p53 Gene Mutations Are Common in Uterine Serous ... - NCBI

6 downloads 45 Views 2MB Size Report
Uterine serous carcinoma (USC) is an uncom- mon but aggressive type ofendometrial cancer associated with rapid progression of disease and a poor prognosis ...
American Journal ofPathology, Vol. 150, No. 1, January 1997 Copyright C) American Society for Investigative Pathology

p53 Gene Mutations Are Common in Uterine Serous Carcinoma and Occur Early in their Pathogenesis

Hironori Tashiro,* Christina Isacson,* Ross Levine,* Robert J. Kurman,*t Kathleen R. Cho,*tt and Lora Hedrick*tt From the Department of Pathology,* Division of Gynecologic

Pathology, the Department of Gynecology and Obstetrics,t and the Department of Oncology,* TheJohns Hopkins University School of Medicine, Baltimore, Maryland

Uterine serous carcinoma (USC) is an uncommon but aggressive type of endometrial cancer associated with rapid progression of disease and a poor prognosis. Both USC and its recently described putative precursor, endometrial intraepithelial carcinoma (EIC), demonstrate strong p53 overexpression by immunohistochemistry, suggesting alteration of the p53 gene in their pathogenesis. In the present study, we evaluated 21 USCs and 9 EICs for mutations in the p53 gene using direct sequence analysis and found that 90% of USCs and 78% of EICs contain mutations. Significantly, mutations were found in 3 cases of EIC without associated invasive carcinoma and identical mutations were detected in cases with synchronous USC and EIC. Strong p53 immunoreactivity was seen in the majority of USCs and EICs and correlated with p53 gene mutation, although lack of reactivity did not always indicate the absence of a gene mutation. Loss of heterozygosity of chromosome 17p was observed in 100% of USCs and in 43% of EICs, demonstrating that loss of the wild-type p53 allele occurs early in the development of serous carcinoma. Overall, our results reveal that p53 mutations are very common in USC and EIC. The presence ofp53 gene mutations in EICfurther suggests that p53 alteration plays an important role early in the pathogenesis of serous carcinoma, possibly accounting for its aggressive biological behavior. (Am J Pathol 1997, 150:177-185)

Endometrial carcinoma, the most common malignancy of the female genital tract, is composed of several different histological types of malignant epithelial tumors.1 The two major histological types of endometrial carcinoma, endometrioid and serous, exhibit unique biological behaviors and arise in different clinical settings. Uterine serous carcinoma (USC) represents approximately 10% of all endometrial carcinomas. However, due to its aggressive behavior, USC causes a disproportionate number of endometrial cancer deaths.2 Unlike endometrioid carcinoma, the most common type of endometrial carcinoma, USC occurs in an older age group, is not associated with hyperestrogenism, and develops in a setting of endometrial atrophy instead of hyperplasia.3 A putative precursor to serous carcinoma, termed endometrial intraepithelial carcinoma (EIC), has recently been described.4 Little is known about the molecular pathogenesis of uterine serous carcinoma. Many of the past molecular studies have failed to adequately distinguish USC from endometrioid carcinoma, perhaps obscuring identification of genetic alterations prevalent in this relatively uncommon type of endometrial carcinoma. When USCs are separated from the endometrioid type of endometrial carcinoma, strong p53 immunoreactivity is seen in greater than 80% of the cases, as well as in the putative EIC precursor lesion.5 In contrast, p53 immunohistochemical studies have demonstrated positive immunostaining in only Supported by funds from the Stetler Research Fund for Women Physicians (C. Isacson), the Richard W. TeLinde Endowment, and National Institutes of Health grant CA66720 (L. Hedrick). L. Hedrick is a recipient of a Passano Physician Scientist Award. H. Tashiro and C. Isacson contributed equally to this work. Accepted for publication August 27, 1996. Dr. Isacson's present address: Department of Pathology, New York Hospital-Cornell Medical Center, New York, New York. Address reprint requests to Dr. Lora Hedrick, Department of Pathology, Division of Gynecologic Pathology, The Johns Hopkins University School of Medicine, Ross Research Building, Room 656, 720 Rutland Avenue, Baltimore, Maryland 21205.

