OUP user menu

p16INK4a Expression Analysis as an Ancillary Tool for Cytologic Diagnosis of Urothelial Carcinoma

Kumiko Nakazawa CT, PhD, Shin-ichi Murata MD, PhD, Tsutomu Yuminamochi CT, Yoshio Ishii CT, Shikine Ohno CT, Tadao Nakazawa MD, PhD, Tetsuo Kondo MD, PhD, Ryohei Katoh MD, PhD
DOI: http://dx.doi.org/10.1309/AJCP61KNVHJVHAFN 776-784 First published online: 1 November 2009

Abstract

We immunochemically studied p16INK4a expression in 116 urine cytologic samples and compared results with 190 histologic samples. The cytologic samples were classified into 4 groups: 1, mild cellular atypia; 2, moderate cellular atypia; 3, severe cellular atypia; and 4, malignancy. Overexpression of p16INK4a was detected in none of 32 cases in group 1, 8 (16%) of 50 cases in group 2, 5 (42%) of 12 cases in group 3, and 11 (50%) of 22 cases in group 4. In addition, by histologic analysis, p16INK4a overexpression was not detected in nonneoplastic urothelium, except for a few cases of reactive atypia, but it was detected in about 50% of urothelial carcinomas. In particular, a high incidence (16/20 [80%]) of p16INK4a overexpression in high-grade carcinomas was noted in cytologic samples. Immunocytologic analysis of p16INK4a expression in cytologic samples is a useful ancillary tool for detection of urothelial carcinoma with infiltrating potential.

Key Words:
  • Urinary bladder
  • Urothelial carcinoma
  • p16INK4a
  • Immunohistochemistry
  • Immunocytochemistry
  • Urine cytology

Urothelial carcinoma of the urinary bladder is one of the most frequent malignancies in industrialized countries.1 Although the incidence of urothelial carcinoma in Japan is lower than in other industrialized countries, it is the 12th and 13th most frequently diagnosed malignancy in men and women, respectively, and more than 6,000 people die of the disease per year.2 Based on the 2004 World Health Organization classification of tumors of the urinary system, urothelial carcinoma has 2 subtypes based on cellular and structural atypia, low- and high-grade types, and 2 subtypes based on invasion, noninfiltrating and infiltrating types.3 Low-grade urothelial carcinoma, which has an incidence of 70% to 80%, is usually noninfiltrating and has an excellent prognosis but shows frequent relapse.3 About 30% of these recurrent tumors may show progression to higher grade with stromal invasion.3,4 High-grade noninfiltrating urothelial carcinoma, including papillary and flat types, namely carcinoma in situ (CIS), often progresses to infiltrating carcinoma. High-grade infiltrating carcinoma usually has a poor prognosis.

Treatment options for urothelial carcinoma vary depending on factors such as subtype, stage, and tumor size. Usually, low-grade noninfiltrating carcinoma is treated by transurethral resection (TUR); high-grade noninfiltrating or early infiltrating carcinoma is treated by TUR followed by bacillus Calmette-Guérin (BCG) therapy and chemotherapy; and total cystectomy with or without chemotherapy is used to treat advanced-stage infiltrating carcinoma.5 From the therapeutic point of view, it is important to detect high-grade malignancy and invasion potential.

Urine cytology is the primary mode of detecting and monitoring urothelial carcinoma. Cytologic specificity is relatively high for the detection of urothelial carcinoma, but sensitivity is low, particularly in low-grade urothelial carcinoma. Diagnosis of urothelial carcinoma by fluorescence in situ hybridization (FISH) has been popular because of its comparable specificity and higher sensitivity than urine cytology.613 The Vysis UroVysion (Quest Diagnostics, Madison, NJ) FISH method can detect deletion of locus 9p21 and polysomy of chromosomes 3, 7, and 17. The locus 9p21 contains the CDK2/p16INK4a gene, which encodes cyclin-dependent kinase (CDK) inhibitor and regulates cell proliferation.

