CLINICAL FEATURES AND IMMUNOEXPRESSION OF p53, MIB-1 AND PROLIFERATING CELL NUCLEAR ANTIGEN IN ADRENAL NEOPLASMS
ANTONIO C. P. MARTINS, ADAUTO J. COLOGNA, SILVIO TUCCI, JR., HAYLTON J. SUAID AND RODRIGO A. R. FALCONI
From the Medical School of Ribeirao, University of Sao Paulo, Sao Paulo, Brazil
ABSTRACT
Purpose: We evaluated the clinical features and immunoreactivity of p53 protein, MIB-1 antigen and proliferating cell nuclear antigen (PCNA) in adrenal neoplasms.
Materials and Methods: A total of 26 patients with adrenocortical adenoma and 24 patients with carcinoma were treated with adrenalectomy. Clinical features and immunohistochemical reactions were compared in adult vs pediatric tumors.
Results: There was a bimodal age distribution of carcinomas and adenomas, with a first peak occurring before age 5 years. The proportion of carcinomas in children (18 of 29) was higher than in adults (6 of 21). Carcinoma and adenoma occurring in children presented more commonly as the virilizing syndrome, while in adults Cushing’s syndrome was more common. All adenomas in adults were p53 negative, while in children 4 of 11 adenomas (36%) were p53 positive. Histolog- ical Weiss criteria were the most reliable pathological features to distinguish adenoma from carcinoma. Other pathological features, including tumor weight, rate of mitotic figures and immunoexpression of p53 protein, MIB-1 antigen and PCNA, exhibited a striking difference in adenomas and carcinomas but none demonstrated sensitivity or specificity of 100%. Of all the computerized tomographic characteristics analyzed, including tumor size, shape, necrosis/hem- orrhage, attenuation and contrast enhancement, only tumor size (greater than 5 cm) showed sensitivity and specificity of 100% in the differential diagnosis. Children and adults with carci- noma had similar curves of survival (p = 0.76). Carcinoma stage and PCNA immunoexpression displayed an association with outcome.
Conclusions: Endocrine syndromes differed in adults and children but other clinical features were similar in both groups. The role of p53 protein, MIB-1 antigen and proliferating cell nuclear antigen in discrimination of adenomas from carcinomas is unclear.
KEY WORDS: proliferating cell nuclear antigen, protein p53, MIB-1 antibody, adrenal gland neoplasms
Due to the low incidence of adrenal tumors, there is a paucity of data on their epidemiology, natural history and response to treatment, despite improvements in clinical de- tection.1 Differentiation between benign and malignant tu- mors is difficult and usually is determined by a combination of parameters such as tumor size, image features of comput- erized tomography (CT) or magnetic resonance imaging and histological criteria. However, sometimes the final diagnosis is dictated by the clinical behavior.1-3
Some reports conclude that adrenal carcinomas in children and adolescents are clinically and histologically different from those occurring in adults.3,4 There are also doubts re- garding whether the molecular markers could help in the differential diagnosis or in predicting outcome. The mutation of p53 suppressor gene is reported to occur in a varying proportion of adrenal carcinomas (27% to 52%), while data for adenomas are conflicting.3,5,6 This disparity generates controversy regarding whether p53 mutations may represent a predictive factor for tumor outcome or a marker for distinc- tion between benign and malignant disease.3,5-9
There are few reports on other molecular markers such as MIB-1 antigen and proliferating cell nuclear antigen (PCNA). Some data reveal that MIB-1 immunoexpression distinguishes adenomas from carcinomas and may or may not be associated with outcome. 3,7,8,10 Other studies have demonstrated that over expression of PCNA is helpful in the differential diagnosis but has no prognostic value. 9-12 The aim of this retrospective study was to evaluate some epide-
miological parameters and the relevance of some clinical pathological features and molecular markers (p53, MIB-1 and PCNA) in the differential diagnosis and clinical course of adrenal neoplasms.
MATERIALS AND METHODS
Of 53 patients with the diagnosis of primary adrenal neo- plasm operated on at our institution between January 1979 and December 1999, 50 were selected for this study. Inclu- sion criteria were adequate clinical pathological data, appro- priate specimens for histological and immunohistochemical analysis, and followup of 5 years or more. Of the 3 patients excluded from the study 1 died of sepsis in the immediate postoperative period and 2 were lost to followup.
