Repeated exposure to fractional CO₂ laser delays squamous cell carcinoma formation and prevents clinical and subclinical photodamage visualized by line‐field confocal optical coherence tomography and histology

Abstract

Objectives

Ablative fractional laser (AFL) is a well‐established modality for treating ultraviolet radiation (UVR)‐induced skin photodamage. We aimed to investigate the potential of AFL to delay squamous cell carcinoma (SCC) formation and prevent photodamage in a preclinical UVR‐induced SCC model.

Materials and Methods

Hairless C3.Cg‐Hr ^hr /TifBomTac mice (n = 50) were exposed to UVR three times weekly throughout the study. UV‐exposed mice were randomized to two groups that received dorsal CO₂ AFL (10 mJ/mb, 10% density) or no treatment. AFL was performed every other week for a total of 16 weeks (nine treatments in total). The primary outcome was time to tumor occurrence. In a subset of mice on Day 150, prevention of clinical photodamage was assessed by examination of skin tightness and dyspigmentation. Concomitantly, assessment of subclinical photoprevention based on normalization of keratinocyte dysplasia, dermo‐fiber morphology (collagen and elastin fibers), and skin thickness, was performed using line‐field confocal optical coherence tomography (LC‐OCT) and histology.

Results

Repeated AFL treatments delayed SCC tumor development compared to UVR control mice by 12, 19, and 30 days for first, second, and third tumors, respectively (p ≤ 0.0017). Compared to UVR controls, AFL prevented photodamage both clinically and subclinically, based on LC‐OCT and histology. In the epidermal layer, AFL imparted photopreventative effects including reduced dyspigmentation and keratinocyte dysplasia (1 vs. 2.5, p = 0.0079) and partial normalization of the epidermal thickness (p < 0.0001). In the dermis, AFL led to twofold greater skin tightness (p = 0.0079), improved dermo‐fiber structure, and dermal thickness (p = 0.0011).

Conclusion

In conclusion, repeated AFL treatments of UVR‐exposed skin significantly delayed SCC tumor formation and prevented clinical and imaging‐assessed subclinical signs of photodamage, indicating a potential for AFL in prevention strategies for SCC and photodamage in high‐risk populations.

INTRODUCTION

Keratinocyte cancer is the most common cancer in fair‐skinned populations, comprising primarily basal cell carcinoma and squamous cell carcinoma (SCC). ¹ The major risk factor is ultraviolet radiation (UVR) from exposure to sunlight. ² A continued accumulation of mutations in keratinocytes results in the development of premalignant actinic keratoses that may subsequently progress to SCC. ³ SCC and its precursor lesions may cause considerable morbidity and substantial increases in incidence rates and even mortality are observed in high‐risk populations that include individuals that are fair‐complexioned, have heavily sun‐damaged skin or are immunocompromised. ⁴ , ⁵ In addition to the increased skin cancer risk, UVR causes photodamage and photoaging such as skin laxity, mottled pigmentation, and altered skin thickness. ⁶ , ⁷ , ⁸ , ⁹ , ¹⁰ Despite current substantial photo‐preventative efforts such as education on the use of sunscreen and avoidance of excessive UVR exposure, SCC incidence rates continue to increase ¹¹ and thus improved prevention strategies to reduce the impact of UVR on skin are needed. Future potential alternatives include systemic pharmacological photo‐prevention ¹² and physical interventions such as ablative fractional laser (AFL), dermabrasion, and chemical peels that are all well‐established for the rejuvenation of photodamaged skin. ¹³ These modalities have also been suggested for the prevention of keratinocyte cancers and its premalignant precursors. ¹⁴ Notably, a single AFL treatment has been shown to induce a sustained reduction of actinic keratoses and a decrease of SCCs in heavily photodamaged aged human skin over a period of 2 years. ¹⁵ In addition, preclinical studies show that AFL can clear keratinocyte cancers and that AFL may potentially suppress the development of new tumors. ¹⁶

In the present study, we aimed to investigate the potential for AFL to delay the formation of SCCs and prevent UVR‐induced photodamage in a well‐established autochthonous murine SCC prevention model.

