Deciphering the Clinical Presentations in LMNA-related Lipodystrophy: Report of 115 Cases and a Systematic Review

Abstract

Context

Lipodystrophy syndromes are a heterogeneous group of rare genetic or acquired disorders characterized by generalized or partial loss of adipose tissue. LMNA-related lipodystrophy syndromes are classified based on the severity and distribution of adipose tissue loss.

Objective

We aimed to annotate all clinical and metabolic features of patients with lipodystrophy syndromes carrying pathogenic LMNA variants and assess potential genotype-phenotype relationships.

Methods

We retrospectively reviewed and analyzed all our cases (n = 115) and all published cases (n = 379) curated from 94 studies in the literature.

Results

The study included 494 patients. The most common variants in our study, R482Q and R482W, were associated with similar metabolic characteristics and complications though those with the R482W variant were younger (aged 33 [24] years vs 44 [25] years; P < .001), had an earlier diabetes diagnosis (aged 27 [18] vs 40 [17] years; P < .001) and had lower body mass index levels (24 [5] vs 25 [4]; P = .037). Dyslipidemia was the earliest biochemical evidence described in 83% of all patients at a median age of 26 (10) years, while diabetes was reported in 61% of cases. Among 39 patients with an episode of acute pancreatitis, the median age at acute pancreatitis diagnosis was 20 (17) years. Patients who were reported to have diabetes had 3.2 times, while those with hypertriglyceridemia had 12.0 times, the odds of having pancreatitis compared to those who did not.

Conclusion

This study reports the largest number of patients with LMNA-related lipodystrophy syndromes to date. Our report helps to quantify the prevalence of the known and rare complications associated with different phenotypes and serves as a comprehensive catalog of all known cases.

Lipodystrophy syndromes are a heterogeneous group of rare genetic or acquired disorders characterized by generalized or partial loss of adipose tissue (1). While earlier reports estimated a prevalence between 1 in 10 million and 1 in 1 million, recent studies have predicted that the clinical prevalence of inherited lipodystrophy syndromes is as high as 1 in 20 000 individuals, while the molecular prevalences of the specific subtypes is even more common, up to 1 in 7000 (2, 3). Pathogenic variants in various genes, including AGPAT1, BSCL2, CAV1, PPARG, AKT2, and LMNA have been associated with the recognizable phenotypes of inherited lipodystrophy syndromes (3).

LMNA-related lipodystrophy syndromes are classified based on the severity and distribution of adipose tissue loss (2). The most common form is familial partial lipodystrophy type 2 (FPLD2), also known as the Dunnigan-type lipodystrophy (1, 4, 5). However, patients carrying different pathogenic LMNA variants may also present as having atypical FPLD2, multisystem FPLD syndrome, non-Dunnigan FPLD, atypical Werner syndrome, atypical progeroid syndrome, mandibuloacral dysplasia, or generalized lipodystrophy syndromes with or without progeroid features (1, 3, 4, 6-8). The LMNA (lamin A/C) gene, which encodes the nuclear lamina proteins lamin A and its isoform lamin C, has been associated with pivotal roles in the maintenance of the nuclear organization and nuclear matrix homeostasis (9, 10). Due to lamins' ubiquitous and variable expression throughout the body, laminopathies can have far-reaching consequences, affecting not just adipose tissue but also skeletal muscle, the cardiovascular system, kidneys, and peripheral nerves (10, 11). The clinical manifestations of LMNA variants can vary widely, even among members of the same family who carry the same variant (12). Patients with FPLD2 show ectopic fat accumulation in the absence of peripheral adipose depots (13). As lipid spills over to ectopic sites driven by adipose tissue loss, a state of low-grade systemic inflammation (14) develops along with several hormonal and metabolic complications, such as hypertriglyceridemia and insulin resistance that can progress to diabetes. This can lead to several debilitating organ involvements, including pancreatitis, cardiomyopathy, early atherosclerosis, nephropathy, retinopathy, hepatic steatosis, and fibrosis, among others (1, 15).

Diverse phenotypical and clinical features of atypical adipose distribution, progeroid features, and multisystemic involvements pose a challenge to clinicians in the diagnosis, follow-up, and treatment of these rare disorders. Hence, several reviews and meta-analyses have focused on the FPLD2 phenotypes (4) and the overall systemic involvements of the LMNA gene (16, 17), while others reviewed the LMNA-related cardiomyopathy diseases or muscular dystrophies (18, 19). Nevertheless, due to the rarity and heterogeneity of this syndrome, additional data compiling all available information on the genotype-phenotype relationships of LMNA-related lipodystrophy syndromes is a missing gap in the literature (1, 3, 20, 21).

