ATR-101, a selective ACAT1 inhibitor, decreases ACTH-stimulated cortisol concentrations in dogs with naturally occurring Cushing’s syndrome

Hyperadrenocorticism, commonly known as Cushing’s syndrome (CS), is an endocrine disorder with many similarities between dogs and humans [1]. In both species, most cases of naturally-occurring disease result from adrenocorticotropic hormone (ACTH) secreting pituitary adenomas, while remaining cases often result from cortisol-secreting adrenal tumors [2, 3]. Regardless of etiology, clinical signs are usually a result of increased circulating cortisol concentrations. Dermatologic manifestations are frequent and can include thin skin, easy bruising, hyperpigmentation, and recurrent infections [4]. The catabolic effects of cortisol often lead to muscle wasting and central obesity resulting in a classic “Cushingoid” appearance. Serious consequences such as hypertension, dyslipidemias, and thromboembolism are reported in both species [2, 5].

In addition to pathophysiologic parallels, similar treatment options exist for both species [6‐8]. However, the actual treatment pursued for pituitary-dependent disease often differs as hypophysectomy and irradiation, both common therapies for humans with pituitary lesions, are infrequently performed in dogs in part due to expense and limited availability. Furthermore, depending on definition used, procedural success of dogs undergoing hypophysectomy for CS could be as low as 60% when considering post-operative mortality, treatment failures, and disease relapse [7, 9]. Direct comparative studies have not been performed, but median survival times do not appear to be substantially different in dogs undergoing hypophysectomy compared to dogs receiving medical therapy alone [9‐11]. As such, most dogs with Cushing’s disease are treated medically with agents that target the adrenal gland. Common therapies include trilostane, which inhibits enzymatic conversion of steroids by 3-beta-hydroxysteroid dehydrogenase, or mitotane, an adrenal cytotoxic agent [8, 12]. In veterinary medicine, the complicated dosing and side effects of mitotane have led to declining use, with a corresponding increase in trilostane treatment [13]. However, variable pharmacologic activity of trilostane has led to differing dosing recommendations with some dogs requiring up to three times daily dosing, and a small number of dogs having minimal response [14, 15]. Monitoring remains frequent, and fatal adrenal gland necrosis has been reported occasionally [15]. In humans, relapse of hypercortisolemia occurs in 15–66% of patients undergoing hypophysectomy, and persistent hypercortisolemia occurs in 14–72% of patients undergoing radiation therapy [2, 16]. Consequently, medical therapy is necessary for a subset of human patients; however, current options have limitations ranging from poor drug tolerability to variable efficacy [16].

Treatment has remained similar in dogs and humans with adrenal-dependent disease. Although adrenalectomy is curative for benign disease, management of malignant adrenocortical carcinoma (ACC) has remained challenging [17, 18]. Median survival times in dogs with detectable metastases are approximately 2–4 months despite medical or surgical intervention [18, 19]. In humans, surgical resection is palliative as the disease is often advanced at the time of diagnosis. Chemotherapeutics evaluated to date have been largely ineffective at reducing tumor burden, and controlling hypercortisolemia remains difficult [17, 20]. No drug is FDA approved for treatment of canine ACC, and the only FDA approved drug for treatment of human ACC is mitotane. However, stable disease or partial remissions only occur in approximately 30% of human patients with malignant disease, and survival times are poor [21]. Furthermore, mitotane toxicities are common and negatively affect quality of life in both humans and dogs [17, 20, 22]. Multi-agent protocols also have been used to treat human ACC; however, the median progression free survival of 5 months is discouraging [23].

Given these limitations, there is a critical need for new, targeted therapies for CS and ACC. ATR-101 (also known as PD132301–2) is a small molecule therapeutic in human clinical development for the treatment of ACC, congenital adrenal hyperplasia, and Cushing’s syndrome (CS). ATR-101 is a potent (IC₅₀ = 0.009 μM) and selective inhibitor of acyl coenzyme A:cholesterol acyltransferase isoform 1 (ACAT1). ACAT1 catalyzes cholesterol ester formation from cholesterol and long-chain fatty acyl-CoA [24], and in the adrenal cortex, is particularly important in creating a reservoir of substrate for steroid biosynthesis. Inhibition of this enzyme by ATR-101 at low doses results in decreases in circulating cortisol levels and other adrenal steroids in a time- and dose-dependent manner due to the lack of esterified cholesterol reservoirs. At high doses, ATR-101 leads to accumulation of free cholesterol in adrenocortical cells, ultimately leading to cellular stress and apoptosis [25‐27]. As such, ATR-101 therapy could be of benefit for multiple diseases associated with adrenal steroid dysregulation [26, 27]. Recent in vivo pharmacologic studies in healthy dogs [25, 27] have provided support for development of ATR-101 in ACC and endocrine diseases such as CS and congenital adrenal hyperplasia given that ATR-101 preferentially distributes to the adrenal glands and selectively inhibits adrenal ACAT1 activity. The frequently encountered gastrointestinal and neurologic side effects of mitotane have not been observed in initial studies of ATR-101 administration to healthy dogs [27]. A recent study has suggested that ACAT1 may also be a target of mitotane [28], but other, yet to be elucidated targets, may account for some of the effects of mitotane, although this has not been thoroughly investigated [29].

