A novel autosomal recessive lipodystrophy syndrome due to homozygous LMNA variant

Abstract

Background

Despite major advances in understanding the molecular basis of various genetic lipodystrophy syndromes, some rare patients still remain unexplained.

Cases

We report a novel autosomal recessive lipodystrophy affecting two sisters aged 17 and 19 years and characterised by early onset intellectual disability, and subsequent development of near-generalised loss of subcutaneous fat with diabetes mellitus, extreme hypertriglyceridemia, hepatic steatosis, short stature, clinodactyly, joint contractures, leiomyoma of uterus and cataracts in childhood. The lipodystrophy was more pronounced in the upper and lower extremities, and there was no associated muscular hypertrophy. Using whole exome sequencing in this consanguineous Hispanic pedigree, we report disease-causing homozygous p.Arg545His LMNA variant in the affected subjects, and confirm the lack of pathogenic variants in other known lipodystrophy genes. The mother and a younger brother were both heterozygous for p.Arg545His LMNA variant and were overweight with acanthosis nigricans without any evidence of lipodystrophy. Our patients are distinct from previously reported autosomal recessive lipodystrophy syndromes and have no overlap with other autosomal recessive laminopathies, including mandibuloacral dysplasia, Emery-Dreifuss muscular dystrophy and Charcot-Marie-Tooth neuropathy.

Conclusion

Our report of this unusual familial generalised lipodystrophy syndrome adds to the pleiotropy associated with biallelic autosomal recessive LMNA variants.

BACKGROUND

Great progress has been made in elucidating the molecular genetic basis of various lipodystrophy syndromes in the last two decades and approximately 20 loci are currently known for autosomal recessive and autosomal dominant lipodystrophies.^1,2 Despite this progress, some rare patients still remain unexplained. Here, we investigated the molecular basis of a novel autosomal recessive lipodystrophy syndrome in two affected sisters belonging to a consanguineous Hispanic pedigree.

CLINICAL CASES

FGL 100.4

The proband, a Hispanic female aged 18.8 years, was born full term with birth weight of 2.8 kg and birth length of 50.8 cm. She had speech delay in early childhood, and developmental profile-II showed intellectual quotient (IQ) of 40. She has verbal and intellectual disability and requires help with activities of daily life. She had menarche at 10 years of age, with regular menstruation thereafter. At age 11 years, she developed diabetes mellitus and was started on insulin and metformin. Her serum islet cell antibody, glutamic acid decarboxylase (GAD) antibody, insulin autoantibody and Islet antigen-2 antibody were negative. Despite high-dose insulin, ~3.8 units/kg/day, she had multiple episodes of hyperglycaemia and ketosis. At 15 years of age, she developed acute pancreatitis, and eruptive xanthomas with serum triglycerides levels >10 000 mg/dL and subsequently had bimonthly hospital admissions for severe hypertriglyceridemia with or without acute pancreatitis. Abdominal ultrasound imaging revealed hepatic steatosis and liver biopsy showed macrovesicular and microvesicular steatosis in 2% of hepatocytes.

Physical examination at 16 years of age showed near-generalised lack of subcutaneous fat, with minimal subcutaneous fat in the face and trunk (figure 1A,B). Her body weight was 32.2 kg (0.1 percentile), height 140.2 cm (0.1 percentile) and body mass index (BMI) was 16.4 kg/m² (2.7 percentile). She had lipemia retinalis, thin lips with perioral pigmentation, dry and rough skin and multiple xanthomas on the extremities (figure 1C), but no acanthosis nigricans (figure 1D). She had hepatomegaly 2 cm below costal margin, bilateral clinodactyly of the third, fourth and fifth toes (figure 1E), along with ankle and knee contractures. She had Tanner IV pubic hair but relatively small breasts. There was no evidence of neuropathy, myopathy, acro-osteolysis, muscular hypertrophy, buffalo hump or double chin.

Figure 1.

