Expression of ~53 in Human Esophageal Carcinoma: An lmmunohistochemical Study With Correlation to Proliferating Cell Nuclear Antigen Expression
HIRONOBU SASANO, MD, SHUKICHI MIYAZAKI, MD, YUJI GOOUKON, MD, TETSURO NISHIHIRA, MD, TAKASHI SAWAI, MD, AND HIROSHI NAGURA, MD Immunolocalization of the nuclear protein p53 tumor suppressor gene product is considered to be one of the best methods of detecting a mutated form of ~53. We have studied p53 immunohistochemically by using monoclonal antibody pAbl801 in 15 cases of esophageal squamous cell carcinoma. Immunoreactive p53 was observed in the nuclei of tumor cells in 4% paraformaldehyde-fixed, frozen sections (12 of 15) and paraffin-embedded sections (11 of 15), but not in routinely processed (10% formalin-fixed) specimens. p53 expression was closely correlated with the malignant phenotype, including dysplasia. p53 was not observed in histologically normal mucosa, except in three cases in which scattered immunoreactivity was observed in parabasal and basal cells. Immunostaining of ki67 and proliferating cellular nuclear antigen on adjacent tissue sections revealed that p53 expression was strongly correlated with ki67 and proliferating cellular nuclear antigen in carcinoma and dysplastic cells, but not in normal mucosa, suggesting involvement of the mutated form of p53 in the cell cycle of malignant cells. Immunohistochemical patterns of p53 were not related significantly to clinicopathologic parameters in the cases examined. Therefore, p53 expression was strongly associated with the proliferation of carcinoma cells but not with that of normal cells in esophageal carcinoma. HUM PATHOL !23:1238-1243. Copyright IQ 1992 by W.B. Saunders Company
p53 is a 53-kd nuclear phosphoprotein, first recognized when it was coprecipitated with the SV40 large T antigen from SV-lO-infected murine cells.’ This nuclear protein is highly conserved in vertebrates and is believed to regulate entry into and progression through the normal cell cycle,’ being induced during the transition from GO to Gl phase in the cell cycle.” In most normal fetal and adult tissues p53 is expressed at low levels and has a very short half-life.” In cultured cells increased levels are associated with a mutated form of the protein” or with stabilization of the protein in a complex with viral antigens.’ It also has been demonstrated that mutant p53 can immortalize cells in vitro. In addition, in conjunction with an activated rus oncogene, p53 can produce a fully transformed phenotype in primary cultured fibroblasts.“,’ Conversely, the wildFrom the Department of Pathology and Surgery, Tohoku University School of Medicine. Sendai, Japan. Accepted for publication January 28, 1992. Key words: p53 in esophageal cancer. Address correspondence and reprint requests to Hironobu Sasano, MD, Department of Pathology. Tohoku University School of Medicine, 2-1 Seiryo-machi, Sendai, Japan 980. Copyright 0 1992 by W.B. Saunders Company 0046~8177/92/2311-0009$5.00/O
1238
type protein has a tumor suppressor action.‘Therefore, increased levels of p53 may indicate increased stability and higher steady-state levels of the protein and may serve as a marker for the mutated form of p53 protein in human malignancies. Esophageal cancer is the sixth most common cancer in males worldwide. This type of cancer is a major cause of death in some areas of Asia, including Japan. Abnormalities of the p53 gene have been studied extensively in a variety of human malignancies. In human esophageal carcinoma frequent loss of heterozygosity of chromosome 17p, in which the p53 gene is located, has been reported by Meltzer et al” and Wagata et al.‘” However, in research on the p53 gene itself, there has been only one study (by Hollstein et al”) that reported frequent mutation of p53. To the best of our knowledge, a study of gene expression, including immunohistochemical approaches, has not been reported in human esophageal malignancies. Therefore, we examined immunolocalization of the p53 nuclear protein in 15 cases of esophageal carcinoma by using various methods of fixation and tissue preparation. In addition, to study the relationship of p53 expression with cell proliferation in human malignancies, we performed immunohistochemical analyses of proliferating cellular nuclear antigen (PCNA) (a 36kd S-phase-associated nuclear protein”) and of ki67 (the monoclonal antibody that recognizes the nuclear antigen in proliferating cells’“) on adjacent tissue sections and compared the findings. MATERIALS Materials:
AND METHODS Esophageal
Cancer
