Testing for causality between systematically identified risk factors and glioma: a Mendelian randomization study

BACKGROUND

Whilst epidemiological studies have provided evidence of associations between certain risk factors and glioma onset, inferring causality has proven challenging. Using Mendelian randomization (MR), we assessed whether associations of 36 reported glioma risk factors showed evidence of a causal relationship.

METHODS

We performed a systematic search of MEDLINE from inception to October 2018 to identify candidate risk factors and conducted a meta-analysis of two glioma genome-wide association studies (5739 cases and 5501 controls) to form our exposure and outcome datasets. MR analyses were performed using genetic variants to proxy for candidate risk factors. We investigated whether risk factors differed by subtype diagnosis (either glioblastoma (n = 3112) or non-glioblastoma (n = 2411)). MR estimates for each risk factor were determined using multiplicative random effects inverse-variance weighting (IVW). Sensitivity analyses investigated potential pleiotropy using MR-Egger regression, the weighted median estimator, and the mode-based estimator. To increase power, trait-specific polygenic risk scores were used to test the association of a genetically predicated increase in each risk factor with glioma onset.

RESULTS

Our systematic search identified 36 risk factors that could be proxied using genetic variants. Using MR, we found evidence that four genetically predicted traits increased risk of glioma, glioblastoma or non-glioblastoma: longer leukocyte telomere length, liability to allergic disease, increased alcohol consumption and liability to childhood extreme obesity (> 3 standard deviations from the mean). Two traits decreased risk of non-glioblastoma cancers: increased low-density lipoprotein cholesterol (LDLc) and triglyceride levels. Our findings were similar across sensitivity analyses that made allowance for pleiotropy (genetic confounding).

CONCLUSIONS

Our comprehensive investigation provides evidence of a causal link between both genetically predicted leukocyte telomere length, allergic disease, alcohol consumption, childhood extreme obesity, and LDLc and triglyceride levels, and glioma. The findings from our study warrant further research to uncover mechanisms that implicate these traits in glioma onset.

Abstract P2-07-06: Cumulative copy number imbalances after neoadjuvant chemotherapy residual breast tumor is an independent predictor of relapse

Background: Identifying breast cancer patients after neoadjuvant chemotherapy (NAC) at greatest risk of recurrence would enhance selection of patients who may benefit from novel adjuvant treatments.

Patients. 243 stage I-III breast cancer patients who underwent NAC with ≥10% residual tumor cellularity were identified from the MD Anderson Cancer Center and Ben Taub General Hospital, Harris County hospital. Tumor DNA was isolated for DNA copy number using OncoScan CNV FFPE, Affymetrix. Median follow-up was 67.8 months. Continuous residual cancer burden (RCB) scores with CNI data were available for 152 cases. To test if CNIs covering large regions were associated with recurrence after adjusting for prognostic variables and study site, data were summed to a chromosome-arm level. Eleven chromosome arms with false discovery rate <0.05 for breast cancer recurrence were identified. A stepwise multivariable model including age at diagnosis, tumor subtype, histologic grade, pre- and post-treatment stage, study site, and the 11 chromosomal arms were used to fit a parsimonious multivariate model for recurrence. Minimizing the Akaike Information Criterion yielded a final model with post-stage and a 5-arm CNI (5A-CNI) indicator including 2q, 3q, 4q, 10p, and 18p. Tumors were classified on 5A-CNI as 0 [no CNI], 1 [1- 2] and 2 [> 2].

Results. The study population included 76 non-Hispanic White, 89 Hispanic, and 68 African American patients with a mean age of 49.1 years. 105 patients were classified as 5A-CNI-0, 97 as 5A-CNI-1 and 41 as 5A-CNI-2. A higher 5A-CNI score was associated with tumor grade, ER-negative tumors (p<0.002) and tumor subtype (p=0.014). For 5A-CNI scores of 0, 1 and 2, recurrence rates of 14%, 34% and 58.5% were observed, respectively. In the final multivariable model adjusted for post-stage, RCB and study site, when compared to 5A-CNI-0, the hazard of recurrence was elevated for 5A-CNI-1 (HR= 2.27 [95% CI, 1.01-5.1]) and 5A-CNI-2 tumors (HR=7.43 [95% CI, 2.85-19.39]). Further, while the sample size is limiting, of 10 patients who were RCB3 and 5A-CNI-2, 9 relapsed (90%) during follow-up compared to only 6 of 43 (14%) of RCB3 patients with 5A-CNI-0 (p<10-6). For patients with RCB1 or 2, relapse did not differ by 5A-CNI score. Neither race nor ethnicity were found to be independently associated with recurrence or tumor subtype. However, African American, followed by Hispanic patients, were more likely than non-Hispanic White patients to be classified as 5A-CNI-2 (p=0.013).