Loss and overexpression of p16INK4a protein can affect carcinogenesis and progression of malignancy from various organs, such as the uterine cervix.1420 Most previous studies that have analyzed p16INK4a in urothelial carcinoma have described loss of expression of p16INK4a in urothelial carcinoma, and only a few have reported overexpression of p16INK4a protein.2131 Previous studies that have analyzed p16INK4a expression in urothelial cancer have used tissue sections, and none has evaluated the validity of p16INK4a expression analysis for urinary cytologic diagnosis. In the present study, we analyzed immunocytochemically p16INK4a expression in atypical cells in urine cytology and compared it with histologic analysis of p16INK4a.

Materials and Methods

Cases

First, we examined 190 histologic lesions from 77 bladder biopsies, 99 TURs, and 14 surgical resections obtained from 149 patients (123 males, 26 females; mean age, 69 years). Multiple lesions from 1 patient and recurrent lesions were included. Then, we analyzed 116 urine cytologic samples from 112 patients (89 males and 23 females; mean age, 69 years). These samples consisted of 107 voided urine, 4 bladder washing, and 5 catheterized urine samples. Samples obtained from patients with recurrent cancer were also included. In 38 cases, p16INK4a expression was immunochemically studied in histologic and cytologic samples. The patients had urogenital symptoms, including hematuria and ureteral calculus. Of the patients, 23 had a history of urothelial carcinoma.

Histologic Diagnosis

Histologic classification was according to the 2004 World Health Organization classification.3 Histologic diagnoses from 190 histologic sections included the following: 20 normal urothelium, 12 reactive atypia after BCG therapy, 11 atypia of unknown significance, 19 dysplasia, 23 urothelial CIS, 37 noninfiltrating low-grade papillary urothelial carcinoma, 20 noninfiltrating high-grade papillary urothelial carcinoma, 29 early infiltrating high-grade urothelial carcinoma (pT1), and 19 advanced infiltrating urothelial carcinoma (pT2-4). The stage of most carcinomas was determined by TUR or from surgical material, while that of some noninfiltrating carcinomas was evaluated by biopsy and clinical findings.

Cytologic Diagnosis

Cytologic samples were classified into 4 groups: 1, mild cellular atypia (probably benign); 2, moderate cellular atypia (indeterminate for malignancy); 3, severe cellular atypia (“suspicious” of malignancy); and 4, malignancy. The cytologic diagnosis of 116 samples comprised 32 in group 1, 50 in group 2, 12 in group 3, and 22 in group 4.

Immunohistochemical and Immunocytochemical Studies

For immunohistochemical analysis, 4-μm-thick tissue sections on silane-coated glass slides were prepared. Urine cytologic samples were stained with Papanicolaou stain after centrifugation (1,500 rpm for 5 minutes) and fixed in YM’s Fluid Fixative (Muto Pure Chemicals, Tokyo, Japan) on silane-coated glass slides. After cytologic diagnosis, the samples were immunostained. p16INK4a immunostaining of the histologic and cytologic samples was performed using a CINtec p16INK4a Research Kit (clone E6H4; DakoCytomation, Glostrup, Denmark) with a DAKO Autostainer Plus (DakoCytomation Colorado, Fort Collins), after a high-temperature antigen unmasking technique using Epitope Retrieval Solution (DakoCytomation, Glostrup) (10 minutes, 95°C).

The intensity of the immunoreactivity detected in the nuclei and/or cytoplasm of atypical cells was scored by the number of positive cells, as follows: negative (0), no discernible staining or fewer than 5% positive cells; 1+, 5% to 20% positive cells; 2+, 21% to 70% positive cells; 3+, more than 71% positive cells with strong intensity, usually in nuclei and the cytoplasm. Histologic and cytologic samples showing 2+ or 3+ immunoreactivity were considered to show overexpression of p16INK4a.