177

178

Tashiro et al

AJPJanuary 1997, Vol. 150, No. 1

approximately 20% of endometrioid carcinomas with a noted absence of p53 alterations in atypical hyperplasia, the precursor of endometrioid carcinoma.6'7 It is well established that missense mutations in the p53 tumor suppressor gene often lead to an increase in the intranuclear concentration of p53 protein, frequently detected as strong p53 immunohistochemical staining. However, several studies of endometrial carcinoma have not found a correlation between positive p53 immunostaining and p53 gene mutations.8'9 To determine whether p53 immunostaining correlates with p53 gene mutation and to assess the relationship of EIC to USC, we analyzed a set of selected cases of USC and EIC for p53 alterations. Expression of p53 was evaluated by immunohistochemistry, and p53 genetic alterations were detected by direct sequencing of exons 5 to 8 of the p53 gene. Chromosome 1 7p loss of heterozygosity (LOH) was assessed with polymerase chain reaction (PCR) analysis of two microsatellite loci located on chromosome 17p. Our findings reveal a very high frequency of p53 gene mutation in both USC and EIC and show that strong p53 immunostaining reflects the presence of p53 gene mutations in these lesions. Furthermore, our molecular analyses support the notion that EIC is a precursor of USC. These results indicate that alterations of the p53 gene occur frequently and relatively early in the pathogenesis of USC and support a molecular genetic pathway for the development of USC distinct from that of the more common endometrioid type of endometrial carcinoma.

Materials and Methods Clinical Specimens Twenty-one USCs, including five cases of USC with synchronous EIC, and four cases of EIC in the absence of invasive serous carcinoma were obtained in a retrospective review of the surgical pathology files of The Johns Hopkins Hospital from 1989 to 1995. This group of cases included only those with paraffin blocks available for analysis. The patients ranged in age from 50 to 80 years old (mean, 66.6). Seventeen of the twenty-one USCs were hysterectomy specimens, three cases included endometrial biopsies or curettages and subsequent hysterectomy specimens, and one case (SE17) was a curettage specimen alone. The four cases of EIC without associated USC were biopsies or curettages with subsequent hysterectomy specimens. All specimens were fixed in 10% neutral buffered formalin and paraffin embed-

ded for histological examination with hematoxylin and eosin (H&E) staining. The cases were reviewed by three pathologists (C. lsacson, R. J. Kurman, and L. Hedrick) and categorized according to published criteria. USC is broadly defined as a carcinoma exhibiting a glandular or papillary architecture lined by cuboidal cells showing marked cytological atypia and nuclear pleomorphism. EIC is characterized by an abrupt replacement of the endometrial surface epithelium and glands by cytologically malignant cells that resemble those of USC. Eight of the twentyone USCs were confined to the uterus at the time of diagnosis and the remainder presented with metastatic intra-abdominal spread of disease in addition to the uterine tumor. Four cases of EIC demonstrated no evidence of invasion in the thoroughly sampled hysterectomy specimens.

p53 Immunohistochemistry An average of two paraffin blocks were selected from each case and cut serially at a thickness of 4 ,um. p53 protein was detected immunohistochemically using either mouse monoclonal antibody p53 (Novocastra Laboratories, Newcastle, UK) or mouse monoclonal antibody DO-7 (Dako, Carpinteria, CA) at a dilution of 1:100 with overnight incubation at room temperature on a Tech Mate 1000 automated stainer (BioTek Solutions, Santa Barbara, CA). Both antibodies recognize an epitope between amino acids 35 and 45 in the amino-terminal region of the wild-type and mutant forms of the p53 protein. Antigen enhancement was performed by immersing the slides in sodium citrate buffer and heating in a steamer for 20 minutes. A conjugated reporter enzyme (peroxidase) was used for detection with a chromogenic substrate, and tissues were counterstained with hematoxylin. A negative control without primary antibody was run with each case. p53 immunoreactivity was regarded as positive when brown staining was localized to the tumor cell nuclei. The p53 staining was given an immunoreactive score based upon the intensity of nuclear staining and quantity of cells stained according to a subjective grading system.10 The staining intensity was divided into four categories: 0, negative; 1, weakly positive; 2, moderately positive; 3, strongly positive, with the most strongly staining case as the upper limit. The quantity of cells stained were scored as follows: 0, no staining; 1, 1 to 10%; 2, 11 to 50%; 3, 51 to 80%; 4, >80% of tumor nuclei stained. When multiple sections of a case were examined, each slide was scored individually and then averaged to reflect a single score. An immunoreactive score was

p53 Mutation in Serous Carcinoma Pathogenesis

179

AJPJanuary 1997, Vol. 150, No. 1

rendered as a product of the scores obtained for staining intensity and quantity. A total score of 8 to 12 was considered strong immunoreactivity, 4 to 6 was moderate, 1 to 3 was weak, and 0 was negative. The immunoreactivity scores were generated by one pathologist (C. Isacson) without knowledge of the p53 mutation results.