MATERIALS AND METHODS

Animals

This study utilized immunocompetent female C3.Cg‐Hr ^hr /TifBomTac hairless mice (n = 55; Taconic). In contrast to SKH‐1 mice which lack pigmentation, these mice produce pigment in response to UVR similarly to human skin. ¹⁷ UVR exposure generally induces a high multiple of SCC tumors in this strain, which is a well‐established preclinical model for evaluating photocarcinogenesis and skin tumor prophylaxis. The mice were housed under standard conditions as previously described. The study was approved by the Danish Animal inspectorate (2019‐15‐0201‐01666) and followed FELASA guidelines.

Mice were exposed three times per week (Mondays, Wednesdays, and Fridays) to 3.5 standard erythema doses (SEDs). The mice were irradiated from above for 23 min and 11 s while in a lidded cage with the freedom to move around. The UVR source consisted of one UV6 tube (Philips) and five Bellarium‐S SA‐1‐12 tubes (Wolff). UVR emission was measured using a double monochromator (Bentham DM150; Bentham Instruments Ltd.) and checked with a dual calibration check source module (Optronic laboratories). The light source emitted 5.9% in the UVB range (erythema weighted 89.9%) and had a maximum wavelength of approximately 365 nm. The UVR spectrum is depicted in Supporting Information: Figure 1. A total of 50 mice received the UVR dosing from the age of 17–21 weeks and until the end of the study. Additional five mice did not receive any UVR (i.e., non‐UVR‐exposed) and were used as age‐matched control for healthy skin.

Study design

The 50 UVR‐exposed mice were divided into two groups of 25 mice, one received AFL treatment in addition to UVR and one served as UVR control with no treatment. The first AFL treatment was performed 30 days after UVR initiation and was repeated every 2 weeks for a total of nine treatments. The first UVR dose after each AFL treatment was omitted for both groups to allow undisturbed skin healing.

Outcome measures are summarized in Table 1. The primary outcome measure was time to development of tumors assessed continuously in all mice. Secondary outcome measures focused on photodamage and included clinical evaluation of skin tightness and dyspigmentation as well as subclinical evaluation of epidermal keratinocyte dysplasia, morphology of dermo‐fibers (a conglomerate of structural components including collagen and elastin fibers), and skin thickness in a subset of mice at Day 150. Photoprevention was defined as substantial improvement of clinical and subclinical signs of photodamage in both epidermis and dermis.

Table 1.

Study design and outcome measures

Outcome measure	Technique	Number of mice per group		Measurements per individual mouse	Total number of measurements
		UVR	UVR + AFL
Primary outcome measure
Tumor formation	Clinical assessment	25	25	‐	‐
Secondary outcome measures on photodamage
Local skin reactions	Clinical assessment	5	5	‐	‐
Dyspigmentation	Blinded clinical assessment	5	5	5	50
Skin tightness	Blinded clinical assessment	5	5	5	50
Subclinical changes	LC‐OCT	5	5	1	15a
Epidermal thickness	LC‐OCT	5	5	6	90a
Dermal and hypodermal thickness	Histology	6	7	5	65

Abbreviations: AFL, ablative fractional laser; LC‐OCT, line‐field confocal optical coherence tomography; UVR, ultraviolet radiation.

^a Includeive age‐matched mice that received neither UVR nor AFL.

Laser intervention

Following anesthesia with HypDorm (fentanyl citrate 0.158 mg/ml, fluanisone 5 mg/ml, midazolam 2.5 mg/ml), a dorsal area of 2 × 3 cm was outlined with black tattoo marks on each mouse. Using a 10,600 nm Ultrapulse® fractional CO2‐laser with a DeepFx handpiece (Lumenis, Inc.), one group received AFL at 10 mJ/microbeam, 10% density covering a dorsal area of 2 × 3 cm. Settings were titrated in a prepilot study to achieve local skin responses of fractional pinpoint pattern with erythema but without wound formation. Clinical photographs were taken before and after AFL treatment (Days 0, 1, and 3 post‐AFL) using a Canon 50D camera with 60 mm macro and ring flash (Canon) and used to document local skin reactions (Table 1).

