Chilblains outbreak during COVID‐19 pandemic: A Type‐I interferonopathy?

To the Editor,

Since early start of coronavirus disease (COVID‐19) pandemic, an outbreak of chilblains‐like lesions was largely recognized worldwide. ¹ , ² , ³ , ⁴ Pernio or chilblains is characterized by erythema and swelling at acral sites, most commonly on the toes and fingers, and typically triggered by cold exposure. ¹ , ² , ³ , ⁴ A form of COVID‐19‐associated chilblains (also called COVID‐toes) has mainly affected the feet of young patients with mild or no other symptoms of COVID‐19. ² The temporal relationship between the appearance of these lesions amidst coronavirus pandemic and their similarities with the skin lesions observed in some monogenic type‐I interferonopathies prompted dermatologists and scientists around the world to hypothesize that these lesions might have been provoked by SARS‐CoV‐2 infection and by an exaggerated up‐regulation of type‐I interferon (IFN) pathway triggered by the virus. ³ Previous studies in skin biopsies of patients with COVID‐toes have demonstrated the presence of SARS‐CoV‐2 by immunohistochemistry and electron microscopy. ⁴ , ⁵ However, a direct link between COVID‐toes and SARS‐CoV‐2 infection could not be made since all tests including nasopharyngeal swab (NPS) polymerase chain reaction (PCR) and serum antibodies against SARS‐CoV‐2 have been negative in most of previously reported patients. ³ , ⁴ , ⁵ To date, the contribution of type‐I IFN response to the viral control and inflammation in COVID‐toes has not yet been fully evaluated.

Herein, we describe a cohort of 28 consecutive children (median age 14 years; range 4–18) with COVID‐toes seen in a Pediatric Dermatology Department in a tertiary care hospital during the COVID‐19 outbreak peak occurring in Barcelona in March–May 2020. The aim of this study was to investigate the role of type‐I IFN signature in peripheral blood of pediatric patients with COVID‐toes and to demonstrate an association between COVID‐toes and SARS‐CoV‐2 infection.

Epidemiological, clinical and laboratory features of enrolled patients are presented in Table 1, Figure 1 and Tables S1. All of them presented with COVID‐toes at different clinical stages at the time of evaluation because of limited hospital access during full lockdown (median from onset 17 days; range 3–53 days). None have prior personal or family background of chilblains. Most lesions resolved spontaneously without residual scars, with the severest cases requiring topical treatment (n: 8; 29.6%) or gabapentin (n: 1; 3.7%) (Table S2). Interestingly, we observed 3 families in which lesions appeared among siblings.

TABLE 1.

Summary of epidemiologic, clinical, laboratory, treatment and outcome of 28 pediatric patients with COVID‐toes

Characteristic	Value
Age
Median (range)	14 years (4 to 18 years)
Gender
Male	11 (39.3%)
Female	17 (60.7%)
Past medical history
Yes	10 (35.7%)
Atopic dermatitis	3 (30%)
No	18 (64.3%)
COVID‐19‐related symptoms
Yes	6 (21.4%)
Preceded	5 (83.3%)
Concomitant	1 (16.7%)
No	22 (78.6%)
Duration of chilblains at evaluation
Median (range)	17 days (3 to 53 days)
Household contact
No suspicious contact	15 (53.6%)
Suspected contact	8 (28.6%)
Confirmed contact	5 (17.9%)
Location of lesions (N: 28)
Hands	11 (39.3%)
Feet	12 (42.9%)
Hands and feet	5 (17.9%)
Characteristics of skin lesions
Symmetrical	23 (82.1%)
Duration (N: 5); median (range)	3.5 weeks (3 to 6 weeks)
Residual scar	0 (0%)
Symptoms
Pruritus	22 (78.6%)
Pain	20 (71.4%)
Movement limitation	15 (53.6%)
Morphological features of the lesions
Erythematous or violaceous plaques	16 (57.1%)
Purpuric macules and papules	11 (39.3%)
Vesicles or blisters	10 (35.7%)
Erosions	1 (3.6%)
Necrosis	1 (3.6%)
Edema	3 (10.7%)
Cyanosis	5 (17.9%)
Treatment (N: 27) a
Yes	8 (29.6%)
No	19 (70.4%)
SARS‐CoV‐2 qRT‐PCR
Positive	0 (0%)
Negative	28 (100%)
SARS‐CoV‐2 Serology
Positive	3 (10.7%)
Negative	25 (89.3%)
SARS‐CoV‐2 Cellular Assays (N: 16/28)
Positive	8 (50%)
IFN‐γ ELISpot (N: 15)	1 (6.7%)
Flow cytometry (N: 16)	8 (50%)
Negative	8 (50%)
Type‐I IFN signature (N: 24/28) b
Positive	7 (29.2%)
Negative	17 (70.8%)
Cellular Assays and IFN signature performed (N: 10/28)
Both positive	4 (40%)
Only cellular assays positive	4 (40%)
Only IFN signature positive	1 (10%)
Both negative	1 (10%)

^a They include topical corticosteroids, anti‐H2, NSAID, oral antibiotics, mupirocin and gabapentin.

