Pathogenic germline variants in patients with features of hereditary renal cell carcinoma: Evidence for further locus heterogeneity

Inherited renal cell carcinoma (RCC) is associated with multiple familial cancer syndromes but most individuals with features of non‐syndromic inherited RCC do not harbor variants in the most commonly tested renal cancer predisposition genes (CPGs). We investigated whether undiagnosed cases might harbor mutations in CPGs that are not routinely tested for by testing 118 individuals with features suggestive of inherited RCC (family history of RCC, two or more primary RCC aged <60 years, or early onset RCC ≤46 years) for the presence of pathogenic variants in a large panel of CPGs. All individuals had been prescreened for pathogenic variants in the major RCC genes. We detected pathogenic or likely pathogenic (P/LP) variants of potential clinical relevance in 16.1% (19/118) of individuals, including P/LP variants in BRIP1 (n = 4), CHEK2 (n = 3), MITF (n = 1), and BRCA1 (n = 1). Though the power to detect rare variants was limited by sample size the frequency of truncating variants in BRIP1, 4/118, was significantly higher than in controls (P = 5.92E‐03). These findings suggest that the application of genetic testing for larger inherited cancer gene panels in patients with indicators of a potential inherited RCC can increase the diagnostic yield for P/LP variants. However, the clinical utility of such a diagnostic strategy requires validation and further evaluation and in particular, confirmation of rarer RCC genotype‐phenotype associations is required.

tested for by testing 118 individuals with features suggestive of inherited RCC (family history of RCC, two or more primary RCC aged <60 years, or early onset RCC ≤46 years) for the presence of pathogenic variants in a large panel of CPGs.All individuals had been prescreened for pathogenic variants in the major RCC genes.We detected pathogenic or likely pathogenic (P/LP) variants of potential clinical relevance in 16.1% (19/118) of individuals, including P/LP variants in BRIP1 (n = 4), CHEK2 (n = 3), MITF (n = 1), and BRCA1 (n = 1).Though the power to detect rare variants was limited by sample size the frequency of truncating variants in BRIP1, 4/118, was significantly higher than in controls (P = 5.92E-03).These findings suggest that the application of genetic testing for larger inherited cancer gene panels in patients with indicators of a potential inherited RCC can increase the diagnostic yield for P/LP variants.However, the clinical utility of such a diagnostic strategy requires validation and further evaluation and in particular, confirmation of rarer RCC genotype-phenotype associations is required.

| INTRODUCTION
Renal cell carcinoma (RCC) is a group of human cancers derived from renal epithelium that comprise a variety of histological and genetic backgrounds.Worldwide, RCCs account for around 2.4% of all malignancies, with a prevalence of about 4.4 per 100 000 individuals and a cumulative lifetime risk (to age 75 years) of approximately 0.5%. 1 Molecular genetic studies have identified multiple genetic causes for RCC predisposition.The best recognized cause of familial RCC is the dominantly inherited familial cancer syndrome von Hippel-Lindau (VHL) disease caused by germline mutations in the VHL tumor suppressor gene. 2,3Inactivating mutations in a number of tumor suppressor genes including VHL, FH, FLCN, SDHB, and BAP1, activating mutations in the MET proto-oncogene and constitutional chromosome 3 translocations are well established causes of inherited predisposition to renal cancers. 4Though it has been suggested that 24% to 33% of individuals with RCC may meet referral criteria for genetic testing, 5 the majority of patients who undergo routine genetic testing for germline variants in the "major inherited RCC genes" (ie, .VHL, FH, FLCN, SDHB, BAP1, MET) do not have detectable pathogenic variants (unpublished observations).
Recently, studies in a number of different human cancer types have identified pathogenic variants in a wider range of cancer predisposition genes (CPGs) than have been traditionally associated with the cancer of interest. 6,7In addition, germline genetic testing of a cohort of individuals with advanced RCC revealed 16% of individuals presented with a pathogenic cancer-associated germline variant, of which only about a third occurred in the widely recognized RCCassociated genes. 8We hypothesized that applying a wider CPG testing strategy to a cohort of affected individuals with features of inherited RCC might increase the diagnostic yield of pathogenic/likely pathogenic (P/LP) variants and we proceeded to investigate a large panel of CPGs in 118 unrelated probands pre-screened for germline mutations in VHL, MET, FLCN, SDHB, FH, and BAP1.

