Human TTC7A mutations mainly disrupt intestinal and thymus epithelia leading to multiple intestinal atresia-combined (T and B cell) immunodeficiency.

In addition to the PID classification of “combined immunodeficiencies with associated or syndromic features,” the TTC7A gene should also be included as an “immunodeficiency affecting cellular and humoral immunity.” Patients with biallelic missense TTC7A mutations not involving tetratricopeptide repeat domains and developing autoimmune disorders have a relatively better prognosis.

For patients with TTC7A mutations other than bi-missense mutations involving tetratricopeptide repeat domains, an advanced prenatal diagnosis and the lack of intestine donors suggest that termination would be relatively humane while the use of Rho A inhibitors remains hypothetical.

In addition to infection, refractory colitis can be an autoimmune disorder in patients with PIDs, especially in those with combined T and B immunodeficiency (CID), Wiskott–Aldrich syndrome, chronic granulomatous diseases, defective IL-10 signaling, and hyper IgE syndromes. We first excluded well-defined immunodeficiency syndromes as demonstrated in patients with Wiskott–Aldrich syndrome by the size and number of platelets and its characteristics. We then evaluated lymphocyte proliferation, superoxide production, and IL-10 signaling.

The Chang Gung Human Investigation Committee approved all carried methods and all experimental protocols in this study, and the patients’ parents or guardians’ informed consent were obtained. We confirmed that all methods were performed in accordance with the relevant guidelines and standard regulations.

To induce lymphocyte proliferation, peripheral blood mononuclear cells (PBMCs; 10⁵/well) were incubated with different concentrations of phytohemagglutinin and pokeweed (PWM) mitogens for 3 days, or the Candida antigen and Bacillus Calmette–Guérin vaccine for 7 days. To measure reactive oxygen species (defined as H₂O₂ production), phorbol myristate acetate-stimulated (1 µM) polymorphonuclear cells (2 × 10⁶/mL) were incubated with the non-fluorescent compound 2′,7′-dichlorodihydrofluorescein diacetate dissolved in ethanol (Sigma, St. Louis, MO, USA), as previously described (42). To evaluate IL-10 signaling, purified PBMCs (5 × 10⁶/mL) were stimulated overnight with 50 ng/mL E. coli lipopolysaccharide (Sigma-Aldrich, St. Louis, MO, USA) alone or with 20 ng/mL recombinant human IL-10 (R&D Systems, Minneapolis, MN, USA). The supernatants were evaluated using a commercially available TNF-α enzyme-linked immunosorbent assay development kit in duplicate using a Tecan Sunrise enzyme-linked immunosorbent assay micro-plate reader (R&D Systems, Minneapolis, MN, USA).

The basic immunological levels, liver and renal functions of these two sibling patients are shown in Table 1. Quantitative analysis of their immunity revealed the lower borderline of naïve CD4 lymphocytes (CD4+CD45RA+) and memory B cells but higher expressions of CD8 memory cells and activated (CD3+HLA-DR+) T lymphocytes in patient 2 who had more severe IBD-like diarrhea at 3 months of age. Both had hypogammaglobulinemia (including lower IgG, IgA, and IgM by age) after 2 months, except for a higher IgE level possibly related to TPN infusion. Functionally, lymphocyte proliferation was lower; however, superoxide production from stimulated polymorphonuclear cells and IL-10 signaling in lipopolysaccharide-treated PBMCs were within the normal range, thereby excluding chronic granulomatous diseases and defective IL-10 signaling (TNF-α suppression).

The candidate genes of CID prone to refractory diarrhea including IL-2 receptor common gamma chain (IL2RG), ZAP70, CD3 gamma, ORAI-1 and ITK, and those also associated with refractory IBD-like colitis such as FOXP3, STAT3, IL10, IL10RA, IL10RB, NEMO, and STAT1 were all sequenced (6–21, 43, 44). Every two oligonucleotide primers were selected to cover the entire coding region. These primer sequences were based on human genome sequences and are available on the Resource of Asian Primary Immunodeficiency Diseases website (45) or upon request.

Because of the wild types of the candidate genes in the first sibling, WGS was performed for index patient 2 and her parents using an Illumina HiSeq 2000 system (Illumina Inc., San Diego, CA, USA). The data were processed using CLC Genomic Workbench 9.5.2 software (Qiagen). Briefly, paired-end reads were removed from low-quality bases (Q < 30, Phred scale) and aligned to a human reference genome (GRCh37/hg19). The mapping parameters were set to default except for mapping length, which was set to 0.9 and mapping similarity, which was also set to 0.9 and only allowed to read the mapped reference genome once. Variants were filtered based on sequencing coverage (removed if the depth was <10) and by comparisons with a common variant database (dbSNP version 138) (46–48).

To confirm the WGS results by Sanger sequencing, total RNA was extracted from the PBMCs or from established lymphoblastoid cell lines using TRIzol (Life Tech. Carlsbad, CA, USA), followed by reverse transcription polymerase chain reaction as previously described (49). The two pairs of primer sequences of the TTC7A gene were based on human genome sequences NM 001288951.1 and designed to cover the whole coding region by three pairs: (TTC7A-328F: CAC TTC TTG GCC GCA CCT TC, TTC7A-1401R: CTT CCT CGA TGT TGT CCT TG; TTC7A-1254F: AGT GAG GAG TGC TAC TGG AG, TTC7A-2320R: GGA GTC TGG AAG CCT TCT GG; TTC7A-2123F: CCT GGA TGT TGT CAA CAT GG, TTC7A-3150R: GCT GCA GAG GCG TTC ACT AA). Simultaneously, the genomic DNA in exon 2 (forward: CCT GAG CCA TTC ATG TAC TG; backward: ACT AGC CAT ACT CCA AGT GC; around 450 nucleotides) and exon 4 (forward: ACT GCT AAC AACC ACC GTC; backward: TAG CAC ACA GGT GAC CTA TG; around 400 nucleotides) was amplified and confirmed based on NT_022184.15.