DNA Extraction DNA for p53 gene analysis and 17p LOH analysis was extracted from paired normal and tumor frozen tissue in 7 cases and from paraffin-embedded tissue in 19 cases. Genomic DNA was isolated from cryostat sections of the frozen tumor tissue samples as previously described to obtain >80% tumor cells.11 The paraffin-embedded DNA extractions were performed with 10-p.m sections from 14 USCs using macrodissection with separate, clean blades to obtain >80% tumor cells in 8 cases and 50 to 75% in 6 cases. The DNA from 2 USCs and the 9 ElCs was extracted using a modified microdissection technique. Either 5- or 10-p.m sections were first deparaffinized and then rehydrated and stained with hematoxylin. Under microscopic visualization, sterile 27.5-gauge needles were used to selectively scrape the intraepithelial lesions from the slides to obtain >90% tumor cells. Both the macro- and microdissected samples were incubated in xylene for 10 minutes and spun in a microcentrifuge, and the pellet was washed with 100% (v/v) ethanol. The samples were then dried, resuspended in 30 to 50 p.1 of 50 mmol/L Tris (pH 9.0) with 200 ng/,ul proteinase K, and covered by 20 p.l of mineral oil. After overnight digestion at 600C, the samples were heated at 1 000C for 10 minutes to inactivate proteinase K. Negative controls were included with each extraction, consisting of a tube without added tissue and a tube containing only scraped paraffin, but were otherwise processed similarly. Internal controls included 3 USCs (SE4, SE8 and SE10) that underwent separate extractions of DNA from both biopsy and hysterectomy samples. Repeat DNA extractions of 12 cases were performed for confirmation of sequence results.

p53 Gene Sequence Analysis The extracted tumor DNA was examined for mutations in the p53 gene using direct DNA sequence analysis. Four intron-based primer sets were designed to amplify exons 5 to 8 of the p53 gene in both the sense and antisense directions as follows: exon 5 (sense) 5'GACTTTCAACTCTGTCTCC3', exon 5 (antisense) 5'GAGCAATCAGTGAG-

GAATC3'; exon 6 (sense) 5'TCCCCAGGCCTCTGATTCC3', exon 6 (antisense) 5'TGACAACCACCCTTAACCC3'; exon 7 (sense) 5'CAAGGCGCACTGGCCTCATC3', exon 7 (antisense) 5'CACAGCAGG CCAGTGTGCAG3'; and exon 8 (sense) 5'GATTTCC TTACTGCCTCTTGC3', exon 8 (antisense) 5'GTGAA TCTGAGGCATAACTGC3'. PCR amplification was performed in 50-,ul reaction volumes containing 2 to 500 ng of genomic DNA, 10 mmol/L Tris/HCI (pH 9.2), 75 mmol/L KCI, 1.5 mmol/L MgCl2, 160 p.mol/L each of dGTP, dATP, dTTP, and dCTP, 0.5 p.mol/L of each primer, and 0.5 U of Taq polymerase (Perkin-Elmer, Norwalk, CT, or Gibco BRL, Gaithersburg, MD). Each PCR reaction was heated to 800C before the dinucleotides were added. Amplification was performed with 40 cycles consisting of 1 minute at 950C, 1 minute at 610C, and 1 minute at 720C followed by a single 5-minute extension at 720C. The annealing temperature for exon 5 was 550C. After PCR amplification, 5 p.l of the reaction volume was run on a 2% agarose gel and visualized with ethidium bromide staining. The PCR products were then phenol-chloroform extracted, precipitated with sodium perchlorate/isopropanol, and sequenced directly with either SequiTherm (EpiCenter Technologies, Madison, WI) or ThermoSequenase (Amersham Life Science, Cleveland, OH) cycle sequencing kits using a corresponding 32P-end-labeled primer as previously described.12 The radiolabeled PCR products were fractionated by electrophoresis on a denaturing 6% polyacrylamide gel containing 8 mol/L urea and visualized by autoradiography.

Microsatellite Analysis for Loss of Heterozygosity Two microsatellite loci on chromosome 17p were amplified by the PCR using MapPairs (Research Genetics, Huntsville, AL), TP5313 and CHRNB1.14 CHRNB1 (17p12-pll) is centromeric to TP53 (17p13.1). Paired normal and tumor tissues were analyzed in all cases using the same conditions previously described.15 As a control, a sample lacking DNA was included for each set of reactions. The radiolabeled products were fractionated on 6% polyacrylamide/8 mol/L urea gels and visualized by autoradiography. LOH of chromosome 17p was identified when there was a relative decrease (>50%) in the intensity of the signal of one allele in the tumor as compared with the matched normal DNA. The tumors were scored as noninformative when only one allele was present in either TP53 alone or in both loci.