Statistical Analysis

Statistical analysis was performed using the χ2 test for independence. A P value of less than .05 was considered as statistically significant.

Results

Expression of p16INK4a in Histologic Sections

Comparison Between Nonneoplastic and Neoplastic Lesions

Table 1 summarizes expression of p16INK4a in urothelial lesions. Among 43 nonneoplastic lesions, including 20 normal urothelium, 12 reactive atypia, and 11 atypia of unknown significance, 21 (49%) cases were negative for p16INK4a and 19 (44%) showed focal immunoreactivity (1+) Image 1 . Of the 12 samples of reactive atypia after BCG therapy, 3 (25%) showed 2+ positive overexpression of p16INK4a Image 2 . No nonneoplastic cases had 3+ positive overexpression of p16INK4a. On the other hand, 105 (70.9%) of 148 neoplastic lesions showed positive immunoreactivity for p16INK4a. Moreover, overexpression (2+ or 3+) of p16INK4a was detected in 8 (42%) of 19 dysplasia cases, 16 (70%) of 23 CISs, 12 (32%) of 37 noninfiltrating low-grade papillary urothelial carcinomas, 9 (45%) of 20 noninfiltrating high-grade papillary urothelial carcinomas, 17 (59%) of 29 early infiltrating high-grade urothelial carcinomas, and 8 (42%) of 19 advanced infiltrating urothelial carcinomas. Overexpression (3+) of p16INK4a was detected only in definite malignant lesions.

View this table:
Table 1
Image 1

Normal epithelium (A, H&E, ×200; B, p16INK4a immunohistologic stain, ×200). Expression of p16INK4a was detected in a few nuclei (score, 1+).

Comparison Between Low- and High-Grade Neoplastic Lesions

In low-grade urothelial carcinoma, 18 (49%) of 37 cases indicated focal and weak immunoreactivity (1+) of p16INK4a Image 3 . In the 12 (32%) of 37 cases of low-grade urothelial carcinoma that showed p16INK4a overexpression, scoring was 2+ for 7 cases (19%) and 3+ for 5 cases (14%). In high-grade noninfiltrating urothelial carcinoma, including 23 CISs and 20 noninfiltrating high-grade papillary urothelial carcinomas, 25 (58%) of 43 cases moderately to strongly overexpressed p16INK4a in both nuclei and cytoplasm (2+ and 3+). Fewer than half of the cases of high-grade advanced infiltrating urothelial carcinoma showed overexpression of p16INK4a (8/19 [42%]) Image 4 . There was a significant difference (P = .02) in incidence of cases with p16INK4a overexpression between high-grade (50/91 [55%]) and low-grade (12/37 [32%]) urothelial carcinoma.

Expression of p16INK4a in Cytologic Samples

Table 2 shows p16INK4a expression in urine cytologic samples. Atypical cells in group 1 revealed mild immunoreactivity but no overexpression of p16INK4a Image 5 . Overexpression of p16INK4a was detected in 8 (16%) of 50 cases in group 2 with moderate atypia (indeterminate for malignancy) Image 6 , 5 (42%) of 12 cases in group 3 with severe atypia (suspicious of malignancy), and 11 (50%) of 22 cases in group 4 (malignancy). The incidence of p16INK4a overexpression among the 4 groups was significantly different (P = .002; χ2 test). Also, the incidence of p16INK4a overexpression between 2 groups was significantly different, except between groups 3 and 4 (groups 1 and 2, P = .02; 1 and 3, P = .0001; 1 and 4, P = .00004; 2 and 3, P = .05; 2 and 4, P = .01; and 3 and 4, P = .69; χ2 test).