In this study, we analyzed all 106 available reports (6-8, 12, 13, 20-120) on cases with various LMNA-related lipodystrophy phenotypes, omitting the distinct and well-defined form of Hutchinson-Gilford progeroid syndrome (100, 121, 122). We also included follow-up data for our 90 previously published cases (13, 20, 21, 92, 110-118), in addition to an unreported 25 patients, to provide a comprehensive and detailed analysis of the genotype-phenotype correlations of patients with lipodystrophy who carry distinct pathogenic LMNA variants.

Materials and Methods

Case Series

We retrospectively collected the clinical data of all patients with lipodystrophy carrying pathogenic LMNA variants from the divisions of metabolism and endocrinology of 4 medical centers: the University of Michigan (USA), the University of São Paulo (Brazil), Dokuz Eylul University (Turkey), and Ege University (Turkey).

Online Search Strategy

We further selected cases for inclusion in the study using web search tools including PubMed, OMIM, and Google Scholar using the following search terms: “LMNA,” “lipodystrophy,” and “partial lipodystrophy” (1997 to present, last accessed June 14, 2023). The electronic search strategy is presented in Fig. 1. Studies eligible for our analysis had to include information on the pathogenic LMNA variant and provide at least one of the following clinical characteristics per individual: body mass index (BMI) and the presence of insulin resistance, diabetes, dyslipidemia, hypertriglyceridemia, hypertension, hepatic steatosis, pancreatitis, nephropathy, or retinopathy.

Figure 1.

Electronic search strategy.

In this study, lipodystrophy phenotypes were classified as FPLD2 and other overlapping phenotypes. FPLD2 was classified as typical and atypical FPLD2 (4). Typical FPLD2 was defined by loss of adipose tissue from the upper and lower extremities and trunk, with an excess of adipose tissue in the face, neck, and supraclavicular region. Atypical FPLD2 was defined when adipose tissue loss was less severe, with lipoatrophy being more evident in the legs (4). Other overlapping phenotypes were considered (i) in the presence of atypical progeroid features, defined when the individual's appearance appeared older than their chronological age, which may have been accompanied by premature gray hair, beaked nose, or pinched facies, and (ii) when skeletal defects were observed, including those with mandibuloacral dysplasia, which is a heterogeneous phenotype, primarily characterized by mandibular and clavicular hypoplasia (8).

Exclusion criteria were articles not written in English, animal studies, and review articles not containing the required detailed LMNA variant-phenotype information. Patients with Hutchinson-Gilford progeroid syndrome were not included in the analysis. To avoid duplications, previous reports on our patients, on whom we collected the follow-up data for this analysis, were also excluded from the literature review. Two reviewers (O.B. and M.C.F.F.) independently screened titles and abstracts to be included in the study.

The following data on the patients' clinical and laboratory characteristics were recorded: sex, age at diagnosis, BMI, pathogenic LMNA variant, and presence and age of onset of additional complications/features including insulin resistance, diabetes, dyslipidemia, hypertriglyceridemia, hypertension, hepatic steatosis, pancreatitis, nephropathy, and retinopathy. Details of circulating lipid concentrations (triglycerides [TGs], total cholesterol, high-density lipoprotein [HDL], low-density lipoprotein [LDL]), glycated hemoglobin A_1c (HbA_1c), leptin, adiponectin, and fibroblast growth factor-21 (FGF-21) concentrations were also noted. Additional data on previously unreported patients regarding complaints of pain, and body composition demonstrating the pattern of adipose tissue loss have also been included. Fat and lean body mass were estimated using dual x-ray absorptiometry. The ratio of percentage fat mass of the trunk to percentage fat mass of the legs (FMR) was also presented.

Diabetes was diagnosed when HbA_1c was greater than or equal to 6.5% or fasting plasma glucose was greater than or equal to 126 mg/dL (123), or with a report of antidiabetic agent use. Dyslipidemia was diagnosed with a total cholesterol level of 200 mg/dL or greater or TGs of 150 mg/dL or greater, or HDL of less than 50 mg/dL for women and less than 40 mg/dL for men, or LDL level of 130 mg/dL or greater, or treatment with lipid-lowering drugs (124). Hypertension was defined when blood pressure was 140/90 mm Hg or greater or the patient was using antihypertensive medications (124). Nephropathy was diagnosed by the presence of diabetic kidney disease with albuminuria of more than 30 mg/g and/or reduced glomerular filtration rate or in the presence of focal segmental glomerulosclerosis or its clinical features (125). Abnormal cardiac function was defined with the presence of valvular defects, conduction abnormalities, or anatomical defects identified via echocardiography or electrocardiogram.

Ethical Approval

This study was approved by the ethical committee of the University of Michigan (institutional review board [IRB] Med approval No. LD-LYNC HUM#00127427) and the research ethics committee of the Clinical Hospital of Ribeirao Preto (IRB Med approval No. CAAE 56714421.3.0000.5440) and was performed according to the principles of the Declaration of Helsinki.