180

Tashiro et al

AJP January 1997, Vol. 150, No. 1

Table 1.

p53 immunostaining and mutation analysis 0, negative

Histology USC (n = 16) USC and EIC* (n = 5) EIC (n = 4) Total USCs (n = 21) Total EICs (n = 9)

p53 total score, number positive (%) 1-3, 4-6, weak moderate

2 (13) 1 (20) 0 (0) 3 (14) 1 (11)

0 0 0 0 0

2 (13) O (0) O (0) 2 (10) 0 (0)

(0) (0) (0)

(0) (0)

8-12, strong

p53 gene mutation, number positive (%)

12 (75) 4 (80) 4 (100) 16 (76) 8 (89)

15 (94) 4 (80)t 3 (75) 19 (90) 7 (78)

*Synchronous USC and EIC. mutations present in separately isolated regions of USC and EIC.

tidentical

Results Immunohistochemical Detection of p53 Protein Sixteen of twenty-one USCs (76%) and eight of nine EICs (89%) demonstrated strong immunoreactivity. The strongly positive cases included four of the five USCs with synchronous EIC. In these cases, both lesions showed a similar staining pattern. All four cases of EIC without associated USC demonstrated strongly intense staining in >80% of the cells. Only two cases of USC were moderately positive. None of Table 2. Case

SE1 SE2 SE3 SE4 SE5 SE6 SE7 SE8 SE9 SE10 SE16 SE17 SE21 SE24 SE28 SE29 SE14 SE14 SE19 SE19 SE26* SE22* SE27 SE27 SE30 SE30 SE31 SE33 SE34 SE35

the cases were weakly positive. Three cases of USC, including one case with synchronous EIC, were nonreactive for p53. Overall summary and individual case results for p53 expression are presented in Tables 1 and 2, respectively. Representative H&Estained sections with corresponding p53 immunostaining are illustrated in Figure 1.

p53 Gene Mutations p53 gene mutations were identified in 19 of 21 USCs (90%) and in 7 of 9 EICs (78%; Table 1). Signifi-

Individual Case Summary of p53 Expression, Mutation Analysis, and Loss of Heterozygosity

Histology

p53 immunostaining Intensity Quantity Score 9

3 4 4 4 3

12 8 12

USC

3 3 2 3 3

USC

0

0

0

USC

3 3

4 4

12 12

USC

0

0

0

USC

2 3 3 2 2 3 3 3 2 3 3 3 3

3 4 4 4 3 4 4 4 4 4 3 4 4

6 12 12 8 6 12 12 12 8 12 9 12 12

EIC

0

0

0

USC

0

0

0

EIC

3 3 3 3 3 3

4 4 4 4 4 4

12 12 12 12 12 12

USC USC USC USC

USC

USC USC

USC USC USC USC USC

EIC USC

EIC USC

EIC

USC

EIC EIC EIC EIC

9

NI, noninformative. *SE26 and SE22 represent two different samples from the

Exon

Codon

5 8 8 8 5

175 273 282 273 175 225 275 280 243 248 248 248 247

7 8 8 7 7 7 7 7

5 5

241 248 220 220 248 248 154 154

8 8 6 8 6

282 282 205 273 195

7 7 6

6 7 7

same

patient.

p53 mutation analysis Nucleotide (aa) change Type Missense Missense Missense Missense Missense 4-bp insertion Missense Missense 1 -bp insertion Missense Missense Missense Missense Wild type Missense Missense Missense Missense Missense Missense Missense Missense Wild type Wild type Missense Missense Missense Missense Missense Wild type

CGC CAC (arg -> his) CGT > TGT (arg -> cys) CGG > TGG (arg cys) CGT TGT (arg -> cys) CGC CAC (arg -> his) -*

LOH 17p Nl + +

TGT > TTT (cys -> phe) AGA AGC (arg ser) CGG CAG (arg-> gIn) CGG > CAG (arg > gIn) CGG >TGG (arg cys) AGG AGT (arg > ser)

TTC (ser -> phe) CAG (arg -> gIn) TAT TGT (tyr cys) TAT TGT (tyr cys) C GG CAG (arg > gIn) C GG CAG (arg gln) GGC > GTC (gly val) GGC > GTC (gly val)

+ + + NI + + Nl +

TCC CGG

+ +

NI

CGG CGG TAT

CGT TAT

--

0GGG (arg gly) GGG (arg > gly) TGT (tyr -> cys) CGAT (arg -> his) TGT (tyr cys)

p53 Mutation in Serous Carcinoma Pathogenesis

181

AJPJanuary 1997, Vol. 150, No. 1

- -7'

%'

D

sr

1

W

'

'..-

^>0 D~4.'. ~~

x4.

