Pharmacology and Pharmacokinetics of the Newer Generation Aromatase Inhibitors¹ Free

Breast cancer is one of the leading causes of death in women in Western countries. It is anticipated that in the year 2002, 205,000 new cases of breast cancer will be diagnosed, and of those, 40,000 women will die from the disease in the United States (1). A significant proportion of patients with both early- and advanced-disease cancer have hormone-sensitive breast cancer, and this subset of patients can be identified by the presence of estrogen receptors (2, 3). Recently, a number of newer agents have become available to treat this subgroup of patients, including the AIs³ and the estrogen receptor down-regulator fulvestrant (Faslodex; Refs. 4, 5, 6, 7, 8).

In the United States, there are currently three AIs available for clinical use: (a) anastrozole; (b) letrozole; and (c) exemestane. AIs lower the estrogen level in postmenopausal women by inhibition of the P450 cytochrome enzyme aromatase, which catalyzes the conversion of androgens to estrogens. These agents are only effective in patients in whom ovarian function has been effectively suppressed either naturally or therapeutically.

Aminoglutethimide (9, 10) and testolactone (11, 12) are members of the earlier generation AIs. Testolactone was the first AI extensively investigated for treatment of advanced breast cancer. It was initially viewed as an androgen, even though virilizing effects were almost completely absent. Aminoglutethimide became established as an AI before the potential of testolactone was fully recognized. The efficacy of aminoglutethimide was evaluated in postmenopausal women and found to be similar to that of adrenolectomy and/or hypophysectomy (9, 10). However, the use of these earlier generation agents was not widespread because of substantial morbidity and lack of selectivity for the aromatase enzyme, which necessitated concomitant steroid supplementation.

The AIs are classified into two types: (a) type I, suicidal or noncompetitive inhibitors; and (b) type II, competitive inhibitors (13, 14). Type I inhibitors are steroidal compounds, and type II inhibitors are nonsteroidal drugs. Both types mimic normal substrates (androgens), competing with the substrate for access to the binding site on the enzyme. After initial binding, the next step differs for the two types: once a noncompetitive inhibitor has bound, the enzyme initiates its typical sequence of hydroxylation, but hydroxylation produces an unbreakable covalent bond between the inhibitor and the enzyme protein. Enzyme activity is thus permanently blocked; even if all unattached inhibitor is removed, enzyme activity can only be restored by new enzyme synthesis. Competitive inhibitors reversibly bind to the active enzyme site, and either no enzyme activity is triggered, or it is without effect. The inhibitor can disassociate from the binding site, allowing renewed competition between the inhibitor and the substrate for binding to the site. As a result, the effectiveness of competitive inhibitors depends on the relative concentrations and affinities of the inhibitor and the substrate. Continued activity requires constant presence of inhibitor.

To compete for binding to the active site, both competitive and noncompetitive inhibitors must necessarily share important structural features with the endogenous substrate. Noncompetitive inhibitors must also share structural features with androgens, allowing them to interact with the catalytic residual on the enzyme protein. This renders them inherently selective. By contrast, most competitive inhibitors interact with the heme iron, a common feature of all cytochrome P450 enzymes. Some may also bind to the highly conserved oxygen binding site in addition to the substrate binding site. Thus, unless the specificity of a competitive inhibitor is reinforced through other structural features, it may block the activity of a variety of cytochrome P450 enzymes, as does aminoglutethimide.

Both anastrozole and letrozole are type II nonsteroidal AIs, whereas exemestane has a steroidal structure and is classified as a type I AI, also known as an aromatase inactivator because it irreversibly binds with and permanently inactivates the enzyme. The clinical relevance of these differences in mechanism of action, if any, remains to be established.