^b The result is positive when 28‐IRG score is ≥1.7.

FIGURE 1.

Skin lesions observed in patients. (A) Bilateral purple to red, chilblain‐like lesions on toes. (B) Central darkening could be seen in some lesions. (C) Violaceous discoloration could also occur on the hands. (D) Lesions could also be present on heels and could become necrotic or bullous.

As expected, SARS‐CoV‐2 detection by RT‐PCR in NPS and serological tests were negative in all patients except for 3 patients who were only positive for IgM (Table S4). Subsequently, T‐cell‐mediated immune response against SARS‐CoV‐2 was evaluated by both ELISpot and flow cytometry in 16 enrolled patients, detecting SARS‐CoV‐2‐reactive CD4⁺ and CD8⁺ T cells in 8 patients (50%) (Table S4). These results demonstrated that a high proportion of patients with COVID‐toes have specific T cells against SARS‐CoV‐2, which can be observed even when PCR and serological tests against SARS‐CoV‐2 were negative. These results are subject to cautious interpretation since T‐cell reactivity against SARS‐CoV‐2 antigen peptide pools has been detected in unexposed adults. ⁶ This may reflect a possible cross‐reactivity among other human seasonal coronaviruses. ⁶

IFN signature has been evaluated in both pediatric mild COVID‐19 (pCOVID‐19) and in children with severe COVID‐19 who developed multisystem inflammatory syndrome in children (MIS‐C). Type‐I IFN signatures were strongly elevated in pCOVID‐19, whereas type‐II IFN response was more prominent in MIS‐C. ⁷ We further investigate the role of type‐I IFN response in the pathogenesis of COVID‐toes. Type‐I IFN is a cytokine family that encodes 13 IFN‐α subtypes, a single IFN‐β and other less studied cytokines including IFN‐ω, IFN‐κ and IFN‐ε. All these proteins signal through a heterodimeric type‐I IFN receptor and induce the expression of hundreds of interferon‐stimulated genes (ISG) that confer protection to viral infections. An excessive up‐regulation of type‐I IFN pathway leads to autoimmune and autoinflammatory manifestations, typically seen in various monogenic type‐I interferonopathies. By contrast, a defective type‐I IFN response results in severe immunodeficiency characterized by life‐threating viral infections. Recent studies have shown that type‐I IFN signaling is highly impaired in patients with severe COVID‐19. ⁸ , ⁹ Bastard et al. identified neutralizing autoantibodies against type‐I IFN in up to 13.7% (135/987) of patients with life‐threatening COVID‐19. ¹⁰ The detection of these antibodies in healthy population increases with age, being detectable in nearly 4% over 70 years old. ¹¹ The results from these studies are usually diverse and can be contradictory, since there are many variables that may influence on the results, such as timing of sampling, nature of tissue investigated and the criteria used to classify disease severity. Interestingly, although defective and delayed type‐I IFN response has been associated with severe COVID‐19, local induction of ISG has been detected in bronchoalveolar lavage of several patients with COVID‐19, suggesting that a robust IFN response can be developed locally. ¹² Similarly, staining for MxA and activation of pJAK1 in lesional tissue of three COVID‐toes patients revealed type‐I IFN induction locally. ⁵ More recently, Frumholtz and colleagues measured blood type‐I IFN signature based on the expression of 6 ISG in 50 adult patients with COVID‐toes. They showed type‐I IFN polarization in COVID‐toes patients compared with HC. ¹³ We assessed whole blood type‐I IFN signature based on the expression of 28 ISG (Table S7) in 24 patients, and we observed that 29.2% (7/24) of enrolled patients displayed increased ISG scores (Table S4 and Figure S1). Of these, 57.1% (4/7) patients had a positive cellular assay, suggesting a high possibility of having been in contact with the virus. Recently, Gehlhausen and colleagues examined 21 individuals who had COVID‐toes during the first wave of the pandemic and found no evidence supporting an association with prior SARS‐CoV‐2 infection in 19 of those patients. ¹⁴ Instead, Gehlhausen et al. attribute COVID‐toes to altered behavior during the lockdown such as not wearing socks/shoes at home. This publication highlights the still questioned existence of a link between chilblains and SARS‐CoV‐2. We next compared the ISG‐score of COVID‐toes with 26 healthy controls (HC) and found no significant differences (Figure 2A). However, when we made clusters of COVID‐toes patients according to the duration of chilblains at evaluation, we observed that patients who started chilblains before 7 days at evaluation tended to present with significantly higher ISG‐score than HC (p < .01) (Figure 2B). Some of the differences between our study and the one published by Frumholtz et al. are the age (adults vs. pediatric), timing of samples' collection (more diverse in our cohort because of lockdown restrictions) and the number of patients included in both studies (50 vs 28). Besides, we have seen that globally, the group of monogenic type‐I interferonopathies had ISG more elevated than do COVID‐toes (p < .0001). The most significant difference was observed in IFI27 gene, which was highly and significantly overexpressed in monogenic type‐I interferonopathies (fold change: 7.58; p‐value: 1.47E−16) but less significantly overexpressed in COVID‐toes (fold change: 3.13; p‐value: 5.12E−05) (Tables S5 and S6).

