Defining the molecular response to trastuzumab, pertuzumab and combination therapy in ovarian cancer

BACKGROUND

Trastuzumab and pertuzumab target the Human Epidermal growth factor Receptor 2 (HER2). Combination therapy has been shown to provide enhanced antitumour activity; however, the downstream signalling to explain how these drugs mediate their response is not clearly understood.

METHODS

Transcriptome profiling was performed after 4 days of trastuzumab, pertuzumab and combination treatment in human ovarian cancer in vivo. Signalling pathways identified were validated and investigated in primary ovarian xenografts at the protein level and across a timeseries.

RESULTS

A greater number and variety of genes were differentially expressed by the combination of antibody therapies compared with either treatment alone. Protein levels of cyclin-dependent kinase inhibitors p21 and p27 were increased in response to both agents and further by the combination; pERK signalling was inhibited by all treatments; but only pertuzumab inhibited pAkt signalling. The expression of proliferation, apoptosis, cell division and cell-cycle markers was distinct in a panel of primary ovarian cancer xenografts, suggesting the heterogeneity of response in ovarian cancer and a need to establish predictive biomarkers.

CONCLUSION

This first comprehensive study of the molecular response to trastuzumab, pertuzumab and combined therapy in vivo highlights both common and distinct downstream effects to agents used alone or in combination, suggesting that complementary pathways may be involved.

Abstract ES8-1: Systems Biology — Can It Explain Everything?

Systems biology has entered the biological lexicon but is only just gaining acceptance within the clinical community. Some reasons for slow uptake include a misunderstanding of the opportunities (and limitations) offered by systems approaches, an absence of real-life ‘success-stories’ of clinical application, and perhaps deficiencies in cross-disciplinary training and multidisciplinary teamwork. Nevertheless, breast cancer, and its management, is complex and multi-parametric and therefore ideally suited to the systems approach. There are a number of opportunities to apply systems biology to real clinical problems, such as drug and biomarker discovery, the integration of clinical and biological data for therapeutic decision-making, and refining pharmacogenetic and pharmacogenomic algorithms for prognosis and prediction in breast cancer patients. In some of these areas there has been success, in others a more complete understanding of systems biology is required in order to tailor new technologies, mathematics, and clinical trial design to successfully utilize systems approaches. In this talk I will discuss where and how systems biology may be useful to the clinician, both now and in the future.

Citation Information: Cancer Res 2010;70(24 Suppl):Abstract nr ES8-1.

Abstract P2-09-12: Phosphoproteomic Pathway Profiling of Breast Cancers for Biomarker and Target Discovery for Personalized Therapy

Introduction: The control and activation of signaling pathways has been shown to influence response to targeted therapy (PTEN and PI3K for trastuzumab, KRAS and MAPK for cetuximab). Single tissue biomarkers (eg HER2-receptor status) are often poor surrogates for the activation state of downstream pathways such as MAPK and PI3K. We have used phosphoproteomic profiling in order to reclassify breast tumors for therapeutic response prediction.

Methods: Eleven cell lines were profiled using phosphoproteome arrays (250 targets) and reclassified using unsupervised clustering. Sensitivity to therapy predicted by pathway activation status was correlated with sensitivity to therapies in vitro measured by the AlamarBlue cell viability assay. Target specificity was assayed using in-cell western blotting. Phosphoprotein clusters were validated in 107 breast tumors by reverse phase protein arrays (RPPA), and the expression of targets validated in independent cohorts of cancers by quantitative immunofluorescence (AQUA) on tissue microarrays.

Results: Phosphoproteomic profiling revealed 3 main clusters of breast tumors with overlap with existing molecular phenotypes (luminal, basal, and mixed) and different profiles of pathway activation (Cluster A: anti-apoptotic, Cluster B: growth factor-dependent, Cluster C: cell cycle-dependent). Good to excellent correlations between pathway activation and in vitro sensitivity to targeted therapies were shown for several pathways and individual molecules within pathways, including components of NFkB, STAT, and cell cycle. These biomarkers were shown to be present in clinical cohorts by RPPA and AQUA, and related to molecular phenotype. Conclusions: Phosphoproteomic profiling can be used to reclassify tumors in order to make a priori predictions about therapeutic targeting. These clusters exist in real clinical specimens and are therefore also promising biomarkers for targeted therapy.