| Subjects
Individuals diagnosed with RCC referred to Regional Genetics Centres for consideration of genetic testing were assessed for eligibility based on the presence of clinical features associated with inherited RCC.Individuals were recruited if they matched one or more of the following criteria: (a) At least one first or second degree relative with RCC, (b) no family history of RCC but two or more separate primary RCC before age 60 years, or (c) diagnosed with RCC at age 45 years or less.
Assignment of groupings based on clinical criteria was carried out hierarchically in the order given, where, for example, a patient with bilateral RCC aged under 45 years with a family history of RCC would be categorized as familial and a patient with bilateral RCC aged under 45 years without a family history of RCC would be categorized as multiple RCC.For four individuals in whom the precise age at diagnosis of RCC was not available the age at genetic testing was used.Individuals with confirmed or likely pathogenic variants in BAP1, FH, FLCN, MET, SDHB and VHL were excluded from the study.All study participants gave written informed consent, and the study was approved by the South Birmingham Research Ethics Committee.

| Molecular genetics studies
DNA was extracted from peripheral blood lymphocytes in a regional genetics laboratory using standard techniques.A total of 100 samples

| Bioinformatics
Further details of bioinformatic protocols and methodology can be found in the Appendix S1.FASTQ files for both case and ICR1000UK exomes 9 were aligned to genome reference GRCh38 using BWA-MEM (version 0.7.15-r1140) with ALT-contig post-processing.PCR duplicates were flagged by SAMtools rmdup (version 1.4.1) and variant calling carried out using GATK unified genotyper (version 3.7-0-gcfedb67).Variants from targeted sequencing panel and exome datasets were called independently and a "virtual" panel applied to the exome variants via vcftools, restricting the reported variants to the Illumina TruSight Cancer sequencing panel target bed intervals (with an additional 3 bp paddingAppendix S1).Full alignment and variant calling pipeline provided in Appendix S1.VCF files were filtered to remove low quality calls and sequencing artefacts using vcftools and in-house bioinformatics pipelines (supplementary Table S1).Lastly, genomic regions were restricted to a total of 67 cancer-related genes sub-selected from the original cancer gene panel as utilized previously 7 which were targeted on the Illumina TruSight Cancer sequencing panel.
In addition, a single variant in MITF (rs149617956) was also assessed in conjunction with the previously described genes due to previous associations with RCC risk 10 (Supplementary Table S2, S3).
Variants passing quality filtering were annotated with ANNOVAR to provide genomic region annotation, variant consequence, functional in silico prediction, reference minor allele frequencies for datasets of 1000 genomes project (1KG) and Exome Aggregation Consortium (ExAC), 11 and reported ClinVar data, where available.Variants were selected by variant consequence, filtered to be rarer than 1% (minor AF < 0.01) in both 1KG and ExAC, in order to exclude common SNPs.In silico predictive metrics provided by ANNOVAR were used to inform potential pathogenicity but were not used as filtering cutoffs for candidate selection.ACMG guidelines 12,13 were applied to all candidate variants to determine clinical significance utilizing InterVar (version 20 180 827).Somatic variant calling was performed jointly using both Strelka2 (version 2.9.10) and Mutect2 (version 3.7-0-gcfedb67) with annotation performed as described for germline variant calls.The data that support the findings of this study are available on request from the corresponding author.The data are not publicly available due to privacy or ethical restrictions.
Structural variant calling was performed using SvABA (version 1.1.3) 14to identify any large indels or structural variants within the same genomic regions described for SNV calling.Full details of structural variant calling process are described in the Appendix S1.

| Statistical analysis
Proportion confidence intervals were calculated using R base function binom.test at CI 95%, Odds ratios were calculated using the oddsratio.fisherfunction in epitools package (version 0.5-10), and two-tailed Fisher's exact tests were calculated using the fisher.test function in base, using R (version 3.5.1).and comprised of 68.6% clear cell RCC, 27.1% papillary RCC, and 4.29% chromophobe RCC.Summary of the distribution of clinical features are given in Table 1 (full details in Supplementary Table S6).

