Accepted Manuscript Title: Detection of Somatic Variants in Peripheral Blood Lymphocytes Using a Next Generation Sequencing Multigene Pan Cancer Panel Author: Bradford Coffee, Hannah C. Cox, John Kidd, Scott Sizemore, Krystal Brown, Susan Manley, Debora Mancini-DiNardo PII: DOI: Reference:
S2210-7762(17)30013-3 http://dx.doi.org/doi: 10.1016/j.cancergen.2017.01.002 CGEN 513
To appear in:
Received date: Revised date: Accepted date:
13-7-2016 21-12-2016 6-1-2017
Please cite this article as: Bradford Coffee, Hannah C. Cox, John Kidd, Scott Sizemore, Krystal Brown, Susan Manley, Debora Mancini-DiNardo, Detection of Somatic Variants in Peripheral Blood Lymphocytes Using a Next Generation Sequencing Multigene Pan Cancer Panel, Cancer Genetics (2017), http://dx.doi.org/doi: 10.1016/j.cancergen.2017.01.002. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Detection of Somatic Variants in Peripheral Blood Lymphocytes Using a Next Generation Sequencing Multigene Pan Cancer Panel Bradford Coffee, PhD,a Hannah C. Cox, PhD,a John Kidd, MS,a Scott Sizemore, BA,a Krystal Brown, PhD,a Susan Manley, MS, CGC, MBA,a Debora Mancini-DiNardo, PhDa a
Myriad Genetic Laboratories, Inc., 320 Wakara Way, Salt Lake City, Utah, USA 84106
Bradford Coffee, PhD: [email protected]
Hannah C. Cox, PhD: [email protected]
John Kidd, MS: [email protected]
Scott Sizemore, BA: [email protected]
Krystal Brown, PhD: [email protected]
Susan Manley, MS, CGC, MBA: [email protected]
Debora Mancini-DiNardo, PhD: [email protected]
Corresponding Author: Bradford Coffee, PhD 320 Wakara Way Salt Lake City, UT 84106 Email: [email protected]
Phone: 801-584-3080 Word Count: 1738
Page 1 of 13
Next generation sequencing detects germline and somatic variants.
Pathogenic variants (PVs) evaluated in individuals undergoing 25-gene panel testing.
Likely somatic PVs detected in 0.06% of individuals, mostly in TP53, CHEK2, and ATM.
Likely somatic PVs accounted for 40% of PVs in TP53; <1% of PVs in most other genes.
Likely somatic PVs more frequently identified in older individuals (p<0.001).
Abstract Next Generation Sequencing (NGS) multigene panels, which are routinely used to assess hereditary cancer risk, can detect both inherited germline variants and somatic variants in cancer-risk genes. We evaluated the frequency and distribution of likely somatic pathogenic and likely pathogenic variants (PVs) detected in >220,000 individuals who underwent clinical testing with a 25-gene panel between September 2013 and March 2016. Likely somatic PVs are defined as variants with NGS read frequencies from 10%-30%. Overall, 137 (0.06%) individuals were identified as carrying likely somatic PVs, most commonly in TP53 (73), CHEK2 (27), and ATM (20). Among this group, a second PV with a NGS read frequency consistent with a germline variant within the same gene or a different gene on the panel was detected in 21 individuals (15.3%), which is similar to the detection rate in our general testing population. Likely somatic PVs accounted for 38.8% of all PVs in TP53. In comparison, likely somatic PVs accounted for <1% of PVs in most other genes. Likely somatic PVs were more frequently identified in older individuals (p<0.001). Additional studies are ongoing to further investigate the incidence and clinical implications of somatic variants, enabling the appropriate medical management for these patients. Keywords: Somatic Variants, Next-Generation Sequencing, Cancer, TP53, ATM, CHEK2
Page 2 of 13
Introduction Next Generation Sequencing (NGS) multigene panels are used to identify germline pathogenic variants (PVs) in cancer susceptibility genes . However, it is also possible for somatic variants to be identified in these same genes. While a portion of these findings may be attributed to classic mosaicism (i.e. acquired during embryonic development and segregating to multiple tissues), other explanations include hematologic disease [2, 3], circulating tumor cells, and clonal hematopoiesis associated with aging [4-7]. Somatic variants are more readily detected with NGS, compared to Sanger sequencing. Importantly, NGS quantitatively determines the proportion of variant bases in a sample down to low read frequencies. Inherited germline mutations have NGS read frequencies that are approximately 50% of the total number of reads of a given position in a DNA sequence. NGS can detect significant deviations from this 50% read frequency, which indicates that a variant is not present in the germline. With the increased use of NGS in hereditary cancer testing, somatic variants will be more frequently detected and may complicate the interpretation of test results. Medical management recommendations from professional societies, such as the National Comprehensive Cancer Network (NCCN), are based on the well-studied risks associated with pathogenic germline variants in cancer risk genes. For example, germline PVs in TP53 are diagnostic of Li-Fraumeni syndrome, which is associated with a high life-time risk of cancer and early onset breast cancer . As such, NCCN guidelines recommend that individuals with a germline PV in TP53 receive increased breast screening and possibly risk-reducing mastectomy. Given the increased demand for genetic testing, it is critical that somatic mosaic variants identified during clinical genetic testing be investigated in order to ensure appropriate patient care and family testing. Here, we evaluate variants with allele frequencies significantly lower than expected for germline variants that were identified during testing with a 25gene hereditary cancer panel.