Table 1.Significant difference in distribution of 5 arm CNI classifier by Race/Ethnicity in Study Sample (p =0.013).5A-CNI012Non-Hispanic Whiten=44; 57.9%n=25; 32.9%n=7; 9.2%Hispanicn=32; 36%n=42; 47.2%n=15; 16.9%African Americann=28; 41.2%n=23; 33.8%n=17; 25%

Conclusion. The 5A-CNI score in post NAC tumor identifies a patient population with very poor prognosis independent of current clinical prognostic factors including RCB. Validation of these findings may lead to a post NAC genomic test that identifies patients who would benefit from additional treatment Further investigation of the nature of the association between the 5A-CNI score and race/ethnicity, which appears independent of tumor subtype, is warranted.

Citation Format: Thompson PA, Brewster A, Tsavachidis S, Armstrong G, Do K-A, Ha M-J, Gutierrez C, Symmans F, Bondy M. Cumulative copy number imbalances after neoadjuvant chemotherapy residual breast tumor is an independent predictor of relapse [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 P2-07-06.

Copy number imbalances between screen- and symptom-detected breast cancers and impact on disease-free survival

Screening mammography results in the increased detection of indolent tumors. We hypothesized that screen- and symptom-detected tumors would show genotypic differences as copy number imbalances (CNI) that, in part, explain differences in the clinical behavior between screen- and symptom-detected breast tumors. We evaluated 850 women aged 40 and above diagnosed with stage I and II breast cancer at the University of Texas MD Anderson Cancer Center between 1985 and 2000 with information available on method of tumor detection (screen vs. symptoms). CNIs in screen- and symptom-detected tumors were identified using high-density molecular inversion probe arrays. Cox proportional modeling was used to estimate the effect of method of tumor detection on disease-free survival after adjusting for age, stage, and the CNIs. The majority of tumors were symptom detected (n = 603) compared with screen detected (n = 247). Copy number gains in chromosomes 2p, 3q, 8q, 11p, and 20q were associated with method of breast cancer detection (P < 0.00001). We estimated that 32% and 63% of the survival advantage of screen detection was accounted for by age, stage, nuclear grade, and Ki67 in women aged 50 to 70 and aged 40 to 87, respectively. In each age category, an additional 20% of the survival advantage was accounted for by CNIs associated with method of detection. Specific CNIs differ between screen- and symptom-detected tumors and explain part of the survival advantage associated with screen-detected tumors. Measurement of tumor genotype has the potential to improve discrimination between indolent and aggressive screen-detected tumors and aids patient and physician decision making about use of surgical and adjuvant treatments.

Evaluation of Changes in Plasma Protein Profiles during Neoadjuvant Chemotherapy in HER2-Positive Breast Cancer Using MALDI-TOF/MS Procedure

Background: Comparison of protein profiles of the plasma before chemotherapy and after completion of neoadjuvant chemotherapy may reveal tumor markers that could be used to monitor tumor response.Patients and Methods: We examined matching pre- and post-treatment serum samples 39 HER2-postive breast cancer patients (n=78 samples) who all received 6 months of preoperative chemotherapy with or without trastuzumab in the context of a randomized clinical trial. Serum was analyzed with an Applied Biosystems 4700 Proteomics Analyzer matrix-assisted laser desorption/ionization time-of-flight (MALDI-TOF) mass spectrometer. Samples were grouped and pooled into those who achieved pathological complete response (pCR, n=20) and those who had residual disease (RD, n=19). We compared matching baseline and post-chemotherapy/pre-surgery samples separately in both response groups and also compared baseline samples between the two response groups.Results: MALDI-TOF analysis revealed a total of 2329 and 3152 peaks in pooled samples of cases with pCR and RD, respectively. A total of 32 peaks were differentially expressed between base line and post-chemotherapy pCR samples and 643 peaks in cases with RD (false discovery rate ≤ 20%). A total of 8 differentially expressed proteins were identified in the before- and after-chemotherapy samples from their peptides after digestion and LC-MALDI-TOF/TOF. These included 4 AFM, C3, hemopexin, SAP in pCR samples and AP1, hemopexin, Complement B, amyloid P component in the RD group.Conclusion: Our study suggests that MALDI mass spectrometry may be used to detect differences in baseline serum profiles of patients who are highly sensitive to chemotherapy and those who are less sensitive. Also, changes occur in the serum during chemotherapy and this may offer the possibility of monitoring response to treatment in the future.

