The Prognostic Significance of Tumor-associated Antigen 22-1-1 Expression in Adenocarcinoma of the Uterine Cervix¹

Previously, we established a new MoAb,³ 22-1-1, against a novel secreted tumor-associated Ag that was produced from mice immunized with SiSo cells that had been derived from adenocarcinoma of the uterine cervix (1). The tissue distribution and biological property of 22-1-1 Ag were already reported (2). An immunohistochemical study revealed that 22-1-1 Ag was expressed in 87.5% of adenocarcinomas of the uterine cervix, 66% of endometrial adenocarcinomas, and 58.8% of ovarian carcinomas. However, it was not detected in normal uterine cervical tissues. Moreover, 22-1-1 Ag was detected in 82.6% of invasive SCCs of the uterine cervix, 16.7% of microinvasive SCCs, and 20% of carcinoma in situ; however, it was not detected in normal uterine squamous epithelium and dysplasias (3). The 22-1-1 Ag expression may relate to tumor cell progression and invasion in SCC of the uterine cervix.

In this study, the relationship between 22-1-1 Ag expression and clinicopathological variables and prognostic significance of 22-1-1 Ag in adenocarcinoma of the uterine cervix were investigated.

Seventy cases with clinical stage I and II pure adenocarcinoma of the cervix were initially treated by hysterectomy at the Department of Gynecology and Obstetrics of Kyushu University Hospital (Kyushu, Japan) from 1979 through 1989. Of 70 cases, 14 cases were excluded for various reasons. Among the 14 excluded cases, there were 3 cases in which pathological material was lost and 11 in which the pertinent paraffin block was not available for immunohistochemical staining after histological review. Therefore, the remaining 56 cases were selected for this study. Fifty-two of 56 patients were treated by radical hysterectomy, and the remaining 4 were treated by simple hysterectomy because the initial and preoperative diagnosis was microinvasive adenocarcinoma of the cervix (stromal invasion measuring less than 3 mm in depth; 4). Pelvic lymphadenectomy and para-aortic lymph node biopsy were performed in 52 and 43 cases, respectively. Patients with involvement of the whole thickness of the cervical wall, the presence of parametrial invasion, and/or pelvic lymph node metastasis were treated with adjuvant external-bean radiation therapy of 41.4–54 Gy to the whole pelvis. Intracavitary brachytherapy was performed if the surgical margin of the vaginal stump was histologically involved or if the free margin measured less than 1 cm. The number of slides examined per case ranged from 9 to 46, with a mean of 21.1 (median, 19).

Histologically, the slides of the hysterectomy specimens were evaluated for the stage and grade of the carcinoma, the depth of myometrial invasion, and the presence or absence of pelvic and/or para-aortic lymph node metastasis, lymph-vascular space invasion, parametrial invasion, and histological type. No patients were lost to follow-up, and the median follow-up duration for survivors is 108 months after the initial surgery (range, 63–199 months).

For immunohistochemical analysis, streptoavidin-biotin methodology (Histofine SAB-PO kit; Nichirei, Tokyo, Japan) was used for formalin-fixed and paraffin-embedded specimens. Four-μm sections were cut from paraffin-emedded blocks. Endogenous peroxidase was blocked with 0.3% hydrogen peroxide in methanol for 30 min. Slides were then washed twice in PBS and incubated with normal rabbit serum diluted in PBS for 15 min. The 22-1-1 hybridoma culture supernatant fluid, diluted 1:20 in PBS, was applied, and the slides were incubated in a humid chamber for 30 min. After two additional washes, sections were incubated with biotinylated rabbit antimouse antibody for 30 min. Sections were washed three times in PBS and incubated with avidin-biotinylated peroxidase complex for 30 min. After three additional washes in PBS, diaminobenzidine tetrahydrochloride working solution was applied. Sections were then counterstained in methylgreen in permount. The entire procedure was performed at room temperature. Negative controls were treated in the same way, but the 22-1-1 MoAb was replaced by mouse IgM. The 22-1-1 Ag was localized with a 1:20 dilution of hybridoma culture supernatant fluid. Tissue sections with less than 5% reactive tumor cells were considered negative, and those with more than 5% reactive cells were defined as positive. In this study, the discriminating points were 25 and 50% as in other reports (5, 6, 7). Five to 25% (weak/focal), 25–50% (moderate), and more than 50% (strong) reactive tumor cells were defined as one, two, and three positive, respectively.

