摘要
Hypobetalipoproteinemia (HBL) is defined by a low level (5th centile) of total apolipoprotein B (apoB) and/or low-density lipoprotein cholesterol (LDL-C) in plasma (1–3). The primary causes of HBL are genetic defects of very low-density lipoprotein (VLDL) or chylomicron production and its secretion pathways. These disorders include familial HBL (FHBL; OMIM 107730), abetalipoproteinemia (ABL; OMIM 200100), chylomicron retention disease (CRD; OMIM 246700) and CRD-related disorders, which are Anderson disease (OMIM 607689) and Marinesco-Sjögren syndrome (MSS; OMIM 246700). Secondary causes include chronic illness, strict vegetarians, intestinal fat malabsorption due to chronic liver disease, biliary obstruction, hyperthyroidism, and pancreatic insufficiency (3). Familial HBL is a codominant disorder caused by mutation in the APOB gene (3p21), which encodes apoB, the main constituent protein of VLDL, LDL-C, and chylomicron (3,4). Heterozygous individuals are mostly asymptomatic or mildly symptomatic with fatty liver and liver dysfunction; the frequency of the FHBL heterozygote is 1:500 to 1:1000 in the white population (3,4). The FHBL homozygote is rare, but affected individuals are severely symptomatic (3,4). Homozygous FHBL, ABL, and CRD share common clinical and biochemical characteristics: fat malabsorption; steatorrhea; failure to thrive; complications of fat-soluble vitamin deficiency such as acanthocytosis, retinitis pigmentosa, and spinocerebellar ataxia; and low levels of total cholesterol, VLDL, LDL-C, and total apoB (3,4); however, acanthocytosis and retinitis pigmentosa are less commonly found among individuals with CRD (2,5). Lipid-filled enterocytes are often seen in all 3 disorders, with variable degree of severity (3,4,6–8). Clinically, it is difficult to clearly distinguish homozygous FHBL from ABL and CRD. Plasma chylomicron and apoprotein levels with its fraction (apoB-100 and -48, apoA-I, -II, and -IV) are essential for the differential diagnosis of these disorders; however, these tests are generally not available in many countries. Homozygous FHBL and ABL individuals have a virtual absence of plasma apoB and its containing lipoproteins (3,8). In contrast to homozygous FHBL and ABL, apoB-100 and plasma triglyceride levels are normal in patients with CRD because of intact hepatic apoB synthesis (1–3). In addition, CRD can be distinguished from ABL by the absence of apoB-48 and chylomicron in plasma after the consumption of a fat-containing meal (1–3). ABL and CRD are autosomal recessive disorders caused by mutations of microsomal triglyceride transfer protein (MTP) gene (4q24) and SAR1B gene (5q31.1), respectively (3,4,8,9). CRD is the rarest disorder of hereditary HBLs, with only 11 mutations of SAR1B in 14 families described (2,9). CASE REPORT The patient was an 11-month-old boy, born at term with a birth weight of 3100 g after an uneventful pregnancy. The parents were nonconsanguineous and had another 5-year-old healthy son. The family history was noncontributory. The patient was exclusively breast-fed during the first month of life, followed by combined breast and bottle feeding. At 4 months of age, he weighed 6.2 kg (50th centile) and began to have loose, greasy, pale yellow stool 2 to 4 times daily. At 11 months of age, the patient was referred to our hospital because of failure to thrive and worsening diarrhea, with bulky and oily stool 4 to 6 times daily. The stool fat was not quantified. Physical examination showed a body weight of 6.3 kg (<3rd centile), length 73 cm (25th centile), and head circumference 42 cm (<3rd centile). The results of general and neurological examinations were within normal limits. Complete blood counts showed hemoglobin 12 g/dL, mean corpuscular volume 81.6 fL, anisocytosis 1+, microcytosis 1+, no acanthocytosis, and normal profiles of white blood cells and platelets. Electrolytes, blood urea nitrogen, creatinine, liver enzymes, albumin, and total protein levels were normal. Blood chemistry determinations demonstrated low levels of cholesterol, apoA, and apoB (Table 1) (10,11). The prothrombin time was mildly prolonged at 16.2 (10–13) seconds, international normalized ratio 1.4 (0.85–1.10). The fat-soluble vitamin levels were markedly low: vitamin A at 15.7 (20–50) μg/dL and vitamin E at 89.8 (650–1200) μg/dL. Examination of the stool revealed numerous fat globules, negative reducing substance, and no blood cells. Nerve conduction velocity yielded normal findings.TABLE 1: Plasma lipid and lipoprotein profiles of patient and family membersUpper endoscopy demonstrated generalized whitish duodenal mucosa, normal esophageal mucosa, and normal gastric mucosa. Histological examination revealed fat-filled duodenocytes without villous atrophy, suggestive of primary HBL. Lipid and lipoprotein profiles of the family members are shown in Table 1. Immunoblot testing to identify apoB and apoA fractions and abnormal apoB forms was unavailable in the country; therefore, only total apoB and apoA were obtained and were used in guiding the need for genetic analysis. For a given codominant transmission of FHBL, plasma apoB levels are expected to be low in both asymptomatic parents, or at least in 1 parent in the case in which the other parent does not harbor an APOB mutation and a de novo mutation arises during gametogenesis; whereas for ABL and CRD, inherited in an autosomal recessive pattern, apoB levels should be normal in carrier parents. On the basis of these rationales and the normal lipoprotein profiles of our patient's parents, we assumed that ABL or CRD was likely in this case. We therefore probed the MTP and SAR1B genes in this patient. Blood for genetic analysis was obtained after written informed consent was given and institutional review board