Diagnosis of Hydatidiform Moles Using p57 ... · Diagnosis of Hydatidiform Moles Using p57...

Am J Clin Pathol 2008;129:749-755 749749 DOI: 10.1309/7XRL378C22W7APBT 749

© American Society for Clinical Pathology

Anatomic Pathology / HER2 and p57 in Molar pregnancy

Diagnosis of Hydatidiform Moles Using p57 Immunohistochemistry and HER2 Fluorescent In Situ Hybridization

Robin D. LeGallo, MD,1 Edward B. Stelow, MD,1 Nilsa C. Ramirez, MD,2 and Kristen A. Atkins, MD1

Key Words: Fluorescent in situ hybridization; FISH; HER2; p57; Hydatidiform mole; Gynecology

DOI: 10.1309/7XRL378C22W7APBT

A b s t r a c tFluorescent in situ hybridization (FISH) for HER2

performed on paraffin-embedded tissue samples and immunohistochemical analysis for p57 may be useful ancillary studies to aid in the diagnosis and classification of hydatidiform moles (HMs). HER2 FISH was validated against 24 paraffin-embedded sections of HMs and hydropic abortions and showed an 85% concordance rate. A morphologic assessment based on 44 cases showed 25% disagreement between original and consensus diagnosis based on H&E-stained slides, all of which involved the differential diagnosis of partial mole and hydropic abortion. Immunohistochemical analysis for p57 and HER2 FISH were performed, and a final diagnosis was assigned by using the results from all ancillary studies. p57 staining was absent in 11 of 13 complete moles, and HER2 was triploid in 8 of 10 partial moles. HER2 FISH and immunohistochemical analysis for p57 are useful ancillary techniques in the evaluation of HM, especially when triploid content is seen.

In the United States, hydatidiform moles (HMs) compli-cate approximately 1 in 1,500 pregnancies and are classified as complete moles (CMs) or partial moles (PMs) based on clinical, morphologic, and genetic differences. Persistent ges-tational trophoblastic disease develops after a CM in 10% to 30% of cases and after a PM in 0.5% to 5%. Choriocarcinoma arises in up to 3% of CMs and is a rare but reported sequelae of PMs.1-3 Despite the difference in risk for persistent disease or metastases, the postevacuation surveillance recommenda-tions are identical with frequent monitoring of serum human chorionic gonadotropin levels until normalization and monthly monitoring for 6 months thereafter.4 Thus, distinguishing a molar pregnancy from a hydropic abortion (HA) and correctly classifying the mole as partial or complete has management and prognostic implications. However, the diagnosis and classification of HM is subject to great interobserver and intra-observer variability, with the greatest difficulty distinguishing PM from HA. Ancillary studies, including cytogenetics, ploidy analysis, and immunohistochemical analysis, use the underly-ing genetics of these entities and sometimes aid in the diagno-sis and classification of HM.

p57kip2 (p57) is the protein product of the paternally imprinted but maternally expressed gene CDKN1C located on chromosome 11p15.5.5 Because CMs lack a maternal genomic component, they are not expected to express imprinted genes that are normally expressed by the maternal allele, and immuno-histochemical analysis for p57 has been shown to be a valuable tool in the diagnosis of a CM.6-10 However, immunohistochemi-cal analysis for p57 cannot distinguish a PM from an HA.

Because fresh tissue samples for cytogenetics are not always available and many laboratories do not perform ploidy analysis on paraffin sections, we studied a paraffin fluorescent

750 Am J Clin Pathol 2008;129:749-755750 DOI: 10.1309/7XRL378C22W7APBT


LeGallo et al / HER2 and p57 in Molar pregnancy

in situ hybridization (FISH) technique coupled with p57 in the diagnosis of HM. HER2 is an oncogene located on chromo-some 17q21.1 that is amplified in a subset of breast carcino-mas. FISH analysis for HER2 using formalin-fixed, paraffin-embedded tissue samples is widely used for amplification analysis, is readily available in many laboratories, and has a quick turnaround time. We postulated that it could also be used as a surrogate marker for ploidy in the evaluation of HMs.