| Variant filtering
A total of 1955 and 237 variants passed quality control filtering requirements (Appendix S1) in the targeted sequencing and virtual panel sets, respectively.After variant filtering (Appendix S1), a total of 159 variants were retained from the targeted sequencing and 25 variants were retained from the virtual panel sets, respectively.Variants present in both sets were merged resulting in a total of 174 variants across the targeted regions.
Analysis of the pathogenic or likely pathogenic (P/LP) variants identified in this set were divided into three categories subpanels based on the clinical associations and inheritance patterns of the affected genes: (a) Category I genes (n = 14) had a known association with syndromic or non-syndromic RCC predisposition, (b) Category II genes (n = 18) were those in which heterozygous pathogenic variants are known to be associated with predisposition to multiple tumor types, and (c) Category III genes (n = 35) which are associated with cancer predisposition when there are biallelic pathogenic variants or those which have been associated with a single non-RCC tumor phenotype.List of targeted genomic regions are listed in supplemental information Table S2.
Of the 174 variants assessed, 16 were classified as pathogenic or likely pathogenic (P-LP) variants (three pathogenic, 13 likely pathogenic), corresponding to four nonsense variants, three frameshift deletions, one frameshift insertions, and eight nonsynonymous substitutions.The 16 variants were observed in 19 cases (16.1%; 95% CI: 9.98-23.0).P/LP variants were equally distributed by count across the inherited subtypes (nine variants in familial, six variants in early onset, and four variants in bilateral/multifocal).All 16 P/LP variants are described in Table 2 and all 19 patients harboring the aforementioned variants in Table 3.
Structural variant calling, performed using SvABA, did not identify any structural variants which passed quality control in the targeted genomic regions.

| Detection of variants in category I: RCC predisposition genes
As expected, no P/LP variants were detected in genes that had previously been analyzed before inclusion in this study (VHL, MET, FLCN, T A B L E 2 Sixteen variants identified as pathogenic or likely pathogenic by ACMG guideline classifications assigned by InterVar Sequencing in the individual's unaffected brother did not reveal the variant.Though this variant had been previously associated with predisposition to RCC and melanoma 10 there was no reported family history of melanoma.

| Detection of variants in category II: Multisite CPGs
Six distinct P/LP variants in three genes in which heterozygous pathogenic variants are known to be associated with predisposition  cancer, found at a rate of 0.19% (28/14526) (Supplementary Table S5).
Ser422Valfs*15) was identified in two individuals, both of whom presented with multifocal RCC in their fifth decade.The frameshift deletion is considered to be pathogenic and has previously been detected in both germline sequencing of breast 17 and prostate cancer. 18,19An additional CHEK2 nonsynonymous variant (NM_007194.The variant occurs as a singleton in the gnomAD data set 15 and is considered to be highly deleterious by multiple in silico predictive tools.

| Analysis of tumors from cases with germline BRIP1 truncating variants
Pathology blocks from RCC from two related patients with a truncating BRIP1 variant (BRIP1 NM_032043: c.2698G>A: p.Arg798*) were available for analysis.The proband (RCC-102) presented with a 63 mm RCC at age 39 years.Histopathological review revealed that the tumor contained some sheets of cells with clear cytoplasm, in keeping with classification as a clear cell RCC.However, in many areas the tumor showed very variable morphology, with a tubulo-papillary architecture and areas where the cells had very abundant eosinophilic cytoplasm (see Figure S2).The tumor cell nuclei were predominantly WHO/ISUP grade 2, but some were interpreted as grade 3.There was no sarcomatoid or rhabdoid morphology.There was no tumor necrosis but there was a marked infiltrate of chronic inflammatory cells within the tumor, including lymphocytes and macrophages.