Page 3 of 13
Materials and Methods This analysis includes 222,098 consecutive individuals who underwent clinical genetic testing with a 25-gene hereditary cancer panel that includes APC, ATM, BARD1, BMPR1A, BRCA1, BRCA2, BRIP1, CDH1, CDK4, CDKN2A, CHEK2, EPCAM, MLH1, MSH2, MSH6, MUTYH, NBN, PALB2, PMS2, PTEN, RAD51C, RAD51D, SMAD4, STK11, and TP53. Sequencing and large rearrangement analyses were performed for all genes on the panel, except EPCAM (large rearrangement only). Tested samples included blood (n=199,550), saliva/buccal (n=20,431), or a different/unspecified (n=2,117) sample type. Individuals were ascertained based on clinical suspicion of hereditary cancer risk. PVs are those variants that receive a laboratory classification of Pathogenic or Likely Pathogenic. Variants with read frequencies between 30% and 70% were considered germline in the absence of contrary evidence. This broad frequency range was selected in order to ensure that all germline variants are identified and is seven standard deviations from the mean frequency of all detected variants . Variants that are likely somatic are conservatively defined here as those with read frequencies between 10% and 30%, where 10% reflects the validated lower limit for this assay . These likely somatic variants represent a substantial allelic imbalance and are highly unlikely to be inherited. All likely somatic variants were confirmed with Sanger sequencing with alternate PCR primers and/or repeat NGS sequencing to rule out allelic imbalances caused by a technical artifact. Variants observed with read frequencies <10% are below the laboratory reporting threshold and are not called or classified for pathogenicity. Samples from individuals with known hematologic malignancies were excluded from testing. Clinical information was obtained from provider completed test request forms. Fisher’s Exact tests were performed to determine the differences between the prevalence of low-level PVs by age.
Page 4 of 13
False Discovery Rate controlling adjustments were used for multiple comparisons. Adjusted p-values <0.05 were considered statistically significant.
Results Overall, 137/222,098 (0.06%) individuals were identified as carrying a likely somatic PV, which accounts for 0.71% of all PVs identified during panel testing (Table 1). This includes 120 (0.06%) of patients who submitted a blood sample and 12 (0.06%) of patients who submitted a saliva/buccal sample for testing. Somatic variants were identified in 5 patients with an unspecified sample type. A second, germline PV was identified in 21 (15.3%) individuals with a likely somatic PV. Three of the 21 individuals had likely somatic and germline PVs in the same gene (TP53-1, ATM-2). Overall, 86.2% of patients with a somatic PV in any gene had a personal history of cancer at the time of testing, which is higher than that observed for all other tested patients (43.9%). The largest proportion of likely somatic PVs was identified in TP53 (73), which accounted for 38.8% of all PVs identified in TP53 (Table 1). After TP53, likely somatic PVs were most common in ATM (20) and CHEK2 (27). Unlike TP53, these variants accounted for only 1.36% and 1.47% of all PVs in ATM and CHEK2, respectively (Table 1). The remaining likely somatic PVs were identified in 9 other genes and accounted for <1% of the total PVs in most genes. Likely somatic variants accounted for a higher proportion of the PVs identified in CDH1 and STK11, which is likely due to the relatively small number of total PVs (<100) identified in those genes. No likely somatic variants were identified in the other genes with <100 PVs. The mean age at testing for individuals with only a likely somatic PV (59.3 years) was older than those with only germline PVs (49.1 years) and those with no PV (48.3 years). The incidence of likely somatic PVs is stratified by age at testing in Figure 1, and shows that the frequency increases with age. This trend was most pronounced for TP53, ATM, and CHEK2, which account for 86.3% of all likely
Page 5 of 13