Citation Information: Cancer Res 2009;69(24 Suppl):Abstract nr 2037.

Evaluation of serum profiles changes after neoadjuvant chemotherapy for breast cancer using MALDI-TOF/MS procedure

e22072

Evaluation of serum profiles changes after neoadjuvant chemotherapy for breast cancer using MALDI-TOF / MS procedure. Background: Response to primary chemotherapy (CT) for breast cancer is heterogeneous among patients and a more tailored treatment would be beneficial in term of reducing exposure to an unnecessary toxicity and optimization of response rates. Mass spectrometry analysis of serum might be helpful in detecting specific changes in response to primary CT. Methods: An applied Biosystems 4700 Proteomics Analyzer matrix-assisted laser desorption/ionization time-of-flight (MALDI-TOF) mass spectrometer was used. A breast cancer cohort of 78 sera samples from 39 HER2 positive patients consisting of matched pretreatment and (6 months) posttreatment samples was used. Blood samples were collected serially before each treatment cycle every 3 weeks of neoadjuvant CT. Samples were divided into those who achieved pathological complete response (pCR, n= 20) and those who had residual disease (RD, n=19). Low-mass differentially expressed peptides were identified using MALDI-TOF/TOF. Results: This procedure yielded a total of 2329 and 3152 peaks respectively, for the responders and non-responders. Biological variation analysis revealed a total of 32 peaks for responders and 643 peaks for non-responders to be differentially regulated with a false discovery rate less than 20%. A total of 8 differentially expressed proteins were identified from their peptides after digestion and LC-MALDI-TOF/TOF. Four in tumors with pCR (AFM, C3, hemopexin, SAP) and four proteins in the RD group were identified (AP1, hemopexin, Complement B, amyloid P component) Conclusions: Our study suggests that MALDI mass spectrometry may be used to predict the tumor response to neoadjuvant chemotherapy. Proteomic analysis may be useful in developing tailored chemotherapy for breast cancer.

No significant financial relationships to disclose.

Rapamycin-regulated gene expression signature determines prognosis in breast cancer

Abstract #3072

Background: Mammalian target of rapamycin (mTOR) is a serine/threonine kinase involved in multiple intracellular signaling pathways promoting tumor growth. mTOR is aberrantly activated in a significant portion of breast cancers and is a promising target for treatment. Rapamycin and its analogues are in clinical trials for breast cancer treatment. Patterns of gene expression (metagenes) may also be used to simulate a biologic process or effects of a drug treatment. In this study, we tested the hypothesis that the gene-expression signature regulated by rapamycin could predict disease outcome for patients with breast cancer.&#x2028; Methods: MDA-MB-468 breast cancer cell line is triple negative and PTEN null. Sensitivity of MDA-MB-468 breast cancer cell line to rapamycin was determined by in vitro growth assays and in vivo tumor studies. Total RNA was extracted from cell lines treated with rapamycin or DMSO for 24 hours and xenografts treated with DMSO or rapamycin for 1 or 22 days. Total RNA was used for expression profiling and hybridized to Affymetrix HG-U133 Plus 2.0 arrays. We used two well-described primary breast cancer datasets of 251 (Miller et al, Proc. Natl. Acad. Sci. USA 2005;102:13550) and 286 (Wang et al, Lancet. 2005;365:671) patients from the public domain and assessed the prognostic capability of rapamycin-regulated gene expression signatures.&#x2028; Results: Colony formation and sulforhodazine B (IC50 &lt; 1 nM) assays, and xenograft animals showed that MDA-MB-468 cells were sensitive to treatment with rapamycin. The comparison of in vitro and in vivo gene expression data identified a 31-gene expression signature that was up- regulated by rapamycin treatment in vitro as well as in vivo (FDR of 0.1). From the 31 genes on HG-U133 Plus 2.0 array, 20 that were included on HG-U133A array were used for analysis. In the Miller dataset, rapamycin metagene index (RMI) did not correlate with tumor size or lymph node status. High (0.75 percentile) RMI was significantly associated with longer survival (P=0.015). On multivariate analysis, RMI (P=0.029), tumor size (P=0.015) and lymph node status (P=0.001) were prognostic. In the Wang dataset, RMI predicted time to disease relapse (P=0.009).&#x2028; Conclusion: Rapamycin-regulated gene expression signature predicts clinical outcome in breast cancer. This supports the central role of mTOR signaling in breast cancer biology and provides further impetus to pursue mTOR-targeted therapies for breast cancer treatment.

Citation Information: Cancer Res 2009;69(2 Suppl):Abstract nr 3072.