| DISCUSSION
We analyzed a cohort of 118 individuals with clinical characteristics suggestive of inherited RCC (but no known genetic cause) for germline variants in 68 cancer-related genes.This gene panel strategy was previously used to analyse a large cohort of patients with multiple primary tumors and in that study we found that there was a significant diagnostic yield of P/LP variants in CPGs for which the tumor phenotype in the relevant patient was atypical. 7e only pathogenic variant identified in a category I gene was a previously described nonsynonymous variant in MITF (c.952G>A: p. E318K).The E318K variant was linked to non-syndromic RCC predisposition in a cohort of individuals presenting with both RCC and melanoma in which variant carriers demonstrated a 5-fold increased risk for melanoma, RCC, or both and functional studies demonstrated MITF upregulation and differential expression of MITF target transcripts. 10,23Subsequently the E318K variant was confirmed to be associated with a melanoma predisposition 24,25 ; however, the association of MITF E318K with RCC predisposition has been less well studied and provides limited support for the association between RCC predisposition and MITF E318K. 26,27In this instance, the identification of MITF E318K in this cohort is difficult to interpret given the limited sample sizes and the identification of only a single carrier.The category I genes also included rarer RCC cancer predisposition genes such as CDC73, PTEN, TSC1, and TSC2 that have been linked to syndromic forms of inherited RCC and we did not identify pathogenic variants in these genes (the cohort had been ascertained via clinical geneticists and so we would have expected syndromic cases to have been identified prior to recruitment).
Several VUS variants were identified in TSC2 and MET.Three variants in MET (NM_000245: c.T2543C: p.V848A, NM_000245: c.G1406C: p.R469P, and NM_000245: c.A1336G: p.I446V) were present at allelic frequencies lower than 8.5E-05, with in silico predictions being variable, but none of the variants fall within the tyrosine kinase domain associated with constitutional activation of c-MET, 28,29 and none had been reported as somatic events in sporadic RCC based on data from the catalogue of somatic mutations in cancer (COSMIC). 30x missense variants were identified in TSC2, associated with tuberous sclerosis complex (MIM: 613254) which predisposes individuals to renal angiomyolipomas and cysts, as well as hybrid or oncocytic RCC in between 2% and 4% of cases. 31,32Histological information was not available for these individuals to assess if they presented with histologies consistent with loss of TSC2.The predicted pathogenicity of these missense variants, as well as the allele rarity, is variable but two variants (NM_000548 c.G4657T: p.G1553C & NM_000548: c.G5117A: p.R1706H) occur within the Rap GTPase activating protein domain implicated in RHEB inhibition 33 and one variant (NM_000548: c.C2476A: p.L826M) arises in a Tuberin-type domain, though its direct function is not known.None of the 6 variants identified in TSC2 had been reported as somatic events in sporadic RCC in COSMIC.All VUS variants are detailed in the Appendix S1.
Previously, segregation analysis of non-syndromic familial RCC was found to be consistent with an autosomal dominant inheritance pattern with incomplete penetrance. 34Together with recent findings that the cancer phenotype of well-established cancer predisposition genes may be wider than initially recognized 6,7,35,36 this raised the possibility that we might find pathogenic variants in category II genes in individuals with features of inherited RCC.We identified pathogenic variants in three category II genes (BRIP1, BRCA1, CHEK2) in 6.8% (8/118 probands), of our cohort (6.8% of familial cases, 9.1% of multi/bilateral cases and 2.3% of early onset cases).Four probands harbored truncating mutations in BRIP1.Pathogenic BRIP1 variants were initially reported to predispose individuals to both breast and ovarian cancers, 37,38 though more recent evidence from epidemiological studies of pathogenic BRIP1 variants in breast cancer have found no association with breast cancer susceptibility. 16,39We note that the potential link between RCC predisposition and pathogenic BRIP1 variants would be strengthened if any of the rare non-truncating BRIP1 variants identified in probands were to be proved to be pathogenic.(n = 1) or BRCA2 (n = 3) germline variants. 41However, in BRCA1 and BRCA2 mutation carriers ascertained through a personal and/or family history of breast/ovarian cancer the risk of RCC had not been reported to be increased. 42ile some inherited RCC cases are caused by genes (eg, VHL, MET, BAP1) which show high somatic alteration rates in sporadic RCC, others inherited RCC genes (eg, FLCN and SDHB) do not display frequent somatic alteration in sporadic RCC.BRCA1, CHEK2, or BRIP1 are frequently somatically altered in sporadic RCC in the TCGA dataset (1.2%, 1%, and 1.1%, respectively) at a rate that would be indicative of common somatic driver genes. 43However, BAP1, BRCA1, BRCA2, BRIP1, and CHEK2 gene products have related functions in DNA repair pathways that may make a common role in RCC predisposition more plausible.BAP1 (BRCA1 associated protein) was originally identified due to direct interactions with the RING finger domain of BRCA1 and functions as a de-ubiquitinating enzyme.BAP1 has been determined to form multiple protein complexes and known functions include removal ubiquitin groups from histone H2A lysine 119 residues to regulate gene expression, 44 modulation of DNA damage repair by de-ubiquitinating BARD1 (which binds to BRCA1), indirectly modulating the efficacy of BRCA1-driven DNA repair pathways, 45 and mediates the recruitment of homologous recombination proteins to DNA damage foci. 46Given the interconnected functions and pathways associated with CHEK2, BRIP1, and BRCA1, it can be hypothesized that germline pathogenic variants in these gene might predispose to a broader range of cancers in a manner similar to that described with BAP1 predisposition syndrome, including RCC.Two P-LP variants were also identified in PMS2 across four individuals though the truncating variant present in three of these individuals occurs in a noncanonical isoform, annotated as an intronic substitution.The PMS2 variants in this study were not independently confirmed and known issues regarding PMS2 pseudogenes 47 make interpretation more complex.
The observation that eight of the nine genes identified in this study with P or LP variants are associated with DNA repair pathways in some capacity could suggest a potential enrichment across all DNA repair pathways but interpretation should be cautious given that cancer panels are biased towards DNA repair pathway components due to frequent alterations in somatic sequencing, and several of these genes only result in cancer presentation under autosomal recessive inheritance, which was not demonstrated here.
Epidemiological studies have reported multiple risk factors including smoking, obesity and hypertension 48,49 but these features were not reported in this study.An interesting further examination of the results described herein is the relationship between what appear to be generalized cancer predisposition genes, or at least rare causes of cancers outside of the canonical cancer spectrum, and impact of environmental factors in the resulting genotype-phenotype correlations.
In summary we found that in a cohort of patients with features associated with inherited predisposition to RCC and no detectable mutation in routinely tested RCC CPGs, extension of testing to a larger CPG panel revealed pathogenic variants in CPGs associated with multiple cancer types in a subset of patients.This finding is consistent with previous studies of patients with early onset or advanced RCC that have been analyzed by larger gene panels and with the results of patients with multiple primary tumors 7,8,43 If patients with germline mutations in DNA repair genes, such as BRCA1, BRCA2, BRIP1, and CHEK2 were shown to benefit from specific targeted therapies there would be a clear case for incorporating a wider genetic testing protocol into clinical care.However, we suggest that before the implementation into routine clinical practice of wider CPG testing for patients with potential non-syndromic inherited RCC proceeds, further studies are required to establish a causal link between RCC predisposition and pathogenic variants in BRIP1, BRCA1, BRCA2, and CHEK2 and to determine more accurately renal cancer risks in patients with pathogenic variants in these genes so that appropriate renal screening protocols for asymptomatic gene carriers can be defined.

