Case Report
Patient 1 (P1) is a 4-year-old female born to parents of Pakistani origin with distant consanguinity (see Supplemental Appendix for complete clinical details). At 6 weeks of age, she developed respiratory failure due to influenza B, from which she recovered. Over the next few months, she developed recurrent respiratory infections requiring supplemental oxygen and treatment with bronchodilators and systemic corticosteroids. At 8 months of age, she was diagnosed with bilateral subdural hygromas, which eventually resolved. At 11 months old, she developed severe respiratory disease due to Enterovirus/Rhinovirus requiring hospitalization and oxygen supplementation, but not mechanical ventilation, from which she recovered. She received all routine immunizations, including live viral vaccines, without adverse events; she never received BCG vaccination. At age 4 years, she developed a rapidly-growing neck mass that invaded the left sternocleidomastoid, mediastinum and retroperitoneum. A mass was also noted in the recto-uterine pouch. Analysis of the mass revealed Burkitt lymphoma, with tumour cells positive for EBER; bacterial, mycobacterial, and fungal cultures yielded no organism. She was treated with R-COPADM-1 protocol (rituximab; vincristine; prednisone; methotrexate; folinic acid; cyclophosphamide; doxorubicin; cytarabine) followed by R-CYM2 (rituximab; methotrexate; folinic acid; cytarabine). During her active chemotherapy, she developed multiple viral infections including Coronavirus OC43/HKU1, rotavirus and norovirus gastroenteritis and enterovirus/rhinovirus URTI. In the context of screening for high-risk exposure, she was found to be PCR-positive for SARS-CoV-2 (B.1.1.529 VOC; Omicron); she manifested only with coryza. Five weeks later, her PCR test reverted to negative. Two weeks later, though, she had new cough and fever, and was SARS-CoV-2 PCR positive again; she did not require oxygen supplementation and resolved without complications. Three months after successful completion of her chemotherapy, she developed a febrile, tonic-clonic seizure. Imaging revealed focal areas of T2/T2 FLAIR hyperintensity and edema, with corresponding restricted diffusion in the cortex of bilateral parieto-occipital and posterior temporal lobes. Cerebrospinal fluid PCR analysis yielded Jamestown Canyon virus; Jamestown Canyon serology was IgM Positive by ELISA and 1:40 by plaque reduction neutralization test. She clinically stabilized but was left with significant neurologic impairment. Given her history and fulminant presentation, further investigations were performed (Immunophenotyping data, Table 1).
Table 1: Laboratory findings
Parameter
|
Reference range
|
June 18 2022***
|
August 20 2022
|
February 22 2023
|
WBC
|
4.7 - 13.5 10ˆ9/L
|
7.9
|
7.6
|
17.0 (H)
|
Hemoglobin
|
105 - 135 g/L
|
114
|
129
|
138 (H)
|
Platelets
|
150 - 450 10ˆ9/L
|
367
|
410
|
333
|
Neutrophils
|
1.5 - 8.5 10ˆ9/L
|
5.8
|
4.9
|
11.1 (H)
|
Lymphocytes
|
1.0 - 5.5 10ˆ9/L
|
1.2
|
1.6
|
3.2
|
Monocytes
|
0.1 - 1.0 10ˆ9/L
|
0.9
|
0.5
|
1.2 (H)
|
Eosinophils
|
0.0 - 0.6 10ˆ9/L
|
0.0
|
0.5
|
1.3 (H)
|
Basophils
|
0.0 - 0.1 10ˆ9/L
|
0.1
|
0.1
|
0.1
|
IgG
|
5.3 - 12.8 g/L
|
3.3 (L)
|
9.5*
|
10.1*
|
IgA
|
0.20 - 1.50 g/L
|
0.33
|
0.34
|
0.20
|
IgM
|
0.50 - 2.00 g/L
|
0.25 (L)
|
0.90
|
0.60
|
IgE
|
<=144.0 ug/L
|
336.0 (H)
|
|
101.0
|
% Total CD3
|
43.0 - 76.0 %
|
82.8 (H)
|
74.2
|
68.0
|
# Total CD3
|
