Accelerated squamous cell carcinoma growth driven by voriconazole‐induced COX‐2 overexpression

Voriconazole (VRCZ) is a widely used antifungal agent, especially for treating invasive aspergillosis and other serious fungal infections.1 However, prolonged use has been associated with photosensitivity and an increased risk of cutaneous squamous cell carcinoma (cSCC).2, 3 Clinically, VRCZ-induced cSCC (VRCZ-cSCC) has been reported to grow more aggressively than typical cSCC.4 However, the mechanisms underlying the accelerated growth of VRCZ-cSCC remain unclear. Cyclooxygenase-2 (COX-2) is a key mediator of inflammation and tumour promotion, contributing to cell proliferation, angiogenesis and immune evasion through prostaglandin E2 (PGE2) signalling.5 COX-2 overexpression has been linked to several malignancies, including colorectal and lung cancers.6, 7 In addition, COX-2 expression is known to be upregulated by ultraviolet (UV) radiation.7, 8 Based on these findings, and considering that VRCZ increases UV sensitivity,2 we hypothesized that VRCZ may exacerbate COX-2–mediated tumour-promoting pathways in the skin. To explore this, we performed immunohistochemistry to detect COX-2 expression on formalin-fixed tissue samples from one case of VRCZ-cSCC, five cases of typical cSCC and two normal skin samples. From each of these eight patient tissue samples, we obtained five high-power fields (400x magnification) for analysis. In total, 40 high-power fields were analysed. The VRCZ-cSCC sample was obtained from the left cheek, while the five typical cSCC samples were derived from other facial areas. The two normal skin samples were collected from abdominal regions. The VRCZ-cSCC sample was derived from a 73-year-old man who had been receiving VRCZ therapy for over 20 years for chronic pulmonary aspergillosis. He presented with a tumour on his left cheek that had rapidly enlarged over a 2-month period (Figure 1a). COX-2-positive areas were quantified using ImageJ (NIH, Bethesda, MD)9 as absolute pixel-based values from these fields, with a greyscale intensity threshold set from 30 to 130 in 8-bit images and directly compared without normalization. Statistical analysis was performed using one-way ANOVA with JMP Pro 17 (SAS Institute Inc., Cary, NC), with p < 0.05 considered significant. We also examined the expression of Ki-67, a well-established marker of cell proliferation, to assess tumour growth dynamics. COX-2 and Ki-67 expression levels were significantly higher in VRCZ-cSCC than in typical cSCC and normal skin. The average COX-2-positive area was 4.18 pixels2/1.0 × 105 in VRCZ-cSCC, compared to 0.95 in typical cSCC and 1.41 in normal skin; Ki-67-positive areas were 2.13, 1.54 and 0.65 pixels2/1.0 × 105, respectively (p < 0.0001 for both) (Figure 1b–d). The analysis using a mixed-effects model demonstrated a trend towards a positive association between COX-2 and Ki-67 expression (β = 0.146, p = 0.069). This relationship is further illustrated in a scatter plot (Figure 1e), which shows the correlation across all cSCC samples while accounting for the nested data structure within individual patients. Voriconazole has been reported to cause chronic photosensitivity. This effect is mediated by its N-oxide metabolites, which absorb UVB and undergo photodegradation, generating reactive oxygen species (ROS) that induce DNA damage and contribute to photocarcinogenesis.2 This UV-dependent mechanism is thought to play a primary role in tumour initiation. In the progression phase, continued voriconazole exposure may sustain high COX-2 expression, thereby promoting tumour growth. In support of this, Ikeya et al. demonstrated that voriconazole upregulates COX-2 expression in keratinocytes through an aryl hydrocarbon receptor-dependent mechanism, even in the absence of UV exposure.9, 10 They demonstrated that this occurs via the classical (genomic) AhR-ARNT-dependent pathway, leading to a robust and sustained transcriptional activation of COX-2. This key finding indicates that while VRCZ metabolites may participate in tumour initiation, voriconazole itself directly promotes tumour development. Collectively, our findings indicate that voriconazole-associated tumour progression in this case may have been mediated, at least in part, by COX-2 overexpression. These findings suggest that VRCZ may influence both the initiation and promotion phases of cSCC development: by enhancing UV sensitivity and DNA damage via its N-oxide metabolites, and by upregulating COX-2 expression to accelerate tumour progression. Although this is a single case study, the observed association between voriconazole exposure and COX-2 overexpression suggests a potential mechanism of tumour progression that warrants further investigation in larger patient cohorts. Targeting COX-2 may represent a potential therapeutic approach in patients with VRCZ-associated cSCC. None. None declared. The patient provided written informed consent for publication of their clinical details and associated images. The data supporting the findings of this study are available from the corresponding author upon reasonable request.