somatic PVs identified here. Individuals who underwent panel testing over the age of 60 had a statistically higher probability of carrying a somatic PV than those tested before the age of 60 (p<0.001). Although this is weighted by the proportion of likely somatic PVs in TP53, ATM, and CHEK2, statistical significance was maintained even when these three genes were excluded (p=0.014). We have also observed that the NGS read frequency of a somatically acquired variant can increase over time and be similar to that of an inherited germline mutation. For example, this was observed for one individual who was tested twice over a span of 3 months. In the first blood sample, the NGS read frequency for a PV in TP53 was approximately 30%. When a second blood sample was tested 3 months later, the NGS read frequency had increased to about 45%, indicating that this variant was somatically acquired. Notably, this individual did not have a personal/family history consistent with LiFraumeni syndrome. Follow-up testing was performed on skin punch samples from several other individuals who have a TP53 PV with a NGS read frequency consistent with germline PV. The PV was absent in the skin from these individuals, indicating that it was somatically acquired (data not shown). Given the frequency of somatic PVs in TP53, all individuals with apparent germline PVs in TP53 are now offered follow-up testing as part of clinical testing. This may include testing of a fibroblast sample obtained from a skin punch biopsy or testing of a blood or saliva sample from a family member. Absence of the PV in fibroblasts has provided evidence that PVs in TP53 detected at allele frequencies within the expected germline range can be somatic. Presence of the variant in a family member demonstrates germline transmission and establishes a diagnosis of Li-Fraumeni syndrome. Both followup testing approaches have provided evidence that a PV is present in the germline or was somatically acquired.
Page 6 of 13
Given the significant phenotypic overlap of many hereditary cancer syndromes, NGS panels enable the simultaneous analysis of multiple genes associated with increased cancer risks [1, 10-14]. This enables medical management decisions to be informed by gene-specific guidelines based on known risks associated with germline variants. However, the quantitative nature of NGS also enables the detection of somatic variants that may complicate genetic test interpretation. The data presented here show that the detection of likely somatic PVs in clinical hereditary cancer testing is rare, occurring in 0.06% of all individuals tested. Likely somatic PVs were most common in TP53, ATM, and CHEK2. Notably, likely somatic PVs accounted for nearly 40% of all PVs identified in TP53, but an extremely small proportion of PVs identified in other genes. In addition, somatic variants were identified in saliva samples at a frequency similar to the proportion of saliva samples in our general testing population (87.1% vs. 89.7%). Saliva is a mixture of fluids from different sources and DNA extracted from saliva can variably be derived from epithelial cells, blood cells, as well as DNA from bronchial and nasal secretions, or other sources . Thus if a PV is detected in a saliva sample, it is unclear from which tissue the PV is derived. We demonstrate that individuals tested over the age of 60 have a higher probability of carrying a likely somatic PV. This is consistent with previous studies estimating that 44% of healthy individuals at 50 years of age have at least one hematopoietic stem cell that carries a randomly acquired PV in TP53 [5, 6]. A previous study of blood-derived DNA collected from healthy individuals also found that somatic variants of any pathogenicity were present in 10% of individuals older than 65 compared to 1% of individuals younger than 50 . Although this suggests that clonal hematopoiesis associated with aging is partially responsible for the somatic variants detected here, multiple mechanisms are likely at play. The cause of these somatic variants cannot be more precisely determined here as patient treatment history is unknown. Additional studies are underway to determine if the mosaicism is due to classic mosaicism or to hematologic disease or circulating tumor cells.
Page 7 of 13
The data presented here show that detection of likely somatic PVs in TP53 appears to be relatively common and read frequencies of confirmed likely somatic variants can be similar to those expected for germline findings. Given the clinical implications of LFS and the frequency of somatic PVs in TP53, it is important to verify germline PVs in this gene prior to making medical management decisions. As a result, it is now standard practice for our laboratory to offer follow-up testing to patients with apparent germline PVs in TP53. This follow-up testing has revealed that a portion of apparent germline PVs in TP53 are actually somatic in origin and additional studies are underway to further evaluate such findings. Collectively, these findings highlight the importance of clinical follow-up to conclusively determine whether a PV in TP53 has been inherited or somatically acquired.
