PTEN protein loss by immunostaining: analytic validation and prognostic indicator for a high risk surgical cohort of prostate cancer patients

Abstract

Purpose: Analytically validated assays to interrogate biomarker status in clinical samples are crucial for personalized medicine. PTEN is a tumor suppressor commonly inactivated in prostate cancer that has been mechanistically linked to disease aggressiveness. Though deletion of PTEN, as detected by cumbersome FISH spot counting assays, is associated with poor prognosis, few studies have validated immunohistochemistry (IHC) assays to determine whether loss of PTEN protein is associated with unfavorable disease.

Experimental design: PTEN IHC was validated by employing formalin fixed and paraffin-embedded isogenic human cell lines containing or lacking intact PTEN alleles. PTEN IHC was 100% sensitive and 97.8% specific for detecting genomic alterations in 58 additional cell lines. PTEN protein loss was then assessed on 376 prostate tumor samples, and PTEN FISH or high resolution single nucleotide polymorphism microarray analysis was done on a subset of these cases.

Results: PTEN protein loss, as assessed as a dichotomous IHC variable, was highly reproducible, correlated strongly with adverse pathologic features (e.g., Gleason score and pathologic stage), detected between 75% and 86% of cases with PTEN genomic loss, and was found at times in the absence of apparent genomic loss. In a cohort of 217 high risk surgically treated patients, PTEN protein loss was associated with decreased time to metastasis.

Conclusion: These studies validate a simple method to interrogate PTEN status in clinical specimens and support the utility of this test in future multicenter studies, clinical trials, and ultimately perhaps for routine clinical care.

©2011 AACR.

Figures

Figure 1. PTEN protein expression by immunohistochemistry in isogenic cell line controls with and without somatic PTEN gene loss

(A) Wildtype HCT116 and (D) DLD-1 colon cancer cells, and (G) MCF-10A breast epithelial cells show PTEN protein expression by immunohistochemistry, while the same cell lines with homozygous PTEN deletion by somatic homologous recombination (PTEN KO), show absent PTEN protein (B, E, H). (C, F) HCT116 and DLD-1 cells with PTEN deletion transiently transfected with CMV-PTEN show high PTEN protein expression in a subset of cells. (I) MCF10A cells with hemizygous deletion of PTEN show levels of PTEN immunostaining intermediate between wildtype and cells with homozygous PTEN deletion.

Figure 2. PTEN protein expression by immunohistochemistry in human prostate specimens

(A) Typical cancer case with uniform retained PTEN protein expression in all malignant glands. (B) Typical cancer case with PTEN protein loss in malignant glands, while adjacent benign glands (arrow) retain PTEN protein expression. (C) Some tumors showed only weak PTEN protein positivity (left panel) but were still readily distinguishable from cases with total PTEN protein loss (right panel). (D) Intra-tumoral heterogeneity for PTEN protein expression is evident in this tumor specimen, where malignant glands with PTEN loss and PTEN protein retention are intermingled. Note the presence of cells with and without PTEN expression within a single malignant gland (arrow). (E) High grade prostatic intraepithelial neoplasia (PIN) with PTEN loss in most of the involved gland, while PTEN protein is retained in a minority of luminal cells (arrow) and all basal cells (arrowhead). An adjacent benign gland expresses PTEN. (F) PTEN protein loss in lymph node metastasis of prostatic carcinoma. While the cytoplasm is negative, some glands show a small amount of staining at the apical plasma membrane (arrow), a finding of uncertain significance.

Figure 3. PTEN protein loss by immunohistochemistry is highly correlated with prostate cancer pathologic stage and grade

(A) PTEN protein is more frequent in higher pathologic stage tumors (p=0.003 by Pearson’s chi square test). (B) PTEN protein loss is more common in higher Gleason grade tumors (p=0.0001 by Pearson’s chi square test). (C) PTEN protein loss is least common in benign prostate tissues and PIN and most common in metastatic prostate tumors (p=0.001 by Pearson’s chi square test).

Figure 4. PTEN protein expression by immunohistochemistry is highly correlated with PTEN genomic loss by FISH

In Case A, PTEN protein is highly expressed in malignant glands, corresponding to a normal PTEN/CEP10 ratio by FISH, with retention of PTEN (red) and CEP10 (green) FISH signals in malignant cells. In Case B, PTEN protein is markedly decreased in malignant glands, with a corresponding PTEN/CEP10 ratio of 0.18. Malignant cells show homozygous loss of the PTEN (red) signal in malignant cells, with retention of the CEP10 (green signal).

Figure 5. PTEN protein loss by immunohistochemistry is associated with poor clinical outcomes in a surgical cohort of high risk prostate cancer patients

(A) The Kaplan-Meier curve shows a significant decrease in metastasis-free survival for patients with PTEN protein loss by immunohistochemistry (p= 0.03). (B) The Kaplan-Meier curve for disease-specific survival shows a non-significant decrease in prostate-cancer specific survival in patients with PTEN protein loss (p = 0.06).