Comparison of p16INK4a Overexpression Between Cytologic and Histologic Diagnoses

Diagnosis in most cases in groups 3 and 4 was confirmed by histologic examination, while only 7 cases in group 2 in which malignancy was suspected clinically were diagnosed histologically. Thirty-eight cases were studied immunochemically for p16INK4a expression using histologic and cytologic samples. Comparison of the cytologic and histologic diagnoses with the immunohistochemical results for the 38 cases is summarized in Table 3 . Of the 38 cases, 29 (76%), namely 16 (80%) of 20 cases with p16INK4a overexpression and 13 (72%) of 18 cases without overexpression by cytology, had consistent results in histologic sections and cytologic samples Image 7 . All 20 cytologic samples with p16INK4a overexpression were confirmed to be neoplastic by histologic diagnosis. Of 7 cytologic cases with negative p16INK4a expression in group 4, 4 (57%) had infiltrating carcinoma.

Image 2

Reactive atypia after bacille Calmette-Guérin therapy (A, H&E, ×200; B, p16INK4a immunohistologic stain, ×200). Expression of p16INK4a was detected with moderate intensity in some nuclei (score, 2+).

Image 3

Low-grade papillary urothelial carcinoma (A, H&E, ×200; B, p16INK4a immunohistologic stain, ×200). Expression of p16INK4a was detected with mild intensity (score, 1+).

Discussion

Immunohistochemical and cytogenetic studies have shown aberrant expression of various cell-cycle–related proteins based on losses and gains of chromosomal regions in bladder cancer.4,2729 Human p16INK4a, the gene of which is located on the 9p21 locus, is one of the CDK inhibitors that regulates the cell cycle and prevents abnormal cell proliferation. Aberration of p16INK4a expression has also been considered as a cause of carcinogenesis and tumor progression in urothelial carcinoma.4,3236 In most previous studies, aberration of p16INK4a expression in urothelial carcinoma has been seen as a loss of p16INK4a expression. Homozygote deletions, methylation of 5′ CpG islands, and mutations of the p16INK4a gene have been reported in 0% to 66%, 15% to 67%, and 0% to 7% of urothelial carcinomas, respectively, as causes of loss of p16INK4a expression.2123,3742 On the other hand, expression of p16INK4a protein has been reported in 11% to 100% of urothelial carcinomas.2131

View this table:
Table 2
Image 4

High-grade papillary urothelial carcinoma (A, H&E, ×200; B, p16INK4a immunohistologic stain, ×200). Overexpression of p16INK4a was diffusely detected with strong intensity in the nuclei and cytoplasm (score, 3+).

Image 5

Voided urine cytologic sample in group 1 (mild cellular atypia; probably benign) (A, Papanicolaou, ×600; B, p16INK4a immunocytologic stain, ×600). Clinical diagnosis was urinary calculus. Expression of p16INK4a was detected in a few nuclei and cytoplasm (score, 1+).

Image 6

Voided urine cytologic sample in group 2 (moderate cellular atypia; indeterminate for malignancy) (A, Papanicolaou, ×600; B, p16INK4a immunocytologic stain, ×600). Histologic diagnosis was urothelial dysplasia. Expression of p16INK4a was detected in many moderately atypical cells (score, 2+).

View this table:
Table 3

In the present study, nonneoplastic lesions such as normal urothelium and atypia of unknown significance showed only weak expression of p16INK4a, whereas 3 (25%) of 12 cases of reactive atypia after BCG therapy and 70 (47.6%) of 147 neoplastic lesions such as dysplasia and urothelial carcinoma revealed overexpression of p16INK4a. In particular, CIS showed a high incidence (70%) of p16INK4a overexpression. It is interesting that advanced infiltrating carcinomas, even high-grade tumors, showed a lower incidence of p16INK4a overexpression than noninfiltrating high-grade tumors.

Image 7

Voided urine cytologic sample in group 4 (malignancy) (A, Papanicolaou, ×600; B, p16INK4a immunocytologic stain, ×600; C, H&E, ×200; D, p16INK4a immunohistologic stain, ×200). Histologic diagnosis was urothelial carcinoma in situ. Overexpression of p16INK4a was detected by cytologic and histologic examination (score, 3+).