All three drugs (anastrozole, letrozole, and exemestane) are given p.o. as once-daily doses. The recommended daily doses are as follows: for anastrozole, 1 mg; for letrozole, 2.5 mg; and for exemestane, 25 mg. The time needed to reach maximal E₂ suppression is 2–4 days for anastrozole and letrozole and 7 days for exemestane (15, 16). In the exemestane study, these measurements were done every 7 days, and it is not known whether maximal estrogen suppression after exemestane may have occurred earlier. Anastrozole and exemestane achieve a steady-state drug level by 7 days (17). In contrast, letrozole has been reported to take 60 days to achieve plasma steady-state levels, which may be reflective of the accumulation that occurs with letrozole over long-term dosing (18). These data suggested a nonlinear relationship after repeated dose administration of letrozole. The half-lives of anastrozole and exemestane are 41 and 27 h (19, 20), respectively, and for letrozole, the half-life has been reported to be up to 4 days (21). All three AIs effectively reduce E₂, E₁, and E₁S: anastrozole by 81–94% (22, 23); letrozole by 88–98% (23); and exemestane by 52–72% (17, 24). In one prospective study, a comparative evaluation was carried out to determine the extent of plasma E₁, E₁S, and E₂ suppression after 6 weeks of treatment with either anastrozole or letrozole in postmenopausal women with metastatic disease (23). There were no significant differences in reduction of plasma E₂ by either agent (84.9% and 87.8%, respectively, for anastrozole and letrozole). Anastrozole and letrozole reduced E₁ levels by 81.0% and 84.3%, respectively; the extent of this suppression was significantly greater in the letrozole group (P = 0.019). Letrozole reduced plasma E₁S levels significantly more than anastrozole (98% versus 93.5%, P = 0.019). The clinical significance of this additional suppression of estrogen by letrozole remains to be defined. There are no comparative studies evaluating exemestane in relation to other agents, but in individual studies, E₂ levels were suppressed by 65% with exemestane (17).

Breast cancer tumor cells and also stromal cells have the potential to synthesize estrogen using local aromatase activity. Local production of estrogen by this mechanism results in a significantly higher concentration of estrogens in tumor than in the circulation in postmenopausal women. Separate studies evaluating these three agents have demonstrated that all three can reduce tumor estrogens (25, 26, 27). These studies were not carried out as comparative trials; thus, it is difficult to assess whether one agent is more effective than another in reducing intratumoral estrogen. With all three agents, reductions within the tumors have been in the range of 97–98%. The initial second-line Phase III studies evaluated different doses of AIs, and one of the questions in those studies was whether higher doses would result in further reduction in intratumoral estrogen and in higher antitumor activity. The data from these studies have not provided convincing evidence that there is a dose-dependent antitumor activity of AIs: the 1 and 10 mg doses of anastrozole had similar antitumor activity (2); and second-line studies of 0.5 and 2.5 mg of letrozole (28) have shown no consistent superiority of the higher dose. There are no comparable prospective studies in which different doses of exemestane have been studied.

All three drugs have been evaluated at concentrations of ≤500 μm for pharmacokinetic activity with the cytochrome P450 system in the human and rat microsomal preparations (29, 30, 31). Anastrozole inhibits, in decreasing order of magnitude, CYP1A2, CYP2C8/9, and CYP3A4 (29). Anastrozole has no effect on CYP2A6 or CYP2D6. Letrozole strongly inhibits CYP2A6, moderately inhibits CYP2C19, and has a low affinity for CYP3A4 (30). Exemestane is metabolized by CYP3A4 (31). There is potential for drug-drug interactions if patients are prescribed concomitant medications that interact with these cytochrome P450 enzymes. Both anastrozole and letrozole have no known drug-drug interaction with cimetidine (a marker for CYP3A4 activity) or warfarin [a marker for CYP3A4 and CYP1A2 activity (29, 30)]. No drug-drug interactions have been formally reported for exemestane, although there is potential for interactions with drugs that affect CYP3A4. There is known drug-drug interaction of tamoxifen with anastrozole and letrozole. Concomitant administration of either anastrozole or letrozole with tamoxifen decreases the plasma level of the AI. Concomitant administration of letrozole and tamoxifen decreased the level of letrozole by 38% (90% confidence interval, 32–43%; Ref. 32). Anastrozole and tamoxifen administrated concomitantly in the ATAC trial lowered the plasma anastrozole level in the combined arm by 27% (90% confidence interval, 20–30%; Ref. 33). Whereas the combined therapy arm of the ATAC trial demonstrated a decrease in the plasma concentration of anastrozole, systemic E₂ suppression was similar to that observed in patients treated with anastrozole alone. Based on the observed drug-drug interaction and the lack of clinical data demonstrating the superiority of the combined arm, this arm of the study is being discontinued.

