Correction to: Relationship between tumor biomarkers and efficacy in MARIANNE, a phase III study of trastuzumab emtansine ± pertuzumab versus trastuzumab plus taxane in HER2-positive advanced breast cancer.

Following publication of the original article [1], the authors reported the following errors in the article.

Relationship between tumor biomarkers and efficacy in MARIANNE, a phase III study of trastuzumab emtansine ± pertuzumab versus trastuzumab plus taxane in HER2-positive advanced breast cancer.

BACKGROUND: The phase III EMILIA and TH3RESA trials demonstrated clinical benefits of trastuzumab emtansine (T-DM1) therapy in patients with previously treated HER2-positive metastatic breast cancer (MBC). Data from these and other trials showed that T-DM1-associated survival benefits were observed across biomarker subgroups tested in these trials. Prespecified, exploratory analyses of the phase III MARIANNE study examined the effects of HER2-related biomarkers on PFS in patients administered T-DM1 in the first-line MBC setting. METHODS: In MARIANNE, patients with previously untreated HER2-positive MBC were randomized (1:1:1) to trastuzumab plus taxane, T-DM1 plus placebo, or T-DM1 plus pertuzumab. Biomarker subgroups included HER2 and HER3 mRNA expression levels (≤median vs. >median), HER2 staining intensity (IHC 3+ vs. 2+ vs. 0/1+), PIK3CA status (mutated vs. non-mutated), PTEN H-score (≤median vs. >median), and PTEN protein expression level (0 vs. 1+ vs. 2+ vs. 3+ vs. 4+). PFS was analyzed descriptively for each subgroup using Kaplan-Meier methodology. Additional exploratory post-hoc analyses evaluated the effects of HER2 heterogeneity. Multivariate analyses were also performed. RESULTS: Median PFS was numerically longer for patients with HER2 mRNA levels >median versus ≤median across treatment arms. In general, there were no predictive biomarkers of benefit for either T-DM1 treatment arm; most hazard ratios were close to 1 with wide confidence intervals that included the value 1. Focal HER2 expression (IHC 3+ or IHC 2+) was present in 3.8% of patients and was associated with numerically shorter PFS in the T-DM1-containing treatment arms versus trastuzumab plus taxane. Compared with non-mutated PIK3CA, mutated PIK3CA was associated with numerically shorter median PFS across treatment groups. Post-hoc multivariate analysis showed HER2 mRNA expression and mutated PIK3CA were prognostic for PFS (P ≤ 0.001 for both biomarkers). CONCLUSIONS: In MARIANNE, biomarkers related to the HER2 pathway did not have predictive value for PFS when comparing T-DM1 (with or without pertuzumab) with trastuzumab plus taxane. However, HER2 mRNA level and PIK3CA mutation status showed prognostic value. Evaluation of other potential biomarkers, including immune markers, is ongoing. TRIAL REGISTRATION: Registration number: NCT01120184 . Date of registration: April 28, 2010 (registered prospectively).

Trastuzumab Emtansine With or Without Pertuzumab Versus Trastuzumab Plus Taxane for Human Epidermal Growth Factor Receptor 2-Positive, Advanced Breast Cancer: Primary Results From the Phase III MARIANNE Study.

Purpose Trastuzumab and pertuzumab are human epidermal growth factor receptor 2 (HER2) -targeted monoclonal antibodies, and trastuzumab emtansine (T-DM1) is an antibody-drug conjugate that combines the properties of trastuzumab with the cytotoxic activity of DM1. T-DM1 demonstrated encouraging efficacy and safety in a phase II study of patients with previously untreated HER2-positive metastatic breast cancer. Combination T-DM1 and pertuzumab showed synergistic activity in cell culture models and had an acceptable safety profile in a phase Ib and II study. Methods In the MARIANNE study, 1,095 patients with centrally assessed, HER2-positive, advanced breast cancer and no prior therapy for advanced disease were randomly assigned 1:1:1 to control (trastuzumab plus taxane), T-DM1 plus placebo, hereafter T-DM1, or T-DM1 plus pertuzumab at standard doses. Primary end point was progression-free survival (PFS), as assessed by independent review. Results T-DM1 and T-DM1 plus pertuzumab showed noninferior PFS compared with trastuzumab plus taxane (median PFS: 13.7 months with trastuzumab plus taxane, 14.1 months with T-DM1, and 15.2 months with T-DM1 plus pertuzumab). Neither experimental arm showed PFS superiority to trastuzumab plus taxane. Response rate was 67.9% in patients who were treated with trastuzumab plus taxane, 59.7% with T-DM1, and 64.2% with T-DM1 plus pertuzumab; median response duration was 12.5 months, 20.7 months, and 21.2 months, respectively. The incidence of grade ≥ 3 adverse events was numerically higher in the control arm (54.1%) versus the T-DM1 arm (45.4%) and T-DM1 plus pertuzumab arm (46.2%). Numerically fewer patients discontinued treatment because of adverse events in the T-DM1 arms, and health-related quality of life was maintained for longer in the T-DM1 arms. Conclusion T-DM1 showed noninferior, but not superior, efficacy and better tolerability than did taxane plus trastuzumab for first-line treatment of HER2-positive, advanced breast cancer.

