EZH2, Proliferation Rate, and Aggressive Tumor Subtypes in Cutaneous Basal Cell Carcinoma

Basal cell carcinoma (BCC) is the most common human malignancy. Although precise data are not available, non-melanoma skin cancer, of which approximately 80% are BCCs, is estimated to affect 2 to 3.5 million Americans annually.¹ While most sporadic BCCs are treated with excision or other destructive modalities, treatment of locally advanced BCC on the face, hands, feet, and genitals can lead to significant morbidity. Although rare, patients with basal cell nevus (Gorlin) syndrome have numerous BCCs that are difficult to control surgically, and metastatic BCC carries a poor prognosis. For individuals with locally advanced BCC or metastatic BCC, hedgehog signaling pathway inhibitors (vismodegib and sonidegib) constitute the only US Food and Drug Administration-approved medical treatment.^2,3 In phase 2 trials, these inhibitors were associated with a 15% to 58% response rate and 14% to 30% serious adverse event rate.^2,3 Identification of markers for aggressive BCC histology could provide tools for predicting and assessing treatment response or may themselves represent actionable targets. EZH2 is a histone methyltransferase of the polycomb repressive complex 2 that is activated in medulloblastoma, a malignancy that, like BCC, is associated with hedgehog pathway signaling.^4,5 To assess whether EZH2 is a marker for more aggressive BCCs, we compared more vs less histologically aggressive BCCs for expression of EZH2 and Ki67, a marker of proliferation linked to EZH2, using an index that has been used to risk-stratify patients with other cancers.⁶

Methods

We conducted a University of Michigan institutional review board–approved retrospective record review to identify 59 patients (60 specimens) with BCC. All biopsy samples were reviewed by a dermatopathologist (M.C.), and Ki67 and EZH2 immunostains were performed and scored in a semiquantitative manner (Ki67 proliferation index, 0%–100%; EZH2 H score, 0–2) as previously described.⁶ t Tests were used to determine means and mean differences; 95% CIs were determined by Pearson correlation coefficient; P values were based on Fisher exact tests. Statistical analyses were performed using R statistical software, version 3.2.2 (R Foundation).

Results

We compared EZH2 and Ki67 expression in 30 cases (29 patients) with a less aggressive BCC histologic subtype (nodular) and 30 cases (30 patients) with aggressive histologic subtypes (morpheaform, infiltrative, and micronodular) by immunohistochemistry (Figure). The results are summarized in the Table. Aggressive histologic subtypes harbored a greater proportion of Ki67-positive cells compared with the less aggressive group (mean [SD], 35.7% [19.1%] vs 23.3% [10.6%]; P = .003, t test). Similarly, aggressive BCCs demonstrated a greater EZH2 H score than the less aggressive group (mean [SD], 1.30 [0.43] vs 0.77 [0.50]; P < .001, t test). Finally, Ki67 positivity and EZH2 H score positively correlated within each group and overall (aggressive, r = 0.53; P = .003; and less aggressive, r = 0.59; P < .001; Pearson product-moment correlation).

Figure.

Ki67 and EZH2 Expression in BCC

Panels A (hematoxylin-eosin), B, and C show the histopathology of an aggressive BCC; in contrast, panels D (hematoxylin-eosin), E, and F show a less aggressive BCC and demonstrate a lower Ki67 proliferation rate and EZH2 expression (original magnification × 40 for all). BCC indicates basal cell carcinoma.

Table.

Ki67 Positivity and EZH2 H Score by BCC Subtype^a

Characteristic	All Tumors	Aggressive	Less Aggressive	Difference of Group Estimates
Sample size	60	30	30
Ki67 positive, %	…	…	…	P = .003
Mean (95% CI)	29.5 (23.2–33.8)	35.7 (28.5–42.8)	23.3 (19.4–27.3)	12.3 (4.3 to 20.4)
EZH2 H score				P < .001
Mean (95% CI)	1.02 (0.88–1.16)	1.27 (1.10–1.43)	0.77 (0.59–0.96)	0.49 (0.25 to 0.74)
Correlation between Ki67 and EZH2 H score	…	…	…	P = .75
r (95% CI)	0.60 (0.41–0.74)	0.53 (0.21–0.75)	0.59 (0.29–0.78)	−0.06 (−0.43 to 0.31)
P value for test of ρ = 0	4e-7	.0026	.0006

Abbreviation: BCC, basal cell carcinoma.

^aThe table presents means and 95% CIs for Ki67 positivity and EZH2 H score overall, by subtype, and for the difference. P values are from independent samples t tests. Mean Ki67 positivity is significantly (P = .003) higher in the aggressive tumor subtype than in the less agressive subtype; similarly for EZH2 H score (P < .001). Ki67 and EZH2 H score correlate within each subgroup and overall (Pearson product moment correlation and 95% Fisher Z confidence interval are given overall and by subtype). All correlations are statistically significantly different from 0. The correlations by subtype are not statistically significantly different from each other (P = .75, Zou test).

Discussion

Our data reveal that EZH2 expression correlates with the proliferation marker Ki67 and with aggressive BCC subtypes. Our findings are consistent with reports of EZH2 overexpression or gain-of-function mutations in medulloblastoma, as well as in a variety of aggressive solid tumors.^5,6 EZH2 methylates lysine 27 on histone H3 and is associated with tumor suppressor silencing in medulloblastoma.⁵ In the last 2 years, a number of phase 1 and 2 clinical trials have been launched investigating the use of EZH2 and Ki67 as diagnostic biomarkers and EZH2 inhibitors for treatment of various lymphomas and solid organ malignancies (https://clinicaltrials.gov/ct2/results?term=ezh2; https://clinic altrials.gov/ct2/results?term=ki67). We identify EZH2 and Ki67 as biomarkers of BCC with aggressive features. This raises the possibility that EZH2 may be an actionable BCC target, especially because future phase 2 trials will prioritize targeting EZH2 in medulloblastoma, a cancer that, like BCC, often results from dysregulation of hedgehog pathway signaling.^4,5 Future studies are needed to determine whether EZH2 and/or Ki67 expression in BCC may enable risk stratification of response to surgical and anti–hedgehog-based therapies, whether hedgehog and EZH2 pathways represent separate or interdependent targetable pathways in BCC, and whether combining antihedgehog and anti-EZH2 drugs would have therapeutic advantage for aggressive BCC.