Integration of Signal Transduction Inhibitors with Endocrine Therapy: An Approach to Overcoming Hormone Resistance in Breast Cancer¹

Current endocrine therapies for breast cancer produce inhibitory effects on tumor growth either by depriving hormone-dependent cells of estrogen sufficient to reduce ER³ activity (i.e., by ovarian ablation or aromatase inhibition) or by blocking receptor action with competitive nonsteroidal or steroidal antiestrogens (i.e., tamoxifen or fulvestrant). Tamoxifen has been the most commonly prescribed drug for breast cancer, and over two decades, its role expanded from treatment for advanced disease, to established adjuvant therapy after surgery for early breast cancer, which prolongs both disease-free and overall survival (1). Not all patients with breast cancer respond to tamoxifen, and of those who do, most eventually acquire resistance to the drug. Although resistance was demonstrated early in the development of endocrine therapies, any understanding of the mechanisms had to await the discovery of the ER and subsequent unraveling of its complex molecular biology (2). Approximately 30–40% of primary breast carcinomas have very low or absent levels of ER, and invariably this is associated with primary (complete) resistance. After an initial response to endocrine therapy with tamoxifen, some ER-positive tumors respond to second-line endocrine therapy at progression and thus have secondary (partial) resistance. Equally, some tumors, although they express the ER protein, are primarily resistant to tamoxifen from the outset. It is unlikely that one single mechanism can explain hormonal resistance in all breast cancer patients (3).

A key finding in clinical material from breast cancer patients who developed acquired tamoxifen resistance was that ER expression was often retained at relapse and that the ER protein was functional, as indicated by its ability to bind DNA at its estrogen response element and initiate gene transcription (4). Much research focused on whether alterations occurred in the structure and function of the ER protein, which may alter its response to tamoxifen. Point mutations in specific regions of ER generated by site-directed mutagenesis can alter the pharmacological response to tamoxifen from an antagonist to a full agonist (5), although little evidence for these critical mutations was found in human breast cancer patients (6). However, a recent report in premalignant proliferative lesions of the breast suggested that a specific point mutation in the hinge domain of ER may produce a hypersensitive receptor, which results in enhanced binding of coactivators in the presence of low E₂ levels (7).

If the ER protein remains expressed and functional in hormone-resistant breast cancer, it was questioned whether changes occurred in the protein complexes or pathways with which ER interacted. ER can interact with Jun/Fos proteins at the AP-1 promoter site of certain genes such as the collagenase gene, and tamoxifen may act as an agonist on AP-1-regulated genes in a tissue-specific manner (8). Elevated AP-1 DNA binding and c-Jun NH₂ terminal kinase activity have been found both in an experimental xenograft model of tamoxifen resistance (9) and in human Tam-R tumors (10). Activation of this pathway in breast cancer may relate to either enhanced Fos protein expression, enhanced MAPK signaling, or oxidative stress (9, 11). Alternatively, the second ER (ERβ) liganded with antiestrogens such as tamoxifen has been shown to activate AP-1-regulated genes (12), and enhanced expression of the ERβ isoform has been reported in endocrine-resistant breast cancer (13).

It is clear that ER signaling cannot be considered in isolation from the remainder of breast cancer cell biology. In particular, steroid hormone and growth factor cross-talk acts to modulate endocrine response in breast cancer (14). Abnormalities in growth factor signaling pathways may account for the endocrine-resistant phenotype and thus may represent a target for new therapies to overcome resistance and enhance clinical response rate. There appear to be at least three different levels at which this cross-talk may occur (Fig. 1). First, peptide growth factor signaling via the type 1 tyrosine kinase growth factor receptor family (EGFR and HER2) may be up-regulated during endocrine treatment and ultimately become harnessed by the cells, allowing them to grow in a hormone-independent manner. ER may still be involved in such signaling, with downstream MAPK (via ERK1/2) phosphorylating and activating ER independent of steroid ligand (15). Secondly, there is evidence that ER may become involved in non-nuclear estrogen-dependent signaling via interaction with the PI3K cell survival pathway (16). Not only can PI3K phosphorylate and activate ER, but ERα has been shown to interact with the p85a regulatory subunit of PI3K within the cell membrane, leading to activation of protein kinase Akt and the subsequent downstream cellular effects of PI3K. It is recognized that major pathways that activate PI3K include the type 1 insulin-like growth factor receptors, which are known to be both activated and under hormonal control in breast cancer (17). Finally, the ER protein may interact with the stress-activated protein kinase/c-Jun-NH₂-terminal kinase pathway either by ER binding with the AP-1 transcription complex of proteins or possibly via direct activation of ER by p38 MAPK (8). Moreover, in addition to activating ER protein directly, recent evidence suggests that kinase-mediated growth factor signaling may also modulate ER activity indirectly by enhancing activity of key coactivators such as AIB1 and p300/cAMP-responsive element-binding protein (CREB)-binding protein (18).

