IRIS study: a phase II study of the steroid sulfatase inhibitor Irosustat when added to an aromatase inhibitor in ER-positive breast cancer patients

IRIS is the first study to explore the efficacy of the sulfatase inhibitor Irosustat in ER-positive metastatic breast cancer in combination with an AI. In this study, we added steroid sulfatase inhibition to the background aromatase inhibition upon which disease progression had occurred. The underlying rationale for the study is based on multiple observations. Firstly, that after estrone is synthesised from androstenedione (Adione), much of it is rapidly sulphated to estrone sulfate (E1S), by the enzyme oestrogen sulfotransferase (EST) [12]. E1S is known to have a plasma concentration 10–20 times higher than those of the unconjugated oestrogens estrone and estradiol, as well as a longer half-life in the plasma than the unconjugated oestrogens and therefore acts as a reservoir for active estrogens [13]. E1S uptake into malignant breast tissue is facilitated by organic anion transporter polypeptide B (OATP) [14]. Secondly, androgens derived from the adrenal cortex are known to have oestrogenic effects [15]. Thirdly, STS, the enzyme responsible for the hydrolysis of steroid sulphates to their unconjugated biologically active forms, is overexpressed in breast cancer and is known to be associated with a worse outcome [4, 5]. Fourthly, STS is upregulated following exposure to an aromatase inhibitor (AI) which is a putative mechanism of resistance to AIs [8]. Fifthly, the additive use of oestrogen synthesis inhibition can result in clinical benefit as seen when an AI is added to ovarian suppression in premenopausal women with breast cancer [16]. We therefore hypothesised that the addition of an STS inhibitor to an AI at disease progression would result in a lowering of peripheral and intratumoral estrone and DHEA, reversal of resistance and clinical benefit.

The principal findings relate to effect on disease status, adverse events and effects on steroid hormones. Firstly, regarding the anti-tumour effects of Irosustat, the study met its primary endpoint of showing evidence of clinical benefit in a second-line setting based on both local and central reviews, with a median PFS of 2.7 months. Although there were no objective responses, a number of patients experienced disease stabilisation for at least 6 months. It is known that the objective response and clinical benefit rates of second-line endocrine treatment are limited. For example, the objective response rate to second-line endocrine therapy in large phase III studies with exemestane or high-dose fulvestrant has been reported to be in the range of 0.4–6.3%, while median PFS in these studies ranged from 2.8 to 3.8 months [17‐19]. Therefore, Irosustat as a single agent added onto the background of an AI has evidence of clinical activity and, with the caveats relating to cross trial comparison and their limitations, does show efficacy which is comparable to both exemestane and high-dose fulvestrant in this setting. Future studies need to explore the clinical activity of dual aromatase and STS blockade as first-line metastatic treatment in a patients likely to be endocrine sensitive, as well selecting for STS expression.

The suppression of circulating estradiol is a key therapeutic strategy in the management of ER-positive breast cancer, and all the women who entered this study had estrone and estradiol levels below the level of detection as the result of AI treatment. However, steroidogenic hormones derived from the adrenal gland have been implicated in AI resistance. DHEAS is the most abundant steroid in the human circulation and has been shown to be significantly elevated in women whose disease failed to respond to an AI in the metastatic setting [7]. DHEA has been shown to stimulate the proliferation of the oestrogen-dependent MCF-7 human breast cancer cell line, as well as a derived oestrogen-independent variant, and to transactivate ER in both lines [20]. In vitro DHEAS has been shown to transactivate both ER and the AR in a dose-dependent manner, with the DHEAS induced AR transactivity abolished by Irosustat [21]. DHEA can itself be converted to androstenediol by 17β-HSD1, as well as to androstenedione via 3β-HSD1 and 17β-HSD2. Androstenediol is known to have potent oestrogenic effects [15], and can stimulate the growth of hormone-dependent breast cancer cells both in vitro and in vivo [22, 23]. Androstenedione has been demonstrated to induce recruitment of ER and SRC3 to gene promoters and drive ER transcription [24] and AI resistance in vitro [25]. Androstenedione was significantly elevated in women on a second-line AI for metastatic breast cancer as compared to women being treated in the adjuvant setting [26]. Furthermore, androstenedione can be converted to 5a-androstane-3b, 17b–diol (3bAdiol) which is itself estrogenic, and can induce growth in breast cancer cells [27]. This compound has also been implicated in endocrine resistance [28]. Clinically, previous studies have also shown that AIs do not affect the levels of androstenediol [29], while serum DHEAS was significantly higher in women who progressed on an AI [7]. Steroids with potent estrogenic properties may provide an escape mechanism for growth within a low estradiol milieu.

Steroidogenic hormone measurements within the current study revealed treatment with Irosustat resulted in significant and predicable rise at 3 months in estrone sulphate and DHEAS, and decrease in DHEA:DHEAS as well as significant reduction in androstenedione at 3 months. These pharmacodynamics data confirm that Irosustat had effected sulfatase inhibition in our patients. Expression of STS and other enzymes involved in oestrogen biosynthesis were assessed predominately in primary breast cancer samples. Given the small numbers, no formal assessment can be made of the possible use of STS as a biomarker of response. However, moderate to strong expression of STS was observed in only 37 and 25% of the primary and recurrent cases, and the preponderance of tumours with low STS expression clearly may have impacted on the efficacy data. Furthermore, the primary tissue was mainly used and this may not reflect the biology of a tumour which has progressed following a prolonged period of oestrogen deprivation. Ideally, in future studies of Irosustat a metastatic biopsy should be mandated prior to study entry [30], and prospectively enrich for a STS high population of tumours.

Irosustat was well tolerated with most adverse effects being grades 1 and 2, and the most common being dry skin, nausea and fatigue. Dry skin is an expected side effect being reported in both phase I studies [10, 11]. This is expected given that hereditary deficiency of STS results in dry, scaly skin as a result of abnormal corneocyte retention and thickening of the stratum corneum [31].

There are several limitations to our study. Primary tumours were utilised for the IHC assessment since metastatic biopsies from the time of study entry were not routinely collected [30], particularly given the evidence that STS is upregulated following exposure to an AI [8]. Since the study was initiated, the development of ESR1 mutations as a resistance mechanism to AIs has been described. As all patients who entered the study had progressed on an AI, it would be expected that 20–30% would harbour ESR1 mutations [32‐34]. The possible effect of ESR1 mutations with regard to the efficacy of Irosustat is unknown. The study strengths are that we undertook local and central review of radiology, with the study meeting its primary endpoint by both assessments.

More recently, it has been demonstrated that the addition of either an mTOR inhibitor or a CDK4/6 inhibitor to an AI can significantly improve the outcomes of women with ER-positive metastatic breast cancer [17, 35]. Given the clinical activity of these targeted therapies is dependent on the dual targeting of ER with endocrine therapy, the development of a more effective ET backbone may lead to enhance efficacy of these biological therapies. Therefore, any future clinical development will need to attempt to address the question of whether an ET backbone which comprises it and an AI is more clinically effective than an AI alone when combined with targeted therapies such an mTOR or CDK4/6 inhibitors.

In summary, this proof of concept study provides evidence for the first time that combining both an AI and a STS inhibitor can have clinical activity and that the combination is safe and well tolerated. Future studies should explore the clinical activity of Irosustat at an early line of treatment in a populations enriched for STS expression.