Abstract
Objective: The study was designed to test the feasibility that lower genital tract cytology is a compatible medium for robust germ line genetic analyses. Method: BRCA1 and/or BRCA2 gene mutational analysis was done on DNA isolated from liquid-based cervical or vaginal cytology taken from 17 consenting women (age 29–65 years) who previously had genetic counseling followed by BRACAnalysis (Myriad Genetics, Salt Lake City, UT) blood analyses. Eleven women had known mutations in either BRCA1 or BRCA2 (cases) and six had no identified mutations (controls) on entry into the study. Anonymized cytology samples were sent to Myriad Genetics with a request for testing that was limited to the degree of genomic testing previously done on the blood samples. Results: One cervicovaginal specimen from a test-positive woman had inadequate cellular content that precluded gene sequencing and therefore was excluded from this analysis. For the 16 women with adequate cytologic specimens, there was 100% concordance for BRCA mutation test results between blood and genital tract cytology (κ = 1.0; 95% confidence interval, 0.51–1.0). Conclusion: We have shown the feasibility of using liquid-based genital tract cytology as an alternative biospecimen to blood for germ line genetic analysis using a clinical approved assay. It needs to be emphasized that any type of testing for BRCA1 or BRCA2 mutation genotype should only be done in the setting of pretest and posttest counseling.
Introduction
Liquid-based cervical cytology sampling for monolayer slide preparation has recently been adopted as a more accurate means of clinical cervical cancer screening (1). The clinical analysis uses only a fraction of the total collected sample. Additional studies such as human papillomavirus testing, which may be helpful in triaging clinical management, can be done on residual sample aliquots (2). The “leftover” cervical cells suspended and preserved in methanol-based liquid Papanicolaou (Pap) smear fixative has recently been described as an excellent source of nucleic acid, which can be used in molecular studies of cervical carcinogenesis (3). This mixture of cervical, vaginal, polymorphonuclear, and lymphocytic cells has also been proposed as a viable source of constitutional DNA for germ line genetic analysis (3). The clinical demand for gene mutation analyses is rapidly increasing to address risk factors such as inherited disorders, thrombophilias, and cancer risk. Finding efficient ways to perform robust genetic testing in alternative clinical samples will be critical in further research studies and future clinical biomarker development strategies. One of the most recent highly validated complex genetic analysis is the BRACAnalysis (Myriad Genetics, Salt Lake City, UT) sequencing testing for BRCA1 and BRCA2 germ line mutations (4, 5). To explore the potential clinical utility of liquid-based Pap smear samples as a source of reliable germ line DNA, we did BRACAnalysis gene mutation analyses in a masked fashion on Preservcyt (ThinPrep, Cytyc Corp., Boxborough, MA) cytology samples from women who had previously undergone BRCA blood testing.
Materials and Methods
This protocol was approved by the institutional review board at our institution. Women were recruited from the clinical practices of the authors (C.Y.M. and J.O.S.) and from the Clinical Cancer Risk Assessment Program (G.E.T.). Eligible women included those who had previously undergone comprehensive genetic counseling and any type of BRCA testing and who were aware of their results. Women were excluded if they did not have appropriate genetic counseling, if they opted for nondisclosure of the results after testing, or if they were unwilling or incapable of undergoing pelvic exam or Pap smear testing.
BRCA1 and/or BRCA2 gene mutational analysis was done on DNA isolated from liquid-based cervical or vaginal cytology taken from 17 consenting women (age 29–65 years) who previously had genetic counseling followed by BRACAnalysis (Myriad Genetics) blood analyses. Eleven women had known mutations in either BRCA1 or BRCA2 (cases) and six had no identified mutations (controls) on entry into the study. Anonymized cytology samples were sent to Myriad Genetics with a request for testing that was limited to the degree of genomic testing previously done on the blood samples. This was done to avoid ethical concerns of possibly discovering a mutation in part of the gene(s) in which we had not previously tested.
Cervicovaginal cytology was done in the standard fashion using a Papette (Wallach Surgical Devices, Inc., Orange, CT) collection broom for ectocervical and vaginal cell sampling followed by an endocervical sample obtained using an endocervical brush. Vaginal cytology was collected using the Papette broom scraped across the vaginal cuff. The collection broom and/or brush were eluted into the Preservcyt media per the manufacturer's instructions and the resultant suspended cytologic specimen was stored at room temperature until analysis. The entire ThinPrep Pap sample (∼15 mL) was centrifuged and the cell pellet was suspended in phosphate buffer solution. The entire suspended pellet was incubated with Qiagen protease (Qiagen, Valencia, CA) and DNA was extracted from the digested cellular material using a Qiagen column. The purified cytology DNA was subjected to sequencing using the PCR-based BRACAnalysis automated sequencing methodology as described previously (4) including sequence determination in both forward and reverse directions.