The cross-talk picture is complicated further by the fact that members of the peptide growth factor receptor family not only homodimerize but also heterodimerize and cooperate with each other to impart different cell signal cascades (19). Drugs targeting EGFR, which in vitro only appear to inhibit erbB1, in vivo may also modify HER2 signaling by interfering with EGFR/HER2 heterodimer signaling. This has been demonstrated in HER2-overexpressing breast cancer cell lines, where the EGFR-specific TKI ZD-1839 inhibited growth and in addition modulated HER3 signaling and prevented its activation of the PI3K/Akt cell survival pathway (20). Targeting EGFR alone, therefore, may have multiple other intracellular effects that are dependent on the cell context and degree of cross-talk that exists with other intracellular signaling pathways, including other peptide growth factor receptors as well as ER.

Recent evidence has suggested that several of these signal transduction pathways that cross-talk with ER become up-regulated or activated in endocrine-resistant breast cancer. Whereas it has been known for some time that enhanced EGFR or HER2 protein expression may be fundamental to growth control in ER-negative breast cancer (21), similar up-regulation of growth factor signaling has been observed in ER-positive breast cancers that are resistant to endocrine therapy and are associated with a poorer prognosis (22). Endocrine-resistant ER-positive MCF-7 cells show increased dependence on EGFR/MAPK-mediated signaling (23), whereas similar enhanced activation of the MAPK pathway was reported as an adaptive resistance mechanism in ER-positive MCF-7 breast cancer cells exposed to long-term estrogen deprivation (24). Although Tam-R MCF-7 breast cancer cells express levels of ER equivalent to hormone-sensitive wild-type MCF-7 cells, these also contain markedly elevated levels of total and activated EGFR/HER2 in association with activated phosphorylated ERK1/2 (25). This group found that increased MAPK-mediated growth may have resulted from greater autocrine TGF-α stimulation because administration of an ER down-regulator (ICI 182,780) significantly reduced ER and TGF-α expression and inhibited proliferative activity in Tam-R cells, although this inhibition could be overcome by the administration of EGF and/or TGF-α (26). They subsequently reported that the EGFR TKI ZD-1839 completely blocked both basal and ligand (EGF or TGF-α)-stimulated activation of ERK1/2 and profoundly inhibited cell growth in Tam-R cells (25). EGFR may not be the only up-regulated growth factor receptor because evidence exists that HER2/HER3 levels, together with PI3K/Akt signaling, may be enhanced in some Tam-R breast cancer cells lines.⁴

Whereas enhanced peptide growth factor signaling may account for hormone-resistant growth, it appears that liganded (tamoxifen or estrogen) ER may still be required to facilitate signal transduction even in endocrine-resistant tumors. We have recently demonstrated that whereas long-term estrogen-deprived MCF-7 cells retain ERα and ERβ expression and express progesterone receptor, they are resistant to the growth-inhibitory effects of tamoxifen and refractory to growth stimulation by E₂ but in contrast remain extremely sensitive to growth inhibition by the pure antiestrogen fulvestrant (Faslodex; Ref. 27). Fulvestrant is a steroidal compound that results in degradation and down-regulation of ER protein. The MAPK pathway was up-regulated in these cells with increased expression of activated phosphorylated ERK1/2 (28), which, as indicated previously (15), may result in steroid-independent phosphorylation and activation of ER signaling. If ER phosphorylation secondary to peptide growth factor/MAPK activation remains the mediator of endocrine-resistant growth in these cells, then ER down-regulation may also be an effective treatment. This has been confirmed in experiments with HER2-transfected ER-positive MCF-7 xenografts that are resistant to tamoxifen, where similar inhibition of growth was seen with either ER down-regulation by fulvestrant or the TKI ZD-1839 (29).