Materials and Methods

Patient SpecimensThe pathology archives were searched for the pathologic

diagnoses of CM and PM and for the clinical diagnosis of rule out HM. The original histologic diagnoses of the 44 cases included 13 CMs, 15 PMs, and 16 nonmolar pregnancies. These cases made up the set used for the morphologic study. Cytogenetic data were available for 14 of the cases, and an additional 10 cases with cytogenetic data were obtained from the Columbus Children’s Hospital, Columbus, OH. This set included 14 triploid, 9 diploid, and 1 tetraploid case and was the basis of validation of FISH for HER2.

HER2 FISHHER2 FISH was performed on all 54 cases. Sections

(3 µm) from paraffin-embedded tissue samples were baked at 58°C for 15 minutes. FISH was performed according to the manufacturer’s automated procedure, Inform, using the Ventana Benchmark instrument (Ventana Medical Systems, Tucson, AZ). Briefly, slides were rinsed, deparaffinized, rinsed, treated with reaction buffer, and incubated at 37°C for 4 minutes. One drop of ISH-protease 3 was applied and rinsed with reaction buffer. After ISH blocker application, 1 drop of antimouse/fluorescein isothiocyanate was added and incubated for 40 minutes (to amplify the signal). The slides were then rinsed with reaction buffer several times and cover slipped. A breast cancer specimen with known HER2/neu gene amplification was used as a positive control sample. All slides were viewed under a fluorescence microscope.

FISH was repeated once in all cases for which a signal was not obtained or was too weak to score. A total of 50 cells were counted, and HER2 FISH was scored as triploid or tetraploid if more than 10% of nuclei showed 3 or 4 signals, respectively. The 24 cases with cytogenetic correlation served as the basis for the validation study.

p57 Immunohistochemical AnalysisA representative slide containing decidual and villous

tissue was submitted for immunohistochemical analysis for p57 (dilution 1:200; NeoMarkers, Fremont, CA). Paraffin

sections were cut at 4 µm, mounted on positively charged glass slides, and baked overnight. Sections were deparaf-finized in xylene and ethanol, and antigen retrieval was performed by placing in DAKO TRS (pH 6.0 citrate buffer; DAKO, Carpinteria, CA) and put in a pressure chamber for 30 seconds. Sections were removed and allowed to cool in buffer for 20 minutes, following which they were rinsed thoroughly in deionized water. Slides were processed on a DAKO Autostainer Plus S3400 (DAKO). Incubation with secondary antibodies was performed, followed by treatment with a 3,3’-diaminobenzidine-containing chromogenic solu-tion and hematoxylin counterstaining. Appropriate positive and negative control experiments were run with each case. p57 staining was considered negative when cytotrophoblasts and villous stromal cells showed less than 10% nuclear immunoreactivity. Maternal decidua and proliferating tro-phoblasts served as internal control samples, and staining was necessary for the results to be considered valid.

Validation Study for HER2/neuThe HER2 FISH results were compared with the cytoge-

netic data, and a concordance percentage was calculated based on the number of concordant samples over the total number of samples.

Morphologic StudyAll H&E-stained slides were reviewed by 3 of us

(E.B.S., K.A.A., and R.D.L.), and a consensus review diag-nosis was made based only on the H&E-stained slides. A final diagnosis was assigned by using the results from all available data, including morphologic features, cytogenetics, HER2 FISH results, and the results of immunohistochemi-cal analysis for p57 according to the algorithm outlined in zFigure 1z.

H&E

Classify

Completemole

Partialmole

Hydropicabortion

+ – Completemole

p57 Partialmole

HER2

Unsure

Unsure

Diploid Triploid

Mole

zFigure 1z Algorithm for the analysis of hydropic villi in the products of conception.

prgmea.com

http://www.prgmea.com



Anatomic Pathology / original article

Results

HER2 FISHFISH was successfully performed on 46 cases. Initially, 9

cases had no signal or a signal too weak to score. On repeated FISH, only 1 of these cases yielded a signal strong enough for analysis, for an overall success rate of 85% (46/54).

Validation StudyOf the 24 cases with cytogenetics data, 14 were triploid,

9 were diploid, and 1 was tetraploid on karyotyping. No signal was seen by HER2 FISH in 4 triploid cases. Of the remaining cases, FISH results were concordant with cytogenetics in 7 of 10 triploid cases, 9 of 9 diploid cases, and 1 of 1 tetraploid case, for an overall concordance rate of 85% (17/20; 95% confidence interval, 62%-97%) zTable 1z.