ACKNOWLEDGEMENTS
were analyzed using Illumina TruSight Cancer Sequencing Panel (Illumina, San Diego, CA) on the Illumina MiSeq platform.Seventy five probands (18 of whom were also analyzed by the Illumina TruSight Cancer Sequencing Panel) had exome sequencing data generated by Illumina TruSeq Exome library preparation on the Illumina HiSeq 4000 or Illumina NextSeq platform.In total 118 probands were analyzed by panel and/or exome sequencing.
The 118 unrelated individuals with RCC eligible for inclusion were subdivided into three clinical subsets: 44 cases with a positive family history and 74 sporadic cases comprising 30 cases with multifocal or bilateral disease and 44 cases with early onset RCC only).Median age of onset across all cases was 42 years (range 10-74) and 52 years (range 29-74) in the familial cases, 48 years (range 31-72) in multifocal/bilateral cases and 33 years (range 10-46) in early onset cases).Histological subtype was available for 70 of 118 cases (59.3%) BAP1) and only a single P/LP variant was identified in a gene previously linked to RCC: a MITF nonsynonymous variant in (NM_000248.3:c.952G>A: p.E318K) was identified in an individual who presented with clear cell RCC at age 74 years and whose son was reported to have presented with clear cell RCC at age 53 years.
4: c.1427C>T: p.Thr476Met) was also identified in one individual with nonspecified RCC at 58 years and had a reported family history.The variant falls within the protein kinase domain of CHEK2 and in vitro studies had reported loss of kinase activity and loss of DNA repair function. 20,21A single individual with early onset papillary RCC at age 40 years was found to carry a BRCA1 frameshift deletion in exon 15 (NM_007300.3:c.4563delA: p.Lys1521Asnfs*5), which was absent in the noncancer gnomAD data set.A PMS2 nonsense variant was identified in three individuals, purported to occur within the fourth amino acid (NM_001322015: c.11C>G: p.Ser4*) but on review was found only to affect noncanonical isoform 14, resulting in an intronic substitution within the canonical isoforms of PMS2.Furthermore, one individual was identified with a PMS2 nonsynonymous variant, occurring within the canonical transcript (NM_000535: c.2066C>T: p.Thr689Ile).The PMS2 nonsynonymous substitution occurs within exon 12 resulting a threonine to isoleucine substitution in a c-terminal dimerization domain.