approval was obtained. Isolation of genomic DNA, polymerase chain reaction, and sequencing were performed according to standard protocols. Exons 1–18 of MTP and exons 1–7 of SAR1B genes were analyzed (primer sequences are available on request). GenBank reference sequences were MTP: NT_016354.18 and NM_000253.1, and SAR1B:NC_000005.8 and NM_016103.2. Polymerase chain reaction–restriction digestion with appropriate enzymes was used as a second method for confirming the mutations identified and for mutation screening in 100 normal control samples. The genetic analyses revealed no pathogenic variant of the MTP gene but 3 reported single nucleotide polymorphisms in 1 normal control sample; they were 453C>T (rs#991811), 891C>G (rs#2306985), and IVS6-116A>G (rs#2306984). SAR1B analysis demonstrated a maternally inherited G→A transition at nucleotide 32, causing an amino acid substitution from glycine to aspartic acid (G11D) in exon 2 and a paternally inherited A→G transition at nucleotide 224, causing an amino acid substitution from aspartic acid to glycine (D75G) in exon 4 (Fig. 1). The asymptomatic brother had wild-type sequence. The G11D and D75G mutations create HphI and AIW26I restriction sites, respectively. Neither mutation was found in the control samples, suggesting highly probable pathogenic alleles.FIG. 1: A, Exon 2, G→A transition at nucleotide 32 resulting in a missense mutation, G11D. B, Exon 4, A→G transition at nucleotide 224 resulting in a missense mutation, D75G. C, Multiple sequence alignment of Sar1a and Sar1b of human, cow, wild boar, pig, mouse, rat, Chinese hamster, and zebrafish, respectively. Notice highly conserved amino acid at residues G11 and D75.DISCUSSION Homozygous/compound heterozygous FHBL, ABL, and CRD share common clinical characteristics, as mentioned previously. The APOB gene contains 29 exons, whereas SAR1B and MTP genes are smaller genes; the cost would be phenomenal if all 3 genes were to be screened in every patient with severe HBL. Tarugi et al (3) suggested plasma lipoprotein profiles by immunoblot to identify the presence of truncated apoBs shorter than apoB-100, which indicates mutations in exons 26–29 of APOB gene. If no truncated apoBs are detected in plasma, then there is no clue to the involvement of the APOB gene, and whole gene sequencing is required. (3) Under circumstance of limited resources and where immunoblot testing is unavailable, applying genetic principles of inheritance patterns in combination with parental lipoprotein profiles can be useful for selecting the genes for genetic analysis. In the present case, the absence of acanthocytosis and retinitis pigmentosa is suggestive of CRD/Anderson disease rather than ABL (5). However, severely low lipid profiles, including undetectable total cholesterol and HDL-C levels, represent unusual findings in CRD/Anderson disease (6,7,9). Both carrier parents had normal, although low-range, apoB levels, whereas the noncarrier brother had a perfectly normal level of apoB. These data suggest that a low apoB level may be used as a possible indicator of SAR1B carrier among asymptomatic individuals from a confirmed CRD family. Mutations of the SAR1B gene, which encodes the intracellular Sar1b protein, result in a defect in chylomicron exocytosis from the enterocytes. The Sar1-GTP promotes the formation of endoplasmic reticulum (ER) to Golgi transport carriers (12,13). By using a computational analysis program, Charcosset et al (2) demonstrated functional domains of Sar1b, deduced from published studies. The N-terminus is involved in the anchoring of Sar1b-GTP complex on the ER membrane, GTP binding, and hydrolysis, and the C-terminus regulates the interactions of Sar1b with the membrane (2). Amino acid residue G11 is located in the N-terminal Sec-12 interacting site (residues 1–19), which is necessary for anchoring to ER membrane, and also involved in the mechanism to deform the membrane (2). The amino acid D75 is situated on the active sites (residues 75–78) for GTP hydrolysis (2). The G11D mutation is a nonconservative amino acid substitution from neutral residue (glycine) to negatively charged residue (aspartic acid), and vice versa for the D75G mutation. These 2 mutations are highly conserved amino acid residues present in Sar1b and Sar1a across various species (human, cow, wild boar, pig, mouse, rat, Chinese hamster, and zebrafish) (Fig. 1C), and human Sar1a, which differs from human Sar1b by only 20 amino acid residues (13). The G11D and D75G mutations of SAR1B have not been reported. Our data are in agreement with previous reports that missense mutations of SAR1B represent the most common cause of CRD/Anderson disease and that most of the missense mutations map to the GDP or GTP binding site of Sar1b (13). Nemeth et al (14) reported 2 CRD patients with low concentrations of plasma total cholesterol, LDL-, and HDL-C and of apolipoprotein A-I and apoB. Normally, apolipoprotein A-I (apoA-I) constitutes approximately 2/3 of total apoA level in plasma, the remaining constituents being apoA-II and apoA-IV (10). ApoA-I is also a major fraction of apoA assembled in the HDL-C molecule (10). Therefore, we assumed that the extremely low level of apoA in the present case led to reduced synthesis of HDL-C. Treatment in the present case included restriction of dietary fats, orally medical formula high in medium-chain triglyceride, fat-soluble vitamin supplementation, and weekly intravenous fat emulsion (at the beginning). At the time of this report, the patient was 2.5 years old with normal growth and psychomotor development. He can tolerate the higher amount of dietary fats; however, long-term follow-up is required. Acknowledgment The authors thank Dr Mahippathorn Chinnapha for assistance with English corrections.