Citation Information: Cancer Res 2010;70(24 Suppl):Abstract nr P2-09-12.

Abstract P5-05-06: The Role of Sprouty 2 in HER2 Signaling in Breast Cancer; Decreased Expression Is Associated with Poor Outcome Including in Trastuzumab-Treated Patients

Purpose: Resistance to trastuzumab is a frequent clinical problem, in part due to the overriding activation state of the MAPK/PI3K signaling pathways. Sprouty-family proteins are negative regulators of MAPK/PI3K signaling, but their role in HER2 signaling and resistance to therapy is unknown.

Experimental design: The association of Sprouty 2 gene expression with grade, HER2-status, and survival was investigated in a meta-analysis of 1107 breast tumors from six published microarray studies. Sprouty regulation in response to HER2/HER3 signaling was studied using qRT-PCR. Changes in expression of Spry2 and feedback inhibition on trastuzumab-resistance were studied using full-length/dominant-negative transfection or chemical inhibition in SKBr3 and BT474 cell lines in cell viability assays. Finally, expression of Spry2 was measured in a cohort of 122 patients treated with trastuzumab by quantitative fluorescence microscopy (AQUA).

Results: Low Spry2 gene expression was associated with high HER2 protein expression. Spry2 was regulated through HER2/HER3 signaling as a delayed early gene, an effect reversed by treatment with the dimerization inhibitor pertuzumab. Overexpression of Spry2 in the SKBr3 trastuzumab-resistant cell line resulted in synergistic inhibition of cell viability with trastuzumab. Restitution of feedback inhibition with the PI3K-inhibitor LY294002 had a similar effect. Low Spry2 expression was associated with increased risk of death (HR = 2.28, 95% CI 1.22 - 4.26; p=0.008) in trastuzumab-treated patients, including in multivariate analysis (Cox regression analysis, p=0.002).

Conclusions: These results suggest that combinations of negative regulators of growth factor signaling may be an effective therapeutic strategy in breast cancers with deficient feedback inhibition.

Citation Information: Cancer Res 2010;70(24 Suppl):Abstract nr P5-05-06.

Abstract P6-07-02: A Systems Analysis of Oscillations in the MAPK Pathway Reveals New Mechanisms of Feed-Forward and Feedback Signaling Control

Oscillations occur in diverse areas of cellular biology, from mechanotransduction to circadian control. Their physiological, pathological, and therapeutic significance in breast cancer remains undetermined, and their existence has yet to be described in the key oncogenic signaling pathways. Mathematical modeling lends itself to detailed study of these complex dynamic phenomena. We observed damped oscillations in pERK with approximately 20 minute periodicity in MCF7 breast carcinoma cell lines after stimulation with heregulin or epidermal growth factor (EGF). In order to establish mechanisms by which these oscillations might occur, we used an ordinary differential equation-based mathematical model of MAPK/PI3K signaling to simulate oscillatory behaviour in MAPK signaling. Both modification of feed-forward (HER2 overexpression) and feedback (abrogation of early gene transcription, EGFR phosphorylation by pERK, crosstalk between pAKT and RAF, and binding of pERK with SOS) loops resulted in the emergence of sustained oscillations in phospho-ERK. When these simulations were tested experimentally, MCF7 cells stably transfected with HER2 (MCF7/HER2-18) showed sustained oscillations in pERK when stimulated with heregulin, which were abrogated by inhibition of HER2 heterodimerisation with pertuzumab. Likewise, suppression of transcriptional feedback with cycloheximide resulted in sustained oscillations in pERK of increased amplitude. We have therefore described oscillations in MAPK signalling for the first time in breast cancer cell lines. Since the duration and amplitude of these oscillations is altered both by addition of the HER2 oncogene, and by inhibition of feedback control, the functional consequences of changes in oscillations may have fundamental consequences both in oncogenesis and therapeutic efficacy.

Citation Information: Cancer Res 2010;70(24 Suppl):Abstract nr P6-07-02.

Abstract P4-02-08: STAT Signaling in Endocrine Sensitive and Resistant Breast Cancer — A Potential Target for Therapy?