Fifteen cases of squamous cell carcinoma of the esophagus were studied. The subjects were all Japanese and were operated on at the Department of Surgery, Tohoku University Hospital, Sendai. Japan, from 1990 to 1991. In four cases invasion was observed up to the submucosa. In all other cases carcinoma was associated with tumor invasion beyond the tunica muscularis (Table 1). Foci of dysplasia were observed in three cases. Nonpathologic or grossly normal esophagus was obtained from sites distant from the carcinoma. Clinical stages, ages, and other clinical and histopathologic parameters of the cases are summarized in Table 1. Collection
and Processing
of Tissue Material
Tissue blocks of both grossly apparent carcinoma and nonpathologic esophagus were trimmed to approximately
~53 EXPRESSION
IN ESOPHAGEAL
CANCER
(Sasano et al)
I\’
RESULTS
Comparison Preparations
Among Different Tissue
In both f’rozen and para~I1-enlbecIdec1 sections irnniunoreactivity was observed in the nuclei but not in the cytoplasnl of the tumor cells. I~~lrnurroreactivity was not present iti interstitial cells, Ivn~ptiowtes, or vascular 1239
HUMAN PATHOLOGY
Volume 23, No. 11 (November
endothelial
cells in the tissue sections of both preparations. In frozen section tissue specimens 12 of 15 were positive for ~53. Immunoreactivity also was observed in the sections directly frozen in OCT compound and postfixed with acetone. In paraffin-embedded sections 11 of 15 examined were stained. As is shown in Table 1, three cases (cases no. 1, 8, and 15) were negative for p53 in both frozen and paraffin-embedded specimens. One case (case no. 14) was positive in frozen sections but not in paraffin-embedded sections. Of 11 cases in which immunoreactivity of p53 was detected in both tissue preparations, seven (cases no. 2, 4, 5, 6, 7, 10, and 13) had a similar percentage of tumor cells positive for p53 in both frozen and paraffin-embedded sections. Routinely processed specimens (ie, fixed in 10% formalin for 48 to 72 hours after gross examination at pathology laboratories) yielded little discernable p53 immunoreactivity.
Nonpathologic
Esophageal
Mucosa
In three cases (cases no. 4, 9, and 11) immunoreactivity against p53 was observed sporadically in both macroscopically and histologically normal esophageal mucosa (Fig 1). The cells positive for p53 were basal and parabasal cells of the mucosa (Fig 1). Cells in the upper two thirds of the epithelium, including keratinizing cells, were negative for p53. Correlation
1992)
including dyskeratosis, positive cells were observed predominantly in the periphery of these nests but not in the central keratinizing areas. Correlation Parameters
With Clinicopathologic
No significant correlation was observed between percentage of ~53 positivity in tumor cells and nuclear grade, histologic grade, mitotic activity, clinical stage, size of the tumor, depth of invasion, or presence of vascular invasion. Comparison With ki67 and Proliferating Cellular Nuclear Antigen lmmunoreactivity In carcinoma localization patterns of p53 were similar to those of ki67 on adjacent frozen sections and to those of PCNA on adjacent paraffin-embedded sections (Fig 4) in the sami cases. However, in nonpathologic esophageal mucosa cells positive for PCNA were observed distinctively in parabasal cells of the epithelium in all the cases esamined, while p53 was negative in all but the three cases clescribed above. Negative
Controls
The tissue sections incubated with 0.01 mol/I, PBS containing bovine serum albumin and normal mouse IgG did not yield the immunoreactivity described above.
With Histologic Types
Immunolocalization of p53 was strongly correlated with the histologic patterns of the cells examined. Prominent immunoreactivity against p53 was observed in atypical cells with a high nucleus to cytoplasm ratio in areas of dysplastic esophageal mucosa but not in surrounding histologically normal mucosa (Fig 2). Most of the infiltrating carcinoma cells forming small islands and/or nests were positive for p53 (Fig 3). In carcinoma forming large cancer nests with central keratinization,
DISCUSSION It has been shown that the 1~53 gene frequently acquires mutations during the development of many I” These mutat%)ns, which cause human malignancies. amino acid substitutions, appear to change the structure of the protein, resulting in a longer half-life, which can be detected as overexpression of p53.“,” It has been shown that overexpression of the p53 protein as deter-
*
1240
FIGURE 1. Case no. II, lmmunostain of ~53 in morphologically normal esophageal mucosa. (Paraffin-embedded section; original magnification X250.)
