Describing patterns of familial cancer risk in subfertile men using population pedigree data.

STUDY QUESTION: Can we simultaneously assess risk for multiple cancers to identify familial multicancer patterns in families of azoospermic and severely oligozoospermic men? SUMMARY ANSWER: Distinct familial cancer patterns were observed in the azoospermia and severe oligozoospermia cohorts, suggesting heterogeneity in familial cancer risk by both type of subfertility and within subfertility type. WHAT IS KNOWN ALREADY: Subfertile men and their relatives show increased risk for certain cancers including testicular, thyroid, and pediatric. STUDY DESIGN, SIZE, DURATION: A retrospective cohort of subfertile men (N = 786) was identified and matched to fertile population controls (N = 5674). Family members out to third-degree relatives were identified for both subfertile men and fertile population controls (N = 337 754). The study period was 1966-2017. Individuals were censored at death or loss to follow-up, loss to follow-up occurred if they left Utah during the study period. PARTICIPANTS/MATERIALS, SETTING, METHODS: Azoospermic (0 × 106/mL) and severely oligozoospermic (<1.5 × 106/mL) men were identified in the Subfertility Health and Assisted Reproduction and the Environment cohort (SHARE). Subfertile men were age- and sex-matched 5:1 to fertile population controls and family members out to third-degree relatives were identified using the Utah Population Database (UPDB). Cancer diagnoses were identified through the Utah Cancer Registry. Families containing ≥10 members with ≥1 year of follow-up 1966-2017 were included (azoospermic: N = 426 families, 21 361 individuals; oligozoospermic: N = 360 families, 18 818 individuals). Unsupervised clustering based on standardized incidence ratios for 34 cancer phenotypes in the families was used to identify familial multicancer patterns; azoospermia and severe oligospermia families were assessed separately. MAIN RESULTS AND THE ROLE OF CHANCE: Compared to control families, significant increases in cancer risks were observed in the azoospermia cohort for five cancer types: bone and joint cancers hazard ratio (HR) = 2.56 (95% CI = 1.48-4.42), soft tissue cancers HR = 1.56 (95% CI = 1.01-2.39), uterine cancers HR = 1.27 (95% CI = 1.03-1.56), Hodgkin lymphomas HR = 1.60 (95% CI = 1.07-2.39), and thyroid cancer HR = 1.54 (95% CI = 1.21-1.97). Among severe oligozoospermia families, increased risk was seen for three cancer types: colon cancer HR = 1.16 (95% CI = 1.01-1.32), bone and joint cancers HR = 2.43 (95% CI = 1.30-4.54), and testis cancer HR = 2.34 (95% CI = 1.60-3.42) along with a significant decrease in esophageal cancer risk HR = 0.39 (95% CI = 0.16-0.97). Thirteen clusters of familial multicancer patterns were identified in families of azoospermic men, 66% of families in the azoospermia cohort showed population-level cancer risks, however, the remaining 12 clusters showed elevated risk for 2-7 cancer types. Several of the clusters with elevated cancer risks also showed increased odds of cancer diagnoses at young ages with six clusters showing increased odds of adolescent and young adult (AYA) diagnosis [odds ratio (OR) = 1.96-2.88] and two clusters showing increased odds of pediatric cancer diagnosis (OR = 3.64-12.63). Within the severe oligozoospermia cohort, 12 distinct familial multicancer clusters were identified. All 12 clusters showed elevated risk for 1-3 cancer types. An increase in odds of cancer diagnoses at young ages was also seen in five of the severe oligozoospermia familial multicancer clusters, three clusters showed increased odds of AYA diagnosis (OR = 2.19-2.78) with an additional two clusters showing increased odds of a pediatric diagnosis (OR = 3.84-9.32). LIMITATIONS, REASONS FOR CAUTION: Although this study has many strengths, including population data for family structure, cancer diagnoses and subfertility, there are limitations. First, semen measures are not available for the sample of fertile men. Second, there is no information on medical comorbidities or lifestyle risk factors such as smoking status, BMI, or environmental exposures. Third, all of the subfertile men included in this study were seen at a fertility clinic for evaluation. These men were therefore a subset of the overall population experiencing fertility problems and likely represent those with the socioeconomic means for evaluation by a physician. WIDER IMPLICATIONS OF THE FINDINGS: This analysis leveraged unique population-level data resources, SHARE and the UPDB, to describe novel multicancer clusters among the families of azoospermic and severely oligozoospermic men. Distinct overall multicancer risk and familial multicancer patterns were observed in the azoospermia and severe oligozoospermia cohorts, suggesting heterogeneity in cancer risk by type of subfertility and within subfertility type. Describing families with similar cancer risk patterns provides a new avenue to increase homogeneity for focused gene discovery and environmental risk factor studies. Such discoveries will lead to more accurate risk predictions and improved counseling for patients and their families. STUDY FUNDING/COMPETING INTEREST(S): This work was funded by GEMS: Genomic approach to connecting Elevated germline Mutation rates with male infertility and Somatic health (Eunice Kennedy Shriver National Institute of Child Health and Human Development (NICHD): R01 HD106112). The authors have no conflicts of interest relevant to this work. TRIAL REGISTRATION NUMBER: N/A.