In addition to TKIs or down-regulation of ER levels, it is clear that other STIs may be effective against these endocrine resistant tumors. In vitro antiestrogen-resistant MCF-7 cells were growth inhibited both by an EGFR TKI (ZD-1839) and by a MEK-1 inhibitor [PD98059 (23)]. Others have shown that in MCF-7 cells stably transfected with HER2 and in BT-474 cells that overexpress HER2, blockade using either a HER2 TKI (AG1478) or a MAPK inhibitor (UO126) abrogated antiestrogen resistance (30). Other STI strategies may also be effective in blocking steroid- or growth factor-induced MAPK activation in breast cancer cells. For example, inhibition of Ras signaling by a FTI or a dominant negative Ras has been shown to significantly inhibit both estrogen- and TGF-α-induced activation of ERK1/2 in MCF-7 breast cancer cells (31). Based on the rationale and preclinical evidence outlined above, STIs might be expected to have clinical activity in hormone-resistant advanced breast cancer, and initial clinical trials in this area have been performed (see below).

The evidence above strongly suggests that enhanced signal transduction pathways may be one of the key adaptive changes accounting for endocrine-resistant growth in breast cancer and that STIs could treat or even prevent endocrine-resistant tumor growth. Several reports have implied that in hormone-sensitive breast cancer, STIs as monotherapy may have a minimal effect on tumor growth and could be less effective than endocrine therapy. This could relate to ER signaling being the dominant determinant of growth in hormone-sensitive cells, with absence of adaptive changes in activation of signal transduction pathways. However, in vitro data suggest that in these hormone-sensitive cells, combined treatment with tamoxifen and the TKI ZD-1839 may provide greater antiproliferative effects and delay hormone-resistant outgrowth by blocking the acquired up-regulation of EGFR/MAPK signaling (32). Such a strategy of combined endocrine/STI therapy could prove more effective than either therapy alone in hormone-sensitive breast cancer and, in particular, could delay the emergence of acquired resistance.

Other STIs having a minimal effect on hormone-sensitive breast cancer also be may more effective if combined with endocrine therapy and thus may provide greater tumor control than endocrine therapy alone. The FTI R115777 inhibits the growth of a number of human breast cancer cells lines in vitro, most of which contain normal wild-type ras genes (33). In vivo, R115777 (tipifarnib, Zarnestra) produced a modest cytostatic effect on MCF-7 xenograft growth, with evidence of induction of apoptosis and enhanced expression of the cell cycle-inhibitory protein p21 (34). In contrast, when R115777 was combined with tamoxifen or estrogen deprivation therapy, combined treatment induced significantly greater tumor regression compared with either endocrine therapy alone.⁵ These in vivo observations support previous findings of synergy between FTIs and tamoxifen in breast cancer cells (35). The mechanism for any interaction may relate to concurrent blockade of cell survival pathways, as suggested by Du et al. (36), who found that a PI3K inhibitor or serum starvation significantly enhanced proapoptotic effects of FTIs. It is well recognized that endocrine therapy may abrogate cell survival pathways by modifying the insulin-like growth factor receptor and PI3K/Akt pathways. Others have suggested that synergy may relate to a preferential cytotoxicity by FTIs against tumor cells induced into a nonproliferation state by tamoxifen (35). These emerging preclinical data have led to a clinical study in breast cancer to investigate the combination of a FTI with endocrine therapy.

For hormone-resistant breast cancer, the strategy of combined STI and endocrine therapy may also be more effective than using STIs alone. Evidence has come from two independent groups who investigated HER2-transfected MCF-7 cells, which are known to be tamoxifen resistant. Kurokawa et al. (30) first reported that signal transduction blockade using a HER2 TKI (AG1478) or a MAPK inhibitor (UO126) may abrogate antiestrogen resistance. In addition, they provided compelling evidence that combined treatment with tamoxifen and either STI was significantly more effective than either therapy alone, not only at inhibiting estrogen-mediated gene transcription and tumor colony survival in vitro but also at delaying tumor xenograft growth in vivo (30). More recently, Massarweh et al. (29) reported that whereas hormone-resistant MCF-7 cells with up-regulated HER2 signaling are sensitive to ZD-1839, combined therapy with ZD1839 and tamoxifen provided maximal growth inhibition and significantly delayed time to disease progression. Similar effects were also seen in this model for estrogen deprivation combined with ZD-1839, which suggests that in the clinic enhanced effects may be seen for the combination of ZD-1839 and aromatase inhibitors.

