Impact of vaccination against severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) on the atopic dermatitis serum proteome

Vaccination against severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) is currently recommended for patients with atopic dermatitis (AD), including those managed with immunomodulatory therapy.^{1 , 2} Dupilumab, an interleukin (IL)-4Rα inhibitor, is most often used in the treatment of moderate-to-severe AD.³ Although available data are limited,^{2 , 4} there is no evidence to suggest that dupilumab may interfere with the immune response to coronavirus disease 2019 (COVID-19) vaccination, and dupilumab has, in fact, been associated with reduced symptom severity and mortality in patients hospitalized with COVID-19.⁵ Conversely, understanding of the effects of COVID-19 vaccines on AD also remains limited, particularly after the 2-week period, with one study reporting transient AD exacerbations in patients taking dupilumab less than or equal to 2 weeks after vaccination.⁶ It is therefore important to characterize the molecular mechanisms underlying the longer-term effects of immunization against SARS-CoV-2 on AD in the context of dupilumab therapy, which we aim to elucidate in this proteomic study of the blood.

Patients were enrolled with institutional review board–approved consent, and serum samples were collected before and at least 14 days after full vaccination (2 doses of Pfizer-BioNTech or Moderna or 1 dose of Janssen/Johnson & Johnson) between June 2020 and February 2022. Inclusion criteria included age above or equal to 12 years and a diagnosis of moderate-to-severe AD (defined as currently or previously on systemic therapy [oral or injected immunomodulatory medications or phototherapy] or candidate for systemic therapy), as previously reported.⁴

Serum samples were centrifuged and stored at −80°C. Aliquots were subsequently processed with the OLINK Proseek multiplex ultrasensitive platform using the following 7 custom panels: inflammation, cardiovascular disease II, cardiovascular disease III, neuroinflammation, cardiometabolic, oncology II, and development (644 measured protein products), similar to prior work.⁷

Analyses were performed using R language (R-project.org) and bioconductor project packages (www.bioconductor.org), as previously described.⁷ Protein expression profiles were fitted to linear models using the R limma framework.⁸ P values were adjusted for multiple hypotheses using the Benjamini-Hochberg procedure to control for false discovery rates (FDRs). Differentially expressed proteins (DEPs) were defined with an absolute fold-change greater than 1.2 and an FDR less than 0.1.

A total of 119 samples were obtained from 68 patients. Treatment groups were defined as follows (n denotes number of samples): limited (topicals/no treatment; prevaccine [Sample 1], n = 8; postvaccine [Sample 2], n = 9) and dupilumab (prevaccine, n = 52; postvaccine, n = 50). Patients on dupilumab had been receiving treatment between 3.3 and 45.7 months at the time of vaccine series completion. No significant differences between treatment groups were found in age, time between completion of vaccine series and sample 2 collection, sex, or vaccination type (Pfizer-BioNTech, Moderna, or Janssen/Johnson & Johnson). In patients for whom clinical severity data were available (investigator's global assessment [IGA] and body surface area [BSA] [%]), average IGA was significantly greater in the limited treatment group (n = 7) than in the dupilumab group (n = 45), whereas BSA was similar between groups. Further demographic information is provided in Table 1 .

Table 1.

