Respiratory viral infections in otherwise healthy humans with inherited IRF7 deficiency

Abstract

Autosomal recessive IRF7 deficiency was previously reported in three patients with single critical influenza or COVID-19 pneumonia episodes. The patients' fibroblasts and plasmacytoid dendritic cells produced no detectable type I and III IFNs, except IFN-β. Having discovered four new patients, we describe the genetic, immunological, and clinical features of seven IRF7-deficient patients from six families and five ancestries. Five were homozygous and two were compound heterozygous for IRF7 variants. Patients typically had one episode of pulmonary viral disease. Age at onset was surprisingly broad, from 6 mo to 50 yr (mean age 29 yr). The respiratory viruses implicated included SARS-CoV-2, influenza virus, respiratory syncytial virus, and adenovirus. Serological analyses indicated previous infections with many common viruses. Cellular analyses revealed strong antiviral immunity and expanded populations of influenza- and SARS-CoV-2-specific memory CD4+ and CD8+ T cells. IRF7-deficient individuals are prone to viral infections of the respiratory tract but are otherwise healthy, potentially due to residual IFN-β and compensatory adaptive immunity.

Graphical abstract

Figure 1.

Patients with biallelic IRF7 variants. (A) Pedigrees of the six kindreds containing seven patients with life-threatening viral infections (P1–P7) bearing rare biallelic IRF7 variants. Solid black symbols indicate patients with critical viral infections. The IRF7 genotype is indicated under each symbol. (B) The plot depicts the population frequency of IRF7 missense and pLoF variants (gnomAD v2.1.1) against CADD score (v1.6, GRCh37). Symbols indicate a total of 463 variants, 4 identified exclusively in patients and 459 present in the gnomAD database. The patient-derived variants reported in this study are highlighted, with pLoF and missense variants colored black and red, respectively. The population-derived homozygous IRF7 missense variants are highlighted in blue. (C) Schematic representation of IRF7. The lower part represents the genomic organization of the IRF7 locus, with black rectangles indicating the exons of the gene according to different transcripts. Below, a track indicates vertebrate nucleotide conservation across the IRF7 locus. The upper part shows the primary protein domain structure of IRF7. The N-terminal portion contains an α-helical DNA binding domain, followed by domains implicated in transactivation, autoinhibition, and regulation, as indicated. The positions of the patient-derived biallelic and population-derived homozygous IRF7 variants are indicated. A blue dotted line indicates the linkage of the IRF7 K179E and Q412R variants. Hmz, homozygous; Htz, heterozygous.

Figure 2.

Expression and activity of novel IRF7 variants. (A) HEK293T cells were transiently transfected with WT or mutant forms of IRF7. IRF7 levels were assessed by Western blotting with antibodies against the N-terminus of IRF7 or an N-terminal FLAG tag. GAPDH was used as a loading control. Representative immunoblots from at least three independent experiments are shown. EV, empty vector. (B) HEK293T cells were transiently transfected with WT or mutant forms of IRF7, together with an IFN-β luciferase reporter and a constitutively expressed reporter. Cells were either left untreated or infected with Sendai virus for 24 h before the assessment of normalized luciferase activity. The significance of differences between variants and the WT (mean ± SEM of n ≥ 3 independent experiments) was determined by two-way ANOVA (*, P < 0.05). Source data are available for this figure: SourceData F2.

Figure 3.