From the evidence presented above, it is clear that an approach to target signal transduction pathways may prove an effective treatment for breast cancer, particularly in the setting of endocrine resistance. An added advantage is that many of the STIs are p.o. active and appear to have an excellent tolerability profile. Clinical trials of STIs are in their infancy, although several different drugs are in development for breast cancer, including EGFR TKIs, FTIs, and mTOR antagonists that target the G₁-S-phase transition of the cell cycle downstream of Akt. Most of the STIs have completed initial Phase I studies to establish the toxicity profile, pharmacokinetic and pharmacodynamic effects, and optimal dose and schedule for further clinical development. The current status of each of these in relation to breast cancer is discussed below.

Several compounds are in development that target the intracellular tyrosine kinase domain of the EGFR and bind to the ATP site of the receptor. Despite the wide diversity of over 500 protein kinases in the human genome, including at least 90 tyrosine kinases, to date these drugs appear both potent and selective. Table 1 indicates the main compounds in development for cancer, with their respective potency against both EGFR and HER2.

ZD-1839 (Iressa) is an anilinoquinazoline that is the most advanced TKI in development. In Phase I studies, diarrhea was the dose-limiting toxicity at 800-1000 mg daily, and a mild (grade 1/2) acne-like skin rash was common in the first 2 months of treatment (37). Phase II studies in over 400 patients with NSCLC have demonstrated that Iressa (250 or 500 mg daily) has activity as monotherapy after failure of previous chemotherapies (38). In addition to objective responses that ranged from 9–18%, a significant number of patients (27–31%) had stable disease, and tumor-related symptoms improved in approximately 40% of patients in both studies, often within the first 2 weeks of treatment. In view of preclinical data suggesting that Iressa may enhance the response to several different chemotherapy agents (39), two randomized Phase III trials of standard chemotherapy with or without Iressa have been completed in over 2000 patients with NSCLC, and results are awaited.

A clear rationale exists to study Iressa in breast cancer cells, based on the response seen in cells resistant to endocrine therapy (25) or in those that overexpress growth factor receptors including HER2 (40). Further evidence in models of EGFR-positive ductal carcinoma in situ cells established as xenografts in athymic mice showed that ZD-1839 induced significant reduction in cell proliferation and increased epithelial cell apoptosis (41). Phase II monotherapy studies in advanced breast cancer have commenced, including patients with Tam-R disease. Several clinical trials of Iressa in combination with hormonal therapy are proposed both in advanced breast cancer and in ER-positive postmenopausal women as neoadjuvant therapy for 3 months before surgery. The latter studies are feasible, given the large safety database in NSCLC patients, together with preclinical data showing greater antiproliferative effects and delay in disease progression in hormone-sensitive breast cancer for combined ZD-1839 and tamoxifen compared with tamoxifen alone (32).

At least four other TKIs are in earlier stages of clinical development, including OSI-744 (Tarceva), which is another p.o. active EGFR-specific TKI with a similar toxicity profile to Iressa (42). Phase II studies of OSI-774 in a continuous 150 mg daily dosing schedule reported clinical activity in head and neck cancer and NSCLC (43, 44). In some of these studies, EGFR overexpression was not a prerequisite for antitumor activity. PKI-166 is a potent and reversible dual inhibitor that targets both EGFR and HER2 and was designed by modeling the ATP-binding site of the EGFR tyrosine kinase and optimizing the chemical classes of the pyrrolo- and pyrazole-pyrimidines (45). It exhibits potent antiproliferative activity against the EGFR-overexpressing cell lines in vitro and in HER2-transformed mammary epithelial cells implanted into nude mice (46). Phase I studies of PKI-166 that examined intermittent or continuous daily dosing schedules have been completed, and nausea, diarrhea, and skin rash were reported as the most frequent toxicities (47). GW-572016 is another potent dual inhibitor of EGFR and HER2 in early stages of clinical development, and initial Phase I studies have shown it to be very well tolerated, with minor gastrointestinal symptoms and rash again the most frequently reported toxicities (48).