Phenotypic features of the affected patients. (A) Lateral and (B) posterior view of patient FGL 100.4. Note the absence of subcutaneous fat in the extremities and buttocks, minimal subcutaneous fat in the trunk, but absence of muscular hypertrophy, or buffalo hump. (C) She had multiple xanthomas in the extremities, especially around elbows. (D) Lack of acanthosis nigricans in the axilla. (E) Clinodactyly of third, fourth and fifth toes bilaterally. (F) Skinfold thickness of FGL 100.4 (black squares) and FGL 100.5 (black circles). The grey bars show 10th to 90th percentile values of normal age-matched females with the median value marked by a horizontal line.²⁰ Both patients have markedly decreased skinfold thickness all over the body indicating near-generalised loss of subcutaneous fat. This appears to be even more pronounced in the upper and lower extremities, and patient FGL 100.5 still has some abdominal and axillary fat. T1-weighted axial MRI scans of patient FGL 100.4. Adipose tissue is seen as areas of bright signal intensity. (G) MRI of the head showing preservation of scalp and orbital fat. MRI of the forearm (H) and thigh (I) showing minimal subcutaneous fat. (J) MRI of the chest showing loss of subcutaneous fat in the thorax and minimal breast tissue. Loss of subcutaneous fat in the abdomen (K) and buttocks (L), but intra-abdominal fat (K) and labial and peri-rectal fat (M) were present in near-normal amounts.

Her skinfold thickness measurements were mostly below the 10th percentile of normal (figure 1F) and total body fat by dual energy X-ray absorptiometry (DEXA) was 22.3% (3rd percentile). Fat in the upper extremities was 25%, in the lower extremities was 17% and in the trunk was 24.1%. Whole body MRI revealed normal subcutaneous fat under the scalp and in orbits (figure 1G), but markedly reduced subcutaneous fat on the arms, forearms (figure 1H), thighs (figure 1I), calves, chest and abdomen (figure 1J and K). Intra-abdominal fat was present in near-normal amount (figure 1K). She had normal fat in the bone marrow, labia majora and peri-rectal region (figure 1L and M). Pelvic MRI scan revealed mild right hydrosalpinx, multiple myometrial and right parauterine fibroids and left ovarian endometrioma. She underwent myomectomy, and pathology showed leiomyoma with ischaemic necrosis and degenerative changes. At 18 years of age, she was diagnosed with bilateral cataracts.

Her fasting blood glucose was 292 mg/dL; total cholesterol >1000 mg/dL, high-density lipoprotein (HDL)-cholesterol, 10 mg/dL; triglycerides 5436 mg/dL; aspartate aminotransferase 60 IU/L; alanine aminotransferase 111 IU/L; haemoglobin A1c 7.2%; leptin level 1.9 ng/mL and adiponectin <2 μg/mL. An SNP array analysis showed a number of segments of allelic homozygosity approximately 577 Mb, highly suggestive of parental relatedness.

FGL 100.5

The sister of the proband aged 17 years was born full term with birth weight of 2.6 kg, and birth length of 50.8 cm. She was diagnosed with intellectual disability since 3 years of age, and currently requires help with activities of daily life. She was diagnosed with hypothyroidism at 10 years age (serum thyroglobulin and thyroid peroxidase antibodies were negative), with ‘small thyroid’ on ultrasound. She was started on levothyroxine 44 μg daily. She had menarche at 9.5 years with irregular menstruation. She had abnormal glucose tolerance with 2 hours postprandial blood glucose value of 273 mg/dL at 16.5 years of age, prompting treatment with metformin 500 mg twice daily. At 16 years 10 months, she had an episode of acute pancreatitis, and insulin therapy was subsequently added at 17 years of age for worsening glycaemic control (haemoglobin A1c 8.6%). Abdominal ultrasound and MRI showed hepatic steatosis. MRI of the pelvis showed anterior uterine pedunculated and subserosal fibroids, ovarian cysts and left hydrosalpinx. Right ovarian cyst was laparoscopically drained, and cytology revealed scattered cellular element without malignancy. At age 17, she was diagnosed with bilateral cataracts.

Physical examination at 16.5 years of age showed absence of subcutaneous fat from the extremities, but near-normal subcutaneous fat in the trunk and face. Her weight was 37.8 kg (0.1 percentile), height 144.6 cm (0.2 percentile) and BMI 18.1 kg/m² (14.5 percentile). She had small mouth, thin lips with perioral pigmentation and dry skin. The liver was palpable 2 cm below the costal margin. She also had bilateral clinodactyly of the third, fourth and fifth toes, dystrophic nails and ankle contractures. She had Tanner 5 pubic hair and breasts.