Results
Of the 17 cytology specimens collected, 6 were vaginal cytology only (women with prior hysterectomy) and 11 were cervicovaginal collections. One cervical cytologic sample had scant cellular content that precluded gene analysis and was excluded from further analysis. For the remaining 16 adequate cytology specimens, both BRCA1 and BRCA2 full-length sequencing was done on five of the six known negative cases and BRCA1 only sequencing was done on one negative case. Full single gene sequencing was done on eight cases targeting the gene with the previous known mutation. Multisite 3 BRACAnalysis designed to detect only mutations 187delAG and 5385insC in BRCA1 and 6174delT in BRCA2 was done on two cases from subjects who had this targeted blood analysis as there sole BRCA testing (Table 1).
Comparative BRCA1/BRCA2 mutation analyses in ThinPrep cytology and blood DNA
Case . | Age, y . | Pap site* . | Gene analyzed . | Cytology result . | Blood result . | |||||
---|---|---|---|---|---|---|---|---|---|---|
Cases | ||||||||||
B1-0001 | 60 | Cervix | BRCA1 | NC† | BRCA1 Q563X | |||||
B1-0002 | 56 | Vagina | BRCA2 | 7855insT | BRCA2 7855insT | |||||
B1-0003 | 29 | Cervix | BRCA1 | 2798del4 | BRCA1 2798del4 | |||||
B1-0005 | 51 | Vagina | Multisite 3 | BRCA1 5385insC | BRCA1 5385insC | |||||
187delAG negative | 187delAG negative | |||||||||
6174delT negative | 6174delT negative | |||||||||
B1-0006 | 58 | Cervix | BRCA1 | IVS5-11T>G | IVS5-11T>G | |||||
Splice site mutation | ||||||||||
B1-0007 | 40 | Cervix | BRCA1 | 187delAG | BRCA1 187delAG | |||||
B1-0010 | 49 | Vagina | Multisite 3 | BRCA2 6174delT | BRCA2 6174delT | |||||
187delAG negative | 187delAG negative | |||||||||
5385insC negative | 5385insC negative | |||||||||
B1-0013 | 62 | Cervix | BRCA1 | 5296del4 | BRCA1 5296del4 | |||||
B1-0014 | 54 | Cervix | BRCA1 | 4184del4 | BRCA1 4184del4 | |||||
B1-0015 | 41 | Cervix | BRCA2 | 1538del4 | BRCA2 1538del4 | |||||
Q713L‡ | Q713L‡ | |||||||||
G3212R‡ | G3212R‡ | |||||||||
B1-00017 | 49 | Cervix | BRCA1 | 5296del4 | BRCA1 5296del4 | |||||
Controls | ||||||||||
B1-004 | 35 | Cervix | BRCA1/BRCA2 | Negative | negative | |||||
B1-008 | 40 | Vagina | BRCA1/BRCA2 | BRCA1 Negative | BRCA1 Negative | |||||
BRCA2 1982M‡; 12490T‡ | BRCA2 1982M‡; 12490T‡ | |||||||||
B1-009 | 56 | Cervix | BRCA1/BRCA2 | Negative | Negative | |||||
B1-0011 | 40 | Vagina | BRCA1 | Negative | Negative | |||||
B1-0012 | 48 | Cervix | BRCA1/BRCA2 | Negative | Negative | |||||
B1-0016 | 65 | Vagina | BRCA1/BRCA2 | Negative | Negative |
Case . | Age, y . | Pap site* . | Gene analyzed . | Cytology result . | Blood result . | |||||
---|---|---|---|---|---|---|---|---|---|---|
Cases | ||||||||||
B1-0001 | 60 | Cervix | BRCA1 | NC† | BRCA1 Q563X | |||||
B1-0002 | 56 | Vagina | BRCA2 | 7855insT | BRCA2 7855insT | |||||
B1-0003 | 29 | Cervix | BRCA1 | 2798del4 | BRCA1 2798del4 | |||||
B1-0005 | 51 | Vagina | Multisite 3 | BRCA1 5385insC | BRCA1 5385insC | |||||
187delAG negative | 187delAG negative | |||||||||
6174delT negative | 6174delT negative | |||||||||
B1-0006 | 58 | Cervix | BRCA1 | IVS5-11T>G | IVS5-11T>G | |||||
Splice site mutation | ||||||||||
B1-0007 | 40 | Cervix | BRCA1 | 187delAG | BRCA1 187delAG | |||||
B1-0010 | 49 | Vagina | Multisite 3 | BRCA2 6174delT | BRCA2 6174delT | |||||
187delAG negative | 187delAG negative | |||||||||
5385insC negative | 5385insC negative | |||||||||
B1-0013 | 62 | Cervix | BRCA1 | 5296del4 | BRCA1 5296del4 | |||||
B1-0014 | 54 | Cervix | BRCA1 | 4184del4 | BRCA1 4184del4 | |||||
B1-0015 | 41 | Cervix | BRCA2 | 1538del4 | BRCA2 1538del4 | |||||
Q713L‡ | Q713L‡ | |||||||||
G3212R‡ | G3212R‡ | |||||||||
B1-00017 | 49 | Cervix | BRCA1 | 5296del4 | BRCA1 5296del4 | |||||
Controls | ||||||||||
B1-004 | 35 | Cervix | BRCA1/BRCA2 | Negative | negative | |||||
B1-008 | 40 | Vagina | BRCA1/BRCA2 | BRCA1 Negative | BRCA1 Negative | |||||
BRCA2 1982M‡; 12490T‡ | BRCA2 1982M‡; 12490T‡ | |||||||||
B1-009 | 56 | Cervix | BRCA1/BRCA2 | Negative | Negative | |||||
B1-0011 | 40 | Vagina | BRCA1 | Negative | Negative | |||||
B1-0012 | 48 | Cervix | BRCA1/BRCA2 | Negative | Negative | |||||
B1-0016 | 65 | Vagina | BRCA1/BRCA2 | Negative | Negative |
Vaginal cuff Pap smears were done in women who had had previous total hysterectomy.