The STAT (for signal transducer and activators of transcription) family of molecules regulate cell growth, survival, and differentiation as important integrators of cytokine and growth factor receptor signaling. STAT3 and STAT5 have been implicated in endocrine resistance in breast cancer. We aimed to identify new modifiers of endocrine sensitivity in breast cancer using functional proteomic profiling of the estrogen/tamoxifen-sensitive MCF-7 cell line and its partially sensitive (LCC1) and fully resistant (LCC9) variants. Using an unsupervised analysis of the expression of 125 paired phosphorylated and non-phosphorylated epitopes in key oncogene and tumour suppressor pathways on antibody arrays, STAT1, STAT3, and their phosphorylated epitopes were differentially expressed between endocrine resistant and parental controls. Differential expression was confirmed by qRT-PCR and western blotting. The STAT1 inhibitor EGCG inhibited cell number in endocrine insensitive and resistant LCC1/LCC9 lines, while the STAT3 inhibitors static and WP1066 were equally effective in endocrine-resistant and parental lines. Reduction in cell number was shown to be due to induction of apoptosis rather than inhibition of proliferation. The effects of the STAT inhibitors were additive, rather than synergistic, when tested in combination with tamoxifen in vitro. Finally, expression of STAT 1 and STAT3 were measured by quantitative immunofluorescence in 54 invasive breast cancers and expression levels of STAT1 and STAT3 transcript analysed in 550 breast cancers from publicly available gene expression datasets (GSE2990, GSE12093, GSE6532). High expression of STAT1, and low expression of STAT3, were associated with a worse distant metastasis free survival (DMFS; RR 1.42 and 0.46; log-rank p=0.042 and P<0.0001, respectively). When treatment with tamoxifen was considered, STAT1 expression was nearly predictive of DMFS (log-rank p=0.067), while STAT3 expression was predictive of DMFS (log-rank P<0.0001). These results suggest that STAT signaling may be important in endocrine resistance, and that STAT inhibitors, particularly STAT1 inhibitors, may be effective therapies in breast cancer even in the resistant setting.

Citation Information: Cancer Res 2010;70(24 Suppl):Abstract nr P4-02-08.

The mammalian target of rapamycin inhibitor everolimus (RAD001) in early breast cancer: results of a pre-operative study

mTOR plays a key role in tumor cell cycle control, proliferation, and survival. RAD001 (everolimus) is a novel macrolide that inhibits mTOR and thus downstream signaling pathways. 31 post-menopausal women with early breast cancer were given 5 mg RAD001 once daily for 14 days prior to surgery. Biopsies were taken at diagnosis and at surgery (post 14 days of treatment) and assessed for immunohistochemical changes in proliferation (Ki67), apoptosis (active caspase-3), p-AKT (s473), p-S6 (s235/236 and s240/244), p-mTOR (s2448), ER, and PR. Five patients did not complete the 2-week treatment period due to adverse events. All adverse events were grade 1 or 2 (NCIC-CTC scale). RAD001 treatment significantly decreased proliferation (geometric mean reduction 74% from baseline (p = 0.019)), particularly in HER-2 positive tumors. High Ki67 pre-treatment correlated with reduction in Ki67, an increase in apoptosis, a reduction in p-AKT (cytoplasmic) and reduction in p-mTOR following treatment. Nuclear expression of p-AKT was significantly reduced with treatment. Tumors that had a reduction in Ki67 with treatment exhibited a significant reduction in cytoplasmic p-AKT. p-S6 staining was significantly reduced independently of Ki67 (p < 0.001 for two sites of phosphorylation). RAD001 5 mg/daily is safe and tolerable in postmenopausal early breast cancer patients and inhibits the mTOR pathway and its downstream effectors, significantly reducing tumor cell proliferation. Tumors with high Ki67, high p-AKT, and HER-2 positivity may be more responsive to mTOR inhibition with RAD001. This is the first study to report results of RAD001 5 mg as a single agent in early breast cancer.

A gene on the HER2 amplicon, C35, is an oncogene in breast cancer whose actions are prevented by inhibition of Syk

Background:

C35 is a 12 kDa membrane-anchored protein endogenously over-expressed in many invasive breast cancers. C35 (C17orf37) is located on the HER2 amplicon, between HER2 and GRB7. The function of over-expressed C35 in invasive breast cancer is unknown.

Methods:

Tissue microarrays containing 122 primary human breast cancer specimens were used to examine the association of C35 with HER2 expression. Cell lines over-expressing C35 were generated and tested for evidence of cell transformation in vitro.

