Novel CARMIL2 Mutations in Patients with Variable Clinical Dermatitis, Infections, and Combined Immunodeficiency

Abstract

Combined immunodeficiencies are a heterogeneous collection of primary immune disorders that exhibit defects in T cell development or function, along with impaired B cell activity even in light of normal B cell maturation. CARMIL2 (RLTPR) is a protein involved in cytoskeletal organization and cell migration, which also plays a role in CD28 co-signaling of T cells. Mutations in this protein have recently been reported to cause a novel primary immunodeficiency disorder with variable phenotypic presentations. Here, we describe seven patients from three unrelated, consanguineous multiplex families that presented with dermatitis, esophagitis, and recurrent skin and chest infections with evidence of combined immunodeficiency. Through the use of whole exome sequencing and autozygome-guided analysis, we uncovered two mutations not previously reported (p.R50T and p.L846Sfs) in CARMIL2. Real-time PCR analysis revealed that the biallelic frameshift mutation is under negative selection, likely due to nonsense-mediated RNA decay and leading to loss of detectable protein upon immunoblotting. Protein loss was also observed for the missense mutation, and 3D modeling suggested a disturbance in structural stability due to an increase in the electrostatic energy for the affected amino acid and surrounding residues. Immunophenotyping revealed that patient T_reg counts were significantly depressed, and that CD4⁺ T cells were heavily skewed towards the naïve status. CD3/CD28 signaling impairment was evidenced by reduced proliferative response to stimulation. This work broadens the allelic heterogeneity associated with CARMIL2 and highlights a deleterious missense alteration located outside the leucine-rich repeat of the protein, where all other missense mutations have been reported to date.

Keywords: CARMIL2; CD28; RLTPR; autosomal recessive; dermatitis; immunodeficiency; infection; whole exome sequencing.

Figure 1

Novel CARMIL2 mutations in patients with combined immunodeficiency. (A) Index case from F3 with disseminated warts infection. (B) Pedigrees of the three consanguineous families in this study. Index cases are indicated by arrows, and genotypes (in red) are provided for all family members we had access to. Patients are identified with the abbreviations used in the text. Exome sequencing was performed on patient F1P3, who is boxed in red. (C) Circular representation of chromosomes indicating regions of homozygosity (ROH) in affected (pale blue) and unaffected (pink) individuals. Dark blue denotes the only region that contains ROH across all patients, which coincides with the CARMIL2 locus. (D) Stacked Venn diagram illustrating the total number of variants observed with exome capture, following inclusion of the indicated filters. (E) Sequence chromatograms of the frameshift and nonsense CARMIL2 mutations, with control sequences for reference.

Figure 2

Characterization of CARMIL2 mutations. (A) Protein sequence alignment reveals that the Arg-50 residue is highly conserved across species. (B) Immunoblot reveals absence of detectable CARMIL2 protein in lymphoblastoid cells from patients with the missense and frameshift mutations. For the latter, no bands were found at the calculated molecular weight of the truncated protein (Figure S3 in Supplementary Material). The beta actin lane serves as a loading control. (C) Real-time RT-PCR data for CARMIL2 expression levels in lymphoblastoid cells from the index cases for families F1 and F3 versus combined data from two normal controls. Asterisks indicate significance levels (*p < 0.05, ***p < 0.001; unpaired Student’s t-test). Error bars indicate SEMs. Data were acquired through two independent experiments conducted in triplicate. (D) The 3D homology models established using the native and mutant amino acids at position 50. (E) Schematic representation of the CARMIL2 protein. Previously reported mutations are listed under the schematic, and those from this study are listed above. Missense mutations are indicated in red, and frameshift in black. The depicted domains include the pleckstrin homology domain (PH), the leucine-rich repeat (LRR), the homodimerization domain (HD), the proline-rich regions (PRR), and the capping protein binding region (CBR). Domain boundaries are approximate and are based on homology with known CARMIL1 domains (12) as well as previous CARMIL2 domain delineation (13). (F) 3D modeling of the N-terminal half of native CARMIL2, with the PH domain in blue and the LRR horseshoe region in green. All previously reported missense mutations (orange) have involved leucine residues within the LRR, whereas the one reported in this study (red) is an arginine that does not show interaction with that domain.

Figure 3

Immunophenotyping of CARMIL2-deficient T cells. (A) Representative plot of viable PBMCs from one control and one patient, gated as shown then stained with CD45RO/CD27 to determine T cell subtype population numbers. Plots show CD4 and CD8 T cells that are naïve (T_N, CD45RO⁻CD27⁺), central memory (T_CM, CD45RO⁺CD27⁺), effector memory (T_EM, CD45RO⁺CD27⁻), and effector (T_Eff, CD45RO⁻CD27⁻). Corresponding percentages are indicated in each quadrant. (B) Summary of CD45RO/CD27 staining data from CARMIL2 patients (n = 7), along with healthy controls (n = 6), separated based on CD4/CD8 subtype. Asterisks indicate significance levels (*p < 0.05, **p < 0.01, and ***p < 0.001) and NS denotes not significant. (C) Summary of CFSE proliferation data on PBMCs from controls and patients, following CD3/CD28 stimulation for 72 h. Representative plots showing T_reg cell counts on a control and a patient, using unstimulated (D) and 72 h-stimulated (E) PBMCs. Viable cells were gated on CD4⁺ and were stained with FOXP3/CD25 and FOXP3/CD127 as indicated. Corresponding percentages are provided in each square. (F) Summary of T_reg cell count data for unstimulated as well as CD3/CD28-stimulated groups. Data are separated based on FOXP3⁺/CD127^low or FOXP3⁺/CD25^high staining.