Morphologic StudyThe results of the cases constituting the morphologic

study are detailed in zTable 2z. There was disagreement between the original and consensus diagnosis in 25% of the cases, all of which involved the differential diagnosis of PM vs HA.

zTable 2zMorphologic Study Results in Hydatidiform Moles

Diagnosis

Case No. Final Original Review p57 Staining HER2 FISH Karyotype Comments

1 CM CM CM – Diploid NA Complete agreement2 CM CM CM – Diploid 46,XX Complete agreement3 CM CM CM – NS NA NS on FISH4 CM CM CM – NS NA NS on FISH5 CM CM CM – Diploid 46,XX Complete agreement6 CM CM CM + NS NA p57 discrepant; NS on FISH7 CM CM CM – Diploid 46,XX Complete agreement8 CM CM CM – Diploid NA Complete agreement9 CM CM CM – Diploid 46,XX Complete agreement10 CM CM CM – Diploid 46,XX Complete agreement11 CM CM CM – Tetraploid 92,XXXX Complete agreement12 CM CM CM + Diploid 46,XX p57 discrepant13 CM CM CM – Triploid NA FISH discrepant14 PM PM PM + Triploid 69,XXX Complete agreement15 PM PM HA + Triploid NA Review discrepant16 PM PM PM + Triploid NA Complete agreement17 PM PM PM + Triploid 69,XXX Complete agreement18 PM PM HA + Triploid NA Review discrepant19 PM PM PM + Triploid NA Complete agreement20 PM PM PM + Diploid NA FISH discrepant21 PM PM HA + Triploid NA Review discrepant22 PM HA PM + Diploid 69,XXX Original discrepant; FISH discrepant23 PM PM PM + Triploid 69,XXX Complete agreement24 HA HA HA + Diploid NA Complete agreement25 HA HA HA + Diploid NA Complete agreement26 HA HA HA + Diploid NA Complete agreement27 HA PM HA + Diploid NA Original discrepant28 HA PM HA + Diploid NA Original discrepant29 HA PM HA + Diploid NA Original discrepant30 HA PM HA + Diploid NA Original discrepant31 HA HA PM + Diploid NA Review discrepant32 HA PM HA + Diploid NA Original discrepant33 HA HA HA + Diploid NA Complete agreement34 HA HA HA + Diploid NA Complete agreement35 HA PM HA + Diploid NA Original discrepant36 HA HA HA + Diploid NA Complete agreement37 HA HA HA + Diploid 46,XY Complete agreement38 HA HA HA + Diploid NA Complete agreement39 HA HA HA + Diploid 46,XX Complete agreement40 HA HA HA + Diploid NA Complete agreement41 HA HA HA + Diploid NA Complete agreement42 HA HA HA + Diploid NA Complete agreement43 HA HA HA + Diploid 46,XY Complete agreement 44 HA HA HA Negative control NS NA NS on FISH; p57 noninformative

CM, complete mole; FISH, fluorescent in situ hybridization; HA, hydropic abortion; NA, not available; NS, no signal; PM, partial mole; +, positive; –, negative.

zTable 1zValidation Data for the HER2 FISH Probe

Cytogenetics HER2 FISH

Diploid, 9 Diploid, 9Triploid, 14 Triploid, 7; diploid, 3; no signal, 4Tetraploid, 1 Tetraploid, 1

FISH, fluorescent in situ hybridization.

prgmea.com





Of the 13 cases assigned a final diagnosis of CM, all were interpreted as CMs in the original and review diagno-ses. There was a complete lack of p57 immunoreactivity in the stromal and cytotrophoblastic cells with positive internal controls in 10 of the 13 cases zImage 1z. One case showed a dual population of villi with differential p57 staining con-sistent with a coexistent CM and nonmolar twin. Two cases (cases 6 and 12) with diagnostic features of a CM showed more than 30% p57 immunoreactivity in villous cytotropho-blasts and stromal cells. Both of these cases were from the same patient and represent the first 2 of 3 recurrent CMs.

HER2 FISH data were available for 10 of the 13 CMs; 3 cases yielded no signal. Of the 13 cases, 8 were interpreted as diploid and 1 as tetraploid, which correlated with the cyto-genetic findings zImage 2z. One case was scored as triploid by FISH (case 13) and had no correlative cytogenetics. This case showed loss of p57 immunoreactivity with strong inter-nal controls and also represented the last of the 3 recurrent moles from the same patient as in cases 6 and 12.