Results:

In primary breast cancers high levels of C35 mRNA expression were associated with HER2 gene amplification. High levels of C35 protein expression were associated with hallmarks of transformation, such as, colony growth in soft agar, invasion into collagen matrix and formation of large acinar structures in three-dimensional (3D) cell cultures. The transformed phenotype was also associated with characteristics of epithelial to mesenchymal transition, such as adoption of spindle cell morphology and down-regulation of epithelial markers, such as E-cadherin and keratin-8. Furthermore, C35-induced transformation in 3D cell cultures was dependent on Syk kinase, a downstream mediator of signalling from the immunoreceptor tyrosine-based activation motif, which is present in C35.

Conclusion:

C35 functions as an oncogene in breast cancer cell lines. Drug targeting of C35 or Syk kinase might be helpful in treating a subset of patients with HER2-amplified breast cancers.

miRNA Profiling of Endocrine-Resistant Breast Tumours

RationalemiRNAs are a relatively recently discovered class of molecules with major implications in cellular biology, including tumorigenesis; yet little is known about their possible involvement in the development of endocrine resistance in breast cancer. The aim of this study was to identify sub-groups of endocrine-resistant tumours with distinctive miRNA profiles.Methods and ResultsFrozen biopsies were obtained from a series of 51 breast cancers, which were either progressing primary tumours, local recurrences or contra-lateral breast lesions developed while patients received endocrine treatment. Hormonal treatment included aromatase inhibitors or tamoxifen, most of the patients reached post-menopausal age, miRNA profiles were obtained using Illumina DASL® Sentrix Array Matrix.Un-supervised hierarchical clustering split 51 studied endocrine resistant tumours into two major groups, distinctive in their miRNA profiles. Several of miRNAs discriminating between the clusters have already been implemented in breast cancer (mir7, mir10a, mir205, mir206, mir 210). Most of the other miRNAs associated with the main tumour clusters are yet to be studied in relation to breast cancer biology (e.g. mir192, mir625, mir941, mir145 and mir512 among others). The miRNA-based tumour clusters were not clearly associated with a specific treatment or type of resistance. The miRNAs most clearly able to discriminate between tumours resistant to AIs and Tamoxifen included mir186, mir196a, mir196b and miR-594:9.1 (p&lt;0.05, fold change &gt;1.2).DiscussionThe results show heterogeneity of miRNA profiles in endocrine resistant breast cancers, for the first time suggesting that endocrine resistant tumours can be split into two major molecular sub-groups basing on their miRNA profiles. The results suggest that (i) miRNA-mediated gene expression regulation may be involved in the mechanisms of endocrine resistance and (ii) miRNA profiling allows different molecular sub-classes of endocrine-resistant tumours to be distinguished.

Citation Information: Cancer Res 2009;69(24 Suppl):Abstract nr 5130.

Gene Expression Profiles of Endocrine Resistant Breast Tumours

AimTo identify molecular sub-classes of endocrine-resistant breast tumours.MethodsWhole genome expression profiles of 55 endocrine-resistant tumours were obtained using Illumina HT-12 BeadChip arrays. Endocrine resistance was defined as either (i) tumour increased in size during hormonal treatment or (ii) appearance of local recurrence or (iii) contra-lateral breast lesion (metastases or new cancers) during hormonal treatment. Most of the treated patients were post-menopausal. Hormonal treatment included either aromatase inhibitors (letrozole, anastrazole, arimidex) or tamoxifen.ResultsUnsupervised hierarchical clustering using 500 most variable genes split tumours into three major clusters of approximately equal size. The genes discriminating between these clusters included oestrogen receptor- and proliferation- associated set of genes (ESR1, TFF1, TFF3, FOXA1, AGR2, AGR3, XBP1, GATA3, PRAD6B, MYB, CCND1), and stroma-related genes (DCN, COL1A1, COL1A2, COL3A1, COL5A2). Initial analysis of the studied tumours was unable to identify genes associated with clinical features. Sub-division of the dataset may be required to identify these genes more clearly. Genes that tended to separate between tumours resistant to AI and Tamoxifen included MYB, GPR137B, RNF19A, FRMD6, ZNF217 and RPS12. More detailed analysis of the data is underway.DiscussionUnderstanding and overcoming of endocrine resistance will improve treatment results for the majority of breast cancer patients. This abstract reports a new gene expression dataset to compare molecular features of biopsies collected from tumours growing while on endocrine treatment. The initial analysis demonstrates diversity of the resistance mechanisms, enhancing development of individual approaches to endocrine-resistant tumours.