~53 EXPRESSION
IN ESOPHAGEAL
CANCER
(Sasano et al)
FIGURE 2. Case no. 2. lmmunostoin of ~53 in esophageal dysplasia. ~53 is observed in the area of dysplasia (D) but not in that of surrounding normal mucosa (N). (Frozen section; original magnification (200.)
IniticL(I
b\ ittttttttnoltis~o~~i~ttiis~t~~ is c-losei\. cx)t.relatect \vith thr prcst7icc of a DNA mutation in the p.53 gene in human epithriial ovarian cancer,” primary lung can(‘Cl. ” ;1tic1ttutttati breast cxttcer cell line>. “I ‘llier~fore. Gtnplr it~~trt~t1~ot~iatolo~i~ ttlethods can provide strong Itttt~tut~oloevid~ttc~ ()I‘ SWI1 ;I 11.53 DNA mutation.” calization c,bfp.53 has two other advantages over ;I con\.etttional nucleic. acid-based approach to detec.ting DNA tttutations. l‘he first is the relatively rapid and simple ~ec httiyue N11‘i1nnlutlostaitting and its straightforwardThe second, mot-c ness in identifying p5% tnrttations.‘X impot-tatlt, advatttage is that the expression of p.53 catt o~rrelatcd with various other histoIx localized and pathologic paratltcters. A nucleic acid-based approach usually invcks (.otita1llitl;ttioti 1,~ not~tnaligt~at~t and cells hecausz the tissue is proc’rssed :I> ttonviahlr c :III(XT
FIGURE 3. Case no. 10. lmmunostain of ~53 in ,esophageal carcinoma. (Paraffin-embedded section; original magnificatic’n k 150.)
1241
a ttixs.
‘l‘hrrrfore, the results aw stt~otigl\ i1iHuet~c~etl by the prrcx7ltagr of cancer cells present in itic specimett examined. This ttta!’ he one 01‘ ttte reasons we detected mor‘e ftXY]tlettt ahtlot-tlla~itiC5 Of JI:i:$ ill hUtlla~t eSOphageal carcinoma than did Hollstein et al.” who perfi~rrnrti pol~nicrase chain rest-tion amplitk atiort and dit7c.t sryttencittg without histo~~athologi~~ cxatnitiations. hcteroge111 addition, expression of ~53 is ~tldwlly tieous attd ~.otiipletely different among different carcinottias, ;I:, shown in the present stud\. Itlltttutto~tistoc.heIrlic.al attalvsis c’>fp53 ma> he the mobt straiglllforward and poleniiallv wkkspreatl means of identif\lng p53 mutations in httntati txtcer specimens. ah was proposed h!, Iggo et al.‘” Flowe\,er, this approach also has a ttumber 01‘drawbacks. 2s was shown in 1he prrsettt stud!.. tissur pteparatioti i\ cluitc critical
HUMAN PATHOLOGY
Volume 23, No. 11 (November
1992)
FIGURE 4. Case no. 10. Immunostain of PCNA in the same case of esophageal carcinoma as in Fig 3. (Original magnification x250.)
