Cutaneotropic Polyomavirus Abundance in the Non-lesional Skin of Keratinocyte Carcinoma Patients: A Cross-sectional Study

To the editor:

Double-stranded DNA virus families Polyomaviridae and Papillomaviridae are dominant among the cutaneous eukaryotic virome detectable in the skin of patients with primary immunodeficiency.¹ Quantitative research on cutaneotropic human polyomaviruses (HPyVs) and their interactions with host immunity are scarce.^{2, 3} T cell immunity plays a critical role in the pathogenesis of keratinocyte carcinoma (KC), including basal cell carcinoma (BCC) and cutaneous squamous cell carcinoma (cSCC).⁴ Herein, we investigated the tripartite relationship of HPyV, KC, and T cell immunity in the skin.

This cross-sectional study was conducted in the Department of Dermatology at Massachusetts General Hospital between February 2021 and June 2022. Discarded non-lesional surgical cone skin tissues generated during Mohs surgery with a pathologically confirmed diagnosis of either BCC or cSCC were acquired. Inclusion was restricted to patients over 18 years old with lesions in the sun-exposed face or scalp areas. Quantitative Real-time Polymerase Chain Reaction (qRT-PCR) was performed on DNA extracted from homogenized tissues to detect the three most prevalent cutaneotropic HPyVs: Merkel cell polyomavirus (MCPyV), HPyV6, and HPyV7. The viral abundance was determined via standard curves and normalized to human GAPDH. Immunofluorescence staining for CD3, CD8α, and CD103 was performed on formalin-fixed paraffin-embedded (FFPE) sections from a subset of patients based on tissue availability. Data analysis was conducted using STATA BE 17, R version 4.2.1, and GraphPad Prism 9. Detailed methods are reported in Supplementary Methods and Supplementary Tables I-III, available via Mendeley at https://data.mendeley.com/datasets/6b8ppw8tmw/1.

A total of 100 non-lesional skin samples were collected (Table I). MCPyV demonstrated a more prominent presence in the non-lesional skin of cSCC patients (31/33, 94%) compared with BCC patients (52/67, 78%, P = 0.049), and the association remained statistically significant after adjustment for age, gender, and immunosuppressive medication status in the multivariate logistic regression model (OR, 6.30; 95% CI, 1.25-31.82; P = 0.026; Table 2). Viral abundance was analyzed in 25 cSCC cases and age-, sex-, and facial anatomical site-matched BCC cases without previous cSCC history (Supplementary Fig 1, available via Mendeley at https://data.mendeley.com/datasets/6b8ppw8tmw/1). Compared with BCC patients, cSCC patients harbored significantly higher MCPyV viral loads in their non-lesional skin (P = 0.038). The association between viral positivity and abundance and KC types was not observed for HPyV6 or HPyV7.

Table I.

Characteristics of the non-lesional skin samples from BCC and cSCC patients.

Non-lesional skin source	No. (%)		P valuea
	BCC patients	cSCC patients
No. of samples	67	33	NA
Age, median (IQR), y	74 (61-82)	78 (70-83)	0.06
Sex			0.10
Male	31 (46)	21 (64)
Female	36 (54)	12 (36)
Lesion location			0.23
Cheek and Chin	24 (36)	19 (58)
Forehead and Temple	30 (45)	11 (33)
Nose	9 (13)	2 (6)
Scalp	4 (6)	1 (3)
Immunosuppressive medication			0.23
Yes	5 (7)	5 (15)

Abbreviations: BCC, basal cell carcinoma; cSCC, cutaneous squamous cell carcinoma; NA, not applicable.

^aAge was compared by a two-tailed Mann-Whitney U test. Categorical variables were compared using either two-tailed Pearson’s χ² tests or two-tailed Fisher’s exact tests for features observed in less than five patients.

Table II.

Association between cutaneotropic HPyV skin positivity and KC subtypes before and after adjusting for age, gender, and current immunosuppressive therapy status.

HPyV Statusa		Non-lesional skin source, n (%)		P value, observedb	ORc	95% CIc	P value, adjusted model
		BCC patients	cSCC patients
Any HPyV	Negative	5 (7)	2 (6)	0.999	1.00	Reference	0.461
	Positive	62 (93)	31 (94)		1.96	0.33 - 11.82
MCPyV	Negative	15 (22)	2 (6)	0.049	1.00	Reference	0.026
	Positive	52 (78)	31 (94)		6.30	1.25 – 31.82
HPyV6	Negative	16 (24)	8 (24)	0.968	1.00	Reference	0.878
	Positive	51 (76)	25 (76)		1.08	0.38 - 3.07
HPyV7	Negative	30 (45)	15 (45)	0.949	1.00	Reference	0.904
	Positive	37 (55)	18 (55)		1.06	0.43 - 2.58

^aThe specimen was considered positive for the target HPyV type if at least 1 of the 3 technical replicates demonstrated a typical amplification curve.

^bTwo-tailed Pearson’s χ² test. Two-tailed Fisher’s exact tests were used for the features observed in less than five patients.

^cORs and 95% CIs estimated after adjusting for age, gender, and immunosuppressive therapy status.

FFPE sections of non-lesional skin from a subgroup of patients without a history of immunosuppression, including 13 BCC and 14 cSCC cases, were examined (Supplementary Table IV and Supplementary Fig 2, available via Mendeley at https://data.mendeley.com/datasets/6b8ppw8tmw/1). Compared with BCC patients, significantly fewer skin-infiltrating CD3⁺ T (P < 0.001), CD8⁺CD3⁺ T (P < 0.001), and CD103⁺CD8⁺CD3⁺ tissue-resident memory T cells (P = 0.003) were identified in cSCC patients.

The disparity of MCPyV abundance between the normal skin of our cSCC and BCC patients could be potentially attributed to the compromised cutaneous immunity that hallmarks the skin microenvironment at risk for cSCC,⁵ consistent with the significantly fewer skin-resident CD8⁺ T cells observed in cSCC patients. Future longitudinal research is warranted to further determine the role of cutaneotropic HPyV activity in cSCC pathogenesis and as the predictor of skin immune state and cancer risk.

Abbreviations:

HPyV

human polyomavirus

MCPyV

Merkel cell polyomavirus

KC

keratinocyte carcinoma

cSCC

cutaneous squamous cell carcinoma

BCC

basal cell carcinoma

FFPE

formalin-fixed paraffin-embedded

qRT-PCR

quantitative real-time polymerase chain reaction

IQR

interquartile range