Time to tumor development

Starting at Day 30 after UVR initiation, mice were observed weekly for tumor formation. The first four tumors detected (1 mm diameter) were mapped individually and followed weekly until they reached a diameter of 4 mm. “Time to tumor” represents the interval from UVR initiation until detection of a tumor that subsequently increased to size 4 mm diameter.

Clinical evaluation of skin tightness and dyspigmentation

Skin tightness and dyspigmentation were clinically assessed by three blinded assessors in a subset of mice (n = 5 randomly selected from each group) on arbitrary scales of 1–5. The median values from each mouse were used for statistical analysis.

Imaging‐assessed subclinical changes of photodamage

Line‐field confocal optical coherence tomography (LC‐OCT) is novel noninvasive preclinical and clinical optical technique capable of capturing cross‐sectional, en‐face, and three‐dimensional (3D) images. ¹⁸ LC‐OCT was used to grade subclinical signs of photodamage on skin from healthy control skin, UVR controls and UVR + AFL‐treated mice (n = 5). These signs included epidermal keratinocyte atypia as measure of epidermal dysplasia and dermo‐fiber morphology as measure for skin laxity, as LC‐OCT does not distinguish between collagen and elastin fibers. In addition, images were obtained from five age‐matched mice, that had not been exposed to UVR.

LC‐OCT combines high‐resolution imaging (isotropic resolution ~1 µm) with high‐penetration depth (500 µm) using a broadband laser centered in the near‐infrared spectrum (~800 nm). Before the acquisition of 3D images, immersion oil was employed to match the reflective index of the skin at the corresponding clinically inspected dorsal area. Epidermal keratinocyte nuclei were assessed for atypia using a modified version of a previous established epidermal atypia score ¹⁹ by averaging scores from cross‐section and en‐face images by two nonblinded assessors. The score specifically assessed the shape irregularity of keratinocyte nuclei. The scale ranged from 0 to 4, the lowest score, 0, indicates no irregularity (i.e., keratinocyte nuclei are small at the basal membrane and flattened suprabasally) while a maximum of 4 score is consistent with full irregularity in the entire epidermis. For the evaluation of dermo‐fiber morphology, en‐face images at a depth just below the dermal‐epidermal junction was investigated. Deeplive software (Damae Medical) was used for the descriptive evaluation of photodamage.

Skin thickness

Measurements of epidermal thickness were performed in mice from both UVR control, AFL treated, and UVR free groups (n = 5 per group) using ImageJ on cross‐section LC‐OCT images. Measurements were performed at six separate randomly selected locations in each mouse epidermis.

To determine the thickness of dermis and hypodermis, skin biopsies from both AFL treated and UVR control mice were sectioned hematoxylin and eosin stained and digitalized on a Motic EasyScan Pro 6 (Motic). Five separate measurements of dermal and hypodermal thickness were performed on each biopsy.

Statistics

Descriptive statistics of data are presented as medians and interquartile ranges (IQRs)and time to tumor is presented as Kaplan–Meier plots. Nonparametric analyses were used to assess statistical significance (level of significance: p < 0.05) using IBM SPSS Statistics version 28 (International Business Machines Corp.) and GraphPad Prism 9 (GraphPad Software).

RESULTS

In response to AFL treatment murine skin showed minimal local skin reactions (Figure 1). Immediately after treatment, a grid of laser channels was visible (Day 0). At Day 1, laser channels showed signs of slight, superficial pin‐point crusting that resolved at Day 3. Skin displayed minimal erythema and no edema in response to AFL treatment.

Figure 1.