in demonstrating p53 immunoreactivity. It is very difficult to demonstrate immunoreactivity in routinely processed specimens in surgical pathology files when using the antibody PAb 180 1, which is considered to be more sensitive on tissue sections than other antibodies. I4 In addition, it is important to study the possibility that the loss of both p53 alleles can occur, which would result in the absence of the p53 protein (ie, negative immunostain despite p53 DNA abnormalities). Thus, immunohistochemistry alone may be insufficient to elucidate the possible role of ~53. However, despite these potential drawbacks, an immunohistochemical approach of p53 detecting overexpression on tissue sections is considered to be of potential value in the study of the involvement of the p53 gene in human malignancies, especially for histopathologists. It is thought that p53 plays an important role in the control of the cell cycle and that mutations can confer a malignant phenotype or promote the growth of tumor cells.‘” The present study shows that immunolocalization of p53 in esophageal squamous cell carcinoma is similar to that of ki67 and PCNA, well-established cell cycle-related antigens on adjacent tissue sections. p53 is not expressed in normal mucosa in which active cell proliferation occurs in the parabasal layer, as is shown by PCNA immunoreactivity. Therefore, at least in human esophageal squamous cell carcinoma, immunoreactivity of p53 as detected by pAbl80 1 (ie, expression of a mutated p53 protein) is strongly correlated with cell proliferation of malignant phenotypes, but not with the proliferation of normal counterparts. The data also demonstrate a strong correlation between p53 overexpression and malignant transformation in esophageal disorders, as has been reported for colon” and breast” lesions. Prominent p53 expression has been observed in cells exhibiting dysplasia. p53 abnormalities generally have not been detected in various normal tissues. 17~‘8~20~22 In the present study the cases of scattered 1242
p53 immunoreactivity observed in morphologically normal esophageal mucosa may indicate that these cells are morphologically normal but are already in the process of malignant transformation in terms of abnormal cell proliferation. It is becoming apparent that abnormalities of the p53 gene occur with high frequency in a wide spectrum of human malignancies.‘” Because of DNA mutations, which occur throughout a significant region of the coding sequence of ~53, the gene can be a broad target for DNA damage in carcinogenesis. I ’ This is consistent with the prominent expression of p53 in dysplastic esophageal mucosa. The question arises as to whether p53 abnormalities have any independent prognostic significance or any correlation with biologic behavior. In the present study p53 expression did not appear to correlate with any established prognostic indicators of esophageal carcinoma; however, due to the short clinical follow-up of the patients studied we have no data that compare postoperative course and p53 immunolocalization. Nevertheless, no significant correlation was reported between p53 abnormalities and prognosis and/or clinical outcome in ovarian,” colorectal,2” or breast’” cancer. For pathologists the practical value of detecting p53 abnormalities is at present limited by the fact that it is extremely difficult to perform immunostaining successfully on routinely processed pathology specimens, and a nucleic acid-based approach is not yet feasible in a majority of clinical laboratories. Even if immunostaining is successfully done on frozen section specimens the weak correlation with prognosis and clinical outcome described above may also discourage widespread application in diagnostic pathology laboratories. However, a strong relationship between p53 expression and malignant transformation, which was observed in the present study as well as in others,‘“*“’ can be of potential value in differentiating premalignant conditions. Therefore,
~53 EXPRESSION
IN ESOPHAGEAL
1243
CANCER
(Sasano et al:)
HIRONOBU SASANO, MD, SHUKICHI MIYAZAKI, MD, YUJI GOOUKON, MD, TETSURO NISHIHIRA, MD, TAKASHI SAWAI, MD, AND HIROSHI NAGURA, MD Immunolocalization of the nuclear protein p53 tumor suppressor gene product is considered to be one of the best methods of detecting a mutated form of ~53. We have studied p53 immunohistochemically by using monoclonal antibody pAbl801 in 15 cases of esophageal squamous cell carcinoma. Immunoreactive p53 was observed in the nuclei of tumor cells in 4% paraformaldehyde-fixed, frozen sections (12 of 15) and paraffin-embedded sections (11 of 15), but not in routinely processed (10% formalin-fixed) specimens. p53 expression was closely correlated with the malignant phenotype, including dysplasia. p53 was not observed in histologically normal mucosa, except in three cases in which scattered immunoreactivity was observed in parabasal and basal cells. Immunostaining of ki67 and proliferating cellular nuclear antigen on adjacent tissue sections revealed that p53 expression was strongly correlated with ki67 and proliferating cellular nuclear antigen in carcinoma and dysplastic cells, but not in normal mucosa, suggesting involvement of the mutated form of p53 in the cell cycle of malignant cells. Immunohistochemical patterns of p53 were not related significantly to clinicopathologic parameters in the cases examined. Therefore, p53 expression was strongly associated with the proliferation of carcinoma cells but not with that of normal cells in esophageal carcinoma. HUM PATHOL !23:1238-1243. Copyright IQ 1992 by W.B. Saunders Company