PM was the final diagnosis in 10 cases. PM was the original diagnosis in 9 cases but the review diagnosis in only 7 cases; the remainder were classified as HA. All 10 cases of PM showed immunopositivity for p57 zImage 3z.

HER2 FISH was successfully performed on all 10 cases and was scored as triploid in 8 and diploid in 2 zImage 4z. Of the 2 diploid FISH cases, 1 case (case 22) was triploid by cytogenetics, and a karyotype was not available on the other case (case 20).

The remainder of the morphologic study set included 21 cases given the final diagnosis of a nonmolar pregnancy (HA).

This was the original diagnosis in 15 cases and the consen-sus review diagnosis in 20 cases, with the remainder in each set being called PM (Table 2). p57 staining was present in 20 of these cases and was noninformative in 1 case owing to degenerative changes and a negative internal control. HER2 FISH was successful in 20 of the cases and in all cases was scored as diploid.

Discussion

Distinctive histologic features characterize HMs. PMs show a dimorphic villous population with an admixture of hydropic and normal villi, the former of which have scal-loped borders, trophoblastic inclusions, and at least focal trophoblastic hyperplasia. There is usually evidence of fetal development in the form of fetal parts, membranes, or villous vessels containing nucleated RBCs. CMs show uniformly enlarged villi with central cisterns, circumfer-ential trophoblastic hyperplasia that often shows cytologic atypia, and an absence of fetal tissue. However, histologic evaluation alone is prone to great interobserver variability, especially between the diagnoses of PM and HA and in iden-tifying first-trimester CMs.11

Cytogenetics and ploidy analysis have greatly aided in the diagnosis of HMs. CMs are usually diploid, with the vast majority showing a 46,XX karyotype and the remainder showing 46,XY. In approximately 10% of cases, CMs have a tetraploid karyotype. The majority of CMs lack a maternal genome and are androgenetic as a result of reduplication of

zImage 1z p57 staining in a complete mole shows lack of staining in the villous cytotrophoblast and stromal cells with positive staining of the trophoblastic hyperplasia (×100).

zImage 2z Two HER2 signals seen in the center cell suggest diploid DNA content.

prgmea.com





a haploid paternal genome and, less commonly, as a result of a dispermic fertilization.12 A small percentage have a biparental genome, and this is the usual case in recurrent and familial forms.13 PMs are characterized by a triploid genome, and most have a 69,XXX or 69,XXY karyotype. This is composed of a haploid maternal genome and a redu-plicated paternal haploid genome or is a result of dispermy.12 A triploid chromosome complement can aid in the diagnosis of a PM, but neither conventional cytogenetics nor ploidy analysis can discriminate a CM from an HA.

FISH analysis on paraffin-embedded tissue samples has many potential technical and interpretive challenges. An absent or weak signal, as occurred in 8 of our cases, may be due to preanalytic or analytic variables.14,15 Suboptimal fixa-tion or embedding can result in an absent or decreased signal. Ideally, specimens should be fixed as quickly as possible after receipt. However, specimens are often evacuation products after missed abortions and have already undergone marked degenerative changes. A preferred fixative for molecular techniques is 10% neutral buffered formalin, which was used in the majority of cases studied. Storage time can also affect signal intensity, but we were not able to see a correlation with this in our cases. Hybridization problems can also occur during the technical phase owing to incomplete unmasking or poor hybridization. These parameters were not adjusted during this study because we were attempting to study the method that has already been optimized for analysis of breast carcinoma.

Of the 10 cases that were triploid by cytogenetics, 3 were diploid by FISH analysis. One of the possibilities is that decidual tissue was scored, rather than villi. Attempts

were made to circle the areas rich in villi before the hybrid-ization procedure, but many cases had intermixed maternal decidua. The FISH signal was not uniformly strong across tissue samples, and often the slide had to be scanned for a signal strong enough to score. An alternative explanation is that this is a truncation artifact due to incomplete representa-tion of nuclei on cross-section.