Page 8 of 13
Acknowledgements Funding: This work was supported by Myriad Genetics, Salt Lake City, UT.
Conflict of Interest All authors are employed by Myriad Genetics and receive salary and stock options as compensation.
Page 9 of 13
1. Easton DF, Pharoah PD, Antoniou AC, et al. Gene-panel sequencing and the prediction of breastcancer risk. N Engl J Med. 2015;372(23):2243-57. 2. Malcikova J, Smardova J, Rocnova L, et al. Monoallelic and biallelic inactivation of TP53 gene in chronic lymphocytic leukemia: selection, impact on survival, and response to DNA damage. Blood. 2009;114(26):5307-14. 3. Welch JS, Ley TJ, Link DC, et al. The origin and evolution of mutations in acute myeloid leukemia. Cell. 2012;150(2):264-78. 4. Genovese G, Kahler AK, Handsaker RE, et al. Clonal hematopoiesis and blood-cancer risk inferred from blood DNA sequence. N Engl J Med. 2014;371(26):2477-87. 5. Wong TN, Ramsingh G, Young AL, et al. Role of TP53 mutations in the origin and evolution of therapy-related acute myeloid leukaemia. Nature. 2015;518(7540):552-5. 6. Jaiswal S, Fontanillas P, Flannick J, et al. Age-related clonal hematopoiesis associated with adverse outcomes. New Engl J Med. 2014;371(26):2488-98. 7. Xie M, Lu C, Wang J, McLellan MD. Age-related mutations associated with clonal hematopoietic expansion and malignancies. Nat Med. 2014;20(12):1472-8. 8. Daly M, Pilarski R, Axilbund JE, et al. Genetic/Familial High-Risk Assessment: Breast and Ovarian. Version 2.2016. NCCN Clinical Practice Guidelines in Oncology [Internet]. 2016 March 3, 2016. Available from: http://www.nccn.org/professionals/physician_gls/pdf/genetics_screening.pdf. 9. Judkins T, Leclair B, Bowles K, et al. Development and analytical validation of a 25-gene next generation sequencing panel that includes the BRCA1 and BRCA2 genes to assess hereditary cancer risk. BMC Cancer. 2015;15(1):215.
Page 10 of 13
10. Kapoor NS, Curcio LD, Blakemore CA, et al. Multigene Panel Testing Detects Equal Rates of Pathogenic BRCA1/2 Mutations and has a Higher Diagnostic Yield Compared to Limited BRCA1/2 Analysis Alone in Patients at Risk for Hereditary Breast Cancer. Ann Surg Oncol. 2015;22(10):3282-8. 11. Saam J, Arnell C, Theisen A, et al. Patients Tested at a Laboratory for Hereditary Cancer Syndromes Show an Overlap for Multiple Syndromes in Their Personal and Familial Cancer Histories. Oncology. 2015;89(5):288-93. 12. Tung N, Battelli C, Allen B, et al. Frequency of mutations in individuals with breast cancer referred for BRCA1 and BRCA2 testing using next-generation sequencing with a 25-gene panel. Cancer. 2015;121(1):25-33. 13. Tung N, Lin NU, Kidd J, et al. Frequency of Germline Mutations in 25 Cancer Susceptibility Genes in a Sequential Series of Patients With Breast Cancer. J Clin Oncol. 2016;34(13):1460-8. 14. Yurgelun MB, Allen B, Kaldate RR, et al. Identification of a Variety of Mutations in Cancer Predisposition Genes in Patients With Suspected Lynch Syndrome. Gastroenterology. 2015;149(3):604-13. 15. Kaufman E, Lamster IB. The diagnostic applications of saliva--a review. Critical Reviews in Oral Biology and Medicine. 2002;13(2):197-212.
Page 11 of 13
Figure 1. Likely somatic PVs stratified by age at testing.
Table 1. Distribution of likely somatic PVs Likely % of PVs Gene Somatic in Gene PVs (N) TP53 73 38.83% CHEK2
No likely somatic PVs were identified in the following genes: BARD1, BMPR1A, BRIP1, CDK4, CDKN2A, EPCAM, MLH1, MSH2, MUTYH, PTEN, RAD51C, RAD51D, SMAD4.
Page 12 of 13
Page 13 of 13