Anastrozole and letrozole have no androgenic, progestrogenic, or estrogenic effects such as weight gain, acne, or hypertrichosis. Exemestane has weak androgenic properties, and its use at higher doses has been associated with steroidal-like side effects such as weight gain and acne (24, 34). These side effects have to be evaluated if exemestane is given on a prolonged basis as in the adjuvant setting.

These three AIs have somewhat different effects on lipid profile. In one study in patients with metastatic disease (n = 952 at study entry), anastrozole showed no marked effect on lipid profile compared with baseline. Whereas these patients had advanced metastatic disease, these data do provide some evidence that anastrozole did not significantly change the lipid profile compared with the baseline studies (35). A smaller study (n = 44) that compared lipids at baseline and after 32 weeks of treatment with anastrozole (36) also did not show any significant alterations in the lipid parameters. Letrozole was evaluated in 20 women with advanced breast cancer who showed a significant increase in total cholesterol and low-density lipoprotein from baseline after 8 and 16 weeks of therapy (37). In a 9-week trial in advanced breast cancer, exemestane resulted in a significant decrease in both total cholesterol and HDL (38). Exemestane also decreased the total triglyceride levels in this study. In another European Organization of Research and Treatment of Cancer study, 24 weeks of exemestane had no impact on the lipid profile (39). Clinical implications of these changes remain to be defined.

Anastrozole data indicate that it has no impact on adrenal steroidogenesis at up to 10 times the clinically recommended dose, thus suggesting that this drug has little activity on other cytochrome P450 enzymes involved in steroid synthesis (40). In one study in healthy postmenopausal women, basal and ACTH stimulation did not differ from baseline levels after 14 days of anastrozole at a 5- or 10-mg daily dose (40). In another study, there was no change in the ACTH-stimulated response after 115 days on therapy (41). Letrozole studies showed a decrease in basal and ACTH-stimulated cortisol synthesis. In one study, patients with advanced breast cancer showed a significant decrease in ACTH-stimulated aldosterone levels after 3 months of letrozole treatment (2.5 or 0.5 mg; Ref. 42). Exemestane in one small study had no impact on cortisol or aldosterone levels with up to 7 days of treatment, and the dose-ranging studies included doses from 0.5 to 800 mg (43). Data from these limited studies demonstrate that anastrozole and exemestane have no effect on steroidogenesis. In contrast, letrozole alters cortisol and aldosterone levels even at therapeutic doses (18). However, the clinical relevance of these differences remains to be defined.

In the preclinical setting, exemestane has been evaluated and shown to prevent bone loss in that experimental system (44). The data from ongoing clinical trials will be of interest. In healthy postmenopausal women, letrozole given 2.5 mg for 6 months produced an increase in bone resorption, as assessed by increased urinary pyridinoline and deoxypyridinoline, over baseline values (45). In another study, 3 months of letrozole resulted in bone resorption as evaluated by measuring the markers of bone metabolism (46).

The preliminary data of the ATAC study do suggest that anastrozole therapy is associated with higher bone-related events, i.e., fractures, compared with the tamoxifen arm of the study (47). At the present time, there are no long-term clinical data regarding the effect of AIs on the bones, except for the ATAC trial. Bisphosphonate therapy, which can prevent or reduce the risk of bone loss, needs to be considered when offering these agents for extended periods of time to patients with early breast cancer. Currently, prospective studies are being conducted to define the benefit of bisphosphonates in this subset of patients.

In a large second-line breast cancer study (n = 713), the safety and efficacy of letrozole and anastrozole were prospectively evaluated in postmenopausal women with advanced breast cancer (48). In this multicenter, open-label, randomized trial, all patients had previously been treated with antiestrogen therapy and were allowed to have one chemotherapy for metastatic disease. Hormonal receptor status was known in 48% of the patients. The primary end point was time to progression, with secondary end points of objective response rate, duration of response, time to treatment failure, and survival. Time to progression, time to treatment failure, and survival were similar in the two arms. In the intent-to-treat analysis, a higher proportion of patients achieved an objective response with letrozole therapy (19.1% for letrozole versus 12.3% for anastrozole, P = 0.014). Similarly, clinical benefit rate was also higher in the letrozole group compared with anastrozole (27% versus 23%; odds ratio, 1.24; P = 0.218). In patients with known receptor-positive disease, the antitumor activity of the two drugs was similar. In patients with unknown receptor status, however, there were major differences in the objective response rates between the two drugs. Double-blind controlled clinical trials are needed to evaluate the safety and efficacy of these agents in hormone receptor-positive postmenopausal women. These studies will define whether there are major differences in the safety and/or efficacy of the three AIs available for clinical use.