The disease was staged according to the proposal of Henley et al.13 All patients underwent appropriate clinical and hor- monal investigation. Primary treatment consisted of total adrenalectomy. Adjuvant chemotherapy with mitotane (o,p- DDD) or a combination of doxorubicin, cyclophosphamide, vincristine and 5-fluorouracil was administered in patients with carcinoma in whom recurrent (2 patients) or metastatic disease (8 patients) was diagnosed. Both patients with recur- rent disease were also treated with reoperation but the tu- mor was unresectable in 1.
Histological slides were revised to confirm the diagnosis according to the criteria of Weiss et al,2 as well as to select representative areas of tumor, and corresponding formalin fixed, paraffin embedded tissue blocks were sliced (4 mm) for p53, MIB-1 or PCNA immunohistochemical analysis. Immu-
Submitted for publication August 9, 2004.
2138
15
Bank
10
can Cam
Ad Toma
5
0
0-5
5-10 10-20 20-30 30-40 40-50 50-60
Agu [your1; FIG. 1
nohistochemical reactions were performed via the avidin- biotin-immunoperoxidase method with antigen retrieval14 using antibody anti-p53 (DO-7), anti-PCNA (PC10) or anti- MIB-1 (M7240). Sections of known squamous cell carcinoma of the penis and primary antibody replaced by normal mouse serum were used as positive and negative controls, respec- tively.
All slides were examined using an automatic analyzer im- aging system. Two investigators working together in a blind analysis scored nuclear reactivity for each tumor marker. The labeling index was established after counting at least 1,000 cells. To find the optimal separation of patients with carcinoma from those with adenoma, we chose a cutoff pro- portion of labeled nuclei after a preliminary receiver charac- teristic curve analysis of each molecular marker. To test the ability to predict outcome, the best cutoff point found was the median value of labeling index of carcinomas from patients who died.
Comparison and/or association of variables was analyzed with the required tests cited in the results section. All anal- yses were conducted using Stata Statistical Software, release 6.0®. The a level was set at p <0.05.
RESULTS
There was a bimodal age distribution of carcinomas and adenomas, with the first peak occurring before age 5 years
and the second in the fifth decade of life for carcinomas and in the third decade for adenomas (fig. 1).
Table 1 summarizes the epidemiological and clinical data. The proportion of carcinomas in children was higher than in adults (Fisher’s exact test p = 0.02, OR 4.09). Difference in tumor weights of adenomas and carcinomas in pediatric and adult patients calculated by the Kruskal-Wallis test was highly significant (p <0.0001) but Dunn’s multiple compari- son tests showed no difference in adult vs pediatric carcino- mas (p >0.05) or adenomas (p >0.05).
Pure virilization confirmed by measuring serum testoster- one, adrenal androgens and urinary 17-ketosteroid occurred in 61% of children and 0% of adults harboring carcinomas. The prevalence of Cushing’s syndrome confirmed by in- creased serum cortisol and urinary 17-hydroxycorticosteroid associated or not associated with virilization was 33% in children and 83% in adults with carcinomas. For adenomas the prevalence of pure virilization and Cushing’s syndrome associated or not associated with virilization was 63% and 36%, respectively, in children and 20% and 60%, respectively, in adults.
Table 2 outlines the results of immunohistochemical anal- ysis, tumor weight and rate of mitotic figures (MFs). All adrenal adenomas in adults were p53 negative, while in children 4 of 11 adenomas were p53 positive. Logistic regres- sion of selected variables MF, MIB-1, PCNA and p53 re- vealed significance only for MF (p = 0.001, OR 99.4) and PCNA (p = 0.02, OR 15.6).
Comparison of the scores of tumor marker positivity in pediatric vs adult carcinomas and adenomas is outlined in table 3. The scores of positivity of these markers were similar in pediatric vs adult adenomas (p >0.05) and in pediatric vs adult carcinomas (p >0.05). In p53 positive carcinomas p53 reactivity demonstrated an association with the rate of mi- totic figures but not with tumor size (table 4).
Table 5 outlines the results of CT scanning. Logistic re- gression showed that only tumor size was an independent factor in the differential diagnosis (p <0.0001, OR 276.0).
