Abstract P3-10-07: A 3-gene DNA methylation signature fails to predict response to bevacizumab in metastatic breast cancer patients treated within the TANIA phase III trial

Background: Biomarkers predicting response to bevacizumab containing therapy in metastatic breast cancer (MBC) are of urgent need. In a retrospective single-institution analysis we have previously shown that a 3-gene methylation signature (MLH1,POLKand TMBIM6) could discriminate between responders and non-responders to a bevacizumab-based therapy in two independent cohorts of patients with MBC with an AUC of 0.94 and 0.86, respectively (Gampenrieder SP et al. Theranostics. 2018. 8(8):2278-2288). Here, we present the validation of these findings within the prospective phase III trial TANIA (Vrdoljak E et al. Ann Oncol. 2016. 27(11):2046-52) randomizing 494 patients with HER2-negative MBC to chemotherapy plus bevacizumab or chemotherapy alone for two consecutive treatment lines (second- and third-line). All patients had already received bevacizumab-containing therapy in the first-line setting.

Patients and methods: DNA isolated from archival FFPE tumor samples was available from 200 patients consenting to optional translational research within the TANIA trial. Out of these, 176 samples were collected prior to first-line bevacizumab therapy and were analyzed retrospectively. Sufficient DNA for methylation analysis was available from 124 patients: 64 treated with chemotherapy plus bevacizumab and 60 treated with chemotherapy alone. All samples were isolated from the primary tumor. Quantitative methylation analysis was performed by pyrosequencing on the PyroMark Q24 Advanced System (Qiagen). PFS and OS analyses were performed in both study arms comparing “predicted responders” (PRED_R) versus “predicted non-responders” (PRED_NR) based either on median dichotomization or according to the cutoffs for individual CpG and the combined 3-CpG methylation logistic regression model.

Results:Out of the 124 evaluable patients, 32 (25.8%) were classified as PRED_R and 92 as RED_NR by the 3-gene methylation signature. PRED_R did not have a significantly different second-line PFS (HR 0.95, 95%CI 0.57-1.57; P = 0.84) or OS (HR 0.91, 95%CI 0.51-1.60; P = 0.73) when treated in the bevacizumab-containing study arm compared to PRED_NR. In addition, PRED_R did not show a longer PFS when treated with bevacizumab compared to PRED_R treated with chemotherapy alone (HR 0.95, 95%CI 0.59-1.54; P = 0.83). Furthermore, there was no difference in third-line PFS and the combination of second- and third-line PFS between PRED_R and PRED-NR in the bevacizumab arm. In the control arm, PRED_NR showed a statistically significant shorter PFS compared to PRED_R (HR 0.50, 95%CI 0.22-0.77; P = 0.006), but not OS (HR 0.95, 95%CI 0.51-1.77; P = 0.86).

Conclusion: Our 3-gene methylation signature was not confirmed as predictive biomarker for bevacizumab efficacy in metastatic breast cancer.

(This research project was partially supported by ROCHE Austria GmbH)

Citation Format: Gampenrieder SP, Angela R, Rinnerthaler G, Hackl H, Steiner M, Pulverer W, Weinhaeusel A, Klinglmayr E, Karl T, Ilic S, Hufnagl C, Hauser-Kronberger C, Egle A, Greil R. A 3-gene DNA methylation signature fails to predict response to bevacizumab in metastatic breast cancer patients treated within the TANIA phase III trial [abstract]. In: Proceedings of the 2018 San Antonio Breast Cancer Symposium; 2018 Dec 4-8; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2019;79(4 Suppl):Abstract nr P3-10-07.

Abstract P6-07-10: DNA methylation signature predicting bevacizumab efficacy in metastatic breast cancer

Background: Biomarkers predicting response to bevacizumab containing therapy in breast cancer are still missing. Since epigenetic modifications can contribute to an aberrant regulation of angiogenesis and to treatment resistance, we investigated the influence of DNA methylation on bevacizumab efficacy.