p53 is a 53-kd nuclear phosphoprotein, first recognized when it was coprecipitated with the SV40 large T antigen from SV-lO-infected murine cells.’ This nuclear protein is highly conserved in vertebrates and is believed to regulate entry into and progression through the normal cell cycle,’ being induced during the transition from GO to Gl phase in the cell cycle.” In most normal fetal and adult tissues p53 is expressed at low levels and has a very short half-life.” In cultured cells increased levels are associated with a mutated form of the protein” or with stabilization of the protein in a complex with viral antigens.’ It also has been demonstrated that mutant p53 can immortalize cells in vitro. In addition, in conjunction with an activated rus oncogene, p53 can produce a fully transformed phenotype in primary cultured fibroblasts.“,’ Conversely, the wildFrom the Department of Pathology and Surgery, Tohoku University School of Medicine. Sendai, Japan. Accepted for publication January 28, 1992. Key words: p53 in esophageal cancer. Address correspondence and reprint requests to Hironobu Sasano, MD, Department of Pathology. Tohoku University School of Medicine, 2-1 Seiryo-machi, Sendai, Japan 980. Copyright 0 1992 by W.B. Saunders Company 0046~8177/92/2311-0009$5.00/O
1238
type protein has a tumor suppressor action.‘Therefore, increased levels of p53 may indicate increased stability and higher steady-state levels of the protein and may serve as a marker for the mutated form of p53 protein in human malignancies. Esophageal cancer is the sixth most common cancer in males worldwide. This type of cancer is a major cause of death in some areas of Asia, including Japan. Abnormalities of the p53 gene have been studied extensively in a variety of human malignancies. In human esophageal carcinoma frequent loss of heterozygosity of chromosome 17p, in which the p53 gene is located, has been reported by Meltzer et al” and Wagata et al.‘” However, in research on the p53 gene itself, there has been only one study (by Hollstein et al”) that reported frequent mutation of p53. To the best of our knowledge, a study of gene expression, including immunohistochemical approaches, has not been reported in human esophageal malignancies. Therefore, we examined immunolocalization of the p53 nuclear protein in 15 cases of esophageal carcinoma by using various methods of fixation and tissue preparation. In addition, to study the relationship of p53 expression with cell proliferation in human malignancies, we performed immunohistochemical analyses of proliferating cellular nuclear antigen (PCNA) (a 36kd S-phase-associated nuclear protein”) and of ki67 (the monoclonal antibody that recognizes the nuclear antigen in proliferating cells’“) on adjacent tissue sections and compared the findings. MATERIALS Materials:
AND METHODS Esophageal
Cancer
Fifteen cases of squamous cell carcinoma of the esophagus were studied. The subjects were all Japanese and were operated on at the Department of Surgery, Tohoku University Hospital, Sendai. Japan, from 1990 to 1991. In four cases invasion was observed up to the submucosa. In all other cases carcinoma was associated with tumor invasion beyond the tunica muscularis (Table 1). Foci of dysplasia were observed in three cases. Nonpathologic or grossly normal esophagus was obtained from sites distant from the carcinoma. Clinical stages, ages, and other clinical and histopathologic parameters of the cases are summarized in Table 1. Collection
and Processing
of Tissue Material
Tissue blocks of both grossly apparent carcinoma and nonpathologic esophagus were trimmed to approximately
~53 EXPRESSION
IN ESOPHAGEAL
CANCER
(Sasano et al)
I\’
RESULTS
Comparison Preparations
Among Different Tissue
In both f’rozen and para~I1-enlbecIdec1 sections irnniunoreactivity was observed in the nuclei but not in the cytoplasnl of the tumor cells. I~~lrnurroreactivity was not present iti interstitial cells, Ivn~ptiowtes, or vascular 1239
HUMAN PATHOLOGY
Volume 23, No. 11 (November
endothelial
cells in the tissue sections of both preparations. In frozen section tissue specimens 12 of 15 were positive for ~53. Immunoreactivity also was observed in the sections directly frozen in OCT compound and postfixed with acetone. In paraffin-embedded sections 11 of 15 examined were stained. As is shown in Table 1, three cases (cases no. 1, 8, and 15) were negative for p53 in both frozen and paraffin-embedded specimens. One case (case no. 14) was positive in frozen sections but not in paraffin-embedded sections. Of 11 cases in which immunoreactivity of p53 was detected in both tissue preparations, seven (cases no. 2, 4, 5, 6, 7, 10, and 13) had a similar percentage of tumor cells positive for p53 in both frozen and paraffin-embedded sections. Routinely processed specimens (ie, fixed in 10% formalin for 48 to 72 hours after gross examination at pathology laboratories) yielded little discernable p53 immunoreactivity.