All 26 adenomas, designated as such by histological crite- ria, had a clinically benign outcome without recurrence or metastasis. Of 18 children with carcinomas 8 (44%) died of the disease within a mean period of 21.5 months (range 8 to 72). The corresponding figures for adults were 2 of 6 (33%) and 13.5 months (range 12 to 15), respectively.
Number of carcinoma stages I, II, III and IV were 5, 6, 5 and 8, respectively. Figure 2 displays the relationship be- tween patient survival and risk factors. Table 6 lists the results of multivariate analysis to determine the relative risk
| TABLE 1. Epidemiological and clinical parameters of patients with adrenal tumors | ||||||||
|---|---|---|---|---|---|---|---|---|
| Adenoma | Ca | |||||||
| Children | Adults | Children | Adults | |||||
| Yrs median age (range) | 2 | (1-5) | 29 | (24-53) | 3.5 | (1-17) | 44 | (27-51) |
| Race: | ||||||||
| White | 10 | 15 | 16 | 5 | ||||
| Black/biracial | 1 | 0 | 2 | 1 | ||||
| No. sex: | ||||||||
| Male | 4 | 1 | 7 | 0 | ||||
| Female | 7 | 14 | 11 | 6 | ||||
| No. Adrenal gland: | ||||||||
| Lt | 10 | 7 | 8 | 2 | ||||
| Rt | 1 | 7 | 10 | 4 | ||||
| Bilat | 0 | 1 | 0 | 0 | ||||
| Median gm tumor wt (range) | 26 | (5-70) | 27 | (10-140) | 265 | (18-1,280) | 391 | (75-2,820) |
| Median cm tumor size (range) | 3.5 | (1.5-6.0) | 3.4 | (2.0-8.7) | 9.0 | (5.2-19.0) | 12.0 | (6.5-19.0) |
| No. endocrine syndrome: | ||||||||
| Absent | 0 | 1 | 1 | 1 | ||||
| Cushing's | 1 | 8 | 2 | 0 | ||||
| Virilization | 7 | 3 | 11 | 0 | ||||
| Cushing's with virilization | 3 | 1 | 4 | 5 | ||||
| Conn's | 0 | 2 | 0 | 0 | ||||
| TABLE 2. Tumor histological category versus risk factors | ||||||
|---|---|---|---|---|---|---|
| Risk Factors | Category (No. adenoma/adenoca) | Statistical Analysis | ||||
| p Value (OR)* | % Sensitivity | % Specificity | Pos Predictive Value | Neg Predictive Value | ||
| Tumor wt (gm): | ||||||
| Greater than 100 | 19/1 | <0.0001 (95.0) | 79.1 | 96.1 | 95.0 | 83.3 |
| 100 or Less | 5/25 | |||||
| Mitotic figures:+ | ||||||
| Greater than 5/50 high power fields | 22/3 | <0.0001 (84.3) | 91.6 | 88.4 | 88.0 | 92.0 |
| 5/50 High power fields or less | 2/23 | |||||
| MIB-1:# | ||||||
| Pos | 21/8 | <0.0001 (15.7) | 87.5 | 69.2 | 72.4 | 85.7 |
| Neg | 3/16 | |||||
| PCNA:§ | ||||||
| Pos | 21/5 | <0.0001 (29.4) | 87.5 | 80.7 | 80.7 | 87.5 |
| Neg | 3/21 | |||||
| p53: | ||||||
| Pos | 15/4 | <0.001 (9.1) | 62.5 | 84.6 | 78.9 | 70.9 |
| Neg | 9/22 | |||||
* Fisher’s exact test.
7 Reduced from ×400.
# Cutoff 5% or greater.
§ Cutoff 15% or greater.