Her skinfold thickness measurements are shown in figure 1F. Her total body fat by DEXA was 26.8% (18 percentile). Fat in the upper extremities was 27.6%, in the lower extremities was 16.5% and in the trunk was 32.2%. Her fasting blood glucose was 146 mg/dL; HDL-cholesterol 37 mg/dL; triglycerides 262 mg/dL; total cholesterol 210 mg/dL; aspartate aminotransferase 45 IU/L; alanine aminotransferase 77 IU/L; haemoglobin A1c 6.7%; leptin level 9.1 ng/dL; adiponectin 2 μg/mL; C-peptide 3.95 ng/mL and fasting insulin 23.4 μIU/mL. Her SNP array also revealed number of segments of allelic homozygosity approximately 864 Mb.

METHODS

A written informed consent was obtained from all participants and their legal guardian. Height and body weight were measured by standard procedures. Skinfold thickness, DEXA and MRI were performed as reported in online supplementary methods.

Genomic DNA was isolated from peripheral blood using the Easy-DNA kit (Invitrogen, Carlsbad, California, USA). The two affected siblings (FGL 100.4 and FGL 100.5) and their unaffected mother (FGL 100.2) (online supplementary data) underwent exome sequencing using the Integrated DNA Technologies xGen Exome Research Panel V.1.0 on the Illumina platform. Sequencing read length was paired-end 2×150 bp. Sequences were aligned to the human reference genome b37, and variants were called using the Genome Analysis Toolkit³ and annotated using SnpEff.⁴ Based on the results of SNP arrays, we first searched for run of homozygosity (ROH) regions shared by the two affected but not by the unaffected mother using BCFtools/RoH,⁵ followed by filtering for rare missense, nonsense, splicing or frameshift homozygous mutations with minor allele frequency (MAF) <0.01 in the 1000 Genomes Project (http://www.inter-nationalgenome.org/), genome aggregation (gnomAD; http://gnomad.broadinstitute.org/) and UK10K Project databases. Variants with the Genomic Evolutionary Rate Profiling (GERP++) score⁶ >2.0 and Combined Annotation Dependent Depletion score⁷ >15 were considered. We considered missense variants predicted to be ‘probably damaging’ by Polymorphism Pheno-typing v2 (HumDiv; http://genetics.bwh.harvard.edu/pph2/). We also used Sanger sequencing to confirm segregation of the candidate variants within the pedigree.

RESULTS

The ROH analysis revealed a total of 27.1 Mb stretches of homozygous segments >3 Mb in length in the two affected sisters, covering 16.7% and 21.9% of genome; but only 0.8% in the unaffected mother (figure 2A). The ROH regions in the affected sisters were much larger than that (~11%) found in the offsprings of first cousin marriage in the Pakistanis and Arabs strongly supporting consanguinity among parents.⁸ There were five variants meeting the filtering criteria (online supplementary table S1), including a homozygous variant chr1:156 107 470G>A (NM_170707:c.1634G>A, NP_733821: p.Arg545His (rs142191737, MAF=0.0000039 in Latinos in gnomAD)) in LMNA in both patients. Since there was a large ROH encompassing LMNA on chromosome 1 shared by the two affected sisters but not by the unaffected mother (figure 2B), high degree of conservation of lamin A/C Arginine 545 residue (figure 2G) and given the previous association of LMNA mutations with various lipodystrophy syndromes,^9–12 we considered the homozygous p.Arg545His LMNA mutation as disease-causing. Sanger sequencing revealed segregation of the LMNA variant with the phenotype in the family (figure 2C,D,E,F). The other four homozygous variants in CD101, POU2F1, GREB1 and ANK1 were considered highly unlikely to cause lipodystrophy (online supplementary table S1). Whole exome sequencing also confirmed the lack of pathogenic variants in other lipodystrophy genes.

Figure 2.

FGL 100 exome sequencing showing runs of homozygosity, location of the mutation in the LMNA gene, sequence electropherogram, pedigree and conservation of residue among different species. (A) Runs of homozygosity (ROH) from exome sequencing data. For each individual, the top line represents the markers with alternate homozygous genotypes and the bottom line represents heterozygous SNP. The rectangles highlight ROH regions; regions private to one subject are highlighted in grey, whereas regions shared by both affected subjects are in red. Location of LMNA gene within ROH is shown. (B) Illustration of LMNA gene containing 12 exons. Boxes indicate exons and the lines in between them indicate the introns. (C) Sequence electropherogram from Sanger sequencing of FGL 100.3, showing normal sequence of coding region of LMNA. (D) Sequence electropherogram from FGL 100.2 and showing heterozygous c.1634G>A (p.Arg545His) LMNA mutation. (E) Sequence electropherogram from FGL 100.4 showing homozygous c.1634G>A (p.Arg545His) LMNA mutation. (F) FGL 100 pedigree. Circles denote females and squares denote males and the numbers inside the symbols denote current age in years. Subjects with homozygous mutation c.1634G>A, p.Arg545His in LMNA are indicated with ‘A/A’, and those with the heterozygous mutation are indicated with ‘G/A’, and with the wild type as ‘G/G’. Vertical arrows indicate subjects for whom DNA was available, and a slanted arrow indicates the proband. Affected individuals are shown as filled black symbols, unaffected subjects as unfilled symbols. (G) The mutated residue arginine at position 545 (shown in red font) is conserved among all the species.