AFL induces minimal local skin reactions. Clinical images of dorsal mouse skin before and after (Days 0–3) to second AFL treatment and untreated UVR control. Close‐up magnifications are shown for Days 0–3. Minimal crusting was observed on Days 0 and 1. AFL, ablative fractional laser; UVR, ultraviolet radiation.

Tumor development

All UVR‐exposed mice (n = 50) developed tumors, with the first tumor occurring between 127 and 176 days post‐UVR initiation (Figure 2). The median time to the first tumor was 160 days for the AFL group compared to 148 days for UVR controls (p = 0.0017). The delay in tumor development was increased for the second tumor (157 vs. 176 days, p < 0.0001) and for the third tumor observed (160 vs. 190 days, p < 0.0001).

Figure 2.

AFL delays tumor formation. Kaplan–Meyer plots displaying survival time to first, second and third tumor appear (A). AFL increased tumor‐free survival by 12 days for first tumor, 19 days for the second tumor, and 30 days for the third tumor. p values for comparing AFL with UVR control are shown on each plot. Clinical images at Day 156 of an AFL‐treated mouse showing no tumors (B) and a UVR control mouse (C) with a single centrally located tumor (arrow). AFL, ablative fractional laser; UVR, ultraviolet radiation.

Evaluation of photodamage

Photodamage was apparent in UVR‐exposed skin upon visual inspection, showing skin laxity and mottled pigmentation (Figure 2C). In comparison, repeated AFL treatments prevented epidermal and dermal signs of photodamage evaluated from several parameters. Clinically assessed dyspigmentation was reduced in AFL‐treated mice to 2 (1.5–2) from 4 (3.5–5) in UVR control mice (Figure 3A, p = 0.0079), resulting in a less mottled appearance of skin pigmentation (Figure 2B,C). Similarly, skin tightness improved twofold versus control [4 (3.5–4.5) vs. 2 (1.5–2), Figure 3B, p = 0.0079].

Figure 3.

AFL decreases dyspigmentation and increases skin tightness. AFL significantly decreased dyspigmentation (A). Further, AFL‐treated mice displayed an increased skin tightness compared to UVR control mice (B). Dyspigmentation and skin tightness were assessed blinded using arbitrary scales of 1–5. AFL, ablative fractional laser; UVR, ultraviolet radiation.

Imaging‐assessed subclinical changes of photodamage

With regard to subclinical skin changes assessed by LC‐OCT imaging (Figure 4), healthy non‐UVR‐exposed skin displayed uniform epidermis with a thin stratum corneum of one to three cell layers with an epidermal atypia score of 0 (0–0). In comparison, UVR‐exposed mice reached a score of 2.5 (2.5–2.75, p = 0.0079), indicating substantial dysplasia of keratinocytes. Repeated AFL treatments resulted in a significantly lower atypia score of 1 (1–1.75, p = 0.0079).

Figure 4.

LC‐OCT reveals a reduction of photodamage‐related features in AFL‐treated mice compared to UVR‐exposed mice. LC‐OCT images of cross‐sections, as well as en‐face images at epidermal and dermal levels, evaluate keratinocyte irregularity and dermal fiber morphology for healthy non‐UVR‐exposed control mice (A), UVR‐exposed mice (B), and UVR‐exposed mice treated with AFL mice (C). In a cross‐section view for healthy skin, the stratum corneum appeared as a hyperreflective (i.e., bright) line, with an associated thin epidermis consisting of one to three cell layers of stratum granulosum and stratum spinosum cells interrupted by ampulliform rudimentary hair follicles termed utriculi (marked in blue). The stratum granulosum layer contained flattened nuclei, while the stratum spinosum layer displayed keratinocyte nuclei with an oval, hyporeflective (i.e., dark) morphology (marked in green). The dermal‐epidermal junction was clearly visible as a thin hyporeflective layer in cross‐section (shown as a dashed yellow line). Dermis contained well‐defined hyperreflective small, structured dermo‐fibers (marked in red) oriented around utriculi in both in dermal en‐face images. AFL, ablative fractional laser; DEJ, dermal‐epidermal junction; LC‐OCT, line‐field confocal optical coherence tomography; UVR, ultraviolet radiation.