To understand the whole disease course, we searched for all reported patients with TTC7A mutations in a Medline search. Survival curves were calculated using Kaplan–Meier analysis to compare their prognoses based on the phenotypes, genetic mutations, autoimmune disorders, and whether or not they received HSCT. The first follow-up day was defined as their birth date because the majority of the patients with TTC7A mutations presented with intestinal obstructions in the prenatal and neonatal stages. The last follow-up day and duration were defined according to those reported in the studies. We defined the multiple intestinal atresia (MIA) phenotype as multiple levels in the gastrointestinal tract from the stomach to anus, and the CID phenotype as T, B, or natural killer lymphopenia, decreased mitogen or antigen lymphoproliferation, or hypogammaglobulinemia. We defined the IBD-like phenotype as frequent diarrhea with more than seven episodes of unformed stools within 24 h for longer than 2 weeks requiring parenteral nutrition in spite of aggressive treatment. All analyses were performed using GraphPad Prism software, and a p value of <0.05 was defined as being statistically significant.

Patient 1 had a clinical diagnosis of CID characterized by lymphopenia, hypogammaglobulinemia, decreased lymphocyte proliferation, line-related infections, and refractory diarrhea. Candidate genes for PIDs were sequenced, but no mutations were identified at that time. WGS would have explained her genetic defect. Patient 2 had similar phenotypes. To identify disease-associated mutations, we performed WGS on genomic DNA from patient 2 and her parents. The results revealed that 10 genes may have contained compound heterozygous mutations that have not been reported in the dbSNP138 database and may affect protein function (Table S1 in Supplementary Material). Among them, TTC7A deficiency has been reported to result in IBD. To verify the mutations in the TTC7A gene, we applied Sanger sequencing on complementary and genomic DNA from the two sisters and their parents. The data showed that both patients had two TTC7A mutations on different alleles (Figure 4). The c.223 G>A mutation in exon 2 causing the Glu75Lys substitution was from the maternal allele, and the c.520-521 del CT mutation in exon 4 causing the fs174 × 27 frameshift, which resulted in the loss of all nine TPR domains was from the paternal allele (based on NM_001288951.1 and NT_022184.16).

A PubMed search for the key words “intestine,” “atresia,” and “TTC7A mutation” excluding studies without genetic identification revealed 47 patients from 23 unrelated and 11 consanguineous families as well as our 2 patients from 1 family (Table S2 in Supplementary Material). A summary of the 5 kinds of presentations revealed MIA alone in 8 patients, IBD alone in 3 patients, the combination of MIA and CID in 16 patients, the combination of IBD and CID in 14 patients, and the complex phenotype of MAI, IBD, and CID in 8 patients. The median survival age of 5 days in the MIA alone phenotype was significantly shorter compared to the combined phenotypes (p = 0.0022 for MIA and CID; and p = 0.001 for IBD and CID) and the complex phenotype (p = 0.0168) (Figure 5). Sepsis, intestinal obstruction, and viral pneumonia were the three leading causes of mortality (Table 2). Notably, the patients with the combination of IBD and CID phenotype all had biallelic missense mutations, and those with other mutations had the isolated MIA, the complex and the majority of the combination of MIA and CID phenotypes.

The 98 mutant alleles in the 49 patients expanded through the coding region and clustered on exon 2 (40 alleles), exon 7 (12 alleles), and exon 20 (10 alleles) including 47 missense, 32 deletion, 9 splicing, 9 nonsense, and 1 insertion mutation. There were two common hotspot mutations. The first was in a large consanguineous family and included 13 patients with c.211 G>A (E71K in exon 2) in 26 alleles who had the combined IBD and CID phenotype. The second came from a French–Canadian family with the founder effect of exon 7 (+3) AAGT deletion in 10 alleles that caused three kinds of phenotypes including the MIA alone phenotype in four homozygous patients, the MIA and CID phenotype, and the complex of MIA, CID and IBD phenotype in one heterozygous patient each.

With regards to the relative frequency of the 34 family cases rather than 49 individuals, there were 31 deletion, 20 missense, 9 nonsense, 7 splicing, and 1 insertion mutation, similarly concentrated on exon 2 (15 alleles), exon 7 (12 alleles), and exon 20 (10 alleles). As well as the founder effect of exon 7 (+3) AAGT deletion in 10 alleles, the 313-6 TATC deletion of exon 2 in 8 alleles replaced E71K in a large consanguineous family to become the most two common hotspot mutations.

Overall, the patients with biallelic missense mutations (p = 0.0168), unaffected TPR domains (p = 0.0311), and developing autoimmune disorders (p = 0.001) had a relatively better prognosis than those without (Figure 6). Although successful HSCT restored immunity to decrease the frequency of infection, improvements in refractory diarrhea were limited and the survival rate was still low in those receiving HSCT, and not significantly better compared to those who did not receive HSCT (p = 0.9084).