Nonpathologic
Esophageal
Mucosa
In three cases (cases no. 4, 9, and 11) immunoreactivity against p53 was observed sporadically in both macroscopically and histologically normal esophageal mucosa (Fig 1). The cells positive for p53 were basal and parabasal cells of the mucosa (Fig 1). Cells in the upper two thirds of the epithelium, including keratinizing cells, were negative for p53. Correlation
1992)
including dyskeratosis, positive cells were observed predominantly in the periphery of these nests but not in the central keratinizing areas. Correlation Parameters
With Clinicopathologic
No significant correlation was observed between percentage of ~53 positivity in tumor cells and nuclear grade, histologic grade, mitotic activity, clinical stage, size of the tumor, depth of invasion, or presence of vascular invasion. Comparison With ki67 and Proliferating Cellular Nuclear Antigen lmmunoreactivity In carcinoma localization patterns of p53 were similar to those of ki67 on adjacent frozen sections and to those of PCNA on adjacent paraffin-embedded sections (Fig 4) in the sami cases. However, in nonpathologic esophageal mucosa cells positive for PCNA were observed distinctively in parabasal cells of the epithelium in all the cases esamined, while p53 was negative in all but the three cases clescribed above. Negative
Controls
The tissue sections incubated with 0.01 mol/I, PBS containing bovine serum albumin and normal mouse IgG did not yield the immunoreactivity described above.
With Histologic Types
Immunolocalization of p53 was strongly correlated with the histologic patterns of the cells examined. Prominent immunoreactivity against p53 was observed in atypical cells with a high nucleus to cytoplasm ratio in areas of dysplastic esophageal mucosa but not in surrounding histologically normal mucosa (Fig 2). Most of the infiltrating carcinoma cells forming small islands and/or nests were positive for p53 (Fig 3). In carcinoma forming large cancer nests with central keratinization,
DISCUSSION It has been shown that the 1~53 gene frequently acquires mutations during the development of many I” These mutat%)ns, which cause human malignancies. amino acid substitutions, appear to change the structure of the protein, resulting in a longer half-life, which can be detected as overexpression of p53.“,” It has been shown that overexpression of the p53 protein as deter-
*
1240
FIGURE 1. Case no. II, lmmunostain of ~53 in morphologically normal esophageal mucosa. (Paraffin-embedded section; original magnification X250.)
~53 EXPRESSION
IN ESOPHAGEAL
CANCER
(Sasano et al)
FIGURE 2. Case no. 2. lmmunostoin of ~53 in esophageal dysplasia. ~53 is observed in the area of dysplasia (D) but not in that of surrounding normal mucosa (N). (Frozen section; original magnification (200.)
IniticL(I
b\ ittttttttnoltis~o~~i~ttiis~t~~ is c-losei\. cx)t.relatect \vith thr prcst7icc of a DNA mutation in the p.53 gene in human epithriial ovarian cancer,” primary lung can(‘Cl. ” ;1tic1ttutttati breast cxttcer cell line>. “I ‘llier~fore. Gtnplr it~~trt~t1~ot~iatolo~i~ ttlethods can provide strong Itttt~tut~oloevid~ttc~ ()I‘ SWI1 ;I 11.53 DNA mutation.” calization c,bfp.53 has two other advantages over ;I con\.etttional nucleic. acid-based approach to detec.ting DNA tttutations. l‘he first is the relatively rapid and simple ~ec httiyue N11‘i1nnlutlostaitting and its straightforwardThe second, mot-c ness in identifying p5% tnrttations.‘X impot-tatlt, advatttage is that the expression of p.53 catt o~rrelatcd with various other histoIx localized and pathologic paratltcters. A nucleic acid-based approach usually invcks (.otita1llitl;ttioti 1,~ not~tnaligt~at~t and cells hecausz the tissue is proc’rssed :I> ttonviahlr c :III(XT
FIGURE 3. Case no. 10. lmmunostain of ~53 in ,esophageal carcinoma. (Paraffin-embedded section; original magnificatic’n k 150.)
1241
a ttixs.