Citation Information: Cancer Res 2009;69(24 Suppl):Abstract nr 5132.

Rapid screening of tissue microarrays for Her-2 FISH testing is a safe, accurate, efficient and economic method enabling provision of an entirely FISH-based Her-2 testing service

Abstract #2070

Introduction: FISH (fluorescence in situ hybridisation) testing is the gold standard method for Her-2 status assessment in breast cancer yet is only employed in about 30% of tests carried out partly due to cost and labour considerations. Previously we described TMA-based (tissue microarray) testing1 to reducing cost constraints and now describe a rapid screening approach to eliminate unequivocally negative and positive cases and reduce time spent testing.&#x2028; Materials and methods: We examined TMAs (6 cores per case) from 89 cases of invasive breast cancer comparing formal FISH scoring with a rapid screening method using systematic examination of the entire core at x100 magnification without formal counting. Each core was screened by two observers and results recorded for each core as positive (P), equivocal (E) or negative (N). Formal counting and rapid screening methods were timed.&#x2028; Results: The results are given in Table 1. Where both observers agreed that every core in a particular case was negative (43 cases) or positive (10 cases) the predictive value was 100%. The three “Mixed” Groups were constituted as follows: (1) a predominance of positive observations, (2) a predominance of negative observations and (3) a predominance of equivocal observations. Comparison of the FISH scores for the five observational groups using Student's T Test showed significant or highly significant differences (Range: p=0.02-p=0.0001). Screening each 6 TMA core case at x100 magnification took 2 minutes and formal counting of each case took 12 minutes.&#x2028; &#x2028; Conclusion: Using stringent criteria we show that rapid screening of TMAs for routine Her-2 FISH testing provides a safe, economical and time-efficient method for the assessment of Her-2 status in breast cancer. The technique ensures that all patients receive gold standard testing. 65% of Her-2 negative cases can be rapid screened and eliminated from formal counting with 100% confidence. Unanimous positive cases can also be appropriately categorised. The remaining cases require formal counting which takes no longer than is normally the case with whole section techniques. The TMAs also provide a valuable resource for future analyses.&#x2028; Reference: 1. Graham AD, Faratian D, Rae F, Thomas J St J. Tissue microarray technology in the routine assessment of Her-2 status in invasive breast cancer: a prospective study of the use of immunohistochemistry and fluorescence in situ hybridization. Histopathology 2008; 52: 847-855.

Citation Information: Cancer Res 2009;69(2 Suppl):Abstract nr 2070.

Effects of letrozole and anastrozole on low ER expressing invasive breast carcinomas: results from a randomised trial

