Crown Princess Mary Cancer Centre, Westmead Hospital, University of Sydney, Sydney, New South Wales, Australia.
For many years, tamoxifen and aromatase inhibitors have been the mainstay of treatment for ER positive breast cancer, although it has been apparent that resistance to these drugs is a limiting factor. We are now at the beginning of a new era, as drugs that block the development of endocrine resistance are becoming available. Pre-clinical research has given us an understanding of some of the molecular mechanisms of endocrine resistance, identifying new targets for drug development. Chief among these are the PI3 kinase/AKT pathway and the cell cycle control mechanism governed by cyclins and cyclin dependent kinases (CDKs). We are now in the process of integrating mTOR inhibitors, PI3kinas inhibitors and CDK 4/6 inhibitors into clinical practice on the back of clinical trial results that show they can prolong the effects of endocrine agents.
Endocrine therapy for breast cancer dates back to 1896, when the Scottish surgeon Beatson demonstrated the effectiveness of oophorectomy.1 He was testing a hypothesis derived from observations on cows in at his farm, and was right but for the wrong reasons.
However we have continued to make clinical progress without always understanding why. Older readers will remember stilboestrol (an estradiol analogue) being effective in metastatic breast cancer, and even now estradiol can induce responses after progression on aromatase inhibitors.2 One should always be a little sceptical therefore, when offered detailed diagrams and explanations of mechanisms of action. Even in the modern era, as Luen and Loi show in the last contribution to this Forum, we initially were mistaken when considering the mechanism of action of trastuzumab.
That said, the model of drug interaction with the estrogen receptor (ER) has served well as an explanation for the mechanisms of action of the selective estrogen receptor modifiers (SERMs) and the aromatase inhibitors (AIs), and the difference in their toxicities. What has been less clear is why the agents become ineffective over time.
More recently, an increased understanding of intra-cellular signalling pathways and confirmatory preclinical work has led to clinical trials of relevance to clinical practice. Benefits have been seen with drugs that manipulate two relatively separate pathways: the PI3K/AKT/mTOR survival pathway and the cell cycle regulating pathway.
It is simplistic but useful to conceptualise two complementary pathways that link the outside environment to the inner workings of the cell. Thus we have our transmembrane receptors (like HER2 and EGFR) becoming activated as a response to external triggers and initiating activation of intracellular molecules along different (although cross-talking) pathways: the ‘proliferative’ pathway (ras, raf, MAP kinase) and the ‘survival’ or anti-apoptosis pathway (PI3 kinase, AKT, mTOR).
These signal transduction pathways allow the cell to characterise with great sensitivity the nature of the external environment and fine tune what the growth (or quiescence) state of the cell is optimal. Clearly in the malignant state, these processes malfunction and the challenge for scientists and healthcare professionals is to try and interrupt overactive signals.
A large amount of preclinical data exist to show that the PI3 kinase, AKT, mTOR pathway is frequently up-regulated in cancers, and mutations (especially of PI3K genes) can be identified.3 This is true of breast cancer generally, although the incidence of mutations varies between subtypes. PI3K mutations are more frequently seen in ER positive disease (about 35%) and HER2 positive disease (about 25%) than in triple negative disease (<10%). In ER positive disease there are preclinical (and now clinical) data to show that resistance to endocrine therapies is at least in part mediated by abnormalities in the PI3K pathway. A logical development therefore is to examine the action of drugs that can inhibit this pathway. (Companion studies are examining resistance to anti-HER2 therapies.)
The first agents to be tested in ER positive breast cancer were mTOR inhibitors, and the clinical proof of principle was demonstrated in the BOLERO 2 trial.4 This was a phase III study comparing everolimus (Afinitor) plus exemestane with exemestane alone in women with metastatic disease after progression on a standard aromatase inhibitor. The results were compelling. Median progression free survival was 6.9 v 2.8 months (investigator assessment) or 10.6 v 4.1 months (central assessment) with HRs of 0.43 and 0.36 respectively, p<0.001, demonstrating that there was a 60% improvement in the time the disease remained under control.
There was however a complication, namely toxicity. Although the toxicity reported was modest (serious toxicity occurred in less than 10% of cases: grade 3 stomatitis 8%, fatigue 4%, pneumonitis 3%), two things are worthy of note. The first is that these were the usual fitter-than-average clinical trial patients, and the second is that grade 3 toxicity is actually quite significant for the patient receiving treatment. On the other hand, grade 1 toxicity may be too easily recorded (56% had some degree of stomatitis), and we often need to learn about toxicity from our own clinical practices. Anecdotally, clinician enthusiasm for using this drug has lessened somewhat since it first became available on the PBS in Australia because of concerns about toxicity.
On balance however, this is an important proof of principle – clearly inhibiting the PI3K pathway does have an impact on endocrine resistance and this mTOR inhibitor confers a meaningful clinical benefit. It is likely that careful patient selection will lead to the greatest good and the least harm.
The ability to identify clinically effective, low toxicity PI3K inhibitors continues, with no such drugs yet approved in Australia. It should be noted that PI3K biology is complex, and it may be some time before the right balance between efficacy and toxicity is found. PI3K is actually a family of kinases. They are heterodimers, with a regulatory (p85) and a catalytic (p110) domain with p110 existing as four isoforms (alpha to delta), all with potentially slightly different biological functions. In breast cancer, most mutations occur in PI3K alpha (PI3KCA), although it is still uncertain whether having a PI3K mutation is actually predictive of response to PI3K inhibitors. Additionally different drugs affect different PI3K isoforms and the clinical relevance of this is not yet clear.