Finally, much interest has surrounded CI-1033, which is a p.o. active 4-anilinoquinazolone derivative that is a potent and irreversible pan-erbB TKI (49). Preclinically, CI-1033 inhibited EGF-dependent tyrosine phosphorylation in both A431 human epidermoid carcinoma cells and MDA-MB-468 human breast cancer cells, and when it was administered p.o., it inhibited the growth of several different human xenografts in vivo including MB-231 breast carcinomas. There have been three Phase I clinical studies to investigate the safety and pharmacokinetics of oral CI-1033, which to date have revealed mild acneiform skin rash, nausea, and diarrhea as the most frequently reported drug-related toxicities. The largest Phase I study investigated different intermittent schedules (28 or 21 days on/7 days off) in addition to continuous daily dosing and showed no drug accumulation during repeat dosing, with a maximum tolerated daily dose of 175–200 mg in the intermittent schedules (50). In one study, repeat tumor biopsies were taken to determine the pharmacodynamic effects of CI-1033 after 7 days. Modulation of erbB1/HER2 phosphorylation was observed, together with inhibition of tumor cell proliferation (Ki-67 staining) and induction of the cell cycle inhibitor p27. No clear association between biomarker modulation and CI-1033 dose was seen (51). This would suggest that in additional studies, a lower biologically active dose should be evaluated in addition to the maximum tolerated daily dose. It remains to be seen whether pan-erbB inhibitors such as CI-1033 offer any advantage over EGFR TKIs, especially because these may indirectly modulate HER2/HER3 signaling (20).

Our current understanding of the signal transduction cascade downstream from growth factor receptor tyrosine kinases has revealed several other key proteins involved in malignant transformation, including the M_r 21,000 guanine nucleotide-binding proteins encoded by the ras proto-oncogene (Fig. 1). Processed Ras proteins localize to the inner plasma membrane and play a critical role in transmission of a variety of extracellular signals from the cell surface, including growth factor receptors such as EGFR and HER2. Whereas human breast carcinomas are known to contain a very low frequency of Ras mutations (<2%), aberrant function of the Ras signal transduction pathway is thought to be common in breast cancer (52). Posttranslational processing of Ras involves transfer of a 15-carbon farnesyl isoprenoid group from farnesyl diphosphate to the COOH-terminal tetrapeptide CAAX sequence, a reaction catalyzed by the farnesyl transferase enzyme. This led to the development of low molecular weight nonpeptide FTIs, the most advanced of which are SCH66336 (lonafarnib, Sarasar) and R115777 [Zarnestra (53)].

SCH66336 is a tricyclic compound that inhibits the growth of several tumor cell lines as well as K-Ras-transformed xenografts in vivo. In human xenograft studies, a wide variety of tumors including colon, bladder, lung, prostate, and pancreas tumors were growth inhibited in a dose-dependent manner, whereas prophylactic administration of SCH66336 delayed both tumor onset and growth (54). In patients with solid tumors, efficacy has been reported in early Phase I clinical studies in a variety of tumor types, including lung and head and neck cancer (55), and confirmation of biological efficacy has been demonstrated by inhibition of prenylation of prelamin A in buccal mucosa cells in treated patients (56).

R115777 is an imidazole-containing heterocyclic compound that is a potent and selective, p.o. active, nonpeptidomimetic inhibitor of the farnesyl transferase enzyme (33). Long-term therapy with R115777 was investigated in two additional Phase I studies, with either twice daily dosing for 21 or 28 days followed by 7 days of rest (57) or continuous oral dosing (58), with reversible myelosuppression being the dose-limiting toxicity. Based on preclinical evidence that R115777 is active in breast cancer models (34), a Phase II study has been completed in 76 women with advanced breast cancer that was no longer endocrine sensitive and who had progressed after previous chemotherapy. Clinical activity, including objective tumor responses in visceral and nonvisceral sites of disease, was seen in 24% of patients, and the drug was well tolerated with only a 14% incidence of grade 3/4 neutropenia using a 300 mg twice daily oral intermittent dosing schedule (21 days on/7 days off; Ref. 59). In view of these preclinical and clinical data, a randomized Phase II study of Zarnestra and letrozole has now started.

