PAX1 is essential for development and function of the human thymus

Abstract

We investigated the molecular and cellular basis of severe combined immunodeficiency (SCID) in six patients with otofaciocervical syndrome type 2 who failed to attain T cell reconstitution after allogeneic hematopoietic stem cell transplantation, despite successful engraftment in three of them. We identified rare biallelic PAX1 rare variants in all patients. We demonstrated that these mutant PAX1 proteins have an altered conformation and flexibility of the paired box domain and reduced transcriptional activity. We generated patient-derived induced pluripotent stem cells and differentiated them into thymic epithelial progenitor cells and found that they have an altered transcriptional profile, including for genes involved in the development of the thymus and other tissues derived from pharyngeal pouches. These results identify biallelic, loss-of-function PAX1 mutations as the cause of a syndromic form of SCID due to altered thymus development.

Fig. 1.. Pedigrees and PAX1 genetic studies.

(A) Pedigrees and results of Sanger sequencing in patients with PAX1 variants and in healthy controls. For both family A and family B, results of Sanger sequencing in the heterozygous parents are also shown. (B) Schematic representation of the PAX1 protein, and location of the variants identified in affected individuals. OTFCS2: Otofaciocervical syndrome type 2.

Fig. 2.. Molecular and functional analysis of PAX1 mutant proteins.

(A) Western blot showing expression of wild-type and mutant human PAX1 proteins upon transient transfection in 293T cells. (B) Left: Intracellular protein localization upon transfection of HA-tagged WT and mutant PAX1 constructs into 293T cells, followed by staining with TRITC anti-HA. Right: counterstaining with DAPI, demonstrating that the mutant PAX1 protein retain nuclear translocation capacity. Scale bar indicates 10 μm. (C) Results of a luciferase reporter assay demonstrating reduced transcriptional activity of mutant PAX1 proteins, corresponding to the PAX1 variants detected in patients. The promotor region of Nkx3–2 was used to drive luciferase expression. Results of six independent experiments (each run in triplicate) are shown (mean ± SEM). P value was calculated with one-way ANOVA and adjusted by Dunnett’s multiple comparisons test. **; P<0.01.

Fig. 3.. In silico analysis of the PAX1 paired-box domain in WT and mutant proteins.

(A) Molecular modeling of the paired box domain of WT and mutant PAX1 proteins, showing presence of two globular domain separated by a linker. Note that the Asparagine residue at position 155 is adjacent to linker domain, and its deletion results in shortening of the last turn of the 3rd alpha helix in the first globular domain of the paired box domain. (B) Molecular superimposition of WT (in light blue) and mutant PAX1 variants after MD simulation, showing that the both the Val147Leu and the Asn155del variants predominantly affect the conformation of the C-terminus globular domain, whereas both globular domains are affected by the Gly166Val variant. (C) Root-mean-square fluctuation values (RMSF) of WT PAX1, and of the Val147Leu, Asn155del, and Gly166Val variants during MD simulations. RMSF are used here as a measure of the flexibility of different regions of the protein during the MD simulations. The Y-axes indicate the magnitude of the fluctuation while the X-axes indicate the specific location of each amino acid within the paired-box domain.

Fig. 4.. Interface analysis of PAX1 paired-box domain and DNA interaction.

Shown in black are nucleotide residues with which the paired box domain of either WT or PAX1 mutant proteins establish interaction. The amino acids contacting nucleotides of target DNA are indicated on the Y-axis for each PAX1 protein. The red and green colors color indicate loss and gain of DNA binding, respectively.

Fig. 5.. Differentially expressed genes between control and patient TEPs

(A) Heatmap of differentially expressed genes between iPS and TEP stage as determined by RNA-seq. Each heatmap shows the top 3,000 genes which were differentially expressed between iPS and TEP cells, with a significance (q-value <0.01) by the 2-group comparison (t-test). Genes whose expression was found to be upregulated at the TEP stage included epithelial cell markers (EPCAM, KRT8, KRT19) as well as several genes (PSEN1, HES1, ASXL1, HOXA3, HAND2, EPHB3, GATA3) which appeared at the leading edge of GSEA of thymus development in panel B. (B and C) GSEA on thymus development gene set by preranked genes according to -Signed log 10 adjusted p value. The adjusted p value was acquired by DEseq2 analysis using normalized read count of RNA seq data. (D) qRT-PCR analysis of FOXN1 and DLL4 expression at TEP stage of differentiation. Results are from five independent experiments for control and P1, and four independent experiments for control and P4, with triplicates in each case (mean ± SEM). The P value was calculated with two-tailed paired t-test. P<0.05 was considered to be significant. (E) Thymus development genes with evidence of differential expression between patient and control cells (adjusted p value <0.1 and concordant pattern of expression in both RNA-seq experiments). For this comparison, we considered genes that were part of the “Thymus development” gene set in MSigDB v 7.0, and in the top 30 FOXN1-target genes reported by (19). The values displayed are the signed -log10 of the adjusted p value for differential expression.