Deep Transcriptome Profiling of Multiple Myeloma Using Quantitative Phenotypes.

BACKGROUND: Transcriptome studies are gaining momentum in genomic epidemiology, and the need to incorporate these data in multivariable models alongside other risk factors brings demands for new approaches. METHODS: Here we describe SPECTRA, an approach to derive quantitative variables that capture the intrinsic variation in gene expression of a tissue type. We applied the SPECTRA approach to bulk RNA sequencing from malignant cells (CD138+) in patients from the Multiple Myeloma Research Foundation CoMMpass study. RESULTS: A set of 39 spectra variables were derived to represent multiple myeloma cells. We used these variables in predictive modeling to determine spectra-based risk scores for overall survival, progression-free survival, and time to treatment failure. Risk scores added predictive value beyond known clinical and expression risk factors and replicated in an external dataset. Spectrum variable S5, a significant predictor for all three outcomes, showed pre-ranked gene set enrichment for the unfolded protein response, a mechanism targeted by proteasome inhibitors which are a common first line agent in multiple myeloma treatment. We further used the 39 spectra variables in descriptive modeling, with significant associations found with tumor cytogenetics, race, gender, and age at diagnosis; factors known to influence multiple myeloma incidence or progression. CONCLUSIONS: Quantitative variables from the SPECTRA approach can predict clinical outcomes in multiple myeloma and provide a new avenue for insight into tumor differences by demographic groups. IMPACT: The SPECTRA approach provides a set of quantitative phenotypes that deeply profile a tissue and allows for more comprehensive modeling of gene expression with other risk factors.

Duo Shared Genomic Segment analysis identifies a genome-wide significant risk locus at 18q21.33 in myeloma pedigrees.

AIM: High-risk pedigrees (HRPs) are a powerful design to map highly penetrant risk genes. We previously described Shared Genomic Segment (SGS) analysis, a mapping method for single large extended pedigrees that also addresses genetic heterogeneity inherent in complex diseases. SGS identifies shared segregating chromosomal regions that may inherit in only a subset of cases. However, single large pedigrees that are individually powerful (at least 15 meioses between studied cases) are scarce. Here, we expand the SGS strategy to incorporate evidence from two extended HRPs by identifying the same segregating risk locus in both pedigrees and allowing for some relaxation in the size of each HRP. METHODS: Duo-SGS is a procedure to combine single-pedigree SGS evidence. It implements statistically rigorous duo-pedigree thresholding to determine genome-wide significance levels that account for optimization across pedigree pairs. Single-pedigree SGS identifies optimal segments shared by case subsets at each locus across the genome, with nominal significance assessed empirically. Duo-SGS combines the statistical evidence for SGS segments at the same genomic location in two pedigrees using Fisher's method. One pedigree is paired with all others and the best duo-SGS evidence at each locus across the genome is established. Genome-wide significance thresholds are determined through distribution-fitting and the Theory of Large Deviations. We applied the duoSGS strategy to eleven extended, myeloma HRPs. RESULTS: We identified one genome-wide significant region at 18q21.33 (0.85 Mb, P = 7.3 × 10-9) which contains one gene, CDH20. Thirteen regions were genome-wide suggestive: 1q42.2, 2p16.1, 3p25.2, 5q21.3, 5q31.1, 6q16.1, 6q26, 7q11.23, 12q24.31, 13q13.3, 18p11.22, 18q22.3 and 19p13.12. CONCLUSION: Our results provide novel risk loci with segregating evidence from multiple HRPs and offer compelling targets and specific segment carriers to focus a future search for functional variants involved in inherited risk formyeloma.

Shared genomic segment analysis in a large high-risk chronic lymphocytic leukemia pedigree implicates CXCR4 in inherited risk.