900-4,500 10ˆ6/L
|
1,385
|
1,391
|
1816
|
CD3/CD4%
|
23.0 - 48.0 %
|
20.5 (L)
|
25.1
|
22.1 (L)
|
CD3/CD4
|
700-2,200 10ˆ6/L
|
343 (L)
|
470 (L)
|
589
|
CD3/CD8%
|
14.0 - 33.0 %
|
55.7 (H)
|
40.7 (H)
|
1054
|
CD3/CD8
|
300-1,600 10ˆ6/L
|
931
|
764
|
0.56 (L)
|
CD4:CD8 Ratio
|
0.9 - 2.9
|
0.4 (L)
|
0.6 (L)
|
21.4
|
% Total CD19
|
14.0 - 44.0 %
|
2.9 (L)**
|
10.2 (L)
|
573.0
|
# Total CD19
|
200.0-2,100.0 10ˆ6/L
|
45.8 (L)**
|
189.8 (L)
|
10.8
|
% Total CD(16+56) Pos CD3 Neg
|
4.0 - 23.0 %
|
12.9
|
16.8
|
287
|
# Total CD(16+56) Pos CD3 Neg
|
100-1,000 10ˆ6/L
|
202
|
313
|
68.0
|
* Post IVIG
** Had received a dose of rituximab 3 months prior
*** Post high dose corticosteroids during acute treatment of encephalitis
(H): Higher than the upper limit of the reference range
(L): Lower than the lower limit of the reference range
Homozygous p.R249* variant in TYK2
Whole exome sequencing (WES) in P1 identified homozygous c.745C>T (p.R249*) in TYK2 (NM_003331). This variant, which has not been previously reported, was confirmed by Sanger sequencing [Figure 1A]. It is predicted to cause a premature termination codon in exon 7 residing in the FERM domain (aa 28-451), which, together with the adjacent SH2-like domain, mediate its interaction with receptors [Figure 1B]; in particular, this region of TYK2 is necessary for signaling through the type I IFN receptor system9. This variant was predicted in silico to be deleterious (e.g. SIFT, PolyPhen2; CADD Phred; MutationTaster) and is extremely rare (gnomAD allele frequency of 4.13x10-6; no homozygotes reported).
To confirm the predicted pathogenicity of the variant, we first assessed its molecular impact on TYK2 expression by immunoblot of EBV-transformed B-lymphoblastoid cell lines (LCLs) from P1 vs. health controls [HC], showing a loss of expression in P1 [Figure 1C]. We confirm this result in an autonomous A549 cell model system in which endogenous TYK2 was knocked out by CRISPR-Cas9-based technology, then transiently transfected with either wild-type (WT) TYK2 allele or the c.745C>T variant (TYK2 p.R249*) [Figure 1D]. Both immunoblotting against the TYK2 protein or against the C-terminal-myc-tag showed no TYK2 protein detected from the c.745C>T allele, confirming its loss of expression.
Kinase Activity of TYK2 p.R249*is impaired in P1’s LCL
We first analysed the capacity for auto- and trans-phosphorylation of TYK2 by immunoblotting in P1 and HC LCLs following stimulation with either IFNa/b, IL-10 or IL-12. In contrast to responses seen in HC, the absence of TYK2 expression, as well as its phosphorylation, were observed in P1's LCLs [Figure 2A].
We then assessed the impact of TYK2R249* on its downstream kinase activity within the JAK/STAT canonical signaling pathway, by flow cytometry-based measurement of phosphorylation of the respective STAT proteins that mediate signaling from IFNa/b, IL-12, IL-23 and IL-10 [Figure 2B]. For all tested stimuli, P1 LCLs show reduced STAT activation relative to controls, consistent with its lack of TYK2 protein expression. To confirm these anomalies in a patient-independent cell model, an A549 cell line with TYK2 knocked out was then complemented with either empty vector (EV), TYK2 WT or p.R249* construct through transient transfection. No increase in phosphorylated STAT1 or STAT2 via IFNa/b stimulation or of STAT4 in response to IL12 was observed in A549 cells with either EV or bearing the p.R249* allele.