Given the similarities in dogs and humans with naturally occurring CS, the dog offers a unique opportunity to perform proof of concept clinical studies for a human therapeutic. In this study, the pharmacokinetics of ATR-101 were determined in dogs with naturally occurring CS. Furthermore, we investigated the biochemical and adrenal hormonal effects of ATR-101 administration over a 2–4 week time period. Adrenal gland histology was evaluated in dogs with adrenal-dependent disease. The results reported herein further support the ongoing development of ATR-101 as a novel agent for treatment of endocrine disorders associated with adrenal steroid dysregulation.

Our results document that orally administered ATR-101 reduces post-ACTH-stimulated cortisol concentrations in dogs with naturally occurring CS. This effect was observed by day 7 of ATR-101 therapy and sustained through study conclusion on day 28. Mild reductions of aldosterone concentrations were observed, but this effect was not thought to be clinically relevant as plasma aldosterone and serum electrolyte concentrations remained normal for the study duration. The minimal adverse effects also compare favorably with the adverse effect profiles of some currently used drugs such as mitotane. Positive hormonal (cortisol) effects were observed in all dogs with pituitary-dependent disease and 2 of 3 dogs with adrenal-dependent disease. These findings are encouraging given current limitations of medical therapy for CS in humans and the poor survival times and the difficulties in medically controlling hypercortisolemia in both humans and dogs with malignant adrenocortical lesions [16, 17, 19, 20, 39, 40]. The histologic changes in adrenal glands from dogs with adrenal-dependent disease further support previous findings in healthy dogs that ATR-101 distributes to the adrenal glands, inhibits adrenal ACAT1 activity, and induces apoptosis [25, 27]. Although this was not a blinded and controlled study, clinical improvements were reported by the majority of dog owners. In the aggregate, these results are supportive of further investigations and the ongoing development of ATR-101.

A limitation hampering human CS drug development is the paucity of disease and the failure to identify a naturally occurring model of disease. Although reported incidence varies, CS is rare in humans with an approximate annual incidence of 2 cases per million per year [41]. If CS is rare, adrenocortical carcinoma is exceedingly rare with approximately 600 new cases diagnosed in the Unites States each year [17]. In contrast, canine CS is a common endocrinologic disorder occurring at an estimated incidence of 1–2 cases per 1000 dogs per year, representing a thousand-fold greater incidence as compared to humans [42]. Despite disparity in disease frequency, the etiology, clinical signs, and biochemical and histologic characteristics are remarkably similar between species [2, 3, 43]. Furthermore, dogs age approximately 5–7 times faster than humans, and progression of CS is also accelerated. These features make the dog appealing for both molecular studies and clinical trials. Although this potential has been recognized by others [1, 4, 43], the clinical trial reported herein is the first study using naturally occurring CS in the dog to aid in clinical proof of concept for human CS. This should establish a valuable precedent in using the dogs with naturally occurring CS for the development of human therapeutics, hopefully resulting in benefits for both species.

Although one goal in both humans and dogs is to reduce circulating cortisol concentrations and alleviate clinical signs associated with this excess, methods for monitoring medical therapy can differ [2, 16, 20, 31]. Reductions in pre-ACTH-stimulated (basal) cortisol levels were observed in 4 dogs following ATR-101 therapy, but overall changes were not significant. The exact reasons for this are unclear, but are likely related to the episodic patterns of cortisol secretion observed in dogs that are different from the circadian or diurnal patterns of cortisol secretion observed in other species [44, 45]. Cortisol levels in normal dogs can decrease below and increase above established resting intervals throughout the day. High variability in basal or non-ACTH-stimulated cortisol levels is also observed in dogs with CS [46]. Because of this, measurements of unstimulated cortisol concentrations are not of great utility in the diagnosis or monitoring of CS in dogs [47]. While repeated 24 h urinary free cortisol determinations are routinely used for monitoring CS therapy in humans, this is impractical for most pet owners [6]. The urine cortisol to creatinine ratio, the most analogous clinical test in veterinary medicine, does not provide a consistently correct assessment of hormonal status for therapeutic monitoring of dogs with CS [20, 48, 49]. As such, the ACTH stimulation test is the current standard for monitoring medical therapy in dogs, and this was the basis for our decision to use this test for monitoring response to ATR-101 [20]. Reductions in 24 h urinary cortisol in humans and reductions in post-ACTH-stimulated cortisol concentrations in dogs do correlate with clinical improvement, and it would be logical to assume that reductions in one monitoring method would be accompanied by reductions in the other. However, one limitation of the current study is that it is unknown how well these monitoring methods correlate with each other. Given that pre-ACTH-stimulated cortisol level do not necessarily correlate to the degree of hypercortisolemia throughout the day, additional monitoring methods might prove beneficial. Although the urine cortisol to creatinine ratio is not as effective as ACTH stimulation testing for monitoring efficacy of medical therapy in dogs, values are known to decrease from baseline values following treatment with trilostane or mitotane [48, 49]. This monitoring method should be considered in future studies to enhance the translational aspect as it is most analogous to 24 h urinary free cortisol determinations utilized in humans. Other potential assessments of the degree of non-ACTH-stimulated hypercortisolism such as salivary, hair, or fecal cortisol measurements have undergone minimal investigation in dogs with CS, and their role in therapeutic monitoring is unknown [50‐52]. In addition to providing another measure of efficacy, these methods might serve as a more direct correlate to the repeated 24 h urinary free cortisol measurements utilized in humans. Future studies in which dogs with naturally occurring CS are used as a human CS model should consider incorporating multiple treatment monitoring methods, including an assessment of resting hypercortisolism.