.

~~~~~~~~~~~~~~~~'

I'

~ ~4~~~~~4JRe ~

2"' ~~~~~~~~~

, W ',' a,

4

f

o'

~ ~

~

~

'^ e.. 't._

~

,,

*we

-

~

~

,9

~

~

U

~

.

~

;

~I~~~~~~-o ~ ~ ~ ~ ~ ~

, ...

>

Figure 1. H&E(left) and immunohistochemical staining witb monoclonal antibody p53 (right). A and B: Uterine serous carcinoma (SE2) witb strong positive staining involving all epithelial cells but not the surrounding stroma. C and D: Uterine serous carcinoma (SE6) with negative staining. This case contained a 4-bp insertion in exon 7. E and F: Endometrial intraepithelial carcinoma (SE33) with strong positive staining ofthe lesion and an abrupt transition with normal epitbelium. Arrows denote region magnified in inset.

cantly, p53 gene mutations were found in 3 of 4 ElCs in the absence of USC, and identical mutations were detected in isolated regions of EIC with synchronous USC (4 of 5 cases). All strongly p53-immunoreactive cases and 1 moderately immunoreactive case contained single-base-pair substitutions in exons 5 to 8 of the p53 gene (Table 2 and Figure 2). Transitions were identified in 15 cases and transversions were identified in 5 cases, and 64% of these mutations were located in previously described mutational hotspots of the p53 gene.16 Identical sequence results were obtained in the 12 cases that underwent two

separate extraction, amplification, and sequence

analysis procedures. Two of the three USCs that lacked p53 staining also contained p53 mutations (SE6 and SE9). These nonreactive cases had insertion mutations. SE6 demonstrated a 4-bp insertion (TAGG) at codon 225, which is located at the exon/intron boundary (splice acceptor) of exon 7 (Figure 2). It is not possible to accurately predict the consequence of the 4-bp insertion, however, given the lack of immunoreactivity, it is likely that the resultant protein is unstable. Analysis of SE9 showed a 1 -bp insertion (G) at codon 243

182

Tashiro et al

AJPJanuary 1997, Vol. 150, No. 1

EtC

USC

N E U I s C c

BM A C G T

A C G T

4-

WT

N

ElC

E I C

C

A

C

G

T

']

WT

USC

N U S C

p53 Mutation in Serous Carcinoma Pathogenesis 183 AJPJanuary 1997, Vol. 150, No. 1

Table 3.

Loss of Heterozygosity of Cbromosome 17p

Histology

Cases with LOH/informative cases (%)

USC USC* and EIC*

14/14 (100) 4/4 (100)

EIC Total USCs Total ECs

2/4 (50) 1/3(33) 18/18 (100) 3/7 (43)

*Synchronous USC and EIC.

within exon 7, which creates a premature stop codon in exon 8 at codon 262. This mutation predicts a truncated, presumably unstable, protein that would not be detected immunohistochemically. Only two cases did not demonstrate p53 mutations in exons 5 to 8. One case (SE24) demonstrated moderate p53 immunostaining and the other was negative for staining in both the USC and synchronous EIC (SE27).

Loss of Heterozygosity of Chromosome 1 7p Individual and summary case results for LOH of chromosome 17p are presented in Tables 2 and 3, respectively. LOH was identified in all 18 informative cases of USC (100%) and in 3 of 7 informative cases of EIC (43%). The informative cases included the 2 USCs without detectable p53 mutation in exons 5 to 8 (SE24 and SE27), suggesting a mutation outside of the exons analyzed. The EIC cases with LOH included 2 with synchronous USC and 1 case of EIC without associated USC. Significantly, there was no LOH in 2 cases of EIC with associated USC that demonstrated LOH in the invasive component. LOH was also absent in 1 case of EIC alone that contained a p53 gene mutation and in 1 case of EIC alone that did not contain a gene mutation but demonstrated strong p53 immunostaining. Representative autoradiographs of the microsatellite marker assays with corresponding DNA sequences are presented in Figure 2.

Discussion Mutation of the p53 tumor suppressor gene has been detected in a diverse array of tumor types and is the most commonly altered gene in human malignancies known to date.17 It is postulated that the absence of

normal p53 protein can provide cells with a selective growth advantage leading to tumor progression. The most common mechanism leading to the inactivation of p53 is intragenic mutation followed by loss of the remaining wild-type allele. Although the p53 gene