In the healthy dermis, dermo‐fibers appeared small and highly structured, closely oriented around rudimentary hair follicles when imaged by LC‐OCT (Figure 4A). UVR exposure resulted in a severely damaged dermis with substantially reduced structure of dermo‐fibers showing as broad undifferentiated fibers (Figure 4B). In contrast, repeated AFL treatment resulted in dermal photoprevention with mice displaying large, structured dermo‐fibers, with a hair‐like elongated morphology oriented preferably around rudimentary hair follicles (Figure 4C).

Skin thickness

UVR increased epidermal thickness from 22 µm (19–25 µm) to 96 µm (88–106 µm, Figure 5A,C, p < 0.0001). AFL partially prevented the UVR‐induced increase in epidermal thickness, reaching 45 µm (40–60 µm, Figure 5B,C, p < 0.0001). Both dermal and hypodermal thicknesses of UVR‐exposed skin were increased significantly by AFL (Figure 5D,E, p ≤ 0.0011).

Figure 5.

AFL partially normalizes epidermal thickness and increases thickness of dermis and hypodermis. Epidermal thickness was evaluated for UVR control mice (A) and AFL‐treated mice (B) in cross‐section LC‐OCT images. Numbers represent sample thickness evaluations measured in micrometers. Thicknesses of epidermis (C), dermis (D), and hypodermis (E) presented in boxplots of median and interquartile range with min–max whiskers. The dotted line in (C) indicates epidermal thickness in age‐matched non‐UVR‐exposed mice. AFL, ablative fractional laser; LC‐OCT, line‐field confocal optical coherence tomography; UVR, ultraviolet radiation.

DISCUSSION

In the present study, we show that AFL delays the formation of SCCs and elicits both epidermal and dermal photoprevention in UVR‐exposed skin. The effect can be achieved with minimal treatment‐related local skin reactions.

We found that repeated AFL treatments delayed tumor formation significantly up to 30 days for the third tumor. Other modalities, including photodynamic therapy, steroids, and black tattoos, serving de‐facto as an intradermal UV‐blocking agent, have also resulted in delays in tumor formation in this model. ²⁰ , ²¹ , ²² However, they have all entailed local skin reactions that resulted in treatment discontinuation (including wound formation, hyperpigmentation, and substantial loss of skin elasticity) or otherwise would make their application in SCC prevention unfeasible (e.g., black tattoos). In comparison, we found negligible local skin reactions in response to AFL, indicating an overall safe treatment. Keratoacanthomas are an uncommon but notable complication of AFL in sun‐damaged human skin that could affect the observed SCC formation rates. ²³ However, we did not find that any mice developed keratoacanthomas during the study. Further, the pattern in tumor formation delays appears unique for AFL treatment. Whereas tumors 1, 2, and 3 are delayed to a similar extent in comparable studies on prophylactic treatments, we found that for AFL the delay increased progressively from tumors 1 to tumor 3. The mechanism behind the progressively increasing protection is not clear and requires additional investigation.

AFL is well‐established for treating chronically sun‐damaged skin. ²⁴ We found that repeated AFL treatments show prevention of clinical and subclinical signs of photodamage both in epidermis, including dyspigmentation, keratinocyte dysplasia, and epidermal thickness, and in dermis, including skin tightness, dermo‐fiber structure, and dermal thickness. On a clinical level, in AFL‐treated mice the UVR‐induced mottled dyspigmentation was prevented and skin laxity was substantially reduced, consistent with epidermal and dermal photoprevention, respectively. ⁷

