Decker, Paul A.

Background: Larger within-patient variability of lipid levels has been associated with an increased risk of cardiovascular disease (CVD). However, measures of lipid variability are not currently used clinically. We investigated the feasibility of calculating lipid variability within a large electronic health record (EHR)-based population cohort and assessed associations with incident CVD. Methods: We identified all individuals ≥40 years of age who resided in Olmsted County, MN on 1/1/2006 (index date) without prior CVD. CVD was defined as myocardial infarction, coronary artery bypass graft surgery, percutaneous coronary intervention or stroke. Patients with ≥3 measurements of total cholesterol, low-density lipoprotein cholesterol (LDL-C), high-density lipoprotein cholesterol (HDL-C), and/or triglycerides during the 5 years before the index date were retained in the analyses. Lipid variability was calculated using variability independent of the mean (VIM). Patients were followed through 9/30/2017 for incident CVD (including CVD death). Cox regression was used to investigate the association between quintiles of lipid VIMs and incident CVD. Results: We identified 18,642 individuals (mean age 60; 55% female) who were free of CVD at baseline and VIM calculated for at least one lipid measurement. After adjustment, those in the highest VIM quintiles of total cholesterol had a 25% increased risk of CVD (Q5 vs. Q1 HR: 1.25, 95% CI: 1.08-1.45; Table). We observed similar results for LDL-C (Q5 vs. Q1 HR: 1.20, 95% CI: 1.04-1.39) and HDL-C (Q5 vs. Q1 HR: 1.25, 95% CI: 1.09-1.43). There was no association between triglyceride variability quintiles and CVD risk. Conclusion: In a large EHR-based population cohort, high variability in total cholesterol, HDL-C and LDL-C was associated with an increased risk of CVD, independently of traditional risk factors, suggesting it may be a target for intervention. Lipid variability can be calculated in the EHR environment but more research is needed to determine its clinical utility.

Abstract Tesevatinib is a potent oral brain penetrant EGFR inhibitor currently being evaluated for glioblastoma therapy. Tesevatinib distribution was assessed in wild-type (WT) and Mdr1a/b(-/-)Bcrp(-/-) triple knockout (TKO) FVB mice after dosing orally or via osmotic minipump; drug–tissue binding was assessed by rapid equilibrium dialysis. Two hours after tesevatinib dosing, brain concentrations in WT and TKO mice were 0.72 and 10.03 μg/g, respectively. Brain-to-plasma ratios (Kp) were 0.53 and 5.73, respectively. With intraperitoneal infusion, brain concentrations were 1.46 and 30.6 μg/g (Kp 1.16 and 25.10), respectively. The brain-to-plasma unbound drug concentration ratios were substantially lower (WT mice, 0.03–0.08; TKO mice, 0.40–1.75). Unbound drug concentrations in brains of WT mice were 0.78 to 1.59 ng/g. In vitro cytotoxicity and EGFR pathway signaling were evaluated using EGFR-amplified patient-derived glioblastoma xenograft models (GBM12, GBM6). In vivo pharmacodynamics and efficacy were assessed using athymic nude mice bearing either intracranial or flank tumors treated by oral gavage. Tesevatinib potently reduced cell viability [IC50 GBM12 = 11 nmol/L (5.5 ng/mL), GBM6 = 102 nmol/L] and suppressed EGFR signaling in vitro. However, tesevatinib efficacy compared with vehicle in intracranial (GBM12, median survival: 23 vs. 18 days, P = 0.003) and flank models (GBM12, median time to outcome: 41 vs. 33 days, P = 0.007; GBM6, 44 vs. 33 days, P = 0.007) was modest and associated with partial inhibition of EGFR signaling. Overall, tesevatinib efficacy in EGFR-amplified PDX GBM models is robust in vitro but relatively modest in vivo, despite a high brain-to-plasma ratio. This discrepancy may be explained by drug-tissue binding and compensatory signaling.