FIGURE 2.

Type‐I IFN signature of COVID‐toes patients (n: 24) compared with healthy controls (HC; n: 26) and monogenic type‐I interferonopathies (disease control, DC; n: 6). Panel A. 28‐IRG score of COVID‐toes patients, HC and DC. Patients with COVID‐toes show no statistically significant differences in IFN Z‐score compared with HC and differ from DC patients. Panel B. 28‐IRG score of COVID‐toes patients clustered according to the duration of chilblains at evaluation, HC and DC. Patients with COVID‐toes evaluated before 7 days of onset of the skin lesions show statistical increase in 28‐IRG score compared with patients who started later than 7 days. Horizontal dotted line represents the cutoff of 28‐IRG score (≥1.7) from which we considered a positive IFN score. 28‐IRG was calculated following previous publication J Interferon Cytokine Res 2018; 38:171–185. Data are presented as scattered dot plots; the horizontal bar represents the mean, and the whiskers represent the standard error of mean. Only statistically significant differences are depicted. ** denotes p < .01. ns, not significant.

Regarding cytokine quantification, we observed that MCP‐1/CCL2 was the most significant cytokine in COVID‐toes compared to HC, followed by CXCL9/MIG and VEGF‐a (Figure S2). High levels of both MCP‐1/CCL2 and VEGF‐a have been associated with the severest cases of COVID‐19 ¹⁵ and have been reported to be significantly high at admission of MIS‐C patients compared to HC, suggesting an involvement of epithelium in the development of MIS‐C during the acute phase. ¹⁶ In contrast, pro‐inflammatory cytokines commonly elevated in COVID‐19 (i.e. IL‐6, TNF‐α) were in normal range in all COVID‐toes as well as levels of CXCL10/IP‐10, which is usually raised in monogenic type‐I interferonopathies and the severest cases of COVID‐19.

Our study shows some limitations that can difficult data interpretation. Regarding IFN signature, we obtained samples at different time points of skin lesions' onset and we employed peripheral blood samples to enable comparisons with monogenic type‐I interferonopathies, instead of using skin biopsy specimens. Regarding cellular assays, we employed frozen PBMC, which may limit cell viability and impair T‐cell responsiveness. Finally, data on immune responses in children exposed to SARS‐CoV‐2 are still limited to enable comparisons with our results. On the other hand, the strengths of our study are that we describe a relatively big and homogeneous cohort of young patients with COVID‐toes, and to our knowledge, this is the first report that examines IFN signature in pediatric patients with COVID‐toes.

Our results suggest that type‐I IFN response is involved in the pathogenesis of COVID‐toes at an early phase of starting skin lesions but may not be as relevant as it is in monogenic type‐I interferonopathies.

FUNDING INFORMATION

This work has been partially funded by Stavros Niarchos Foundation (SNF), Banco Santander and other private donors of Kids Corona. This study was supported by PI19/01567 grant from Instituto de Salud Carlos III (ISCIII) and co‐financed by the European Union (A M‐V), projects PI18/00223, FI19/00208 and PI21/00211, integrated in the Plan Nacional de I+D+I and co‐financed by the ISCIII – Subdirección General de Evaluación y Fomento de la Investigación Sanitaria – and FEDER (LA), by a 2022 Carmen de Torres grant from Fundació Sant Joan de Déu, RTI2018‐096824‐B‐C21 grant from the Spanish Ministry of Science, Innovation and Universities co‐financed by European Regional Development Fund (JIA) and CERCA Programme/Generalitat de Catalunya.

CONFLICT OF INTEREST

The authors have no conflicts of interest relevant to this article to disclose.

PEER REVIEW

The peer review history for this article is available at https://publons.com/publon/10.1111/pai.13860.

Supporting information

AppendixS1

FigureS1

FigureS2