Humans with inherited MyD88 and IRAK-4 deficiencies are predisposed to hypoxemic COVID-19 pneumonia

Abstract

X-linked recessive deficiency of TLR7, a MyD88- and IRAK-4-dependent endosomal ssRNA sensor, impairs SARS-CoV-2 recognition and type I IFN production in plasmacytoid dendritic cells (pDCs), thereby underlying hypoxemic COVID-19 pneumonia with high penetrance. We report 22 unvaccinated patients with autosomal recessive MyD88 or IRAK-4 deficiency infected with SARS-CoV-2 (mean age: 10.9 yr; 2 mo to 24 yr), originating from 17 kindreds from eight countries on three continents. 16 patients were hospitalized: six with moderate, four with severe, and six with critical pneumonia, one of whom died. The risk of hypoxemic pneumonia increased with age. The risk of invasive mechanical ventilation was also much greater than in age-matched controls from the general population (OR: 74.7, 95% CI: 26.8-207.8, P < 0.001). The patients' susceptibility to SARS-CoV-2 can be attributed to impaired TLR7-dependent type I IFN production by pDCs, which do not sense SARS-CoV-2 correctly. Patients with inherited MyD88 or IRAK-4 deficiency were long thought to be selectively vulnerable to pyogenic bacteria, but also have a high risk of hypoxemic COVID-19 pneumonia.

Graphical abstract

Figure 1.

AR MyD88 and IRAK-4 deficiencies and SARS-CoV-2 infection in 17 kindreds. Pedigrees of the 17 kindreds containing 7 IRAK-4– and 15 MyD88-deficient patients with SARS-CoV-2 infection (P1–P22 are shown). Patients are identified by the number within the symbol for the individual concerned. The mutations are indicated above each pedigree, and the genotype of each individual is identified below the symbol (M, mutation; ND, no data). Kindred N contains five healthy sisters, indicated by the number “5” within the circle symbol.

Figure 2.

Severity of SARS-CoV-2 infection in patients with MyD88 or IRAK-4 deficiencies associated with risk factors. (A) Severity of the infections associated with viral variants. (B) Severity of infection as a function of sex. (C) Severity of infection as a function of age.

Figure 3.

Susceptibility to severe COVID-19 in patients with MyD88 or IRAK-4 deficiency. (A and B) Severity (A) and OR (B) of SARS-CoV-2 infection in MyD88/IRAK-4–deficient patients relative to the age-matched controls from the NC3 cohort infected with the same viral variants. *, P < 0.001. (C and D) Severity (C) and OR (D) of SARS-CoV-2 infection in heterozygous relatives of MyD88- or IRAK-4–deficient patients relative to the Spanish general population between the ages of 20 and 49 yr after the first wave (April 2020).

Figure 4.

Inflammation markers in MyD88/IRAK-4-deficient patients during acute infections. (A) Left: CRP on admission (colored dots), highest level detected (upper bar), and lowest level detected (lower bar) in hospitalized patients with MyD88 or IRAK-4 deficiency. Right: CRP on admission in patients with TLR7 deficiency, hospitalized members of the general population <21 yr of age (Bourgeois et al., 2021), or members of the general population admitted to the ICU (Pierce et al., 2020). Dashed red line: normal range of CRP concentration (<10 mg/dl). (B and C) Ferritin (B) and LDH (C) concentrations in patients with MyD88 or IRAK-4 deficiency (left panels), or TLR7 deficiency, hospitalized members of the general population <21 yr of age (Bourgeois et al., 2021), or members of the general population admitted to the ICU (Pierce et al., 2020; right panels). Dashed red lines: normal range of ferritin (11–336 ng/ml) and LDH (<280 U/liter) concentrations. (D) Left: ANC on admission (colored dots), highest level detected (upper bar), and lowest level detected (lower bar) in hospitalized patients with MyD88 or IRAK-4 deficiency. Right: ANC on admission in patients with TLR7 deficiency, hospitalized members of the general population <21 yr of age (Bourgeois et al., 2021), or members of the general population admitted to the ICU (Pierce et al., 2020). Dashed red line: normal range of ANC (1,500–8,000 cells/mm³).

Figure 5.

Transcriptome analysis of whole-blood samples from SARS-CoV-2–infected individuals. (A) Single-sample gene set enrichment analysis (Hänzelmann et al., 2013) was used to evaluate the IFN-α response, the IFN-γ response, TNF-α signaling through NF-κB, IL-6 JAK-STAT3 signaling, and the inflammatory response. There was one sample for each time-point and patient, and the assay was performed once for each sample. Dot heatmap representing pathway enrichment scores for individual samples. The enrichment score is represented by a colored spot, with red indicating an increase in abundance and blue indicating a decrease in abundance. The intensity of the spots reflects the enrichment score. (B) Time-dependent consistent changes in transcript abundance for type I IFN (red), ISGs with known antiviral functions (purple), other ISGs (yellow), and protein-coding genes (gray) are represented on a scatter plot for IRAK4- and MyD88-deficient patients and a non-infected healthy control. In parentheses, D indicates the number of days after positive RT-PCR for controls and the days after symptom onset for patients. Red arrows indicate the day of treatment with mAbs (casirivimab and imdevimab).