Abstract Introduction Tumefactive multiple sclerosis (MS) and glioblastoma (GBM) can be difficult to differentiate. Clinical presentations are nonspecific and may occur in patients with both diseases. Magnetic resonance imaging (MRI) scans are routinely used to generate an initial differential diagnosis. Although imaging features can help distinguish these entities, there are no pathognomonic signs, neither on traditional nor advanced MRI. Moreover, the ability to differentiate these entities depends on the radiologists’ experience. Ensuring the correct diagnosis has important therapeutic and prognostic implications, as the management of these diseases is very different. To date, a brain biopsy often represents the only option to obtain a diagnosis, but it may be inconclusive and could add unnecessary morbidity. We hypothesize that a machine-learning tool based on routinely performed MRI scans could guarantee a more reliable, fast, and reproducible diagnosis. Methods Only patients with a histological diagnosis were included in the study. Patients that underwent any form of radiation therapy before the surgery were excluded to avoid confounding factors due to treatment-related effects. A total of 99 MS and 257 GBMs were included, and preoperative T2-weighted images were collected. For each patient, the lesions' segmentation was performed, including both the contrast-enhancing and non-enhancing areas of the lesions. The segmentation masks were used to extract radiomic features from the T2-weighted sequences. All MRI images were standardized to the same mean and standard deviation before feature extraction and resampled to a voxel size of 1.0 × 1.0 × 1.0 mm. Features extraction was performed using Pyradiomics (v. 3.0.1) and included first order, shape, and texture features, totaling 100 features. The classifier consisted of a support vector machine (SVM) model. Patients were randomly assigned to the training and test sets with an 80-20% proportion for each class. To select the best hyperparameters for the SVM classifier, a grid-search analysis was performed on the training set using a nested cross-validation scheme (k=5). All experiments were performed using Python (v. 3.5) and the Scikit-learn library (v. 0.23.1). Results The mean accuracy and area under the ROC curve for the best set of hyperparameters on the validation set were 0.83 and 0.86, respectively. The weighted-average precision, recall, and F1 scores on the test set were 0.78, 0.76, and 0.77, respectively. Conclusions Radiomic features extracted from a routinely acquired MRI sequence showed promising results in differentiating tumefactive multiple sclerosis from glioblastoma, opening the possibility to obtain a reliable diagnosis without the need for a brain biopsy. Future experiments Future experiments will focus on using post-contrast T1-weighted images for the radiomic features extractions, testing multiple machine learning models and features selection algorithms, and including more diagnostic classes (i.e., metastasis, lymphomas). Citation Format: Gian Marco Conte, Jeanette E. Eckel-Passow, W. Oliver Tobin, Paul Decker, Daniel H. Lachance, Robert B. Jenkins, Bradley J. Erickson. Differentiation of tumefactive multiple sclerosis and glioblastoma using radiomics features extracted from magnetic resonance imaging and machine learning [abstract]. In: Proceedings of the AACR Virtual Special Conference on Artificial Intelligence, Diagnosis, and Imaging; 2021 Jan 13-14. Philadelphia (PA): AACR; Clin Cancer Res 2021;27(5_Suppl):Abstract nr PO-023.

Inflammatory breast cancer is a highly aggressive form of breast cancer that forms clusters of tumor emboli in dermal lymphatics and readily metastasizes. These cancers express high levels of E-cadherin, the major mediator of adherens junctions, which enhances formation of tumor emboli. Previous studies suggest that E-cadherin promotes cancer when the balance between apical and basolateral cadherin complexes is disrupted. Here, we used immunohistochemistry of inflammatory breast cancer patient samples and analysis of cell lines to determine the expression of PLEKHA7, an apical adherens junction protein. We used viral transduction to re-express PLEKHA7 in inflammatory breast cancer cells and examined their aggressiveness in 2D and 3D cultures and in vivo. We determined that PLEKHA7 was deregulated in inflammatory breast cancer, demonstrating improper localization or lost expression in most patient samples and very low expression in cell lines. Re-expressing PLEKHA7 suppressed proliferation, anchorage independent growth, spheroid viability, and tumor growth in vivo. The data indicate that PLEKHA7 is frequently deregulated and acts to suppress inflammatory breast cancer. The data also promote the need for future inquiry into the imbalance between apical and basolateral cadherin complexes as driving forces in inflammatory breast cancer.