Abstract #1065

Introduction:&#x2028; Changes in proliferation as measured by Ki67 occur within 14 days of starting treatment with an aromatase inhibitor and have been shown to be predictors of long term outcome. The biological effect of different aromatase inhibitors on low ER expressing cancers has not been studied. This study compared changes in proliferation in breast cancers expressing low levels of ER (Allred score 2-5) following 14 days of treatment with anastrozole or letrozole.&#x2028; Materials and methods:&#x2028; Thirty six postmenopausal women with invasive breast cancer, classified as estrogen receptor (ER) poor by an Allred score of 2-5 on initial core biopsy, were enrolled into a randomised pre-operative trial of 14 days treatment with either 2.5 mg of letrozole or 1 mg of anastrozole. Paired biopsy tissue was available for all patients: [15 who received anastrozole and 21 letrozole] with sufficient invasive cancer present for analysis. Assessment included ER (Allred score) and proliferation (% tumor cells Ki67 positive) by immunohistochemistry. Results are presented as means (SEM) and medians (range). Due to the non-normality of the data, pre- and post-treatment Ki67 values were transformed on the log scale. Change between pre- and post- treatment scores was calculated as the difference in the logged scores; analysis is by t-test and Fisher's Exact Test. All tests are two-sided. Results were back-transformed in order to preserve the original scale.&#x2028; Results:&#x2028; Proliferation measured by Ki67:&#x2028; Anastrozole reduced tumour cell proliferation from baseline in 10/15 cancers from a mean of 31.0% (7.1) to 20.8% (6.1). The median and mean reductions in Ki67 were 37% (range -94.0% to 92.6%) and 38% (13.9) from baseline. This equates to an average fall of 2.6 times from baseline, p=0.006, by analysis of the transformed data.&#x2028; Letrozole reduced proliferation from baseline in 17/21 cancers from a mean of 39.0% (7.3) to 25.3% (6.4). The median and mean reductions were 67% (range -115.6% to 99.6%) and 46% (10.6) from baseline. This equates to an average fall of 3.13 times from baseline, p=0.0002, by analysis of the transformed data.&#x2028; There was no evidence of a significant difference between drugs in reduction of proliferation, the number of cases showing a fall in proliferation, or the absolute Ki67 index at 14 days. Only 3/15 (anastrozole) and 2/21 (letrozole) tumours had post-treatment absolute Ki67 indices of ≤1%. There were no significant differences in changes in proliferation between tumours with ER score 2-3 vs score 4-5.&#x2028; Conclusions:&#x2028; 1. Both anastrozole and letrozole significantly reduced proliferation in invasive cancers with low ER expression.&#x2028; 2. There were no evident significant differences between drugs.&#x2028; 3. These data suggest that aromatase inhibitors are of value even in low-ER expressing cancers, either on their own or perhaps as part of combined therapy.&#x2028; 4. Analysis of more patients and of biological response according to molecular phenotype is underway.

Citation Information: Cancer Res 2009;69(2 Suppl):Abstract nr 1065.

A review of the pathology and management of uterine papillary serous carcinoma and correlation with outcome

Uterine papillary serous carcinoma (UPSC) accounts for 10% of endometrial carcinomas but a higher proportion of deaths due to its aggressive nature and poor response to chemotherapy and radiotherapy. In order to add to the knowledge of UPSC in the literature and to review our local practices, we examined the pathology, medical records, and management of all cases of UPSC (67 patients) treated in South East Scotland over a 10-year period and also evaluated the prognostic significance of the percentage of UPSC in endometrial pipelle and hysterectomy specimens. Although only 63% of initial diagnostic biopsies were reported to contain UPSC, rereview of the cases revealed UPSC in 98.5% of the preoperative biopsies. The percentage of UPSC in the tumors did not affect the outcome. Stage, positive omentum, and treatment with external-beam +/- intracavitary radiotherapy were significantly correlated with overall survival and progression-free survival by univariate analysis, but only stage (P < 0.01) was correlated with outcome on multivariate analysis. Chemotherapy did not affect outcome. UPSC may be difficult to diagnose in preoperative biopsies, particularly when present as part of a mixed tumor. Even a small percentage of UPSC in a diagnostic biopsy or hysterectomy specimen is correlated with a poor prognosis. This study emphasizes the need of a cooperative, prospective study on this distinct uterine carcinoma.

A review of the pathology and management of uterine papillary serous carcinoma and correlation with outcome

Uterine papillary serous carcinoma (UPSC) accounts for 10% of endometrial carcinomas but a higher proportion of deaths due to its aggressive nature and poor response to chemotherapy and radiotherapy. In order to add to the knowledge of UPSC in the literature and to review our local practices, we examined the pathology, medical records, and management of all cases of UPSC (67 patients) treated in South East Scotland over a 10-year period and also evaluated the prognostic significance of the percentage of UPSC in endometrial pipelle and hysterectomy specimens. Although only 63% of initial diagnostic biopsies were reported to contain UPSC, rereview of the cases revealed UPSC in 98.5% of the preoperative biopsies. The percentage of UPSC in the tumors did not affect the outcome. Stage, positive omentum, and treatment with external-beam +/− intracavitary radiotherapy were significantly correlated with overall survival and progression-free survival by univariate analysis, but only stage (P < 0.01) was correlated with outcome on multivariate analysis. Chemotherapy did not affect outcome. UPSC may be difficult to diagnose in preoperative biopsies, particularly when present as part of a mixed tumor. Even a small percentage of UPSC in a diagnostic biopsy or hysterectomy specimen is correlated with a poor prognosis. This study emphasizes the need of a cooperative, prospective study on this distinct uterine carcinoma.