consists of a total of 11 exons, approximately 90% of mutations occur in exons 5 to 8.17 In this study we identified mutations in this portion of the p53 gene in 19 of 21 USCs (90%) and in 7 of 9 ElCs (78%). Missense mutations were identified in a total of 20 separate cases and an additional 2 cases contained base pair insertions. The 2 cases of USC and 1 of the cases of EIC that lacked mutations showed 17p LOH, suggesting that mutations outside of exons 5 to 8 may be present. These findings demonstrate that most, if not all, of the cases we analyzed contained p53 mutations. To our knowledge, this is one of the highest reported frequencies of p53 gene mutation in any single type of human tumor. Our results suggest that alteration of the p53 gene plays a central role in the pathogenesis of USC. The identification of concordant mutations in the separately analyzed regions of EIC with associated USC (four cases) provides evidence that EIC is related to the development of USC. In addition, the presence of mutations in three of four cases of EIC without associated USC suggests that p53 gene mutation is a relatively early event in the pathogenesis of USC. In many other tumor systems, p53 gene mutation appears to be a late event, such as in colon carcinogenesis in which p53 mutations occur in the transition from late adenoma to invasive carcinoma.18 In the more common endometrioid type of endometrial carcinoma, p53 mutation and overexpression has been reported to occur in a relatively low percentage of tumors (approximately 20%), and mutations are absent in atypical hyperplasia, the precursor to endometrioid carcinoma.6719 Our findings in USC have similarities to those reported in esophageal and lung cancer tumorigenesis, in which p53 mutations have been identified in preinvasive lesions as well as in invasive carcinomas.20-23

We confirmed the frequency of intense p53 immunostaining previously described in USCs and EICs and show a strong correlation with p53 gene mutations. Previous studies of endometrial carcinoma, including USCs, have not found such a strong correlation.89 These differences may be due to different diagnostic criteria for USC, the inherent subjectivity

Figure 2. p53 gene sequence and corresponding loss of heterozygosity analysis. A: Uterine serous carcinoma with EIC(SE30) showing identicalpoint mutations (C to G) at codon 282 in exon 8. There is loss of the wild-type allele in the transition from EIC to USC. B: Endometrial intraepithelial carcinoma (SE34) uith a point mutation (A to C) at codon 195 in e-xon 6and LOH. C: Uterine serous carcinoma (SE6) with a 4-bp insertion (TAGG) in exon 7 and LOH.

184

Tashiro et al

AJP January 1997, Vol. 150, No. 1

of interpreting p53 immunohistochemical stains, and/or the various methods used for the detection of mutations. In the present study of carefully selected cases of USC, the majority of tumors stained very intensely and uniformly. This pattern of staining correlates most closely with the presence of p53 mutations. In contrast, positive but less intense, variable staining generally does not correlate with the presence of mutations. In our study, all strongly staining tumors and one moderately staining tumor demonstrated missense mutations. Of the three tumors with no staining, two had mutations that were 1- and 4-bp insertions, respectively. Both mutations presumably result in unstable proteins that are unlikely to be detectable by immunohistochemistry. In addition, previous studies have used single-strand conformation polymorphism analysis to screen for mutations. The sensitivity of single-strand conformation polymorphism for the detection of mutations is influenced by many factors and requires careful optimization, with reported sensitivities ranging from 35 to 100%.24 Overall, our results show that strong p53 immunoreactivity is an excellent predictor of p53 gene mutation in USCs and ElCs, although lack of staining may not necessarily reflect the absence of mutations in these tumors. Alteration of the p53 gene often occurs in a twostep mechanism with subtle mutation of one allele followed by loss of the second allele such that only the mutant protein is synthesized.25 We evaluated chromosome 1 7p, where the p53 gene is located, for LOH using microsatellite loci and identified loss of one allele in all informative serous carcinomas. Although this analysis does not determine which allele undergoes loss, the sequence analysis on corresponding cases with mutations suggests that the wild-type alleles are the targets of the LOH. Only one case of USC, one case of USC with synchronous EIC, and one case of EIC alone