increased in a more than dose-proportional manner. The C_max and AUC exposure to vadadustat on days 178 and 360 increased compared with day 1 but were similar between the 2 later time points. Day 178 and 360 accumulation ratios ranged from 1.83 to 3.37 for AUC_0-t. Sex ratios (female/male) ranged between 0.520 and 1.23 for AUC_0-t. Sex-combined day 360 vadadustat C_max and AUC_0-t exposures at the 2, 7, and 20 mg/kg/d dose levels were 2990, 14,000, and 36,400 ng/mL and 7260, 53,200, and 173,500 ng·h/mL, respectively.

Table 2.

Summary of AUC_0-24h and C_max exposure levels for vadadustat in rat plasma.

Dose (mg/kg/d)	Sex	Vadadustat (mean concentration)
		Cmax (ng/mL)	AUC0-24h (ng·h/mL)
Day 1
2	Male	1370	4570
	Female	1450	2720
7	Male	6950	19,800
	Female	6780	10,300
20	Male	21,500	105,000
	Female	42,000	117,000
Day 178
2	Male	3090	8750
	Female	2190	6740
7	Male	13,800	45,300
	Female	12,500	39,000
20	Male	37,000	214,000
	Female	45,400	194,000
Day 360
2	Male	3570	8630
	Female	2400	5900
7	Male	12,900	58,100
	Female	15,100	48,200
20	Male	33,400	156,000
	Female	39,300	191,000

Abbreviations: AUC, area under the plasma concentration–time curve; C_max, maximum observed peak plasma concentration.

Carcinogenicity study animals surviving until scheduled euthanasia were weighed, samples for evaluation of clinical pathology parameters were collected (in mice only), and then the animals were euthanized by isoflurane inhalation followed by exsanguination. All carcinogenicity study animals were subjected to a complete necropsy examination. A full list of the tissues collected is provided in Supplementary Table 4. Toxicokinetic animals surviving until scheduled euthanasia were euthanized by carbon dioxide inhalation (for hematology-assigned rats, this was after blood collection). In vadadustat-treated mice, males showed a statistically significant increase both in the incidence of lung tumors (combined carcinoma/adenoma) in the 15 mg/kg/d group (5/25) compared with the vehicle control (1/25) group and in lung tumors (combined carcinoma/adenoma, alveolar bronchiolar) compared with 15 mg/kg/d (5/25) and 50 mg/kg/d (5/25) in the water control (0/25) (but all were within historic controls at Charles River Laboratories [CRL], Spencerville, Ohio). The incidence of splenic hemangiosarcoma in males and females was higher than that of the historical control range for CRL-Spencerville’s male mice but was not considered related to vadadustat, as it was not statistically significant and was lower than the published incidence rates. An increase in squamous cell carcinoma in the nonglandular stomach of both 5 mg/kg/d and 15 mg/kg/d males (1/25, 4.0%) was not statistically significant and was within published historical control incidences (0.0%-4.0%) for this strain of transgenic mouse. In vadadustat-treated females, there was a statistically significant increasing trend in the incidence of adenoma and adenocarcinoma/adenoma combination in the Harderian glands (also within historic controls at CRL). All other incidences of neoplasms were within the historical incidences for the control hemizygous Tg.rasH2 mice. A summary of neoplastic findings and tumor-related causes of death can be found in Supplementary Tables 5 and 6. Expectedly, administration of NMU led to an increase in neoplasms (data not shown).

In rats, no carcinogenic effects related to vadadustat administration were observed at any dose level in this study. There were no statistically significant tumor findings among males. In vadadustat-treated females, there were 3 tumor findings that were statistically significantly different, including increasing trends in the incidence of malignant pheochromocytoma in the adrenal gland (water control [P = .0077] and vehicle control [P = .0057]) and of hepatocellular adenoma in the liver (water control [P = .0374] and vehicle control [P = .0297]). In both cases, there was no statistically significant difference between the incidence when evaluated by a 1-sided comparison with the combined water and vehicle control groups as well as an absence of vadadustat-related preneoplastic changes or benign tumors in the adrenal medulla and liver. In addition, benign pheochromocytoma was present in the adrenal gland of females in both control groups and the vadadustat-administered group, but these were not statistically significantly different. Other tumor incidences were generally randomly distributed in control and treated groups and/or did not exhibit a dose-related pattern.

The objective of these nonclinical studies was to determine the potential carcinogenicity of vadadustat in 2 rodent models. In both studies, vadadustat was tested at the maximally tolerated dose for at least 6 months in CByB6F1/Tg.rasH2 transgenic mice and 85 weeks in Sprague-Dawley rats. In the present studies, there was no evidence of increased tumor incidence outside the historical or published incidences in the control animals presently utilized. These results support clinical findings from two phase 3 clinical trials with vadadustat in which cancer incidence was similar across treatment groups. ⁶ Collectively, these data suggest that vadadustat has a low potential for causing carcinogenicity.