Only a single additional
variant in BRIP1 (NM_032043: c.C1207T: p. R403W) was identified as at least a VUS.This variant was enriched in comparison to the gnomAD noncancer population (P = 7.0E-04), but singleton variants in lower sample sizes are more difficult to accurately assess.A recent study assessed the functional impact of several novel or rare nonsynonymous variants40 and, though none of these variants were present in our cohort, it highlights the potential for nontruncating variants to contribute to cancer predisposition and need for thorough functional evaluation of variants of uncertain significance.Pathogenic variants were also detected in two other DNA repair genes, BRCA1 (n = 1) and CHEK2 (n = 3 This research was funded by the European Research Council (Advanced Researcher Award), NIHR (Senior Investigator Award and Cambridge NIHR Biomedical Research Centre), Cancer Research UK Cambridge Cancer Centre and National Institute for Health Research (NIHR, grant number RG65966).We thank NIHR BioResource volunteers for their participation, and gratefully acknowledge NIHR Bio-Resource centres, NHS Trusts and staff for their contribution.We thank the National Institute for Health Research and NHS Blood and Transplant.The views expressed are those of the authors and not necessarily those of the NHS, NIHR or Department of Health.The University of Cambridge has received salary support in respect of EM from the NHS in the East of England through the Clinical Academic Reserve.ERW is supported by the Manchester NIHR Biomedical Research Centre (IS-BRC-1215-20007).We acknowledge support from the NIHR UK Rare Genetic Disease Research Consortium.This work made use of data and samples generated by the 1958 Birth Cohort (NCDS), which is managed by the Centre for Longitudinal Studies at the UCL Institute of Education, funded by the Economic and Social Research Council (grant number ES/M001660/1).Access to these resources was enabled via the Wellcome Trust & MRC: 58FORWARDS grant [108439/Z/15/Z] (The 1958 Birth Cohort: Fostering new Opportunities for Research via Wider Access to Reliable Summary of clinical features of individuals with suspected inherited RCC where available Abbreviation: RCC, renal cell carcinoma.SMITH ET AL. Nineteen RCC samples carrying variants identified as pathogenic or likely pathogenic by ACMG guideline classifications assigned by InterVar Statistical association of truncating variant carrier status between the case set 1958 birth control, gnomAD exomes, and ExAC non-TCGA (Fisher's exact test with FDR correction) heterozygous truncating variants in BRIP1 (two cases with NM_032043.3:c.1871C>A:p.Ser624*, and one each with NM_032043.3:c.1161dupA:p.Gln388Thrfs*7, and NM_032043.3:c.2392C>T:p.Arg798*) (Supplementary Table S4).The four probands consisted of two familial cases and two multifocal/bilateral cases.Age at diagnosis of RCC was 54, 64, 46, and 39 years and these patients presented with papillary, two nonspecified, and clear cell RCC, respectively (Table 3; Individuals RCC-043, RCC-074, RCC-031, RCC-102).DNA from an affected family member (second-degree relative) was available for one of the familial cases (RCC-102) and the affected relative (who developed clear cell RCC at age 57 years) harbored the BRIP1 nonsense variant (NM_032043.3:c.2698G>A: p.Arg798*) identified in the proband (see Supplementary Figure S1).39.3 and P = 1.09E-04,OR = 17.4,95% CI: 4.61-46.3).This association is still present in ExAC non-TCGA and gnomAD exome datasets after false discovery rate correction (Table 4).Finally, statistical comparison to data published by Easton et al 16 also demonstrated a statistical enrichment in this series (P = 1.21E-04,OR = 18.2, 95% CI: 4.55-53.1)when compared to truncating variants in BRIP1 in breast T A B L E 3 T A B L E 4 7.While in this study we did not demonstrate statistical enrichment of CHEK2 P-LP variants in our cohort of individuals with features of inherited RCC, joint assessment of the frequency of P-LP variants in CHEK2 in this case series and our cohort of individuals with multiple primary tumor-associated RCC that we analyzed with a similar cancer predisposition gene panel strategy,7