Patients with the newly discovered IRF7 variants do not produce IRF7 protein or IFN-α upon pDC stimulation. (A) Protein levels for IRF7 in PBMCs with and without IFN-β stimulation for 24 h; comparison of patients and healthy controls (HC). Actin staining was used as a loading control. Representative immunoblots (IB) from single independent experiments per patient are shown. (B and C) Frequency of IFN-α– and TNF-producing pDCs (live Lin^–CD11c^–HLA-DR⁺CD303⁺CD123⁺) after 6 h of stimulation with imiquimod for fresh (B) and thawed (C) PBMCs. (D and E) Frequency of IFN-α– and TNF-producing pDCs (live Lin^–CD11c^–HLA-DR⁺CD303⁺CD123⁺) after 6 h of stimulation with CpG ODN for fresh (D) and thawed (E) PBMCs. For P5 and P6, cells were assessed at two independent time points 6 mo apart. (A–E) All data are presented as the frequency of responding cells minus the frequency of the corresponding unstimulated controls. Box plots are bound by the 25th and 75th percentiles. The median is marked, and the whiskers indicate the minimum and maximum. Individual values are plotted. Unpaired t tests were performed to compare healthy controls (HC; n = 8–9) with patients. *, P < 0.05; ****, P < 0.0001. Source data are available for this figure: SourceData F3.

Figure S1.

IFN and inflammatory responses in IRF7-deficient pDCs challenged with SARS-CoV-2 and IAV. RNA-seq analysis of isolated pDCs from P2 and a healthy control (HC) either unstimulated or cultured with SARS-CoV-2 or IAV. (A–C) Genes in the annotated interest groups with expression >2.5-fold higher or lower in P2 vs. HC after viral culture were plotted as a heatmap of expression z-score (A) and expression fold-change in P2 vs. HC (B and C). Data are representative of a single experiment and reanalyzed from Zhang et al. (2020).

Figure S2.

Patients with newly discovered deleterious IRF7 variants produce normal levels of TNF after the TLR3-mediated stimulation of BDCA3⁺ DCs. (A and B) Frequency of TNF-producing BDCA3⁺ DCs (live Lin^–CD123^–HLA-DR⁺CD141⁺) after 6 h of stimulation with poly(I:C), for fresh (A) and thawed (B) PBMCs. All data are presented as the frequency of responding cells minus that of the respective unstimulated controls. Box plots are bound by the 25th and 75th percentiles. The median is marked, and the whiskers indicate the maximum and minimum. Individual values are plotted. Unpaired t tests were performed to compare healthy controls (HC; n = 8–9) and patients. FSC, forward scatter.

Figure 4.

IRF7-deficient patients have enhanced CD4⁺ T cell responses to influenza and coronaviruses. (A and B) Frequency of IFN-γ–producing memory CD4⁺ (live Lin^–CD3⁺CD4^+CD8–CCR7^–CD95⁺; A) or CD8⁺ (live Lin^–CD3^+CD8+CD4^–CCR7^–CD95⁺; B) T cells after stimulation with peptides or PMA + ionomycin for 6 h. Unstimulated controls are shown on the left of the graph. Box plots are bound by the 25th and 75th percentiles. The median is marked, and the whiskers indicate the maximum and minimum. Individual values are plotted. Unpaired t tests were performed to compare healthy controls (HC; n = 9–14) and patients (P5 and P6). *, P < 0.05; **, P < 0.01; ***, P < 0.001; ****, P < 0.0001.

Figure S3.

Large numbers of influenza-specific CD4⁺ T cell blasts in IRF7-deficient patients. (A and B) Whole-blood CD4⁺ (CD3⁺CD4⁺) and CD8⁺ (CD3⁺CD4^–) T cell blast responses to influenza split-virus vaccine (A) or SARS-CoV-2–inactivated virus (B) after 6 d of stimulation; comparison of healthy controls (HC, n = 91 [A] or n = 46 [B]) with patients. For P6, cells were assessed at two independent time points 1 yr apart. Box plots are bound by the 25th and 75th percentiles with the median marked. The whiskers indicate the 5th and 95th percentiles, and outlier values are plotted. Unpaired t tests with Welch’s correction were performed to compare the HC and patient groups. ****, P < 0.0001. (C–E) VirScan assay showing the presence of antibodies against viruses in the serum samples from patients and controls, with the analysis focusing on IAV (D) and coronaviruses (E). Data are representative of a single experiment. Adj Sp_score, adjusted species score; spp_RF, significant species response frequency; AbbPep, abbreviation of peptide (SpeciesName_ProteinName_UniprotID_start_end).