Cutoff 10% or greater.
| TABLE 3. Tumor marker positivity in children and adults | |||
|---|---|---|---|
| % Ca Mean + SE (median) | % Adenoma Mean ± SE (median) | p Value* | |
| Children:+ | |||
| p53 | 18.6 ± 1.9 (18.5) | 10.3 ±2.3 (7.0) | <0.05 |
| MIB-1 | 14.4 ± 2.8 (18.0) | 2.5 ± 1.3 (0.4) | <0.01 |
| PCNA | 21.6 ± 13.5 (22.6) | 8.7 ± 1.8 (8.4) | <0.01 |
| Mitotic figures/50 high power fields# | 22.3 ± 4.1 (19.0) | 2.6 ± 0.7 (2.0) | <0.01 |
| Adults:+ | |||
| p53 | 15.5 ± 2.7 (13.0) | 5.7 ± 0.7 (5.5) | <0.05 |
| MIB-1 | 17.8 ±5.8 (16.0) | 0.2 ± 0.06(0.2) | <0.01 |
| PCNA | 22.5 ±2.2 (20.2) | 6.9 ± 1.8 (3.8) | <0.01 |
| Mitotic figures/50 high power fields# | 13.6 ± 2.5 (12.5) | 0.5 ± 0.2 (0.0) | <0.01 |
* Kruskal-Wallis and Dunn’s multiple comparison tests.
+ Percentage of stained nuclei measured for p53, MIB-1 and PCNA.
# Reduced from ×400.
| Pt No. | p53 Labeled Nuclei/ 1,000 Cells | Mitotic Figures/ 50 High Power Fields* | CT Tumor Size (cm) |
|---|---|---|---|
| 1 | 250 | 32 | 8 |
| 2 | 180 | 23 | 19 |
| 3 | 270 | 50 | 14 |
| 4 | 330 | 68 | 13 |
| 5 | 170 | 4 | 6 |
| 6 | 190 | 32 | 12 |
| 7 | 240 | 7 | 9 |
| 8 | 280 | 17 | 9 |
| 9 | 266 | 21 | 15 |
| 10 | 244 | 23 | 6 |
| 11 | 212 | 13 | 5 |
| 12 | 130 | 43 | 8 |
| 13 | 140 | 11 | 12 |
| 14 | 230 | 8 | 10 |
| 15 | 250 | 21 | 19 |
Multiple regression was used to compare p53 versus mitotic figures (p = 0.02, r2 = 0.33) and p53 versus tumor size (p = 0.7, r2 = 0.008). * Reduced from ×400.
of both factors associated with survival in univariate analy- sis.
DISCUSSION
Our data reveal that there is a bimodal pattern of age distribution not only for adrenocortical carcinomas, but also
for adenomas. The bimodal age prevalence of carcinomas was reported previously by others. 4, 15
Adenomas and carcinomas were more frequent in white females than in males but sex prevalence has varied widely in reviewed series.1,3,4 Adrenal carcinomas and adenomas occurring in children present more commonly as the viriliz- ing syndrome, while in adulthood Cushing’s syndrome pre- dominates. These findings confirm those of previous stud- ies. 1,4
Despite the widespread understanding of the unreliability of differential diagnosis of adrenal tumors, all 26 cases cat- egorized histopathologically as adenomas had a good clinical outcome. This result means that the criteria of Weiss et al2 worked as well in children as in adults. These findings are in accordance with those of a recent study3 and in disagreement with others. 4, 16
All adenomas in adults were p53 negative, which confirms previous reports.3,5,17 However, while adrenocortical adeno- mas can exhibit p53 mutation,6 we detected p53 over expres- sion only in children. In p53 positive carcinomas there was a significant correlation between p53 reactivity and MF but not with tumor size. MF reflects cell proliferation status and affects nuclear DNA content. In this series DNA ploidy was not investigated, but data published elsewhere have demon- strated that there is an association between p53 reactivity and ploidy, and that aneuploid tumors are frequently asso- ciated with a poor prognosis.16 Flow cytometric DNA content analysis appears to be as effective a predictor of outcome as size and histological analysis, and may be of value when the morphological features are ambiguous.18
| TABLE 5. CT characteristics of adenomas and carcinomas | ||||||
|---|---|---|---|---|---|---|
| Characteristic | No. Ca/Adenoma | p Value (OR) | % Sensitivity | % Specificity | % Pos Predictive Value | % Neg Predictive Value |
| Tumor size (cm): | ||||||
| Greater than 5 | 24/3 | <0.0001 (329.0) | 100.0 | 88.4 | 88.9 | 100.0 |
| 5 or Less | 0/23 | |||||
| Necrosis/hemorrhage: | ||||||
| Yes | 10/1 | |||||
| No | 14/25 | 0.001 (17.8) | 41.6 | 96.1 | 90.9 | 64.1 |
| Attenuation :* | ||||||
| Yes | 6/20 | |||||
| No | 18/6 | <0.0005 (10.0) | 75.0 | 76.9 | 75.0 | 76.9 |
| Contrast enhancement: | ||||||
| Heterogeneous | 18/4 | <0.0001 (16.5) | 46.6 | 84.6 | 81.8 | 78.5 |
| Homogeneous | 6/22 | |||||
| Shape: | ||||||
| Irregular | 16/1 | |||||
| Oval/round | 8/25 | <0.0001 (50.0) | 66.6 | 96.1 | 94.1 | 75.7 |
Comparison and/or association of variables was analyzed using Fisher’s exact test.