‘l‘hrrrfore, the results aw stt~otigl\ i1iHuet~c~etl by the prrcx7ltagr of cancer cells present in itic specimett examined. This ttta!’ he one 01‘ ttte reasons we detected mor‘e ftXY]tlettt ahtlot-tlla~itiC5 Of JI:i:$ ill hUtlla~t eSOphageal carcinoma than did Hollstein et al.” who perfi~rrnrti pol~nicrase chain rest-tion amplitk atiort and dit7c.t sryttencittg without histo~~athologi~~ cxatnitiations. hcteroge111 addition, expression of ~53 is ~tldwlly tieous attd ~.otiipletely different among different carcinottias, ;I:, shown in the present stud\. Itlltttutto~tistoc.heIrlic.al attalvsis c’>fp53 ma> he the mobt straiglllforward and poleniiallv wkkspreatl means of identif\lng p53 mutations in httntati txtcer specimens. ah was proposed h!, Iggo et al.‘” Flowe\,er, this approach also has a ttumber 01‘drawbacks. 2s was shown in 1he prrsettt stud!.. tissur pteparatioti i\ cluitc critical
HUMAN PATHOLOGY
Volume 23, No. 11 (November
1992)
FIGURE 4. Case no. 10. Immunostain of PCNA in the same case of esophageal carcinoma as in Fig 3. (Original magnification x250.)
in demonstrating p53 immunoreactivity. It is very difficult to demonstrate immunoreactivity in routinely processed specimens in surgical pathology files when using the antibody PAb 180 1, which is considered to be more sensitive on tissue sections than other antibodies. I4 In addition, it is important to study the possibility that the loss of both p53 alleles can occur, which would result in the absence of the p53 protein (ie, negative immunostain despite p53 DNA abnormalities). Thus, immunohistochemistry alone may be insufficient to elucidate the possible role of ~53. However, despite these potential drawbacks, an immunohistochemical approach of p53 detecting overexpression on tissue sections is considered to be of potential value in the study of the involvement of the p53 gene in human malignancies, especially for histopathologists. It is thought that p53 plays an important role in the control of the cell cycle and that mutations can confer a malignant phenotype or promote the growth of tumor cells.‘” The present study shows that immunolocalization of p53 in esophageal squamous cell carcinoma is similar to that of ki67 and PCNA, well-established cell cycle-related antigens on adjacent tissue sections. p53 is not expressed in normal mucosa in which active cell proliferation occurs in the parabasal layer, as is shown by PCNA immunoreactivity. Therefore, at least in human esophageal squamous cell carcinoma, immunoreactivity of p53 as detected by pAbl80 1 (ie, expression of a mutated p53 protein) is strongly correlated with cell proliferation of malignant phenotypes, but not with the proliferation of normal counterparts. The data also demonstrate a strong correlation between p53 overexpression and malignant transformation in esophageal disorders, as has been reported for colon” and breast” lesions. Prominent p53 expression has been observed in cells exhibiting dysplasia. p53 abnormalities generally have not been detected in various normal tissues. 17~‘8~20~22 In the present study the cases of scattered 1242
p53 immunoreactivity observed in morphologically normal esophageal mucosa may indicate that these cells are morphologically normal but are already in the process of malignant transformation in terms of abnormal cell proliferation. It is becoming apparent that abnormalities of the p53 gene occur with high frequency in a wide spectrum of human malignancies.‘” Because of DNA mutations, which occur throughout a significant region of the coding sequence of ~53, the gene can be a broad target for DNA damage in carcinogenesis. I ’ This is consistent with the prominent expression of p53 in dysplastic esophageal mucosa. The question arises as to whether p53 abnormalities have any independent prognostic significance or any correlation with biologic behavior. In the present study p53 expression did not appear to correlate with any established prognostic indicators of esophageal carcinoma; however, due to the short clinical follow-up of the patients studied we have no data that compare postoperative course and p53 immunolocalization. Nevertheless, no significant correlation was reported between p53 abnormalities and prognosis and/or clinical outcome in ovarian,” colorectal,2” or breast’” cancer. For pathologists the practical value of detecting p53 abnormalities is at present limited by the fact that it is extremely difficult to perform immunostaining successfully on routinely processed pathology specimens, and a nucleic acid-based approach is not yet feasible in a majority of clinical laboratories. Even if immunostaining is successfully done on frozen section specimens the weak correlation with prognosis and clinical outcome described above may also discourage widespread application in diagnostic pathology laboratories. However, a strong relationship between p53 expression and malignant transformation, which was observed in the present study as well as in others,‘“*“’ can be of potential value in differentiating premalignant conditions. Therefore,
~53 EXPRESSION
IN ESOPHAGEAL
1243
CANCER
(Sasano et al:)