Substantial gaps remain in the evidence base for prenatal tobacco use interventions among Indigenous women. Using a cluster randomized controlled trial (RCT), the Healthy Pregnancies Project evaluated a community-level intervention for Alaska Native (AN) women in rural western Alaska. Sixteen villages were randomly assigned to usual care (control, n = 8 villages) or usual care plus a community-level intervention delivered by local AN “Native Sisters” (n = 8 villages). Outcomes were tobacco use rate at delivery and at 2 and 6 months postpartum, with biochemical confirmation obtained at 6 months. The program had high reach, enrolling 73% of all eligible women screened. Of the 352 participants, 67% used tobacco at baseline. No significant differences emerged between study groups on follow-up in tobacco use rates. More intervention than control participants made a quit attempt at 2 months postpartum (70% vs. 51%, respectively, p = 0.012). Participants in both study groups reported the program helped to raise awareness of healthy pregnancies in the study villages. This trial supports the reach of a community-level intervention, but not its efficacy for reducing tobacco use during pregnancy or postpartum. Efforts to sustain early quit attempts appear warranted. The community involvement, and reported impact on raising awareness of the importance of healthy pregnancies, supports the value of the research program in this community.

Abstract Background Face-to-face tobacco cessation has had limited reach and efficacy in Alaska Native (AN) communities. We describe our two-phased approach to develop content for Connecting Alaska Native People to Quit Smoking, a Facebook group intervention to reduce barriers to evidence-based smoking cessation treatment for AN people in Alaska. Methods Phase 1 included semi-structured telephone interviews with 30 AN people who smoke and ten stakeholders. They provided feedback on existing content from the Centers for Disease Control and Prevention Tips campaign and AN digital stories. Phase 2 included an online survey with a new group of 40 AN smokers who provided feedback on existing content via a measure of perceived effectiveness and cultural relevance. Results Phase I results revealed participants evaluated content based upon story strength, relevance to AN culture, emotional appeal, relatability to AN people, and favorite video. No single posting was rated highly across all themes. All perceived effectiveness (PE) and cultural relevance median scores fell between 3.5 and 4.4 (range 1–5). PE scores varied across participant demographic groups. Conclusions Content embodying characteristics perceived to be most appealing, effective, and culturally relevant were selected for the private Facebook group content library with refinements made to incorporate images of AN people engaged in AN activities. PE scores indicate a need for a wide variety of content that moderators could pull from when conducting the intervention. Implications Social media content targeting specific population sectors, such as American Indian/AN people for tobacco cessation needs to be culturally tailored. Our approach provides a model others can follow to determine what is appealing, relevant, and effective messaging. Clinical Trial Registration NCT 03645941.

Abstract Background: The four most commonly-mutated genes in clear cell renal cell carcinoma (ccRCC) tumors are BAP1, PBRM1, SETD2 and VHL. And, there are currently 14 known RCC germline variants that have been reproducibly shown to be associated with RCC risk. However, the association of germline genetics with tumor genetics and clinical aggressiveness are unknown.Methods: We analyzed 420 ccRCC patients from The Cancer Genome Atlas (TCGA). Molecular subtype was determined based on acquired mutations in BAP1, PBRM1, SETD2 and VHL. Aggressive subtype was defined clinically using Mayo SSIGN score and molecularly using the ccA/ccB gene expression subtype. Publically-available Hi-C data were used to link germline risk variants with candidate target genes. Results: The 8q24 variant rs35252396 was significantly associated with VHL mutation status (OR=1.6, p=0.0037) and SSIGN score (OR=1.9, p=0.00094), after adjusting for multiple comparisons. We observed that, while some germline variants have interactions with nearby genes, some variants demonstrate long-range interactions with target genes.Conclusions: These data further demonstrate the link between rs35252396, HIF pathway and ccRCC clinical aggressiveness, providing a more comprehensive picture of how germline genetics and tumor genetics interact with respect to tumor development and progression.