Dr. Mitch Dowsett: I think there has to be the same sort of concern about these pharmacological comparisons as with the Phase II trials. We think of these biochemical measurements as absolute, but they’re not: the prime example of that is the levels of estrogen suppression reported in different studies using different labs. You can look at subtle differences, 85%, 88%, and 90%, but the differences represent the limits of the assays. We have assays which go down to 3 pmol/liter, which is something like 15% of the starting level. These studies just use different methods. They also have different approaches as to when they’ve made the measurements, how they have made the measurements. To get good pharmacological data, they ought to be collected within the context of randomized studies.

Dr. Aman Buzdar: You are absolutely right. I was asked to show the data, and these are the only randomized data available. I think there is no question that we need to do prospective blinded studies so that neither the investigator nor the lab has any influence.

Dr. Dowsett: I think clinicians in particular have a temptation to see biochemical data as absolutes that can be compared in a way that they would not attempt with Phase II clinical trial data.

Dr. James Ingle: There are two bits of information prior to the head-to-head study that was presented by Carson Rose at American Society of Clinical Oncology (48). One is the data from Dr. A. Brodie and Dr. B. J. Long suggesting that letrozole is more potent, at least in their model system (Q. Lu et al., Breast Cancer Res. Treat., 57: 183–192, 1999), and then we had the Geisler study from Dr. Dowsett’s laboratory that indicated that there was more suppression with letrozole than anastrozole (J. Geisler et al., J. Clin. Oncol., 20: 751–757, 2002). But is there any value to the extra little increment of estrogen suppression?

Dr. Steven Come: Are these one-time measurements or are they done serially over time in a system? What happens over time might be significant, just as it is in the models. We’re using what we see at one point in time to predict what the long-term outcome is going to be for the mouse or the patient.

Dr. Dowsett: A long time ago these measurements were done in Dick Santen’s and our studies on estrogen levels at relapse on aminoglutethimide [Santen, R. J. Overall experience with aminoglutethimide in the management of advanced breast cancer. In: R. W. Elsdon-Dew, I. M. Jackson, and G. F. B. Birdwood (eds), Aminoglutethimide: An Alternative Endocrine Therapy for Breast Carcinoma, p. 3. London: Academic Press, 1982; M. Dowsett et al., J. Clin. Endocrinol. Metab., 58: 99–104, 1984]. There was a sniff of a difference in estrogen levels, which you wouldn’t have said could explain the relapse of the patients. If we think of these now as hypersensitive cells, we can begin to interpret those findings. There is some discussion in the ATAC steering committee of collecting blood samples from patients long term to make sure these patients really have got the same degree of suppression.

Dr. Buzdar: Actually, there was a paper comparing estrogen suppression in responders and nonresponders to AIs. The degree of estrogen suppression was identical, so there must be some other mechanism.

Dr. Kent Osborne: I think we are wasting our time looking at blood estrogen levels. The action is in the tumor, where the estrogen levels are high in some of these women because of local aromatase.

Dr. Dowsett: I’m not sure I agree with that entirely. In terms of the proportional importance of intratumoral aromatase, the studies from Bill Miller and Mike Dixon, where they show the proportion of estrogen that gets into the tumor from the outside and the proportion that is synthesized within, demonstrate that there are some tumors in which there is no synthesis within the tumor (reviewed in Breast Cancer Res. Treat., 49 (Suppl. 1): S27−S32, 1998). Some of those patients respond to AIs. There are other patients in whom perhaps 80–90% of the estrogen comes from within the tumor. Overall, it’s about 50/50, but it varies much between tumors.

Dr. Brian Long: A positive correlation between aromatase expression and ER expression in the tumor has not been demonstrated.

Dr. Buzdar: If the tumor has very low aromatase activity and still you see the same degree of response, it means that perhaps the degree of suppression is not that much important.