Buparlisib (BKM120) for example, is a pan PI3K inhibitor, inhibiting all isoforms of p110. It clearly has activity in metastatic ER positive breast cancer, and neoadjuvant studies are ongoing. The most definitive study to date is the BELLE 2 study, the results of which were first presented in December 2015 at the San Antonio Breast Cancer Symposium.5
This phase III trial randomised over 1000 women with metastatic ER positive breast cancer and progression after aromatase inhibitors to fulvestrant (at the appropriate 500mg dose) plus or minus buparlisib. The statistical analysis of this trial is complicated to a degree because the investigators had stipulated a number of co-primary endpoints, based on well justified biological criteria. In short, when the whole cohort was examined there was a modest benefit in progression free survival (PFS) (median 6.9 v 5 months, HR 0.78, p=0.001).
Toxicity was significant, with transaminitis, hyperglycaemia and psychiatric problems of concern. As with everolimus, the actual risk of grade 3 events was not high, but this is probably not going to be an easy drug to use, with patient selection (for example not diabetic, no history of depression) critical.
However the most interesting aspect of this study was the translational subset of patients with identifiable PI3K mutations detected in circulating tumour DNA (ctDNA). In this group of 200 patients, there appeared to be a more pronounced benefit for those with a ctDNA-detected mutation, and no benefit in those with detectable ctDNA but no mutation. It is therefore possible that with improvements in technology we may be able to accurately and cheaply identify those who stand to gain a lot and may accept some toxicity, as well as those who can avoid toxicity and be offered a different drug.
There thus remains substantial interest in developing drugs that interact with this pathway in ER positive (and HER2 positive) disease. It may be for example that a more specific p110alpha inhibiting PI3K inhibitor such as alpesilib could avoid some toxicity, but maintain efficacy. Additionally, drugs that target both AKT and mTOR are in development. The biology of these approaches seems sound and initial efficacy signals are positive, but the drugs we have to date might be regarded as our own era’s stilboestrol, and we are waiting for a tamoxifen.
Meanwhile the science of cell cycle control has been evolving. Nobel prize-winning work in the 1980s demonstrated that in all cells from yeast to mammals, a system of “checkpoint” activators mediates progression through the cell cycle, ensuring orderly activity in the cell and appropriate, accurate DNA copying and mitosis. Thus cyclin dependent kinases (CDKs), periodically activated by specific cyclin proteins, determine for example if a cell will enter the G1 phase of the cell cycle or remain quiescent in G0. Ultimately all the pathways mentioned above impinge upon this central process, and to the extent that cancer is a disease of abnormal proliferation, the cyclin/CDK system is critical.
Since those early findings aberrant CDK activity in cancer cells has been described, and in particular in ER positive breast cancer cells, the cyclin D1/CDK4 partnership has been implicated. In the test tube, artificially increasing the amount of cyclin D1 (thus over-activating CDK4) leads to endocrine resistance,6 raising the possibility that impeding CDK4 activity might usefully overcome endocrine resistance in the clinical setting.
This knowledge has given rise to the era of CDK inhibitors which are now being tested in various clinical scenarios.7 The flagship studies so far have been the PALOMA 1 and PALOMA 3 trials of palbociclib, a CDK4 and CDK6 inhibitor.8,9 In the first phase II study women with untreated metastatic disease were treated with letrozole alone versus with palbociclib and in the phase III study, women treated and progressing on previous aromatase inhibitors were randomised to fulvestrant plus or minus palbociclib.
In both instances, results were quite remarkable, with PFS in the first line cohorts 20 v 10 months (HR 0.49, p=0.0004) and in the second line 9.2 v 3.8 months (HR 0.42, p<0.001). This suggests first that this drug is very active and second, that it may be equally active whether given before or after resistance has developed. This may have an impact on the sequence in which all of these ‘enablers’ might best be used, but clearly much has yet to be learned, including identifying predictors of sensitivity and resistance. Early data suggest that PI3KCA mutations do not affect the efficacy of palbociclib, but that cyclin D status may have this effect.
Finally, unlike the mTOR and PI3K inhibitors tested so far, toxicity seems to be a lot more manageable with this and other CDKIs. Neutropaenia is common but infection rare, and there is some fatigue. This may then be a drug that could be used in well women in the adjuvant setting, and an international adjuvant trial is being activated (the PALLAS study, coordinated in Australia by the ANZ Breast Cancer Trials Group).
After decades of limited progress, the ER positive space is now one of very active scientific enquiry and there are already some arrivals from the bench to the bedside – in fact make that ‘side’, as most of our patients with metastatic breast cancer now are well and active and not in bed.
Clearly there is still a lot to learn about endocrine resistance and how best to target treatment, but the first steps have been taken and we are moving into the adjuvant space. It is not unrealistic to expect that some of these drugs and many of their descendants will have a significant role in improving surgical outcomes in the neoadjuvant setting, reducing the incidence of metastatic disease by successful adjuvant use and helping to turn metastatic disease more and more into a long term chronic condition rather than a fatal one.
Lastly, we should also always be ready to challenge orthodoxy, bearing in mind that perhaps like Beatson, we could be right for the wrong reasons.