Abstract Background: Genome-wide association studies (GWAS) revealed that 25 regions in 24 genes are associated with adult diffuse glioma development. These regions were identified by performing GWAS of glioma overall and by pathology (GBM and nonGBM). The 2016 WHO Classification of Tumors of the Central Nervous System utilizes two somatic alterations to molecularly-classify adult diffuse glioma: IDH mutation and 1p/19q codeletion. We hypothesized that germline variants may increase susceptibility to, or interact with, these somatic alterations to accelerate the development of specific molecular subtypes of glioma. We further hypothesize that germline variants associated with IDH-mutated glioma might be associated with other IDH-mutated tumors, namely, cholangiocarcinoma, acute myeloid leukemia (AML) and melanoma. Methods: We performed a GWAS by glioma molecular subtype - as defined by presence or absence of IDH somatic mutation and 1p/19q codeletion. A total of 1320 glioma cases and 1889 controls were used in the discovery set, and 799 glioma cases and 808 controls in the validation set. A meta-analysis was performed with a genome-wide p-value threshold of 5 × 10−8. GTEx data were used to perform an expression quantitative trait loci (eQTL) analysis. For germline variants that were significantly associated with IDH-mutated glioma, we evaluated pleiotropy with cholangiocarcinoma, AML and melanoma using TCGA and Mayo Biobank controls. Results: Variants in or near D2HGDH on chromosome 2 were genome-wide significant in IDH-mutated glioma (meta p-value = 2.82 × 10−10). TCGA reported that the D2HGDH region was commonly deleted in IDH-mutated gliomas that do not have 1p/19q codeletion. In TCGA data for IDH-mutated, non-codeleted glioma, we observed that the D2HGDH variant was inversely associated with tumor deletions of D2HGDH (odds ratio=0.57, p-value=0.015). The eQTL analyses demonstrated significant associations between D2HGDH germline variant and expression of D2HGDH (p=2.2 × 10−11). Further stratifying IDH-mutated glioma by 1p/19q codeletion status, one variant near FAM20C on chromosome 7 was genome-wide significant in gliomas that have IDH mutation and 1p/19q codeletion (meta p-value=9.56 × 10−9). Analyses are currently underway to evaluate pleiotropy of these IDH-mutated glioma germline variants with other IDH-mutated tumors including cholangiocarcinoma, AML and melanoma. Variants in or near GMEB2 on chromosome 20 were genome-wide significant in IDH wild-type glioma (meta p-value=2.60 × 10−10). The most significant variant in the GMEB2 region remained significant after adjustment for the known RTEL1 glioma risk variant nearby on chromosome 20 (p=0.029). Conclusions: We identified and validated novel germline variants in two genes that are associated with etiology of IDH-mutated and one gene that is associated with IDH wild-type adult diffuse glioma. Citation Format: Jeanette E. Eckel-Passow, Kristen L. Drucker, Thomas M. Kollmeyer, Matthew L. Kosel, Paul A. Decker, Annette M. Molinaro, Terri Rice, Corrine E. Praska, Lauren E. Clark, Alissa A. Caron, Alexej Abyzov, Anthony Batzler, Jun S. Song, Melike Pekmezci, Helen M. Hansen, Lucie S. McCoy, Paige M. Bracci, Joseph Wiemels, John K. Wiencke, Stephen Francis, Terence C. Burns, Caterina Giannini, Daniel H. Lachance, Margaret Wrensch, Robert B. Jenkins. Adult diffuse glioma GWAS by molecular subtype identifies variants in D2HGDH, FAM20C and GMEB2 [abstract]. In: Proceedings of the Annual Meeting of the American Association for Cancer Research 2020; 2020 Apr 27-28 and Jun 22-24. Philadelphia (PA): AACR; Cancer Res 2020;80(16 Suppl):Abstract nr 1193.