Not enough cellularity for DNA analyses.
BRCA2 germ line polymorphisms.
Of the five fully sequenced negative controls, four had no variations detected in BRCA1 or BRCA2 from the cytology DNA. The remaining case (B1-008) showed two BRCA2 polymorphisms (1982M and 12490T), which have not been shown to be disease related but were also identified in the matched blood analysis (Table 1). The two cases positive for the Ashkenazi screen in the blood analysis confirmed the same site-specific mutations using the cytologic specimens (BRCA2 6174delT and BRCA1 5385insC; Table 1). All other BRCA1 or BRCA2 disease-related mutations and nondisease polymorphisms were accurately detected in the respective matched cytology samples (Table 1). The quality of sequence signal was similar between the DNA isolated from blood and the matched cervical cytology (Figure 1).
The fidelity of mutation detection in the cervical DNA specimen (A) as compared with the blood lymphocyte DNA specimen (B) from the same patient. A and B. Top sequence, patient DNA compared with normal BRCA1 sequence (bottom sequence). Arrow, inserted C residue at position 5,385. Top sequence, to the right of the arrow (A and B), heterozygous for a shift in the reading frame.
The fidelity of mutation detection in the cervical DNA specimen (A) as compared with the blood lymphocyte DNA specimen (B) from the same patient. A and B. Top sequence, patient DNA compared with normal BRCA1 sequence (bottom sequence). Arrow, inserted C residue at position 5,385. Top sequence, to the right of the arrow (A and B), heterozygous for a shift in the reading frame.
Discussion
Our group showed 100% concordance (κ = 1.0; 95% confidence interval, 0.51–1.0) in detecting BRCA mutations and polymorphisms using the clinically approved BRACAnalysis sequencing of DNA isolated from whole blood compared with that isolated from matched alcohol preserved lower genital tract cytology. All but one sample had sufficient cellular content for adequate DNA yields for the numerous PCR reactions required for full BRCA gene sequencing. We were surprised and quite encouraged to find that all vaginal cytology collections and nearly all women who were postmenopausal had enough cellular material to perform this robust analysis, as we were concerned that total cellular content would be too low in postmenopausal women and in women without a cervix. Nevertheless, the occasional inadequate specimen should be considered a limitation to this approach over blood and a larger study with greater samples would show a more realistic rate of inadequate cellular content and determine a patient profile for whom this may apply. Identifying the latter would be very important for research studies that may introduce a collection bias if cytology was used for genetic studies. Missing test results will also need to be considered in the context of clinical test performance.
The presence of normal vaginal bacterial flora, cervical mucus, and blood was adequately removed in the DNA isolation and purification steps. These potential contaminants did not have a negative impact on the PCR and sequence analyses. One must consider other potential contaminants present in some Pap smear samples that could potentially alter the results of germ line DNA testing. It is conceivable that sperm deposited in the cervix and vagina within hours prior to a Pap collection could alter germ line DNA analyses, particularly genetic polymorphism analyses, and cause false-positive test results. We did not elicit this history in this small feasibility study. This potential confounder, however, would need to be considered in larger trials.
It must be made clear that we are not advocating that clinical practice adopt BRACAnalysis using cervical cytology for routine analysis. Although we have shown with certainty the feasibility of this application, a large validation trial and regulatory approval would be necessary prior to clinical use. We stress that any type of testing for BRCA1 or BRCA2 mutation genotype should only be done in the setting of pretest and posttest counseling. Of greater interest, we cannot underscore the importance of this finding as it relates to future work in cervical cancer biomarker development and risk assessment for other cancers and disease. The robustness of genetic testing on cervical cytology could be used in future cancer screening, such as combining cervical and breast screening. As more is learned about modifier gene polymorphisms for risk assessment in disease and cancer, applicability of germ line genetic testing to the “leftovers” of a clinically routine, acceptable minimally invasive procedure may have tremendous clinical impact.
Grant support: Presented at the Western Association of Gynecologic Oncology Meeting, June 25–28, 2003, Steamboat Springs, CO.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Acknowledgments
We thank Phil Castle, Ph.D., M.P.H., Investigator, Hormonal and Reproductive Epidemiology Branch, Division of Cancer Epidemiology and Genetics, National Cancer Institute, NIH DHHS, Bethesda, MD for his generous review of this article and Amy Deffenbaugh (Myriad Genetics) for her technical assistance and research coordination.