Transcriptomics hit the target: Monitoring of ligand-activated and stress response pathways for chemical testing.

High content omic methods provide a deep insight into cellular events occurring upon chemical exposure of a cell population or tissue. However, this improvement in analytic precision is not yet matched by a thorough understanding of molecular mechanisms that would allow an optimal interpretation of these biological changes. For transcriptomics (TCX), one type of molecular effects that can be assessed already is the modulation of the transcriptional activity of a transcription factor (TF). As more ChIP-seq datasets reporting genes specifically bound by a TF become publicly available for mining, the generation of target gene lists of TFs of toxicological relevance becomes possible, based on actual protein-DNA interaction and modulation of gene expression. In this study, we generated target gene signatures for Nrf2, ATF4, XBP1, p53, HIF1a, AhR and PPAR gamma and tracked TF modulation in a large collection of in vitro TCX datasets from renal and hepatic cell models exposed to clinical nephro- and hepato-toxins. The result is a global monitoring of TF modulation with great promise as a mechanistically based tool for chemical hazard identification.

Application of RPTEC/TERT1 cells for investigation of repeat dose nephrotoxicity: A transcriptomic study.

The kidney is a major target organ for toxicity. Incidence of chronic kidney disease (CKD) is increasing at an alarming rate due to factors such as increasing population age and increased prevalence of heart disease and diabetes. There is a major effort ongoing to develop superior predictive models of renal injury and early renal biomarkers that can predict onset of CKD. In the EU FP7 funded project, Predict-IV, we investigated the human renal proximal tubule cells line, RPTEC/TERT1 for their applicability to long term nephrotoxic mechanistic studies. To this end, we used a tiered strategy to optimise dosing regimes for 9 nephrotoxins. Our final testing protocol utilised differentiated RPTEC/TERT1 cells cultured on filter inserts treated with compounds at both the apical and basolateral side, at concentrations not exceeding IC10, for 14 days in a 24 h repeat application. Transepithelial electrical resistance and supernatant lactate were measured over the duration of the experiments and genome wide transcriptomic profiles were assayed at day 1, 3 and 14. The effect of hypoxia was investigated for a subset of compounds. The transcriptomic data were analysed to investigate compound-specific effects, global responses and mechanistically informative signatures. In addition, several potential clinically useful renal injury biomarkers were identified.

Synergistic combination of valproic acid and oncolytic parvovirus H-1PV as a potential therapy against cervical and pancreatic carcinomas.

The rat parvovirus H-1PV has oncolytic and tumour-suppressive properties potentially exploitable in cancer therapy. This possibility is being explored and results are encouraging, but it is necessary to improve the oncotoxicity of the virus. Here we show that this can be achieved by co-treating cancer cells with H-1PV and histone deacetylase inhibitors (HDACIs) such as valproic acid (VPA). We demonstrate that these agents act synergistically to kill a range of human cervical carcinoma and pancreatic carcinoma cell lines by inducing oxidative stress, DNA damage and apoptosis. Strikingly, in rat and mouse xenograft models, H-1PV/VPA co-treatment strongly inhibits tumour growth promoting complete tumour remission in all co-treated animals. At the molecular level, we found acetylation of the parvovirus nonstructural protein NS1 at residues K85 and K257 to modulate NS1-mediated transcription and cytotoxicity, both of which are enhanced by VPA treatment. These results warrant clinical evaluation of H-1PV/VPA co-treatment against cervical and pancreatic ductal carcinomas.

Transcriptomic hepatotoxicity signature of chlorpromazine after short- and long-term exposure in primary human sandwich cultures.