This study had no cases with triploid or hyperdiploid FISH results in a morphologically nonmolar pregnancy, but this could conceivably happen in cases of digynic triploidy or rare cases of trisomy 17, which is present in fewer than 1% of chromosomally abnormal spontaneous abortions.16,17 In addition, HER2 overexpression assessed by immunohis-tochemical analysis has been reported in HM and has been proposed as a predictor of persistent disease.18 None of our cases tested showed HER2 signals exceeding the ploidy determined by cytogenetics, suggesting that there was no evidence of amplification. This may be sampling error but suggests that the mode of HER2 overexpression is not by genomic amplification.

The present study acknowledges the difficulty in dis-criminating PM from HA because all discrepant cases were assigned to one of these groups on final diagnosis. For the majority of cases, the addition of ploidy results enabled us to separate the HA and PM groups. We did not encounter the problem of distinguishing CM from PM or HA, in large part owing to the results of immunohistochemical analysis for p57 result, but also likely owing to underrepresentation of first-trimester gestations, when HA and CM show the most morphologic overlap.

zImage 3z p57 staining in a partial mole shows strong nuclear staining in the villous cytotrophoblast and stromal cells (×100).

zImage 4z Three HER2 signals are seen in the center cell, suggesting triploid DNA content.

prgmea.com





There were 5 cases that did not easily stratify into a diagnostic group and were considered discrepant by one of the ancillary studies. Two cases (cases 20 and 22) with a final diagnosis of PM were diploid by FISH, one of which had a triploid karyotype by cytogenetics. This constitutes 20% of all PM diagnoses and is similar to the concordance rate in our validation study. There has been some debate whether nontriploid PM exists, and this is unlikely.19 Scoring of sig-nals in decidual tissue remains a possibility, as mentioned in the validation study discussion. The histologic interpretation of the case without cytogenetics data was PM in the original and review diagnoses and was assessed a third time with the ploidy information available, and all pathologists were in agreement with PM as the diagnosis.

Cases 6, 12, and 13 are of interest because they rep-resent repeated molar gestations in the same patient. All 3 cases were diagnosed as CM in the original and histologic review and assigned a final diagnosis of CM. However, all 3 cases had discrepancies in the p57 or HER2 FISH results. Recurrent molar pregnancies are rare and, in some cases, are familial with an autosomal recessive pattern of inheritance. The genetic locus has been mapped to chromosome 19q13.4, and recently, mutations in NALP7 have been identified in families affected by recurrent molar pregnancies.20 The genotype of recurrent CMs, unlike sporadic moles, is bipa-rental with a haploid maternal and paternal contribution.13 The pathogenesis is still considered one of imprinting, and an abnormal methylation pattern preferentially affecting imprinted genes has been shown. A single study assessing recurrent moles in 7 patients, including a pair of siblings, showed loss of staining of p57 in all cases.21 In the pres-ent study, 2 of 3 cases (cases 28 and 43) of recurrent CM showed immunoreactivity for p57 in the cytotrophoblasts and villous stromal cells in essentially all villi. Both of these cases were diploid by HER2 FISH, and 1 was confirmed diploid by karyotype but parental contribution was not assessed. Case 13 showed a lack of p57 staining with appro-priate staining of the small amount of decidua present on the slide and only focal staining of the proliferating extravillous trophoblasts. HER2 FISH was interpreted as triploid in this case, and a karyotype was not available. The discrepancies in this subset of cases are without a convincing explanation but raise the possibility of other mechanisms underlying recurrent moles.

A combination of immunohistochemical analysis for p57 and HER2 FISH (or potentially other probes) will aid in the accuracy in identifying molar pregnancies. An absence of p57 staining is specific for CM. When the differential diagnosis is PM vs HA, HER2 FISH or potentially other FISH probes can assist in the classification when ample viable tissue is present, particularly when triploid content is seen.

From the Departments of Pathology, 1University of Virginia, Charlottesville; and 2Columbus Children’s Hospital, Columbus, OH.

Address reprint requests to Dr LeGallo: Dept of Pathology, University of Virginia, 1215 Lee St, Charlottesville, VA 22908.

References 1. Berkowitz RS, Goldstein DP. Chorionic tumors. N Engl J Med.

1996;335:1740-1748.