* Attenuation of homogeneous mass with smooth border 10 Hounsfield units or less.
100
100
Print Survival
4300g
J
50
Children
50
-2009
Adulta
ĐÒN: H - 11
#=0.11; HR= 44
0
0
50
100
150
200
250
0
0
50
100
150
200
2:50
Mint #
Mtarihi
100
100
Parunt lunival
I
I - II
50
III - IV
50
P53 LI-200
# = 0.77; HA= 122
P53 LI 4200
- 2002 HH ĐỘ ẨM
0
0
0
50
100
150
200
250
0
50
100
150
200
250
Mintha
Mantı
100
PCLA LI +248
100
HĐ1UMÆ
Parcant Bury wl
HIL -đt
50
50
J
PCLA LI >248
ĐỒ MỊN -ẢNH
# = 0.02; HA= 317
0
0
0
50
100
150
200
250
0
50
100
150
200
250
Monthi
Mint #
| TABLE 6. Results of multivariate analysis of influence of risk factors on survival | ||
|---|---|---|
| Tumor Stage | PCNA | |
| p Value | 0.04 | 0.02 |
| Hazards ratio | 8.6 | 12.0 |
| 95% CI | 1.0, 70.1 | 1.4, 96.6 |
| Cox regression analysis used to evaluate risk. | ||
CT showed that tumor diameter greater than 5 cm differ- entiates benign from malignant tumors, with sensitivity of 100% and specificity of 88.4%. Other CT features revealed a significant difference in adenomas and carcinomas in univar- iate analysis but not in multivariate analysis, which means that they are dependent and less relevant than tumor size. These findings are in agreement with previous reports.1,19
Tumor weight and immunoexpression of p53 protein, MIB-1 antigen and PCNA exhibited a striking difference between adenomas and carcinomas. However, none of these measures demonstrated sensitivity or specificity of 100% in the differential diagnosis. Taking into account the discrep- ancies reported previously3,6,7,9-11 and our own results, the role of such parameters in the differential diagnosis is un- clear.
In contrast to the published data,3,4 we failed to show that carcinomas in children have less aggressive biological behav- ior than in adults. The most reliable parameters to predict outcome were tumor stage and PCNA immunostaining. Tu- mor stage was reported previously as a risk factor associated with prognosis but data on PCNA are conflicting. 1,4, 9, 11, 12, 20 Immunoexpression of MIB-1 antigen and p53 protein had no prognostic value, and the literature on this issue is contro- versial.3,6-9
CONCLUSIONS
Virilization predominated in children, while Cushing’s syndrome was more prevalent in adults. Survival of children and adults with adrenal carcinoma was similar. Tumor size was the most reliable feature of CT to differentiate benign from malignant disease. The role of p53, MIB-1 and PCNA in the differential diagnosis of adrenal masses was unclear. Tumor stage and PCNA over expression displayed an asso- ciation with survival.