In the current study, reductions in post-ACTH-stimulated cortisol concentrations were accompanied by subjective owner-reported clinical improvements in many dogs, although complete symptom resolution was not achieved in any of the dogs. The reasons for incomplete symptom resolution are likely two-fold. First, many clinical signs, especially physical changes and dermatologic abnormalities, can take several months to resolve [20]. Second, the observed reductions in post-ACTH-stimulated cortisol concentrations were heterogeneous and not of the magnitude typically associated with resolution of clinical signs [20, 31]. The lack of further decreases in cortisol concentrations despite an increased ATR-101 dose commencing on day 8 is in contrast to a previous study of ATR-101 in healthy dogs in which dose-dependent decreases in ACTH-stimulated cortisol concentrations were observed [27]. The reasons for the heterogeneity of post-ACTH cortisol responses and the discrepancy between studies are unclear; however, it is plausible that adrenocortical ACAT1 expression could differ between normal dogs and dogs with CS, or even among individual dogs with CS [28]. Similarly, the utilization of only 2 set dosages administered at only 1 frequency in dogs with varying types and severities of CS may account for some variation. The timing of ACTH stimulation testing may also be important as the optimal time for testing in trilostane treated dogs is 4 to 6 h post-drug administration [20]. Regardless, medical therapy of CS in dogs is often individualized as variable dosages and dosing frequencies are utilized in clinical settings [14]. Extended treatment duration, higher ATR-101 dose, or more frequent dosing potentially could result in further hormonal and clinical improvement. Additional studies are needed to evaluate these hypotheses.

Clinical signs of liver disease were not observed in the current study, and biochemical parameters associated with liver function such as bilirubin, glucose, and urea nitrogen remained normal. However, the increases in liver enzyme activity are noteworthy. Dogs with CS usually have increased liver enzyme activity at diagnosis [3], but the continual increases over time despite reductions in post-ACTH-stimulated cortisol concentrations were unexpected. ACAT1 is known to be expressed in numerous tissues including Kupffer cells within the liver; however, expression of ACAT1 is markedly higher in the adrenal cortex which is likely one reason for the adrenal selective effects of ATR-101 [27, 53]. In a previous in-vitro study, ATR-101 did not induce toxic effects on hepatocytes unless the cells were pretreated with agents to block glycolysis or inhibit cytochrome P450-mediated metabolism [54]. In an unpublished toxicity study conducted by two authors of this report (MB and SH), ATR-101 administration to healthy dogs resulted in mild increases in ALP and ALT activities. Histologically, non-glycogen, non-lipid containing vacuolar hepatic change was observed which was not accompanied by necro-inflammatory activity. Both abnormal enzyme activity and histologic changes resolved upon discontinuation of ATR-101. Given that ATR-101 is metabolized by the liver, the increased enzyme activity likely represents an adaptive response to hepatic drug metabolism [55], but the cause for these changes was not determined. The reported reductions in hematocrit in dogs from our study are unlikely to be of clinical significance as they remained within normal reference intervals for the study duration and rebounded by study completion. The reductions in albumin were minor and not deemed to be related to ATR-101 administration or clinically relevant as they also remained within normal reference intervals and appeared to plateau by study completion.

ATR-101 is well-tolerated and reduces post-ACTH-stimulated cortisol concentrations in dogs with naturally occurring CS. Given the striking similarities between CS in dogs and humans, naturally occurring canine CS appears to be a suitable model for human studies. The results reported herein further support the ongoing development of ATR-101 for treatment of disorders associated with excess adrenal steroid production in humans, such as CS, congenital adrenal hyperplasia, and ACC. Additional investigations are required to optimize dosing strategies and further evaluate the effects of prolonged drug administration on efficacy and safety.