DISCUSSION

We report a novel autosomal recessive lipodystrophy syndrome characterised by near-generalised loss of subcutaneous fat with pubertal-onset diabetes, extreme hypertriglyceridemia and hepatic steatosis, short stature, intellectual disability, thin lips with perioral pigmentation, dry skin, clindodactyly of toes, joint contractures, cataracts, uterine fibroids, ovarian cysts and hydrosalpinx. Based on the phenotype, we have labelled this syndrome as autosomal recessive familial generalised lipodystrophy.

Our patients are distinct from previously reported autosomal recessive lipodystrophy syndromes. For example, they did not develop lipodystrophy at birth and lacked increased lean mass as seen in congenital generalised lipodystrophy; they did not have mandibular hypoplasia, acro-osteolysis, alopecia and sclerodermatous skin as observed in mandibuloacral dysplasia; did not have panniculitis and recurrent fever seen in autoinflammatory lipodystrophies; did not have lipomatosis or myopathy seen in those with LIPE mutations; had no increase in dorsocervical, submandibular, pubic and vulvar fat seen in the only patient with PCYT1A mutations and had near-generalised instead of partial lipodystrophy seen in the patient with CIDEC mutation.^13,14 Furthermore, our patients had no overlap with other autosomal recessive laminopathies, including mandibuloacral dysplasia, Emery-Dreifuss muscular dystrophy, Charcot-Marie-Tooth neuropathy and other rare syndromes (online supplementary figure S1).

Interestingly, intellectual impairment has not been reported previously in association with LMNA mutations (https://www.ncbi.nlm.nih.gov/clinvar/?term=lmna). However, since there were no chromosomal deletions on SNP arrays and exome sequencing, and none of the other genes with rare homozygous variants in the affected sisters has a known association with intellectual impairment (online supplementary table S1), there is a strong likelihood that the homozygous p.R545H LMNA mutation caused developmental delay. The arginine residue at position 545 lies in the immunoglobulin fold of globular tail domain (residues 436–552) of lamins A and C.^15–17 Such immunoglobulin folds are known for protein-protein interaction, which could be disrupted with the substitution of arginine for histidine at position 545.

Chan et al¹⁸ reported heterozygous p.Arg545His LMNA variant in a female aged 51 years with lipoatrophy of limbs, and central adiposity suggestive of familial partial lipodystrophy-Dunnigan variety (FPLD) phenotype with normal echocardiogram, however, there was no documentation of reduced extremity subcutaneous fat by patient photographs, skinfolds thickness measurements, DEXA or MRI. Also, there was no documentation of segregation of the variant with FPLD phenotype among any other family member. van Rijsingen et al¹⁹ also reported the same variant in a patient with cardiac involvement, however, no further details about the phenotype were available. In contrast to the previous anecdotal reports of limb lipoatrophy and cardiac disease in two subjects with heterozygous p.Arg545His LMNA mutation,^18,19 we did not observe any lipodystrophy or cardiac problems in the mother and a brother of the proband, who were heterozygous for the mutation (online supplementary data). In contrast, both had marked generalised obesity with BMIs of 40.2 and 36.9 kg/m², respectively. These observations support the autosomal recessive pattern of inheritance in this pedigree.

Our patients are unique compared with previously reported cases with biallelic LMNA variants with partial lipodystrophy (online supplementary table 3). Our patients had generalised lipodystrophy, lack of muscular hypertrophy, lack of excess subcutaneous fat accumulation in the face and dorsocervical regions, clinodactyly, joint contractures, uterine leiomyomas and cataracts supporting our contention that this is a novel syndrome. Furthermore, two of the three case reports are suggestive of an autosomal co-dominant transmission compared with our patients who have autosomal recessive transmission.

In conclusion, our report extends the spectrum of autosomal recessive lipodystrophies and adds to the pleiotropy associated with biallelic recessive LMNA mutations.