Statistical analysis included Mann-Whitney U test, and the Spearman rank correlation test. The Spearman rank correlation coefficients were used to express the association between 22-1-1 Ag expression and clinicopathological variables. P < 0.05 was regarded as statistically significant. Univariate and multivariate analyses of the prognostic significance of 22-1-1 Ag expression and other variables were performed using the statistical package BMDP 1 L and 2 L on an International Business Machines (IBM) System 4381 Computer (Armonk, NY). Survival rates were estimated by the Kaplan-Meier method, and log-rank tests were performed to test the equality of the survivor functions in univariate analyses. Cox’s proportional hazards model was used: (a) to evaluate the effect of 22-1-1 Ag expression on OS, adjusting for the effects of other variables; and (b) to multivariately estimate survivor functions. We tentatively defined 5, 25, or 50% as a cutoff for “low” and “high” expression. The 25% was the most significantly prognostic effect in OS by multivariate analysis, and, finally, 25% was chosen as the cutoff.

The 22-1-1 Ag was negative in 7, 1+ in 14, 2+ in 26, and 3+ in 9 cases. The 22-1-1 Ag existed both in the cytoplasm and on the membrane of cancer cells (Fig. 1). Moreover, 22-1-1 Ag was weakly detected in the glandular lumen of the tumor cells.

The relationship between expression of 22-1-1 Ag and clinicopathological variables in adenocarcinoma of the uterine cervix was shown in Table 1. There was no correlation between 22-1-1 Ag expression and age, surgical stage, tumor grade, myometrial invasion, lymph-vascular space invasion, lymph node metastasis, and parametrial invasion.

The estimated 5-year OS of patients with low 22-1-1 Ag expression (−/+) and high 22-1-1 Ag expression (++/+++) were 90.5 and 71.4%, respectively. Patients with high 22-1-1 Ag expression had significantly worse OS than those with low 22-1-1 Ag expression (log-rank test, P = 0.0193; Table 2 and Fig. 2). In addition, lymph-node metastasis (P = 0.0301), age (P = 0.0417), and clinical stage (P = 0/0499) were significantly related to OS in the univariate analysis (Table 2).

Multivariate analysis for OS using Cox’s proportional hazard model revealed a prognostic significance in expression of 22-1-1 Ag, clinical stage, age, and tumor grade (Table 3).

To assess intra- and interobserver variation in evaluation of the degree of 22-1-1 expression in this study, we selected 23 of the 56 cases at random. To assess intraobserver variation, one investigator (T. Kak.) evaluated the degree of 22-1-1 expression without previous knowledge of the counts. Linear regression showed that the first evaluation of the degree of expression correlated highly with the second one (r = 0.931; n = 23; P < 0.0001). To assess interobserver variation, two investigators (T. Kak. and K. So.) evaluated the degree of 22-1-1 expression without previous knowledge of the counts. Linear regression showed that one investigator’s degree correlated highly with the other (r = 0.967; n = 23; P < 0.0001).

The 22-1-1 MoAb was generated from mice immunized with SiSo cells that had been derived from adenocarcinoma of the uterine cervix (1). The 22-1-1 MoAb strongly stained adenocarcinomas of the cervix (87.5%), endometrial adenocarcinomas (66.0%), ovarian adenocarcinomas (58.8%), and SCCs of the uterine cervix (87.7%; Ref. 2). In nongynecological cancers, 22-1-1 Ag was detected in esophageal SCCs, gastric adenocarcinomas, colon adenocarcinomas, and pancreatic adenocarcinomas, indicating that the 22-1-1 Ag exists widely in various cancer tissues.

