SPAG5 as a prognostic biomarker and chemotherapy sensitivity predictor in breast cancer: a retrospective, integrated genomic, transcriptomic, and protein analysis

Tools

Abdel-Fatah, T.M.A., Agarwal, D., Liu, D.X., Russell, R., Rueda, O.M., Liu, K., Xu, B., Moseley, P.M., Green, A.R., Pockley, A.G. ORCID: 0000-0001-9593-6431, Rees, R.C. ORCID: 0000-0002-4574-4746, Caldas, C., Ellis, I.O., Ball, G.R. ORCID: 0000-0001-5828-7129 and Chan, S.Y.T., 2016. SPAG5 as a prognostic biomarker and chemotherapy sensitivity predictor in breast cancer: a retrospective, integrated genomic, transcriptomic, and protein analysis. The Lancet Oncology, 17 (7), pp. 1004-1018. ISSN 1470-2045

Preview

Text
PubSub5535_Ball.pdf - Post-print
Download (644kB) | Preview

Official URL: http://doi.org/10.1016/S1470-2045(16)00174-1

Abstract

Background: Although the use of proliferation markers/profiles has been recommended when choosing the appropriate systemic-treatment for breast cancer (BC), the best molecular-marker/test that should be used needs to be identified.

Methods: To identify factors that drive proliferation and its associated features in BC an artificial neural network (ANN) based integrative data-mining methodology was applied to three cohorts [(Nottingham-discovery (ND), Uppsala and METABRIC (Molecular Taxonomy of Breast Cancer International Consortium)]. The most prominent genes in the resulting interactome-map were then identified. Given that SPAG5 was associated with many features of proliferation, featured prominently in the interactome-map and has a fundamental role in mitotic-progression,, we hypothesized that it could be a better indicator of proliferation activity. (BC). Subsequently to test if it could provide a more accurate guide for the delivery of systemic therapies in BC, we investigated the clinico-pathological utility of SPAG5: gene copy number aberrations (CNAs); mRNA and protein expression, in over 10,000 BCs. Integrated analysis of SPAG5-gene CNAs, transcript and protein expression was conducted in the ND cohort (n=171) and validated in the METABRIC cohort (n=1980). In addition, the associations of SPAG5 CNAs, transcript and/or protein with BC specific survival (BCSS), disease free survival (DFS) and/or distant relapse free survival (DRFS) were analysed in multiple cohorts including Uppsala (n=249), METABRIC, three-untreated lymph node (LN) negative cohorts (n=684), a combined multicentre clinical data set (n=5439), Nottingham historical early-stage-primary BC (Nottingham-HES-BC; n=1650), Nottingham oestrogen receptor (ER) negative BC (n=697), Nottingham anthracycline-Neoadjuvant-chemotherapy (Nottingham-AC-Neo-ACT; n=200), and MD Anderson Cancer Centre Taxane/anthracycline (MDACC-T/AC-Neo-ACT; n=508) cohorts. The association of SPAG5 transcript and protein expression with pathological response rate (pCR) were also tested in [MDACC-T/AC-Neo-ACT (n=508) and the phase II trial NCT00455533; n=253)] and [Nottingham-AC-Neo-ACT (n=200)] cohorts; respectively.

Findings: SPAG5 gene gain/amplification at the Ch17q11·2 locus was found in 10.4% of BC (206/1980 (; METABRIC) and was reported in 19·4% of PAM50-HER2 (46/237) and 17·8% of PAM50-LumB (87/488). SPAG5-CNA gain/amplification and high SPAG5-transcript and SPAG5-protein were associated with increased risk of death from BC [Uppsala; (HR (CI 95%): 1·50 (1·18-1·92); p=0·00010, METABRIC; (HR (CI 95%): 1·68 (1·40-2·01) p<0·0001), and Nottingham-HSE-BC; (HR (CI 95%): 1·68 (1·32-2·12), p<0·0001); respectively]. Multivariable Cox regression models, including other validated-prognostic factors, (Uppsala: age, size, LN-stage, genomic grade index (GGI), ER, TP53 mutation and MKi67; METABRIC: age, size, LN-stage, histologic-grade, ER, progesterone receptor (PR), human epidermal growth factor receptor 2 (HER2), hormone-therapy, chemotherapy, interaction term of SPAG5 and both chemotherapyy and hormonotherapy; Desmedt-untreated LN- cohort: ER, Nottingham prognostic index (NPI), 76-gene prognostic signature (Veridex) and Adjuvant-Online (AOL); Nottingham-HES-BC: menopausal status, size, LN- stage, histologic-grade, ER, PR, HER2, ki67, hormone-therapy, chemotherapy, interaction term of SPAG5 and both chemotherapy[y and hormonotherapy), showed that high SPAG5-transcript and high SPAG5-protein were associated with shorter BCSS [Uppsala: (HR (CI 95%): 1·62 (1·03-2·53) p=0·036); METABRIC: (HR (CI 95%): 1·27 (1·02-1·58) p=0·034); Desmedt-untreated LN- cohort: (HR (CI 95%): 2·34 (1·24-4·42) p=0·0090), and Nottingham-HES-BC (HR (CI 95%): 1·73 (1·23-2·46) p=0·0020); respectively]. In ER-negative-BC with high SPAG5-protein, administration of anthracycline-adjuvant-chemotherapy had reduced the risk of death by 60% compared to chemotherapy-naive (HR (95% CI): 0·37 (0·20-0·60); p=0·0010). A multivariable Cox regression analysis, which included other validated prognostic factors for chemotherapy (e.g., menopausal status, size, lymph node stage, histologic grade, ER, PR, HER2, Bcl2, chemotherapy, interaction term of SPAG5 and both chemotherapy[y), revealed that SPAG5-transcript+ was independently associated with decreased risk of DRFS after receiving Taxane/anthracycline-Neo-ACT [MDACC-T/AC-Neo-ACT: (HR (CI 95%): 0·68 (0·48-0·97); p=0·0070)]. In multivariable logistic regression analysis, both SPAG5-transcript+ and SPAG5-protein+ and were independent predictors for higher pCR after combination-cytotoxic chemotherapy [MDACC-T/AC-Neo-ACT: (OR (95% CI) 1·71 (1·07-2·74); p=0·024), and Nottingham-AC-Neo-AC: (OR (95% CI): 8·75 (2·42-31); p=0·0010); respectively].

Interpretation: SPAG5 is a novel amplified gene on Ch17q11.2 in PAM50-LumB and PAM-HER2 BC, and its transcript and protein products are independent prognostic and predictive biomarkers, with potential clinical utility as a biomarker for combination cytotoxic chemotherapy sensitivity, especially in ER- BC.

Item Type:

Journal article

Alternative Title:

SPAG5 is novel actionable oncogene that predicts survival benefit from anthracycline therapy in breast cancer patients [running title]

Publication Title:

The Lancet Oncology

Creators:

Abdel-Fatah, T.M.A., Agarwal, D., Liu, D.X., Russell, R., Rueda, O.M., Liu, K., Xu, B., Moseley, P.M., Green, A.R., Pockley, A.G., Rees, R.C., Caldas, C., Ellis, I.O., Ball, G.R. and Chan, S.Y.T.

Publisher:

Elsevier

Date:

July 2016

Volume:

Number:

ISSN:

1470-2045

Identifiers: