Materials and Methods for the 
Endometrial Type Collection

    A type collection of   endometrial tissues studied by histologic morphology (computer morphometric analysis and expert pathologist review) and genetic analysis (clonal composition) is presented in the EIN Type Collection section of this Website.  A brief summary of Methods used for each type of analysis is given here, and representative examples of the type of clonality assays performed (not necessarily from tissues in the Type Collection) are shown in Figures 1-3 below.  The Type Collection hysterectomy series began with 20 cases previously published with accompanying detailed methods, experimental rationale, and illustrated results^5,8. New cases were then added using X chromosome inactivation⁵ and microsatellite marker analysis⁸ methods identical to those in the cited reports. All molecular results reported are based on findings successfully reproduced in duplicate independent experiments. All experiments were performed after review and approval of the Human Studies Committee at Brigham and Women's Hospital.

Case Selection: 158 archival paraffin embedded hysterectomy specimens containing both endometrioid endometrial adenocarcinoma and regions of non-malignant endometrium were selected from the Division of Women's and Perinatal Pathology at Brigham and Women's Hospital, Boston, MA. Carcinomas were tested for the presence of microsatellite instability ("MI", as described below), and all 23 hysterectomies with MI cancers were entered. The first 41 hysterectomies with microsatellite stable cancers and a heterozygous genotype at the X-linked marker used for X inactivation studies (human androgen receptor gene, HUMARA) were also entered. 100 non-malignant, non-polypoid, native endometrial tissue samples were present within these 64 hysterectomies, spanning a wide range of histologic subtypes. 93 of these 100 candidate regions yielded interpretable results from clonality (Table 1), morphometric (Figure 2), and pathologist assessment (Figure 1) thereby constituting the dataset reported under Results. Two tissues were excluded because of scantiness upon resectioning. Four tissues failed clonality analysis by X inactivation: two yielded no PCR products after HhaI digestion, one was non-informative due to excessive skewing of X inactivation within polyclonal control tissues⁶, and one tissue had an equivocal extent⁶ of skewing. One area was omitted as unsuitable for computerized morphometric analysis due to decidualized stroma characteristic of progesterone changes known to invalidate ^1,3 computerized morphometric analysis.

DNA Isolation and Clonal Analysis:   DNA from paraffin sections was isolated by selective ultraviolet irradiation, proteinase digestion, organic extraction, and ethanol precipitation as previously described⁵. Paired tumor-normal DNAs from each hysterectomy were screened with primers for two tetranucleotide repeat (D3S2387, D5S1505) loci previously found to contain novel alleles in the majority (100%) of MI endometrial adenocarcinomas⁸. Those tumors with at least one novel allele in the screen underwent expanded testing with a panel of 10 microsatellites⁸ (PCR primers⁸ from Research Genetics, Huntsville, AL were used to amplify loci: D1S518; D2S1384; D2S1399; D3S2387; D3S2459; D4S1627; D5S1505; D5S816; D8S1132; D21S1435). Tumors with novel alleles in two or more of 10 studied microsatellites were scored as microsatellite unstable, MI, and the full 10 loci studied in a similar fashion in DNAs isolated from non-malignant areas of the endometrium. Non-malignant areas were scored as monoclonal if at least one novel allele was clonally present relative to normal reference myometrium. DNA from normal myometrium of all microsatellite stable cases was amplified at the HUMARA trinucleotide repeat locus using primers AR-a/b⁹, and cases with alleles resolved by a minimum of 3 mm on non-denaturing PAGE judged to be informative for this genetic marker of X chromosome inactivation. X inactivation analysis of matched microsatellite stable tumor, normal myometrium, and non-malignant endometrium was performed using HUMARA PCR (Primers AR-a/b⁹) of HhaI predigested and undigested DNAs. ³²P-TTP labeled PCR products were generated substituting 7-deaza-2'-dGTP for dGTP during amplification⁷ to reduce allele-specific biased amplification, and products resolved by non-denaturing 8% polyacrylamide gel electrophoresis followed by autoradiography. Scoring of paired normal/tumor results was based on visual assessment of band patterns⁵ previously validated⁶ in a study comparing results of quantitative and visual analysis of HUMARA data.

Figure 1: Non-random X inactivation (HUMARA assay) in two non-cancerous endometrial lesions:   AR-a/b PCR assay as described in Materials and Methods of two different (Panels A and B) atypical endometrial hyperplasias shows non-random X chromosome inactivation, indicating monoclonality. DNA from lesional (Lanes 3,4) and normal (lanes 1,2; proliferative endometrium in Panel A, myometrium in panel B) tissue was isolated from the same paraffin block by selective ultraviolet irradiation and amplified with primers AR-a/b either without (lanes 1,3) or with (lanes 2,4) predigestion by HhaI. The two androgen receptor alleles were resolved by polyacrylamide electrophoresis and autoradiography. HhaI predigestion of the normal polyclonal tissue controls does not alter the signal ratio (Lane 2), unlike digestion of the lesional DNA (lane 4) which substantially attenuates one of the alleles. Autoradiographic exposure for panels A and B, 9 and 18 hours respectively. Slides stained with hematoxylin and eosin, and photographed at same magnification, bar measures 100µm. Reprinted, with permission from Reference 9.

 

Figure 2: Structurally altered HUMARA alleles in microsatellite unstable endometrial cancers. Non-denaturing polyacrylamide electrophoresis shows a heterozygous Human Androgen Receptor (HUMARA) trinucleotide genotype in normal myometrium (Lane 1) as two discrete bands. Companion endometrial adenocarcinoma (Lane 2) from the same hysterectomy demonstrates novel alleles as extra bands not seen in the normal tissues. Since the HUMARA locus used for clonality testing by X chromosome inactivation is itself a trinucleotide microsatellite, those tissues with microsatellite instability cannot utilize this marker system to infer clonal status. This has led to use of microsatellites themselves as markers for clonality, as shown in Figure 3.

 

Figure 3: Microsatellite Instability in monoclonal endometrial precancers and cancer.  Microsatellite allelotypes for normal myometrium (M), atypical endometrial hyperplasia (E), and endometrial adenocarcinoma (T) within two blocks (1-2) of a single uterus (94-114) were used to infer stages of neoplastic progression.   A. Representative autoradiograms (5 hour exposure) of PCR products from repetitive tetranucleotide markers T 4.1 resolved by non-denaturing polyacrylamide electrophoresis shows two constitutive alleles in normal heterozygous myometrium (M1 and M2). For each marker, two additional "novel" alleles appear variously in pathologic tissues (hyperplasias E1-2 and carcinomas T1-2). B. A total of 20 novel alleles from 11 tested loci were identified, and the allelotypes mapped to different tissue samples ("myo"metrium, endometrial "hyper"plasia, or "cancer"). An additional 20 alleles were constitutive in all samples, and excluded from the matrix because of their non-informativeness. The number of allelic differences from reference myometrium, _, is shown for each sample. C. Maximum parsimony analysis using the program PAUP was used to reconstruct divergence of hyperplasia and tumor allelotypes from the normal myometrial allelotype. Unresolved polytomies are collapsed to a single branch (e.g., E1/T2/T3) in the tree. The separate "E2" branch is supported by 71% of bootstrap replicates, and a common branch for E1/T2/T3 by 60% of bootstrap replicates. The branching pattern defined by the tree is reflected by horizontal lines in the data matrix. Histology of areas sampled, with progressive increases in extent of cytologic atypia and glandular complexity from proximal (E2) to distal (E1/T2) positions in the lineage tree is shown below (1cm equals 50µm). 

Histology Review:   93 non-malignant endometrial tissues of known clonality were represented as 6-24 mm² circle-delimited regions on histological slides. These were reviewed by each of four experienced subspeciality gynecologic pathologists (GLM, CPC, RMR, and AF) on two separate days, blinded to the results of clonal analysis. Pathologist review took two forms. First, each area was classified using standard WHO diagnostic categories of atypical endometrial hyperplasia (ah), complex endometrial hyperplasia (ch) and simple endometrial hyperplasia (sh). Diagnoses were placed into one of the following categories: carcinoma, hyperplastic, or other (includes cycling, atrophic and reactive endometrium). Areas judged to be hyperplastic were subclassified according to WHO criteria using architectural and cytologic features as follows: atypical hyperplasia (AH), complex (non-atypical) hyperplasia (CH), or simple (non-atypical) hyperplasia (SH). Pathologists were also asked to provide clinical recommendations of patient management, based upon their perception of cancer likelihood conferred by the indicated endometrial tissue, if left in situ. These recommendations ranged from no followup necessary, to required lesion ablation.

Computerized morphometric analysis:   Computerized morphometric analysis of corresponding delineated regions on H&E stained sections was performed with the QProdit 6.1 system (Leica, Cambridge, UK) as previously described^2,4. For each lesion the D-score was calculated, incorporating volume percentage stroma (VPS), standard deviation of shortest nuclear axis (SDSNA), and gland outer surface density (OUTSD), and then classed as probable monoclonal (D<0, n=48), unknown (0<D<1, n=3), or probable polyclonal (D>1, n=42) based on the previously developed outcome-predictive formula D= 0.6229 + (0.0439 x VPS) - (3.9934 x Ln(SDSNA)) - (0.1592 x OUTSD)^2,4.

REFERENCES:

1. Ausems EW, van der Kamp JK, Baak JP. Nuclear morphometry in the determination of the prognosis of marked atypical endometrial hyperplasia. Int J Gynecol Pathol 1985; 4:180-185.

2. Baak JPA, Nauta J, Wisse-Brekelmans E, Bezemer P. Architectural and nuclear morphometrical features together are more important prognosticators in endometrial hyperplasias than nuclear morphometrical features alone. J Pathol 1988; 154:335-341.

3. Baak JPA, van Diest PJ. The Female Reproductive Tract. In: Baak JPA, ed. Manual of Quantitative Pathology in Cancer Diagnosis and Prognosis. Heidelberg:Springer-Verlag, 1991:327-356.

4. Dunton C, Baak J, Palazzo J, van Diest P, McHugh M, Widra E. Use of computerized morphometric analyses of endometrial hyperplasias in the prediction of coexistent cancer. Am J Obstet Gynecol 1996; 174:1518-1521.

5. Jovanovic AS, Boynton KA, Mutter GL. Uteri of women with endometrial carcinoma contain a histopathologic spectrum of monoclonal putative precancers, some with microsatellite instability. Cancer Res 1996; 56:1917-1921.

6. Mutter GL, Boynton KA. X chromosome inactivation in the normal female genital tract: Implications for identification of neoplasia. Cancer Res 1995; 55:5080-5084.

7. Mutter GL, Boynton KA. PCR bias in amplification of androgen receptor alleles, a trinucleotide repeat marker used in clonality studies. Nucleic Acids Res 1995; 23:1411-1418.

8. Mutter GL, Boynton KA, Faquin WC, Ruiz RE, Jovanovic AS. Allelotype mapping of unstable microsatellites establishes direct lineage continuity between endometrial precancers and cancer. Cancer Res 1996; 56:4483-4486.

Copyright 1998-2008 by George L. Mutter, MD.  All Rights Reserved