AIM: Chronic lymphocytic leukemia (CLL) has been shown to cluster in families. First-degree relatives of individuals with CLL have an ~8 fold increased risk of developing the malignancy. Strong heritability suggests pedigree studies will have good power to localize pathogenic genes. However, CLL is relatively rare and heterogeneous, complicating ascertainment and analyses. Our goal was to identify CLL risk loci using unique resources available in Utah and methods to address intra-familial heterogeneity. METHODS: We identified a six-generation high-risk CLL pedigree using the Utah Population Database. This pedigree contains 24 CLL cases connected by a common ancestor. We ascertained and genotyped eight CLL cases using a high-density SNP array, and then performed shared genomic segment (SGS) analysis - a method designed for extended high-risk pedigrees that accounts for heterogeneity. RESULTS: We identified a genome-wide significant region (P = 1.9 × 10-7, LOD-equivalent 5.6) at 2q22.1. The 0.9 Mb region was inherited through 26 meioses and shared by seven of the eight genotyped cases. It sits within a ~6.25 Mb locus identified in a previous linkage study of 206 small CLL families. Our narrow region intersects two genes, including CXCR4 which is highly expressed in CLL cells and implicated in maintenance and progression. CONCLUSION: SGS analysis of an extended high-risk CLL pedigree identified the most significant evidence to-date for a 0.9 Mb CLL disease locus at 2q22.1, harboring CXCR4. This discovery contributes to a growing literature implicating CXCR4 in inherited risk to CLL. Investigation of the segregating haplotype in the pedigree will be valuable for elucidating risk variant(s).

Sequencing at lymphoid neoplasm susceptibility loci maps six myeloma risk genes.

Inherited genetic risk factors play a role in multiple myeloma (MM), yet considerable missing heritability exists. Rare risk variants at genome-wide association study (GWAS) loci are a new avenue to explore. Pleiotropy between lymphoid neoplasms (LNs) has been suggested in family history and genetic studies, but no studies have interrogated sequencing for pleiotropic genes or rare risk variants. Sequencing genetically enriched cases can help discover rarer variants. We analyzed exome sequencing in familial or early-onset MM cases to identify rare, functionally relevant variants near GWAS loci for a range of LNs. A total of 149 distinct and significant LN GWAS loci have been published. We identified six recurrent, rare, potentially deleterious variants within 5 kb of significant GWAS single nucleotide polymorphisms in 75 MM cases. Mutations were observed in BTNL2, EOMES, TNFRSF13B, IRF8, ACOXL and TSPAN32. All six genes replicated in an independent set of 255 early-onset MM or familial MM or precursor cases. Expansion of our analyses to the full length of these six genes resulted in a list of 39 rare and deleterious variants, seven of which segregated in MM families. Three genes also had significant rare variant burden in 733 sporadic MM cases compared with 935 control individuals: IRF8 (P = 1.0 × 10-6), EOMES (P = 6.0 × 10-6) and BTNL2 (P = 2.1 × 10-3). Together, our results implicate six genes in MM risk, provide support for genetic pleiotropy between LN subtypes and demonstrate the utility of sequencing genetically enriched cases to identify functionally relevant variants near GWAS loci.

Family Study Designs Informed by Tumor Heterogeneity and Multi-Cancer Pleiotropies: The Power of the Utah Population Database.

BACKGROUND: Previously, family-based designs and high-risk pedigrees have illustrated value for the discovery of high- and intermediate-risk germline breast cancer susceptibility genes. However, genetic heterogeneity is a major obstacle hindering progress. New strategies and analytic approaches will be necessary to make further advances. One opportunity with the potential to address heterogeneity via improved characterization of disease is the growing availability of multisource databases. Specific to advances involving family-based designs are resources that include family structure, such as the Utah Population Database (UPDB). To illustrate the broad utility and potential power of multisource databases, we describe two different novel family-based approaches to reduce heterogeneity in the UPDB. METHODS: Our first approach focuses on using pedigree-informed breast tumor phenotypes in gene mapping. Our second approach focuses on the identification of families with similar pleiotropies. We use a novel network-inspired clustering technique to explore multi-cancer signatures for high-risk breast cancer families. RESULTS: Our first approach identifies a genome-wide significant breast cancer locus at 2q13 [P = 1.6 × 10-8, logarithm of the odds (LOD) equivalent 6.64]. In the region, IL1A and IL1B are of particular interest, key cytokine genes involved in inflammation. Our second approach identifies five multi-cancer risk patterns. These clusters include expected coaggregations (such as breast cancer with prostate cancer, ovarian cancer, and melanoma), and also identify novel patterns, including coaggregation with uterine, thyroid, and bladder cancers. CONCLUSIONS: Our results suggest pedigree-informed tumor phenotypes can map genes for breast cancer, and that various different cancer pleiotropies exist for high-risk breast cancer pedigrees. IMPACT: Both methods illustrate the potential for decreasing etiologic heterogeneity that large, population-based multisource databases can provide.See all articles in this CEBP Focus section, "Modernizing Population Science."