Patients and methods: A genome-wide methylation profiling using the Illumina Infinium HumanMethylation450 BeadChip was performed in archival FFPE specimen of 36 patients with HER2-negative metastatic breast cancer treated with chemotherapy in combination with bevacizumab as first-line therapy (learning set). Based on objective response and progression-free survival (PFS) and considering ER expression, patients were divided in responders (R) and non-responders (NR). Differentially methylated gene loci (methylation variable position = MVP) and differentially methylated regions (DMR) between R and NR were identified. Only significant sites with a strong change in methylation levels (Δβ>0.15 or Δβ<-0.15), an area under the curve of at least 0.85 by logistic regression analysis, and/or sites in proximity to genes functionally involved in angiogenesis and carcinogenesis were selected and further validated. Validation was performed in 81 bevacizumab treated breast cancer patients (validation set) and in 15 patients treated with chemotherapy only (control set) using targeted bisulfite sequencing. Methylated gene loci were considered predictive if there was a significant association with outcome (PFS) in the validation set but not in the control set using Spearman rank correlation, Cox regression, and logrank test.

Results: In the learning set 435 MVPs (p<0.001) and 144 DMRs (adjusted combined p<0.0001 with at least 3 significant MVPs p<0.05 in the same region) showed significantly different methylation between R and NR. In R 152 sites were hypermethylated with a median change of methylation (Δβ) of 0.13; 283 sites were hypomethylated with a median Δβ of -0.09. A methylation signature of 48 genes was specified allowing a good separation between responders and non-responders (odds ratio=101, p<0.0001; data presented at the 38th SABCS 2015, P3-07-43). At least one methylated cytosine in close proximity to 24 of these genes showed a significant association with PFS in the validation set but was not (or in the other direction) in the control set. Based on these data the methylation signature predicting long lasting response to bevacizumab could be reduced to 24 genes including several genes involved in angiogenesis and carcinogenesis, respectively (FLT1 also known as VEGFR-1, MLH1, GNAS, APC, DKK3, WNT2B and COL4A1).

Conclusion: A 24-gene methylation signature can reproducible discriminate between responders and non-responders to a bevacizumab-based therapy in breast cancer and could help to identify patients deriving greater benefit from anti-VEGF agents.

Citation Format: Gampenrieder SP, Rinnerthaler G, Pulverer W, Weinhäusel A, Hufnagl C, Hackl H, Hauser-Kronberger C, Mlineritsch B, Greil R. DNA methylation signature predicting bevacizumab efficacy in metastatic breast cancer [abstract]. In: Proceedings of the 2016 San Antonio Breast Cancer Symposium; 2016 Dec 6-10; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2017;77(4 Suppl):Abstract nr P6-07-10.

Abstract P3-07-43: DNA methylation patterns correlating with outcome in patients treated with first-line bevacizumab for metastatic breast cancer

Background: Biomarkers predicting response to bevacizumab containing therapy in breast cancer are of urgent need for a personalized treatment approach. DNA methylation is involved in regulation of angiogenesis and development of treatment resistance and could therefore provide predictive information for bevacizumab efficacy.

Patients and methods: A genome-wide methylation profiling using the Illumina Infinium HumanMethylation450 BeadChip was performed in FFPE specimen of 36 patients with HER2-negative metastatic breast cancer treated with chemotherapy in combination with bevacizumab as first-line therapy (discovery set). Twenty-eight (78%) samples came from primary tumor and 8 (22%) from metastasis (2 lung mets., 1 pleural met., 1 liver met., 2 soft tissue mets., 1 ovarial met, 1 bone marrow infiltration). Based on progression-free survival (PFS) and breast cancer subtype (ER+ vs. triple-negative) patients were divided in responder (R) and non-responder (NR). By biostatistical methods differences in the methylation pattern between R and NR were detected. The 48 most interesting genes (e.g. because of there involvement in angiogenesis or carcinogenesis) showing a differential methylation status between R and NR are currently validated in two further metastatic breast cancer cohorts treated with (main set) and without bevacizumab (validation set), respectively. These validated results will be presented at the meeting.

Citation Format: Gampenrieder SP, Rinnerthaler G, Weinhäusel A, Pulverer W, Hugnagl C, Hackl H, Romeder F, Monzo Fuentes C, Hauser-Kronberger C, Mlineritsch B, Greil R. DNA methylation patterns correlating with outcome in patients treated with first-line bevacizumab for metastatic breast cancer. [abstract]. In: Proceedings of the Thirty-Eighth Annual CTRC-AACR San Antonio Breast Cancer Symposium: 2015 Dec 8-12; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2016;76(4 Suppl):Abstract nr P3-07-43.