LC‐OCT imaging allowed us to evaluate skin changes on a subclinical level. In UVR‐exposed mice, photodamage was clearly identified in the epidermis through a high keratinocyte dysplasia score and in the dermis as substantially reduced structure of dermo‐fibers. These findings are consistent with the clinically observed UVR‐induced development of SCC and skin laxity. ²⁵ , ²⁶ In contrast, AFL reduced keratinocyte dysplasia and reintroduced some dermo‐fiber structure, consistent with delayed SCC formation and increased skin tightness, respectively. Thus, AFL appears to induce partial subclinical photoprevention in both epidermis and dermis even if the skin phenotype of non‐UVR‐treated mice was not fully retained. Overall, our results on imaging‐assessed subclinical photodamage are consistent with a previous study using noninvasive imaging to evaluate the effect of picosecond alexandrite laser on UVR‐induced photodamage in a preclinical setting. ²⁷

Epidermal thickness was substantially increased in the UVR control group compared to skin from non‐UVR‐exposed mice. This is consistent with the literature revealing that UVR‐induced photodamage leads to increased epidermal thickness and reduced dermal and hypodermal thicknesses. ⁶ , ⁹ , ¹⁰ In response to AFL, we observed decreased epidermal thickness and improved thickness of dermis and hypodermis, indicating that AFL diminishes the impact of UVR on skin thickness. Notably, other studies show that AFL increases skin thickness in humans. ¹⁷ , ²⁸

Together, our results suggest that alongside photoprevention, AFL might have a prophylactic potential against SCC formation in UVR‐damaged skin. This corresponds well with a previous preclinical study in which AFL treatment concurrent with UVR exposure reduced the formation of skin tumors, although the model they used does not develop pigment. ¹⁶ , ²² Furthermore, an ongoing clinical study appears to support our findings as Spandau et al. have observed diminished levels of actinic keratoses and keratinocyte cancers up to 2 years post‐AFL treatment in heavily sun‐damaged skin. ¹⁵

The exact mechanism of action behind the delayed SCC formation observed remains to be clarified. We have previously shown that in this SCC model AFL caused partial tumor clearance and increased intratumoral immune infiltration including CD4 + and CD8 + T‐cells, ²⁹ suggesting that improved immune surveillance may play a part. Spandau et al. argued that the AFL‐induced reduction of keratinocyte formation rates seen in humans is caused by a restoration of IGF‐1 levels otherwise diminished in aged human dermis. However, whether this mechanism is relevant for our results is unclear as IGF‐1 expression in murine skin differs significantly from that in humans. ¹⁵ , ³⁰ Alternatively, the repeated ablation and regeneration of dysplastic skin could lead to the destruction of premalignant cells and reduce the carcinogenic potential.

In conclusion, we found that repeated AFL treatments significantly delayed the formation of SCC tumors and achieved clinical and subclinical photoprevention in UVR‐exposed pigmented murine skin with minimal local skin reactions. These results indicate a potential for AFL in prevention strategies for SCC and photodamage in high‐risk populations.

AUTHOR CONTRIBUTIONS

Conceptualization: Uffe H. Olesen, Catharina M. Lerche, and Merete Haedersdal. Methodology, formal analysis, and visualization: Uffe H. Olesen, Catharina M. Lerche, Kevin Jacobsen. Funding acquisition and resources: Merete Haedersdal and Catharina M. Lerche. Investigation: Uffe H. Olesen, Catharina M. Lerche, and Kevin Jacobsen. Writing‐original draft: Uffe H. Olesen. Supervision: Merete Haedersdal. Writing — review and editing: Uffe H. Olesen, Catharina M. Lerche, and Merete Haedersdal.

CONFLICT OF INTEREST

The authors declare no conflict of interest.

Supporting information

Supplementary information.

Olesen UH, Jacobsen K, Lerche CM, Haedersdal M. Repeated exposure to fractional CO₂ laser delays squamous cell carcinoma formation and prevents clinical and subclinical photodamage visualized by line‐field confocal optical coherence tomography and histology. Lasers Surg Med. 2023;55:73–81. 10.1002/lsm.23613