Harnessing Population Pedigree Data and Machine Learning Methods to Identify Patterns of Familial Bladder Cancer Risk.

BACKGROUND: Relatives of patients with bladder cancer have been shown to be at increased risk for kidney, lung, thyroid, and cervical cancer after correcting for smoking-related behaviors that may concentrate in some families. We demonstrate a novel approach to simultaneously assess risks for multiple cancers to identify distinct multicancer configurations (multiple different cancer types that cluster in relatives) surrounding patients with familial bladder cancer. METHODS: This study takes advantage of a unique population-level data resource, the Utah Population Database (UPDB), containing vast genealogy and statewide cancer data. Familial risk is measured using standardized incidence risk (SIR) ratios that account for sex, age, birth cohort, and person-years of the pedigree members. RESULTS: We identify 1,023 families with a significantly higher bladder cancer rate than population controls (familial bladder cancer). Familial SIRs are then calculated across 25 cancer types, and a weighted Gower distance with K-medoids clustering is used to identify familial multicancer configurations (FMC). We found five FMCs, each exhibiting a different pattern of cancer aggregation. Of the 25 cancer types studied, kidney and prostate cancers were most commonly enriched in the familial bladder cancer clusters. Laryngeal, lung, stomach, acute lymphocytic leukemia, Hodgkin disease, soft-tissue carcinoma, esophageal, breast, lung, uterine, thyroid, and melanoma cancers were the other cancer types with increased incidence in familial bladder cancer families. CONCLUSIONS: This study identified five familial bladder cancer FMCs showing unique risk patterns for cancers of other organs, suggesting phenotypic heterogeneity familial bladder cancer. IMPACT: FMC configurations could permit better definitions of cancer phenotypes (subtypes or multicancer) for gene discovery and environmental risk factor studies.

Elevated IgM and abnormal free light chain ratio are increased in relatives from high-risk chronic lymphocytic leukemia pedigrees.

Abnormal serum immunoglobulin (Ig) free light chains (FLC) are established biomarkers of early disease in multiple B-cell lymphoid malignancies, including chronic lymphocytic leukemia (CLL). Heavy chains have also been shown to be biomarkers in plasma cell disorders. An unanswered question is whether these Ig biomarkers are heritable, i.e., influenced by germline factors. CLL is heritable but highly heterogeneous. Heritable biomarkers could elucidate steps of disease pathogenesis that are affected by germline factors, and may help partition heterogeneity and identify genetic pleiotropies across malignancies. Relatives in CLL pedigrees present an opportunity to identify heritable biomarkers. We compared FLCs and heavy chains between relatives in 23 high-risk CLL pedigrees and population controls. Elevated IgM (eIgM) and abnormal FLC (aFLC) ratio was significantly increased in relatives, suggesting that these Ig biomarkers are heritable and could offer risk stratification in pedigree relatives. Within high-risk CLL pedigrees, B-cell lymphoid malignancies were five times more prevalent in close relatives of individuals with eIgM, prostate cancer was three times more prevalent in relatives of individuals with aFLC, and monoclonal B-cell lymphocytosis increased surrounding individuals with normal Ig levels. These different clustering patterns suggest Ig biomarkers have the potential to partition genetic heterogeneity in CLL and provide insight into distinct heritable pleiotropies associated with CLL.

Re-interpretation of PAM50 gene expression as quantitative tumor dimensions shows utility for clinical trials: application to prognosis and response to paclitaxel in breast cancer.