Abstract The Mayo Clinic collection of patient-derived xenograft (PDX) models has been developed from the direct implantation of fresh glioblastoma samples into immunodeficient mice. Over 95 models have been extensively characterized with regards to histology, invasion, flank and orthotopic growth rates, and molecular profiling. Evaluation of orthotopic tumors demonstrated clear infiltration of tumor cells into the brain parenchymal in 90% of PDXs with contralateral hemisphere involvement in 64%. Extensive molecular profiling confirmed that our collection of PDXs captures the genetic heterogeneity of GBM, with the majority of known alterations represented at a frequency similar to the TCGA including gain of chromosome 7 and loss of chromosome 10. TERT promoter mutations were the most frequent overall (86%) followed by homozygous deletion of CDKN2A (70%). Multiple alterations were identified in the p53 pathway, including TP53 mutations (36%) and MDM2/4 (10%/2%) amplification. PI3-kinase pathway alterations mainly involved PTEN (48%) while MAP kinase pathway alterations were present in a subset of PDXs (NF1 alterations 17%, BRAF 4%). IDH mutations were found in only two models. Gene expression analysis revealed that our PDX collection includes all GBM subtypes with classical being the most represented (46% vs mesenchymal 31% and proneural 23%). MGMT promoter methylation was also observed in 45% of PDXs. The high fidelity of PDX models to their human counterpart makes these models ideal for testing standard of care and novel targeted therapies. To assess if PDXs are predictive of patient treatment response, 37 orthotopic PDX lines were treated with radiation (RT), temozolomide (TMZ), or concurrent RT/TMZ. Across all PDXs, RT and TMZ demonstrated median survival extension of 90 and 46 days relative to vehicle. Concurrent RT/TMZ further extended survival by 146 days. MGMT promoter methylation was strongly associated with response to TMZ-based strategies but not RT alone.This observation suggests that our PDX collection possesses clinically relevant predictive value. To date, we have characterized treatment responses to over 135 regimens including both GBM standards of care and targeted strategies in our PDX models. Recent lab efforts strive to define the effect of blood brain barrier disruption on drug delivery. With over 20 years of experience, our group has extensive expertise in the development of clinically relevant study designs focused on examining efficacy, tolerability, and pharmacokinetic/pharmacodynamic assessments. Since 2016, we have shared over 5200 individual samples (tumor tissue, cells, DNA/RNA, protein extracts, tissue microarray slides) in both internal and external collaborators. Our models have been utilized in over 41 NIH funded grants, including 33 funded RO1s, as well as by the pharmaceutical community with 50 working relationships to date (20 contracts, 30 collaborations). Finally, our PDX collection has been used in over 120 peer-reviewed published manuscripts. Citation Format: Danielle Burgenske, Ann C. Mladek, Rachael A. Vaubel, Shulan Tian, Mark A. Schroeder, Zeng Hu, Brett L. Carlson, Paul A. Decker, Jeanette E. Eckel-Passow, Jann N. Sarkaria. Evaluation of novel therapeutics using models from the Mayo Clinic GBM patient-derived xenograft (PDX) collection [abstract]. In: Proceedings of the Annual Meeting of the American Association for Cancer Research 2020; 2020 Apr 27-28 and Jun 22-24. Philadelphia (PA): AACR; Cancer Res 2020;80(16 Suppl):Abstract nr 1113.

Abstract Background: Inflammatory breast cancer is a highly aggressive form of breast cancer that robustly forms clusters of tumor emboli in dermal lymphatics and readily metastasizes. Inflammatory breast cancers express high levels of E-cadherin, the major protein of adherens junctions, which may enhance the ability of tumor cells to form such clusters and contribute to metastasis. Seemingly contradictory, E-cadherin has both tumor-suppressing and tumor-promoting roles in cancer; previous studies suggest that this depends on the balance between apical and basolateral cadherin-catenin complexes. Methods: In the present study, we use immunohistochemistry of inflammatory breast cancer patient samples and biochemical analysis of cell lines to determine the expression of PLEKHA7, an apical adherens junction protein. We use viral transduction to ectopically express PLEKHA7 in the SUM149 inflammatory breast cancer cell line. The effect of PLEKHA7 on the aggressiveness of inflammatory breast cancer in 2D, 3D and in-vivo were examined. Results: We determined that PLEKHA7 was deregulated in inflammatory breast cancer, demonstrating improper localization or lost expression in a strong majority of patient samples and very low expression in cell line models. We found that re-expressing PLEKHA7 is sufficient to suppress proliferation, anchorage independent growth, spheroid viability, and tumor growth in-vivo. We also observed a negative-selection pressure within the xenograft tumors to lose PLEKHA7 function or expression.Conclusions: The data indicate that PLEKHA7 is frequently deregulated and acts as a suppressor of inflammatory breast cancer. They also suggest that the resulting imbalance between apical and basolateral cadherin-catenin complexes contributes to growth, survival and emboli-forming capacities of inflammatory breast cancer.