We further evaluated the p38-mediated non-canonical signaling pathway for type I IFN by assessing phosphorylation of p38 in A549 cell lines, given the high basal phospho-p38 (p-p38) activity in LCL. In the absence of TYK2 or in the presence of TYK2R249*, the phosphorylation of p38 is very low in response to IFN stimulation [Figure 2D].
These results indicate that the identified c.745C>T variant in TYK2 results in complete loss of all its known kinase activity in various human cytokine signaling pathways.
Scaffolding function of TYK2 on the expression of associated cytokine receptors
The FERM and SH2 domains of TYK2 also possess scaffolding function2, 10, 11, and previous patients with TYK2 deficiency have decreased cell surface expression of the cytokine receptor subunits, IFN-αR1, IFN-αR2, IL-10R2, and IL-12Rβ13. The absence of the SH2 domain and part of the FERM domain in the p.R249* mutant may similarly abolish TYK2's interaction with these molecules, thus decreasing their stability and cell surface expression. To evaluate this, we measured the surface expression and intracellular levels of IFN-αR1, IFN-αR2, IL-10R2, and IL-12Rβ1 on LCL by flow cytometry [Figure 3]. As previously shown, cell surface expression of IFN-αR1, IFN-αR2, IL-10R2, and IL-12Rβ1 receptor subunits were significantly reduced in P1’s LCL relative to controls [Figure 3, left column], as were their intracellular levels [Figure 3, right column], confirming that the p.R249* allele causes loss of TYK2's scaffolding function. In contrast, cell surface and intracellular levels of IFN-αR2 were comparable between P1 and controls.
Impact of TYK2-deficient patient P1 on ISGs expression
Type I interferon stimulation leads to the expression of IFN-stimulated genes (ISGs), which may be functionally categorized as "robust" (primarily exerting antiviral effects) and "tunable" (primarily mediating proinflammatory or antiproliferative effects)12, 13. Given the severely diminished activation of TYK2/STAT1/STAT2 following IFNa/b-stimulation of P1's LCLs, we analyzed the impact of TYK2R249* on expression of select ISG by RT-qPCR of "robust" (i.e. OAS2, ISG15, MX1) and "tunable" (i.e. IL-8, CXCL11, IRF1) genes [Figure 4A]. In this LCL model, we observed a clear decrease in the expression of the antiviral and antiproliferative ISGs, except for IRF1. We confirmed this observation in A549 TYK2-/- cells complemented with WT or mutant TYK2 [Figure 4B]. Thus, the c.745C>T mutation in TYK2 impairs ISG responses.
Attempts for pharmacological correction in vitro
Given the dire clinical susceptibility to common viruses in this young child, we investigated whether the mutant TYK2 could be pharmacologically circumvented. G418, an aminoglycoside, has previously been shown to mediate ribosomal in-frame read-through of nonsense mutations14-16, while 4-phenylbutyrate (4PBA) is a chemical chaperone that can rescue the expression of mutant proteins normally expressed at the plasma membrane by improving their intracellular trafficking17-19. Neither agent alone, or in combination, restored TYK2 expression in P1's LCL (Supplementary figure S1). Since IFN-g stimulation of its cognate receptor can activate JAK1 and STAT1, which are shared with the IFN-a/b signaling pathway, we sought to determine if IFN-g may offset the diminished ISGs caused by TYK2R249*. Of the genes tested, IFN-g stimulation compensated some of the tunable ISGs (IRF1, IL-8), but none of the robust, antiviral ISGs (Supplementary figure S2).