demonstrated the wild-type DNA sequence in the four exons analyzed. Interestingly, LOH of chromosome 17p was seen in most of these cases, implying the presence of a p53 mutation in an exon not analyzed. Although LOH of 17p was detected in all informative cases of USC, LOH was found in only three of seven informative cases of EIC. This suggests that loss of the wild-type p53 allele occurs subsequent to p53 gene mutation and indicates that 17p LOH is not necessarily a marker of invasive disease. Normal p53 function is linked to cell cycle control including mediation of growth arrest and apoptosis in cells after DNA damage.26-29 The atrophic uterine epithelium in which these tumors arise may be a factor linked to the early acquisition of p53 muta-

tions. Recently, it has been proposed that hypoxia selects for cells with defects in apoptosis.30 It is possible that the relatively hypoxic environment of the atrophic epithelium may select for those cells that have acquired p53 mutations, thus allowing their expansion. Loss of wild-type p53 has been shown in vitro to result in gene amplification and abnormal amplification of centrosomes.A32 Aneuploidy is commonly found in USCs and may reflect the fact that p53 gene mutations occur at a relatively early step in their pathogenesis.33 Our results support that early p53 gene mutation is involved in the pathogenesis of serous carcinoma, thus creating an opportune environment for additional, as of yet uncharacterized, molecular genetic events. In conclusion, USC represents an aggressive form of endometrial cancer associated with strong p53 immunoreactivity and a high frequency of p53 gene mutations. Furthermore, the identification of p53 mutations in EIC strongly supports the view that these lesions are precursors of USC. These findings are consistent with a molecular genetic pathway of USC that is distinct from that of the more common endometrioid type of endometrial carcinoma. It is reasonable to speculate that early mutation of the p53 gene may be an important determinant of the aggressive biological behavior of USC resulting in the poor outcome of patients with this disease.

Acknowledgments We thank Paul B. Gaudin for the DNA extraction of several serous carcinomas and Wes Gage and George Pettis for performing the immunohistochemistry.

References 1. Parker S, Tong T, Bolden S, Wingo PA: Cancer Statis-

tics, 1996. CA Cancer J Clin 1996, 46:5-27 2. Carcangiu ML, Chambers JT: Uterine papillary serous carcinoma: a study of 108 cases with emphasis on the

prognostic significance of associated endometrioid carcinoma, absence of invasion, and concomitant ovarian carcinoma. Gynecol Oncol 1992, 47:298-305 3. Kurman RJ, Zaino RJ, Norris HJ: Endometrial carcinoma. Blaustein's Pathology of the Female Genital Tract. Edited by RJ Kurman. New York, Springer-Verlag, 1994, pp 439-486 4. Ambros RA, Sherman ME, Zahn CM, Bitterman P, Kurman RJ: Endometrial intraepithelial carcinoma: a distinctive lesion specifically associated with tumors displaying serous differentiation. Hum Pathol 1995, 26:1260-1267 5. Sherman ME, Bur ME, Kurman RJ: p53 in endometrial cancer and its putative precursors: evidence for di-

p53 Mutation in Serous Carcinoma Pathogenesis 185 AJPJanuary 1997, Vol. 150, No. 1

6.

7.

8.

9.

10.

11. 12.

13.

14.

15.

16. 17.

18.

19.

verse pathways of tumorigenesis. Hum Pathol 1995, 26:1268-1274 Kohler MF, Berchuck A, Davidoff AM, Humphrey PA, Dodge RK, Iglehart JD, Soper JT, Clarke-Pearson DL, Bast RC, Marks JR: Overexpression and mutation of p53 in endometrial carcinoma. Cancer Res 1992, 52: 1622-1627 Kohler MF, Nishii H, Humphrey PA, Sasaki H, Marks J, Bast RC, Clarke-Pearson DL, Boyd J, Berchuck A: Mutation of the p53 tumor-suppressor gene is not a feature of endometrial hyperplasia. Am J Obstet Gynecol 1993, 169:690-694 Ambros RA, Ross JS, Kallakury BVS, Malfetano J, Kim Y, Hwang J, Breese K, Figge J: p53 gene status in endometrial carcinomas showing diffuse positivity for p53 protein by immunohistochemical analysis. Mod Pathol 1995, 8:441-445 King SA, Adas AA, LiVolsi VA, Takahashi H, Behbakht K, McGovern P, Benjamin I, Rubin SC, Boyd J: Expression and mutation analysis of the p53 gene in uterine papillary serous carcinoma. Cancer 1995, 75:2700-2705 Remmele W, Schicketanz K-H: Immunohistochemical determination of estrogen and progesterone receptor content in human breast cancer: computer-assisted image