COVID-19 chilblain-like lesion: immunohistochemical demonstration of SARS-CoV-2 spike protein in blood vessel endothelium and sweat gland epithelium in a polymerase chain reaction-negative patient.

dear editor, A 35‐year‐old woman with no significant previous medical history presented at the emergency department of Fundación Jiménez Díaz Hospital (Madrid, Spain) on 14 April 2020 with acral purpuric lesions of 3 weeks’ duration. These had started as oedematous, erythematous areas and had appeared coincidentally with fever and coughing, which lasted only 2 days. There was no history of diarrhoea, ageusia, hyposmia or dyspnoea. The patient stated that the skin of her feet acquired a bluish discoloration when she was standing for some time. On examination, arterial pedal pulses were not palpable, and violaceous areas were seen on the dorsa of several toes on both feet (Figure 1a). An ultrasound study ruled out vascular thrombosis. A nasopharyngeal swab reverse‐transcriptase polymerase chain reaction study for SARS‐CoV‐2 RNA and an immunochromatographic assay kit for serum IgG and IgM antibodies performed the same day were negative.

Figure 1

(a) Clinical photograph of chilblain‐like lesions. (b) Perivascular and periadnexal lymphocytic inflammation, with focal evidence of thrombosis in a small vessel (arrow) (haematoxylin and eosin, original magnification × 200). (c, d) Immunohistochemical study of SARS‐CoV‐2 spike protein showing granular staining in endothelial cells (arrows) and sweat gland cells (d, arrowhead). Inset in (d): positive control (cells in lung tissue from an autopsy of a patient with COVID‐19). Original magnification (c) × 400, (d) × 200, (d, inset) × 400.

A skin biopsy obtained while the patients was in the emergency department showed mild perivascular and periadnexal lymphocytic inflammation and focal thrombosis of a small vessel (Figure 1b). There was mild erythrocyte extravasation and occasional evidence of endothelial damage, but no overt vasculitis was identified. Employing a commercially available antibody (SARS‐CoV/SARS‐CoV‐2 spike 1A9; GeneTex, Inc., Irvine, CA, USA), optimized in our laboratory, viral spike protein was distinctly detected by conventional immunohistochemistry as fine‐to‐coarse, bright red granular deposits in the cytoplasm of cutaneous dermal vessels and eccrine cells (both secretory and excretory) (Figure 1c, d). Appropriate negative and positive controls were performed: adult tonsil, a chilblain specimen from early 2019, biopsies from non‐COVID‐19‐related inflammatory dermatoses and substitution of the antibody with saline serum as negative controls, and a post mortem lung specimen from a patient with COVID‐19 as a positive control (Figure 1d, inset).

Perivascular deposits of C5b9, C3 and C1q were seen on direct immunofluorescence study of formalin‐fixed, paraffin‐embedded tissue. Treatment with low‐molecular‐weight heparin and low‐dose aspirin was started; the lesions subsequently healed, and at the time of latest follow‐up (7 June) the patient had recovered completely and pedal pulses were palpable. A second serological study (2 months after the start of the disease), this time by enzyme‐linked immunosorbent assay, was negative for IgG and IgM against SARS‐CoV‐2.

The ongoing COVID‐19 pandemic due to SARS‐CoV‐2, initially regarded as a primarily pulmonary disorder, has evolved into a multisystemic disease reflecting the tropism of the virus and the inflammatory and thrombotic immunological response. Cutaneous manifestations of COVID‐19 have gained the attention of dermatologists worldwide, and six patterns of involvement have been recognized: urticarial rash, morbilliform–maculopapular, papulovesicular, chilblain‐like, livedo reticularis/livedo racemosa like and purpuric–vasculitic. Chilblain‐like acral lesions are being increasingly reported as possibly related to COVID‐19, but a causal relationship is hard to establish, as in many cases (like in ours) no serological or microbiological evidence of SARS‐CoV‐2 infection is detected. This might be due to a swift response from the innate immune system, or to an antibody response different from that of patients without this cutaneous presentation. Nevertheless, systemic or cutaneous endothelial damage might initiate local or systemic thrombotic phenomena, so that involvement of the endothelium is likely to be of pathogenic significance.

Other than in respiratory and alveolar epithelial cells or in alveolar macrophages, the virus has so far been identified by electron microscopy and/or immunohistochemistry in lung alveolar capillaries, endothelial cells of kidney glomeruli, brain blood vessels, colonic mucosa and skin. Our immunohistochemical study represents a straightforward means of linking SARS‐CoV‐2 infection and endothelium.

The presence of the virus in eccrine glands suggests sweat as a source of contagion, but this should be interpreted with caution. Given the similarities of SARS‐CoV and SARS‐CoV‐2, it is worth mentioning that in a 2004 study of four autopsied patients with SARS, Ding et al. found SARS‐CoV nucleoprotein and RNA by immunohistochemistry and in situ hybridization, respectively, in a wide array of tissues, including sweat glands, intestine and kidney. They speculated accordingly the possibility of virus transmission through faeces, urine and sweat. A number of viruses, like hepatitis C virus, are known to replicate in sweat glands and keratinocytes; this could be investigated in sweat obtained by pilocarpine stimulation.