Another approach has been to develop small molecule inhibitors of Raf-I kinase, which interacts at the plasma membrane with activated Ras and initiates the MAPK cascade (Fig. 1). In particular, BAY 43-9006 demonstrated promising activity against various human cancer cell lines and tumor xenografts in vivo, including those with either mutant or wild-type Ras but aberrant growth factor (EGFR/HER2) expression (60). Early results from Phase I studies of various intermittent dosing schedules have suggested that BAY 43-9006 is well tolerated (61). Small molecule inhibitors of MEK and MAPK have also been developed but have yet to enter large-scale clinical development.

Cyclins are downstream proteins expressed at different points in the cell cycle that regulate the progression of cells through various checkpoints. Cyclin activity is regulated by CDKs, which form various protein-protein complexes, and these in turn are regulated by small proteins known as CDK inhibitors. Many of the latter have been considered tumor suppressor genes, which, if lost, may contribute to uncontrolled progression through the cell cycle (e.g., loss of p16 in melanoma). Therapeutic strategies have been conceived to abrogate cell cycle checkpoints, in particular by developing pharmacological inhibitors of CDKs. Two compounds (flavopiridol and 7-hydroxystaurosporine, or UCN-01) derived from microbial and plant sources have CDK inhibitor activity, have shown antitumor activity in various preclinical models including breast cancer, and have entered early clinical development (62).

Other signaling pathways important in breast cancer include the insulin-like growth factor pathway, which activates the PI3K family of lipid kinases, which in turn phosphorylate second messenger phosphoinositides and contribute to cell survival through suppression of apoptosis (Fig. 1). Subsequent activation of Akt and mTOR facilitates progression through the cell cycle by enhancing synthesis and stabilization of key cell cycle proteins including cyclin D1 and c-myc. The rapamycin ester CCI-779 targets mTOR function and has been found to inhibit proliferation of a panel of breast cancer cell lines, including those that were either estrogen dependent or overexpressed HER2 (63). There was a good correlation in breast cancer cells between activation of the Akt pathway, including loss of the regulatory PTEN tumor suppressor gene, and sensitivity to CCI-779, and a Phase II clinical trial in 110 patients with advanced breast cancer has been completed. Clinical efficacy was reported in patients who had failed previous chemotherapy and endocrine treatments, with objective tumor responses (including liver, lung, and chest wall metastases) seen at both low and high doses delivered by a weekly i.v. schedule (64). The most frequent toxicities were mild mucositis, rash, and temporary elevations in transaminases. An oral form of CCI-779 has now been developed, and, based on preclinical evidence for improved efficacy in combination with endocrine therapy, a Phase II trial in breast cancer of oral CCI-779 with aromatase inhibitors is planned.

Unlike conventional cytotoxic chemotherapy used in routine clinical practice, many of these new therapies represent noncytotoxic modulators that target specific identified molecular abnormalities associated with malignancy. Because a major limitation of conventional endocrine and chemotherapy in breast cancer is the emergence of drug resistance, these therapies offer a novel approach because some may circumvent resistance to conventional therapies. Most of the STIs in development are p.o. active and well tolerated, such that integration with current conventional therapies should be feasible quite quickly. However, there is much debate regarding the optimal development strategy for this new class of anticancer agent. In particular, large-scale Phase III trials are only likely to proceed on the background of good tolerability data combined with some hint of clinical activity in breast cancer. Therefore, in addition to preclinical data, many have felt that early evidence of efficacy for STIs in breast cancer is required from Phase I/II trials.