The mechanism of p16INK4a overexpression can be explained in 2 ways. One is that increased p16INK4a gene expression caused by polyploidy of chromosome 9 or amplification of the 9p21 locus can augment directly protein expression. Another explanation is the self-regulation that accompanies abnormally high levels of cell proliferation, namely, p16INK4a protein with a very long half-life accumulates in cells and down-regulates proliferation.43 Self-regulation of p16INK4a overexpression is best known in uterine cervical tumors with human papillomavirus infection. The binding of E6 and E7 oncoproteins of the human papillomavirus to p53 and Rb proteins inhibits the suppressor function of cell proliferation and, consequently, induces overexpression of p16INK4a for self-regulation of cell proliferation.1420,44 In urothelial carcinoma, one study reported polysomy of chromosome 9 as the cause of p16INK4a overexpression.21 However, another study using FISH analysis showed a very low incidence of polysomy of chromosome 9 and a high incidence of deletion of the 9p21 locus in infiltrating carcinomas (data not shown). Based on the preceding discussion, we conclude that the main cause of p16INK4a overexpression in reactive atypia after BCG therapy and urothelial carcinoma is self-regulation of abnormally high cell proliferation, and the lower incidence of p16INK4a overexpression in infiltrating cases results from deletion of the 9p21 locus.

No previous study has evaluated the validity of p16INK4a expression analysis for urinary cytologic diagnosis. In the present study, the sensitivity of detection of p16INK4a expression in cytologic samples was similar to that in histologic sections. Weak expression but not overexpression of p16INK4a was found in cells in group 1 (mild cellular atypia), while p16INK4a overexpression was found in 16%, 42%, and 50% of cases in groups 2 (indeterminate for malignancy), 3 (suspicious of malignancy), and 4 (malignancy), respectively. Unfortunately, we could not compare histologic diagnosis and p16INK4a expression in most cases in group 2; however, 5 cases with p16INK4a overexpression in group 2, which were histologically examined, were all malignant. All cytologic samples with p16INK4a overexpression were confirmed to be neoplastic by histologic diagnosis. We consider that overexpression of p16INK4a in indeterminate or suspicious cases is a useful finding for confirming malignancy in cytology. Also, cytologic samples with p16INK4a overexpression had a high incidence (80%) of high-grade urothelial carcinoma. High-grade urothelial carcinoma did not always show a sufficient number of cells with marked atypia in cytologic samples, and it was often difficult to make a differential cytologic diagnosis of benign or low-grade lesions as shown in Table 3; therefore, p16INK4a overexpression in indeterminate or suspicious cases was a useful marker of high-grade malignancy. Moreover, 57% of cytologic cases with negative p16INK4a expression in group 4 probably resulted from deletion of the 9p21 locus and were infiltrating carcinomas. This suggests that malignancy with marked cellular atypia and negative p16INK4a expression in cytologic samples may be an indicator of stromal invasion.

From a prognostic viewpoint, it is important to detect, as early as possible, high-grade malignant lesions in the urinary bladder that have stromal invasion or high potential of invasion. Immunocytologic analysis of p16INK4a expression in cytologic samples is a useful ancillary tool for detecting urothelial carcinoma and infiltrating potential.

Footnotes

  • Supported by grant 20930013 from the Japanese Ministry of Education, Sports, Science and Culture, Tokyo.

References

  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
  20. 20.
  21. 21.
  22. 22.
  23. 23.
  24. 24.
  25. 25.
  26. 26.
  27. 27.
  28. 28.
  29. 29.
  30. 30.
  31. 31.
  32. 32.
  33. 33.
  34. 34.
  35. 35.
  36. 36.
  37. 37.
  38. 38.
  39. 39.
  40. 40.
  41. 41.
  42. 42.
  43. 43.
  44. 44.
View Abstract