BACKGROUND: We recently showed PAM50 gene expression data can be represented by five quantitative, orthogonal, multi-gene breast tumor traits. These novel tumor 'dimensions' were superior to categorical intrinsic subtypes for clustering in high-risk breast cancer pedigrees, indicating potential to represent underlying genetic susceptibilities and biological pathways. Here we explore the prognostic and predictive utility of these dimensions in a sub-study of GEICAM/9906, a Phase III randomized prospective clinical trial of paclitaxel in breast cancer. METHODS: Tumor dimensions, PC1-PC5, were calculated using pre-defined coefficients. Univariable and multivariable Cox proportional hazards (PH) models for disease-free survival (DFS) were used to identify associations between quantitative dimensions and prognosis or response to the addition of paclitaxel. Results were illustrated using Kaplan-Meier curves. RESULTS: Dimensions PC1 and PC5 were associated with DFS (Cox PH p = 6.7 [Formula: see text] 10-7 and p = 0.036), remaining significant after correction for standard clinical-pathological prognostic characteristics. Both dimensions were selected in the optimal multivariable model, together with nodal status and tumor size (Cox PH p = 1.4 [Formula: see text] 10-12). Interactions with treatment were identified for PC3 and PC4. Response to paclitaxel was restricted to tumors with low PC3 and PC4 (log-rank p = 0.0021). Women with tumors high for PC3 or PC4 showed no survival advantage. CONCLUSIONS: Our proof-of-concept application of quantitative dimensions illustrated novel findings and clinical utility beyond standard clinical-pathological characteristics and categorical intrinsic subtypes for prognosis and predicting chemotherapy response. Consideration of expression data as quantitative tumor dimensions offers new potential to identify clinically important patient subsets in clinical trials and advance precision medicine.

Reparameterization of PAM50 Expression Identifies Novel Breast Tumor Dimensions and Leads to Discovery of a Genome-Wide Significant Breast Cancer Locus at 12q15.

Background: Breast tumor subtyping has failed to provide impact in susceptibility genetics. The PAM50 assay categorizes breast tumors into: Luminal A, Luminal B, HER2-enriched and Basal-like. However, tumors are often more complex than simple categorization can describe. The identification of heritable tumor characteristics has potential to decrease heterogeneity and increase power for gene finding.Methods: We used 911 sporadic breast tumors with PAM50 expression data to derive tumor dimensions using principal components (PC). Dimensions in 238 tumors from high-risk pedigrees were compared with the sporadic tumors. Proof-of-concept gene mapping, informed by tumor dimension, was performed using Shared Genomic Segment (SGS) analysis.Results: Five dimensions (PC1-5) explained the majority of the PAM50 expression variance: three captured intrinsic subtype, two were novel (PC3, PC5). All five replicated in 745 TCGA tumors. Both novel dimensions were significantly enriched in the high-risk pedigrees (intrinsic subtypes were not). SGS gene-mapping in a pedigree identified a 0.5 Mb genome-wide significant region at 12q15 This region segregated through 32 meioses to 8 breast cancer cases with extreme PC3 tumors (P = 2.6 × 10-8).Conclusions: PC analysis of PAM50 gene expression revealed multiple independent, quantitative measures of tumor diversity. These tumor dimensions show evidence for heritability and potential as powerful traits for gene mapping.Impact: Our study suggests a new approach to describe tumor expression diversity, provides new avenues for germline studies, and proposes a new breast cancer locus. Similar reparameterization of expression patterns may inform other studies attempting to model the effects of tumor heterogeneity. Cancer Epidemiol Biomarkers Prev; 27(6); 644-52. ©2018 AACR.

Novel pedigree analysis implicates DNA repair and chromatin remodeling in multiple myeloma risk.

The high-risk pedigree (HRP) design is an established strategy to discover rare, highly-penetrant, Mendelian-like causal variants. Its success, however, in complex traits has been modest, largely due to challenges of genetic heterogeneity and complex inheritance models. We describe a HRP strategy that addresses intra-familial heterogeneity, and identifies inherited segments important for mapping regulatory risk. We apply this new Shared Genomic Segment (SGS) method in 11 extended, Utah, multiple myeloma (MM) HRPs, and subsequent exome sequencing in SGS regions of interest in 1063 MM / MGUS (monoclonal gammopathy of undetermined significance-a precursor to MM) cases and 964 controls from a jointly-called collaborative resource, including cases from the initial 11 HRPs. One genome-wide significant 1.8 Mb shared segment was found at 6q16. Exome sequencing in this region revealed predicted deleterious variants in USP45 (p.Gln691* and p.Gln621Glu), a gene known to influence DNA repair through endonuclease regulation. Additionally, a 1.2 Mb segment at 1p36.11 is inherited in two Utah HRPs, with coding variants identified in ARID1A (p.Ser90Gly and p.Met890Val), a key gene in the SWI/SNF chromatin remodeling complex. Our results provide compelling statistical and genetic evidence for segregating risk variants for MM. In addition, we demonstrate a novel strategy to use large HRPs for risk-variant discovery more generally in complex traits.