Drug-induced liver injury is the most frequent reason for market withdrawal of approved drugs, and is difficult to predict in animal models. Here, we analyzed transcriptomic data derived from short- and long-term cultured primary human hepatocytes (PHH) exposed to the well known human hepatotoxin chlorpromazine (CPZ). Samples were collected from five PHH cultures after short-term (1 and 3 days) and long-term (14 days) repeat daily treatment with 0.1 or 0.2 µM CPZ, corresponding to C(max). Two PHH cultures were additionally treated with 1 µM CPZ, and the three others with 0.02 µM CPZ. Differences in the total number of gene changes were seen between donors and throughout treatment. Specific transcriptomic hepatotoxicity signatures were created for CPZ and consisted of inflammation/hepatitis, cholestasis, and liver proliferation in all five donors, as well as fibrosis and steatosis, which were observed in four of five donors. Necrosis was present in three of five donors, and an indicative signature of cirrhosis was observed after long-term 14-day repeat treatment, also in three of five donors. The inter-donor variability in the inflammatory response to CPZ treatment was associated with variability in the strength of the response of the transcriptomic hepatotoxicity signatures, suggesting that features of inflammation could be related to the idiosyncratic hepatotoxic effects of CPZ in humans.

Design of a testing strategy using non-animal based test methods: lessons learnt from the ACuteTox project.

In the framework of toxicology, a testing strategy can be viewed as a series of steps which are taken to come to a final prediction about a characteristic of a compound under study. The testing strategy is performed as a single-step procedure, usually called a test battery, using simultaneously all information collected on different endpoints, or as tiered approach in which a decision tree is followed. Design of a testing strategy involves statistical considerations, such as the development of a statistical prediction model. During the EU FP6 ACuteTox project, several prediction models were proposed on the basis of statistical classification algorithms which we illustrate here. The final choice of testing strategies was not based on statistical considerations alone. However, without thorough statistical evaluations a testing strategy cannot be identified. We present here a number of observations made from the statistical viewpoint which relate to the development of testing strategies. The points we make were derived from problems we had to deal with during the evaluation of this large research project. A central issue during the development of a prediction model is the danger of overfitting. Procedures are presented to deal with this challenge.

Evaluation of aggregating brain cell cultures for the detection of acute organ-specific toxicity.

As part of the ACuteTox project aimed at the development of non-animal testing strategies for predicting human acute oral toxicity, aggregating brain cell cultures (AGGR) were examined for their capability to detect organ-specific toxicity. Previous multicenter evaluations of in vitro cytotoxicity showed that some 20% of the tested chemicals exhibited significantly lower in vitro toxicity as expected from in vivo toxicity data. This was supposed to be due to toxicity at supracellular (organ or system) levels. To examine the capability of AGGR to alert for potential organ-specific toxicants, concentration-response studies were carried out in AGGR for 86 chemicals, taking as endpoints the mRNA expression levels of four selected genes. The lowest observed effect concentration (LOEC) determined for each chemical was compared with the IC20 reported for the 3T3/NRU cytotoxicity assay. A LOEC lower than IC20 by at least a factor of 5 was taken to alert for organ-specific toxicity. The results showed that the frequency of alerts increased with the level of toxicity observed in AGGR. Among the chemicals identified as alert were many compounds known for their organ-specific toxicity. These findings suggest that AGGR are suitable for the detection of organ-specific toxicity and that they could, in conjunction with the 3T3/NRU cytotoxicity assay, improve the predictive capacity of in vitro toxicity testing.

Data management in large-scale collaborative toxicity studies: how to file experimental data for automated statistical analysis.

High-throughput screening approaches are carried out for the toxicity assessment of a large number of chemical compounds. In such large-scale in vitro toxicity studies several hundred or thousand concentration-response experiments are conducted. The automated evaluation of concentration-response data using statistical analysis scripts saves time and yields more consistent results in comparison to data analysis performed by the use of menu-driven statistical software. Automated statistical analysis requires that concentration-response data are available in a standardised data format across all compounds. To obtain consistent data formats, a standardised data management workflow must be established, including guidelines for data storage, data handling and data extraction. In this paper two procedures for data management within large-scale toxicological projects are proposed. Both procedures are based on Microsoft Excel files as the researcher's primary data format and use a computer programme to automate the handling of data files. The first procedure assumes that data collection has not yet started whereas the second procedure can be used when data files already exist. Successful implementation of the two approaches into the European project ACuteTox is illustrated.