As reported previously, 22-1-1 Ag was shown to be distinct from YH206, GA733, CA125, CEA, and sialyl Lewis molecules in flow cytometric analysis data (2). The tissue distribution of 22-1-1 Ag was also different from those of CA125 and CA19-9 Ags and CEA. There have been no previous reports about a M_r 78,000 tumor-associated Ag secreted in vaginal discharges of uterine cancer patients that has these characteristics. Moreover, there is a possibility that 22-1-1 Ag expression is related to a neoplastic progression. In our previous report (3), immunohistochemical study showed that 22-1-1 Ag was not found in uterine cervical squamous dysplasias. However, 22-1-1 Ag was detected in SCCs. It expressed in 20% of carcinoma in situ, 16.7% of microinvasive carcinomas, and 82.6% of invasive SCCs. The 22-1-1 MoAb reacted weakly with carcinoma in situ, but it strongly stained invasive part of microinvasive carcinomas. These evidences indicate that 22-1-1 Ag is a novel tumor-associated Ag.

In this study concerning the 22-1-1 Ag expression in adenocarcinoma of the cervix, patients with a high 22-1-1 Ag expression had significantly worse OS than those with a low 22-1-1 Ag expression (P = 0.0193). Multivariate analysis showed that 22-1-1 Ag expression correlated significantly and independently with OS. The 22-1-1 Ag expression of adenocarcinoma of the cervix is also significantly correlated with the degree of myometrial invasion. In adenocarcinoma of the cervix, the expression of 22-1-1 Ag may be an important prognostic factor.

Adenocarcinoma of the uterine cervix is less common than SCC among cervical cancers, and in recent years, its incidence has been reported to be gradually increasing (8, 9, 10). However, the etiology, carcinogenesis, and biological characteristics of this adenocarcinoma still remain unclear. Moreover, cervical adenocarcinomas showed lower sensitivity to radiotherapy and chemotherapy as compared with SCCs of the uterine cervix (11, 12).

In adenocarcinoma of the cervix, the clinical efficacy of diagnosis and monitoring the clinical course, and the prognostic value of CA 125, SCC, and CA 19-9 Ags and CEA have been investigated (13, 14, 15, 16, 17, 18, 19, 20, 21, 22). Duk et al. (13) found correlation between elevated serum CA 125 level and stage. Leminen (14) mentioned that elevated CA 125 levels were associated with histological grade. Duk et al. (15) also reported that elevated pretreatment serum CA 125 levels were indicative of lymph node metastasis, and the 5-year survival for patients with elevated levels was 52 versus 96% if pretreatment values were normal in stage IB cervical adenocarcinoma.

Immunohistochemical positive reaction for CEA was seen in 80% of cervical adenocarcinomas (20). Kjorstad et al. (21) reported that patients with initial low serum CEA level had a 5-year survival of 90% in contrast to 11% when pretreatment values were high.

The mucin samples from cervical adenocarcinomas contained extremely high levels of CA 19-9 (16), but there is no correlation between the CA 19-9 levels and clinical status in patients with cervical cancer (22). The 22-1-1 MoAb reactivity was compared with that of CA125, CA19-9 and CEA MoAbs in uterine cancers. As reported previously, 22-1-1 Ag was expressed comparatively broader and stronger in uterine cancers than other tumor-associated Ags (2).

The c-erbB-2 proto-oncogene, also called neu, NGL, and HER-2, is a gene that encodes a growth factor receptor-like molecule with tyrosine kinase activity and has a structure similar to that of the epidermal growth factor receptor. In adenocarcinoma of the uterine cervix, overexpression of c-erbB-2 protein was associated with higher stage and poor prognosis (23).

p21^WAF1/CIP1, which is an inhibitor of G₁ cyclin-dependent kinases, suppresses the cell cycle and inhibits DNA synthesis. Expression of p21^WAF1/CIP1 correlated with a favorable prognosis for patients with adenocarcinoma of the cervix (24).

In preliminary experiments, 22-1-1 Ag was detected in the cervical mucin of uterine cervical cancer patients but not in that of normal healthy donors (2), which indicates that 22-1-1 MoAb can be used in the diagnosis of uterine cancer with cervical mucin. This clinical application of 22-1-1 MoAb is now under investigation.

The novel tumor-associated Ag, 22-1-1 Ag, was derived from adenocarcinoma of the uterine cervix, a different molecule from the tumor-associated Ags and proto-oncogene mentioned previously. The correlation between 22-1-1 Ag expression and prognosis is, therefore, of great interest for more patients, but additional investigations, including detection of 22-1-1 Ag in the serum of cancer patients and 22-1-1 Ag expression in other malignancies, are needed.