During emergency treatment, a 26-month-old sick female infant (height 80.5 cm and weight 9 kg) was sent to our hospital for treatment for a fever accompanied by a thermal spike at 40°C and vomiting. The patient had slight abdominal distension, poor food intake, and drowsiness, and was negative for the Babinski, Brudzinski, and Kernig signs. From seven months old, the patient had been repeatedly hospitalized in our department because of fever and abnormal liver function, with the highest serum-level of glutamic pyruvic transaminase (1629 U/L). Pathological slices showed cloudy swelling in the liver cells and swelling of the renal tubule epithelium.
On hospital admission, blood tests were performed to assess liver and kidney functions. The results showed that the levels of glutamic pyruvic transaminase, aspartate aminotransferase, glutamyl transpeptidase, total bilirubin, direct bilirubin, total bile acid, and adenosine deaminase were higher than the reference values, whereas the levels of alkaline phosphatase and indirect bilirubin were lower than the reference values in the serum (Table 1). During the evaluation of kidney function, the creatinine level (27.10 µmol/L) was lower than the reference range, whereas the uric acid level (461.20 µmol/L) was higher than the reference range. The ammonia level in the blood was 57.50 µmol/L, which was much higher than the reference value. In addition, the alpha-fetoprotein antibody and cytomegalovirus antibody levels in the serum of the patient were higher than the reference values, with levels reaching 16.85 g/ml and 12.922 AU/ml, respectively. Other autoantibodies related to hepatitis in the serum were also tested. The anti-AMA-M2 and anti-M2-3E antibodies tested positive in the patient’s serum (Table S1). Regarding the overall immune function, the levels of immunoglobulin κ-light chain, immunoglobulin λ-light chain, and immunoglobulin G were 5.40 g/L, 2.54 g/L, and 22.75 g/L, respectively, all of which were higher than the reference range.
Table 1
Blood test results for assessing liver and kidney function.
Parameters | Feb 25 Values | Feb 27 Values | Reference |
Glutamic pyruvic transaminase | 7011 U/L | 2875 U/L | 5–40 U/L |
Aspartate aminotransferase | 8103 U/L | 901 U/L | 5–35 U/L |
Glutamyltranspeptidase | 45 U/L | 74 U/L | 7–45 U/L |
Alkaline phosphatase | 274 KU/L | 314 KU/L | 35–135 KU/L |
Total bilirubin | 42.1 µmol/L | 54.4 µmol/L | 1.70–17.1 |
Direct bilirubin | 37.9 µmol/L | 52.5 µmol/L | 0–7.0 µmol/L |
Indirect bilirubin | 4.2 µmol/L | 1.9 µmol/L | 3.0–16.0 µmol/L |
Total bile acid | 374 µmol/L | 426.90 µmol/L | 0–20 µmol/L |
Adenosine deaminase | 45.9 U/L | 41.6 U/L | 5.0–25.0 U/L |
Note that Feb 25 and Feb 27 values were tested on February 25th and 27th, respectively. |
In addition to liver function evaluation, we also assessed the blood coagulation of the patient. The results showed that the D-dimer, prothrombin time, fibrinogen degradation products, and international normalized ratio (INR = 2.77), which is a mean value of risk stratification among patients with liver disease, were higher than the reference ranges (Table 2). As the main clinical presentation of this patient comprised hyperammonemia, hepatic dysfunction, elevated liver biochemical values, and coagulopathy, we identified this patient as having ALF.
Table 2
Blood test results for assessing coagulation-related parameters.
Parameters | Values | Reference |
Fibrinogen | 2.60 g/L | 2.00–4.00 g/L |
Thrombin time | 20.1 s | 14.0–21.0 s |
Prothrombin time | 29.8 s | 10.0–16.0 s |
International normalized ratio | 2.77 | 0.82–1.50 |
Partial thromboplastin time | 31.4 s | 20.0–40.0 s |
D-dimer | 9,672.0 µg / L | 0–1,000 µg / L |
Anti-thrombin III | 394.4% | 70.0–140.0% |
Fibrinogen degradation products | 18.23 µg/mL | 0.00–5.00 µg/mL |
We further investigated whether the ALF in the patient was caused by a gene mutation. Using whole-exome sequencing, we found that NBAS was the only potential gene mutated in the patient. The NBAS mutations were novel and likely pathogenic, containing heterozygote frameshift insertion mutation c.938_939delGC and missense mutation c.1342T > C. Moreover, the missense mutation c.1342T > C lay in the N-terminal regions of NBAS. Sanger sequencing showed that the mother of the proband carried the heterozygous missense mutation c.1342T > C and the father of the proband carried the heterozygous frameshift insertion mutation c.938_939delGC (Fig. 1). According to Sanger sequencing, the primer pair 5ʹ-GAGAAGAGCTTGCGGTGGAT-3ʹ and 5ʹ-CCAGTGTCTTCGGTACCTGC-3ʹ was used for PCR amplification of the DNA segments overlapping the c.938_939delGC mutation site, whereas the primer pair 5ʹ-GAGAAGAGCTTGCGGTGGAT-3ʹ and 5ʹ-CCAGTGTCTTCGGTACCTGC-3ʹ was used for PCR amplification of the DNA segments overlapping the c.1342T > C mutation site. The heterozygous frameshift insertion mutation c.938_939delGC from the paternal allele would cause translation failure of functional NBAS proteins, and the phenotype of the proband would result in the heterozygous missense mutation c.1342T > C from the maternal allele. The heterozygous missense mutation c.1342T > C correlated with the amino acid mutation C448R in the NBAS protein, which is highly conserved in multiple species (Fig. 1). Conversely, the p.C448R mutation is localized at the edge of the β-propeller of the NBAS protein.2
Because T-lymphocyte subsets could be important in the pathogenesis of ALF,9, 15 we further analyzed changes in the peripheral T lymphocytes of the proband using flow cytometry. The results showed that the subtypes of lymphocytes carrying antigens of CD45+, CD3+CD45+ (total T cells), CD3+CD4+CD45+ (T helper lymphocytes), and CD3+CD8+CD45+ (cytotoxic T lymphocytes) were higher in the proband than in children without ALF (Fig. 2A,2B). The ratio of CD3+CD4+ Th cells and CD3+CD8+ cytotoxic T lymphocytes in the proband was significantly lower than that in children without ALF (Fig. 2C).
To uncover the potential mechanism by which the mutation contributes to ALF progression, we investigated whether the mutation affected NBAS mRNA and protein expression. The results showed that ectopic expression of the NBAS C448R construct significantly induced less mRNA (Fig. 3A) and protein expression (Fig. 3B) than that of the wild-type construct. We also found that ectopic expression of NBAS C448R increased reactive oxygen species in 293T cells (Fig. 4A). Forced expression of NBAS C448R induced apoptosis in Jurkat cells (an immortalized T-lymphocyte cell line) (Fig. 4B,4C). These results indicate that NBAS C448R could induce oxidative stress, which contributes to changes in T cell homeostasis (Fig. 2). NBAS regulates the stability of ER-associated transcripts and the ER unfolded protein response,16; thus, we hypothesized that NBAS C448R could induce ER stress. Western blotting revealed that overexpression of NBAS C448R significantly increased the expression of ER stress-related genes, including ATF4, XBP1, CHOP, and BIP (Fig. 4D).17
Together, these preliminarily results showed that the C448R mutation in NBAS, as a novel compound heterozygous mutation that is highly conserved in multiple species, could affect the expression of functional NBAS mRNA and protein, leading to the induction of reactive oxygen species stress and ER stress in T lymphocytes, which contributes to ALF progression.