analysis (QIC score) vs. subjective grading (IRS). Path Res Pract 1993, 189:862-866 Fearon ER, Hamilton SR, Vogelstein B: Clonal analysis of human colorectal tumors. Science 1987, 238:193-197 Katabuchi H, van Rees B, Lambers AR, Ronnett BM, Blazes MS, Leach FS, Cho KR, Hedrick L: Mutations in DNA mismatch genes are not responsible for microsatellite instability in most sporadic endometrial carcinomas. Cancer Res 1995, 55:5556-5560 Jones MH, Nakamura Y: Detection of loss of heterozygosity at the human TP53 locus using a dinucleotide repeat polymorphism. Genes Chromosomes & Cancer 1992, 5:89-90 Guzzetta G, Franco B, Trask BJ, Zhang H, SaucedoCardenas 0, Montes de Oca-Luna R, Greenburg F, Chinault AC, Lupski JR, Patel PI: Somatic cell hybrids, sequence-tagged sites, simple repeat polymorphisms, and yeast artificial chromosomes for physical and genetic mapping of proximal 17p. Genomics 1992, 13:551-559 Burks RT, Kessis TD, Cho KR, Hedrick L: Microsatellite instability in endometrial carcinoma. Oncogene 1994, 9:1163-1166 Hollstein M, Sidransky D, Vogelstein B, Harris CC: p53 mutations in human cancers. Science 1991, 253:49-53 Harris CC: 1995 Deichmann lecture-p53 tumor suppressor gene: at the crossroads of molecular carcinogenesis, molecular epidemiology, and cancer risk assessment. Toxicol Lett 1995, 82/83:1-7 Vogelstein B, Fearon ER, Hamilton SR, Kern SE, Preisinger AC, Leppert M, Nakamura Y, White R, Smits AMM, Bos JL: Genetic alterations during colorectaltumor development. N Engl J Med 1988, 319:525-532 Kihana T, Hamada K, Inoue Y, Yano N, Iketani H, Murao S-i, Ukita M, Matsuura S: Mutation and allelic

20.

21.

22.

23.

24.

25.

26.

27.

28.

29. 30.

31.

32.

33.

loss of the p53 gene in endometrial carcinoma. Cancer 1995, 76:72-78 Sundaresan V, Heppell-Parton A, Coleman N, Miozzo M, Sozzi G, Ball R, Cary N, Hasleton P, Fowler W, Rabbitts P: Somatic genetic changes in lung cancer and precancerous lesions. Ann Oncol 1995, 6:S27-S32 Sozzi G, Miozzo M, Donghi R, Pilotti S, Cariani CT, Pastorino U, Della Porta G, Pierotti MA: Deletions of 17p and p53 mutations in preneoplastic lesions of the lung. Cancer Res 1992, 52:6079-6082 Bennett WP, Hollstein MC, Metcalf RA, Welsh JA, He A, Zhu S-M, Kusters I, Resau JH, Trump BF, Lane DP, Harris CC: p53 mutation and protein accumulation during multistage human esophageal carcinogenesis. Cancer Res 1992, 52:6092-6097 Wang LD, Zhou Q, Hong, J-Y, Qiu S-L, Yang CS: p53 protein accumulation and gene mutations in multifocal esophageal precancerous lesions from symptom free subjects in a high incidence area for esophageal carcinoma in Henan, China. Cancer 1996, 7:1244-1249 Orita M, Suzuki Y, Sekiya T, Hayashi K: A rapid and sensitive detection of point mutations and genetic polymorphism using polymerase chain reaction. Genomics 1989, 5:874-879 Baker SJ, Preisinger AC, Jessup JM, Paraskeva C, Markowitz S, Willson JKV, Hamilton S, Vogelstein B: p53 gene mutations occur in combination with 17p allelic deletions as late events in colorectal tumorigenesis. Cancer Res 1990, 50:7717-7722 Kuerbitz SJ, Plunkett BS, Walsh WV, Kastan MB: Wildtype p53 is a cell cycle checkpoint determinant following irradiation. Proc Natl Acad Sci USA 1992, 89:7491-7495 Kastan MB, Zhan Q, El-Diery WS, Carrier F, Jacks T, Walsh WV, Plunkett BS, Vogelstein B, Fornace AJ: A mammalian cell cycle checkpoint pathway utilizing p53 and GADD45 is defective in ataxia-telangiectasia. Cell 1992, 71:587-597 Lowe SW, Ruley E, Jacks T, Housman DE: p53-dependent apoptosis modulates the cytotoxicity of anticancer agents. Cell 1993, 74:957-967 Lowe SW, Schmitt EM, Smith SW, Osborne BA, Jacks T: p53 is required for radiation-induced apoptosis in mouse thymocytes. Nature 1993, 362:847-849 Graeber TG, Osmanian C, Jacks T, Housman DE, Koch CJ, Lowe SW, Giacci AJ: Hypoxia-mediated selection of cells with diminished apoptotic potential in solid tumors. Nature 1996, 379:88-91 Livingstone LR, White A, Sprouse J, Livanos E, Jacks T, Tlsty TD: Altered cell cycle arrest and gene amplification potential accompany loss of wild-type p53. Cell 1992, 70:923-935 Fukasawa K, Choi T, Kuriyama R, Rulong S, Vande Woude GF: Abnormal centrosome amplification in the absence of p53. Science 1996, 271:1744-1747 Prat J, Oliva E, Lerma E, Vaquero M, Matias-Guiu X: Uterine papillary serous adenocarcinoma: a 10-case study of p53 and c-erbB-2 expression and DNA content. Cancer 1994, 74:1778-1783