Patient Demographics

Parameter	Limited		Dupilumab		P value
	n = 10		n = 58
Age (y)
Mean (SD)	37.5 (14.4)		43.4 (19.3)		.28
Range	19-61
Time between sample collection (mo)
Mean (SD)	6.5 (3.7)		7.2 (2.9)		.58
Range	1-12
Time between vaccine series completion and sample 2 collection (mo)
Mean (SD)	3.8 (1.9)		3 (2.7)		.31
Range	0.7-6		0.1-11.6
Duration of treatment at vaccine series completion (mo)
Mean (SD)	N/A		21.6 (9.5)
Range	N/A		3.3-45.7
Duration of treatment at sample 2 (mo)
Mean (SD)	N/A		26.3 (11.7)
Range	N/A		6-57
IGAa
Mean (SD)	2.9 (0.7)		1.1 (0.8)		1.73E-04
Range	2-4		0-3
BSA (%)a
Mean (SD)	10.9 (11.2)		4.8 (5.2)		.21
Range	1-35		0-28
Sex	n	%	n	%
Female	6	60.0%	25	43.1%	.49
Male	4	40.0%	33	56.9%
Vaccine type	n	%	n	%
Pfizer	8	80.0%	38	65.5%	.13
Moderna	1	10.0%	19	32.8%
Janssen/Johnson & Johnson	1	10.0%	1	1.7%
Race	n	%	n	%
White	7	70.0%	36	62.1%	.59
Asian	0	0.0%	3	5.2%
Black	1	10.0%	12	20.7%
Other	2	20.0%	7	12.1%
Sample size by treatment group	n	%	n	%
Prevaccination	8	47%	52	51.0%	.92
Postvaccination	9	53%	50	49.0%

NOTE. P values for age, IGA, BSA, and time between sample collection determined by unpaired t tests. P values for sex, vaccination type, race, and sample size by treatment groups were calculated with Fisher's exact tests. “Other” race indicates unknown race or that patient declined to specify.

^aIGA and BSA were available for 7 patients in the limited group and 45 patients in the dupilumab-dupilumab group. Subgroup analysis with patients with clinical severity data, adjusting for IGA and BSA, was consistent with unadjusted analysis with the whole cohort.

A total of 2 DEPs (IL-4R, IL-4) were found in intra- and inter-treatment group comparisons. No markers were significantly modulated with vaccination in either treatment group (Fig 1 ), though direction of change was noted to be positive postvaccination for more markers in the limited group than in the dupilumab group. There were also a small number of significant prevaccination intertreatment group comparisons: IL-4 and IL-4R were up-regulated in patients receiving dupilumab relative to limited therapy (Fig 1), and IL-4R was also found to be up-regulated in the dupilumab group relative to the limited group postvaccination. A subgroup analysis with patients with clinical severity scores, adjusted for IGA and BSA, was consistent with these results.

Figure 1.

Expression of a subset of previously published immune markers^7,10 by treatment group and vaccination status for the limited group (n = 10) and dupilumab group (n = 58). Protein expression profiles were fitted to linear regression models, and P values were adjusted for multiple comparisons using the Benjamini-Hochberg procedure. Asterisk denotes FDR < 0.05; double asterisks denote FDR < 0.01. FDR, false discovery rate; post, postvaccination; pre, prevaccination.

To summarize, patients on dupilumab, relative to patients on topical or no therapeutics, expressed significantly more IL-4R and IL-4 prevaccination and more IL-4R postvaccination, and no markers, including other T_H2 markers (eg, IL-4, IL-13, IL-10, CCL7, CCL18), were significantly modulated relative to baseline in either treatment group after full vaccination. The increased expression of IL-4R in the dupilumab-treated group relative to the limited-therapy group may, in fact, be an effect of dupilumab itself as internalization of the IL-4R receptor in immune cells⁹ and a compensatory up-regulation of serum IL-4R with prolonged inhibition. The absence of otherwise significant immune modulation in the serum at least 2 weeks after vaccination suggests that COVID-19 vaccination does not seem to induce proteomic changes relevant to AD pathogenesis, regardless of treatment. Prior work in AD has revealed that pronounced inflammatory protein expression in the skin is accompanied by more modest modulation in the blood,⁷ which may, in part, explain the similarity in limited and dupilumab serum proteomic phenotypes despite differences in baseline severity scores. Additional studies evaluating changes in the skin in the context of COVID-19 vaccination, therefore, may capture more proteomic changes. In conclusion, our report provides molecular context supportive of vaccination against SARS-CoV-2 in patients with AD, without discontinuation of dupilumab. Future studies may expand on our findings with comparisons with more patients, healthy controls, and the inclusion of skin samples.