REFERENCES
1. Ng, L. and Libertino, J. M .: Adrenocortical carcinoma: diagnosis, evaluation and treatment. J Urol, 169: 5, 2003
2. Weiss, L. M., Medeiros, L. J. and Vickery, A. L., Jr .: Pathological features of prognostic significance in adrenocortical carci- noma. Am J Surg Pathol, 13: 202, 1989
3. Wieneke, J. A., Thompson, L. D. and Heffess, C. S .: Adrenal cortical neoplasms in the pediatric population: a clinicopatho- logic and immunophenotypic analysis of 83 patients. Am J Surg Pathol, 27: 867, 2003
4. Wajchenberg, B. L., Albergaria Pereira, M. A., Mendonca, B. B.,
Latronico, A. C., Campos-Carneiro, P., Alves, V. A. et al: Ad- renocortical carcinoma: clinical and laboratory observations. Cancer, 88: 711, 2000
5. Fogt, F., Vargas, M. P., Zhang, Z. and Merino, M. J .: Utilization of molecular genetics in the differentiation between adrenal cortical adenomas and carcinomas. Hum Pathol, 29: 518, 1998
6. McNicol, A. M., Nolan, C. E., Struthers, A. J., Farquharson, M. A., Hermans, J. and Haak, H. R .: Expression of p53 in adrenocortical tumours: clinicopathological correlations. J Pathol, 181: 146, 1997
7. Vargas, M. P., Vargas, H. I., Kleiner, D. E. and Merino, M. J .: Adrenocortical neoplasms: role of prognostic markers MIB-1, P53, and RB. Am J Surg Pathol, 21: 556, 1997
8. Goldblum, J. R., Shannon, R., Kaldjian, E. P., Thiny, M., Davenport, R., Thompson, N. et al: Immunhistochemical as- sessment of proliferative activity in adrenocortical neoplasms. Mod Pathol, 6: 663, 1993
9. Bergada, I., Venara, M., Maglio, S., Ciaccio, M., Diez, B., Bergada, C. et al: Functional adrenal cortical tumors in pedi- atric patients: a clinicopathologic and immunohistochemical study of a long term follow-up series. Cancer, 77: 771, 1996
10. Edgren, M., Eriksson, B., Wilander, E., Westlin, J. E., Nilsson, S. and Oberg, K .: Biological characteristics of adrenocortical car- cinoma: a study of p53, IGF, EGF-r, Ki-67 and PCNA in 17 adrenocortical carcinomas. Anticancer Res, 17: 1303, 1997
11. Ghnassia, J. P., Wagner, M., Gasser, B. and Walter, P. R .: Adrenal cortical tumors. Prognostic evaluation of a series of 12 cases using anti-PCNA antibodies. Ann Pathol, 13: 312, 1993
12. Damiani, D., Della Manna, T., Aquino, L. G., Dichtchekenian, V., Avancini, M., Alves, F. et al: Proliferating cell nuclear antigen immunoreaction in adrenal tumors. Tumori, 81: 273, 1995
13. Henley, D. J., van Heerden, J. A., Grant, C. S., Carney, J. A. and Carpenter, P. C .: Adrenal cortical carcinoma-a continuing challenge. Surgery, 94: 926, 1983
14. Taylor, C. R., Shi, S. R., Chaiwun, B., Young, L., Imam, S. A. and Cote, R. J .: Strategies for improving the immunohistochemical staining of various intranuclear prognostic markers in formalin-paraffin sections: androgen receptor, estrogen recep- tor, progesterone receptor, p53 protein, proliferating cell nu- clear antigen and Ki-67 antigen revealed by antigen retrieval techniques. Hum Pathol, 25: 263, 1994
15. Liou, L. S. and Kay, R .: Adrenocortical carcinoma in children. Review and recent innovations. Urol Clin North Am, 27: 403, 2000
16. Venara, M., Sanchez Marull, R., Bergada, I., Gamboni, M. and Chemes, H .: Functional adrenal cortical tumors in childhood: a study of ploidy, p53-protein and nucleolar organizer regions (AgNORs) as prognostic markers. J Pediatr Endocrinol Metab, 11: 597, 1998
17. Barzon, L., Chilosi, M., Fallo, F., Martignoni, G., Montagna, L., Palu, G. et al: Molecular analysis of CDKN1C and TP53 in sporadic adrenal tumors. Eur J Endocrinol, 145: 207, 2001
18. Amberson, J. B., Vaughan, E. D., Jr., Gray, G. F. and Naus, G. J .: Flow cytometric analysis of nuclear DNA from adrenocortical neoplasms. A retrospective study using paraffin-embedded tis- sue. Cancer, 59: 2091, 1987
19. Fassnachat, M., Kenn, W. and Allolio, B .: Adrenal tumors: how to establish malignancy? J Endocrinol Invest, 27: 387, 2004
20. Kasperlik-Zaluska, A. A., Migdalska, B. M., Zgliczynski, S. and Makowska, A. M .: Adrenocortical carcinoma. A clinical study and treatment results of 52 patients. Cancer, 75: 2587, 1995