The current monotherapy Phase II studies of the FTI R115777 in 76 women (59) and the mTOR antagonist CCI-779 in 110 women with metastatic breast cancer (64) were both in populations of patients who either had endocrine-resistant breast cancer or had failed previous chemotherapy treatments. Clinical activity with objective tumor regressions was reported in both trials, together with a low toxicity profile. Based on preclinical data that demonstrated enhanced antitumor activity for R115777 and CCI-779 in combination with endocrine therapy (compared with either treatment alone), randomized Phase II trials in combination with an aromatase inhibitor are due to start soon. For the TKIs, there are little or no monotherapy data in breast cancer patients, although trials in combination with endocrine therapy are proposed, including neoadjuvant studies where biological parameters of response for the combination can be assessed more easily. If the preclinical data are predictive of what may occur in the clinic, then combined STI and endocrine therapy could prove substantially more effective at controlling tumor growth and delaying the emergence of resistance. Validation of biomarkers such as changes in cell proliferation or apoptosis, together with other pharmacodynamic end points such as ERK phosphorylation, is clearly important in this setting. Much could be learned about the potential benefit of endocrine therapy with or without STI in a neoadjuvant trial by studying changes in signal transduction pathways in surgically removed treated tumor.

In terms of optimal clinical trial design, it is important that the correct question be asked if we are to learn the true benefit of STIs for breast cancer therapy. In hormone-resistant disease, modest clinical activity as demonstrated by an objective tumor response rate of approximately 15–20% may impress some clinicians (but not all). It is important to recognize the cytostatic nature of these drugs; prolonged stable disease (i.e., for at least 6 months) may be indicative of biological activity for these new agents. However, the emerging data suggest that combination therapy with endocrine agents may prove a more effective strategy, and depending on the clinical scenario, the efficacy end points, size, duration, and cost of these clinical trials may vary. If the early promise is fulfilled, then this approach could have a substantial impact on the systemic management of breast cancer.

Dr. Mitch Dowsett: A really key issue is the frequency of these observations that we and a few other people have made of up-regulation of the HER2 and EGFR pathways in resistance to estrogen deprivation or tamoxifen. To my knowledge, there has been no confirmation in a clinical sample set thus far as HER2 and its diagnostically overexpressed level is concerned in Tam-R tissue. We don’t see any increase in HER2.

Dr. Stephen R. D. Johnston: The HER2 model is artificial in that you transfected it, but let me ask Dr. Osborne: in your acquired resistance model, where you’ve done all of the AP-1 studies in the past, have you seen much confirmatory evidence of activation of ERK1/2 and increased EGFR levels?

Dr. Kent Osborne: We do not see much of an activation of the ERKs, but we consistently see activation of c-Jun-NH₂-terminal kinase and p38. That’s why we’re so interested in that pathway. We’ve not looked at phospho-Akt yet. When you think about how a damaged cell might try to change something rapidly to survive, the easiest way for that cell to do so would be to activate the gene product through phosphorylation as opposed to increasing expression of the gene. For instance, we don’t see any change in p38 or in c-Jun-NH₂-terminal kinase levels. We do see lots of phospho-protein compared to the controls, so the cell doesn’t change gene expression. It simply changes posttranslational modifications of the proteins. It’s just recently that the antibodies have become available to look at the activated forms of the proteins, and that’s why there are no data.

Dr. Adrian Lee: But I think we’ll find there are other models showing you can get estrogen resistance and tamoxifen resistance without up-regulation of that HER2 pathway. There would always be multiple mechanisms for resistance, and it’s a case of studying multiple markers to find them.

Dr. Carlos Arteaga: In the experiment that you showed, Zarnestra did not block E₂-induced tumors but blocked estrogen-deprived or tamoxifen-treated tumors. So activation of ras is dispensable in E₂-induced tumors, but it’s indispensable or required in the estrogen-deprived state. Is that correct?

Dr. Johnston: Perhaps if you block ras with a FTI in a cell with intact survival pathways, that cell will sit there quite happily; it won’t die, but perhaps it might not grow. To me, the principle is that you need two hits, and that’s why I think Dr. Arteaga’s data with the dominant negatives through Akt and the dominant negatives to MEK point to the same thing. Blocking both pathways is much more effective than blocking either pathway alone. The issue is how much does endocrine therapy really work through modulating the insulin-like growth factor cell survival pathway? Dr. Lee has produced data to show that that’s certainly one of the major ways it may work. But we know it obviously has other effects on ER-mediated gene transcription directly. We really don’t know the answer, and what we’re trying to do in the xenograft material is actually tease that out. So we’re currently measuring phosphorylation of all of the different signals.