2. Seckl MJ, Fisher RA, Salerno G, et al. Choriocarcinoma and partial hydatidiform moles. Lancet. 2000;356:36-39.

3. Soper JT. Gestational trophoblastic disease. Obstet Gynecol. 2006;108:176-187.

4. Soper JT, Mutch DG, Schink JC, American College of Obstetricians and Gynecologists. Diagnosis and treatment of gestational trophoblastic disease: ACOG Practice Bulletin No. 53. Gynecol Oncol. 2004;93:575-585.

5. Chilosi M, Piazzola E, Lestani M, et al. Differential expression of p57kip2, a maternally imprinted CDK inhibitor, in normal human placenta and gestational trophoblastic disease. Lab Invest. 1998;78:269-276.

6. Merchant SH, Amin MB, Viswanatha DS, et al. p57KIP2 immunohistochemistry in early molar pregnancies: emphasis on its complementary role in the differential diagnosis of hydropic abortuses. Hum Pathol. 2005;36:180-186.

7. Castrillon DH, Sun D, Weremowicz S, et al. Discrimination of complete hydatidiform mole from its mimics by immunohistochemistry of the paternally imprinted gene product p57KIP2. Am J Surg Pathol. 2001;25:1225-1230.

8. Fukunaga M. Immunohistochemical characterization of p57(KIP2) expression in early hydatidiform moles. Hum Pathol. 2002;33:1188-1192.

9. Fukunaga M. Immunohistochemical characterization of p57Kip2 expression in tetraploid hydropic placentas. Arch Pathol Lab Med. 2004;128:897-900.

10. Sebire NJ, Lindsay I. p57KIP2 immunostaining in the diagnosis of complete versus partial hydatidiform moles [letter]. Histopathology. 2006;48:873-874.

11. Fukunaga M, Katabuchi H, Nagasaka T, et al. Interobserver and intraobserver variability in the diagnosis of hydatidiform mole. Am J Surg Pathol. 2005;29:942-947.

12. Slim R, Mehio A. The genetics of hydatidiform moles: new lights on an ancient disease. Clin Genet. 2007;71:25-34.

13. Van den Veyver IB, Al-Hussaini TK. Biparental hydatidiform moles: a maternal effect mutation affecting imprinting in the offspring. Hum Reprod Update. 2006;12:233-242.

14. Petersen BL, Sorensen MC, Pedersen S, et al. Fluorescence in situ hybridization on formalin-fixed and paraffin-embedded tissue: optimizing the method. Appl Immunohistochem Mol Morphol. 2004;12:259-265.

15. Srinivasan M, Sedmak D, Jewell S. Effect of fixatives and tissue processing on the content and integrity of nucleic acids. Am J Pathol. 2002;161:1961-1971.

16. Nagaishi M, Yamamoto T, Iinuma K, et al. Chromosome abnormalities identified in 347 spontaneous abortions collected in Japan. J Obstet Gynaecol Res. 2004;30:237-241.

17. Philipp T, Philipp K, Reiner A, et al. Embryoscopic and cytogenetic analysis of 233 missed abortions: factors involved in the pathogenesis of developmental defects of early failed pregnancies. Hum Reprod. 2003;18:1724-1732.




20. Murdoch S, Djuric U, Mazhar B, et al. Mutations in NALP7 cause recurrent hydatidiform moles and reproductive wastage in humans. Nat Genet. 2006;38:300-302.

21. Fisher RA, Hodges MD, Rees HC, et al. The maternally transcribed gene p57(KIP2) (CDNK1C) is abnormally expressed in both androgenetic and biparental complete hydatidiform moles. Hum Mol Genet. 2002;11:3267-3272.

18. Yazaki-Sun S, Daher S, de Souza Ishigai MM, et al. Correlation of c-erbB-2 oncogene and p53 tumor suppressor gene with malignant transformation of hydatidiform mole. J Obstet Gynaecol Res. 2006;32:265-272.

19. Genest DR, Ruiz RE, Weremowicz S, et al. Do nontriploid partial hydatidiform moles exist? a histologic and flow cytometric reevaluation of nontriploid specimens. J Reprod Med. 2002;47:363-368.

Diagnosis of Hydatidiform Moles Using p57 ... · Diagnosis of Hydatidiform Moles Using p57...

Documents

Transcript of Diagnosis of Hydatidiform Moles Using p57 ... · Diagnosis of Hydatidiform Moles Using p57...