Statistical Analysis

Results were analyzed using SPSS v.24 for Windows. The data are summarized as medians (interquartile range) unless otherwise stated. The numbers (N) of the patients affected (X), and the total number of patients (Y) whose data were available for analyses were given as (N, X/Y). The Mann-Whitney U test was used for comparison between groups. Categorical variables were analyzed by χ² or Fisher exact test. A P value of less than .05 was considered statistically significant.

Results

Characteristics of 25 Previously Unreported Patients

Detailed clinical and biochemical characteristics of all patients are presented in Table 1. Twenty-five (19 women, 76%) patients from 6 unrelated families were followed up with a median duration of 15 (15) years (range, 1-32 years) were evaluated. Six of these patients were male. The median age at diagnosis was 34 (21) years (range, 5-52 years). The patients carrying the R644C variant demonstrated an atypical FPLD2 phenotype, while the rest of the patients demonstrated the classic FPLD2 phenotype. Three (12%) patients had complaints of gastroesophageal reflux, 1 (4%) had bowel problems, 1 (4%) had back pain, 1 (4%) had muscle pain, and 2 (8%) had joint pain. Seven patients harbored R644C (same family), while 18 carried the R482W (5 different families) variant. Patients with a BMI level of 25 (range, 20–31) had an FMR of 1.4 (1) (range, 1.3-2.0), all of whom had an FMR above the 95% CI of reference median values (126). Seventeen patients (68%) had diabetes diagnosed at a median age of 36 (19) years with a median HbA_1c of 8.8% (4), while 2 patients had prediabetes. Five (20%) patients had retinopathy, 5 (20%) had nephropathy, and 12 (48%) patients had nonalcoholic fatty liver disease. None of the patients had myopathy. Twenty-three (92%) patients had dyslipidemia, of whom 22 also had hypertriglyceridemia. Median TG, cholesterol, HDL, and LDL concentrations were 264 (153), 198 (80), 38 (9), and 105 (70) mg/dL, respectively. Three (13%) patients had at least 1 pancreatitis attack, diagnosed at ages 32 (R644C), 32 (R482W), and 48 (R482W) years. The patients carrying R644C (n = 7) and R482W (n = 18) had similar ages at diagnosis, BMI values, and concentrations of cholesterol, HDL, TGs, LDL, and HbA_1c.

Table 1.

Clinical characteristics of our previously unreported patients

Pt	Variant	M/F	Age, y	BMI	FMR	Diabetes	Diabetes age, y	Dys lipidemia	Pancreatitis	Hepato steatosis	HbA1c, %	LDL, mg/dL	HDL, mg/dL	TGs, mg/dL	Cholesterol, mg/dL	Retinopathy	Nephro pathy	Cardiac abnormality
1.1	R644C	F	33	26	1.8	Y	32	Y	Ya	1	11.7	na	60	675	336	Y	Y	N
1.2	R644C	F	44	31	1.8	Y	44	Y	N	1	5.6	90	37	105	148	Y	Y	Yf
1.3	R644C	F	39	24	1.4	Y	46	Y	N	1	9.4	116	35	275	207	N	N	N
1.4	R644C	M	32	na	na	N	—	Y	N	na	6.0	148	46	76	209	N	N	N
1.5	R644C	M	44	na	1.4	Y	44	Y	N	na	5.4	168	43	88	229	N	N	N
1.6	R644C	M	45	29	na	Y	45	Y	N	na	na	141	35	376	251	N	N	N
1.7	R644C	M	na	na	na	N	—	N	N	na	na	100	46	33	153	N	N	N
2.1	R482W	F	52	23	2.0	Y	56	Y	N	na	6.2	157	42	239	248	N	N	N
2.2	R482W	F	14	21	1.4	N	—	Y	N	na	5.1	46	45	354	162	N	N	N
3.1	R482W	M	49	22	na	Y	35	Y	Yb	1d	na	49	38	65	100	Y	Y	N
3.2	R482W	F	23	20	1.3	Y	23	Y	N	1	10.4	137	28	334	232	na	na	N
3.3	R482W	F	40	25	1.8	Y	47	Y	N	1e	5.8	71	38	314	171	N	N	N
3.4	R482W	F	45	27	1.4	Y	46	Y	N	1	9.4	137	45	278	238	N	Y	Yg
3.5	R482W	F	5	25	1.7	N	—	Y	N	N	5.5	77	38	341	183	N	N	N
3.6	R482W	F	42	21	1.1	Y	43	Y	N	1e	7.8	93	23	224	161	N	Y	N
3.7	R482W	M	28	31	na	Y	28	Y	N	1	5.1		29	1487	275	N	N	N
3.8	R482W	F	20	na	1.4	N	—	Y	N	1	4.5	63	47	183	147	N	na	N
4.1	R482W	F	20	27	1.6	Y	20	Y	N	1	9.7	106	40	264	198	N	Y	N
4.2	R482W	F	12	27	1.7	Y	21	Y	N	1	5.8	41	35	276	131	na	na	N
4.3	R482W	F	31	na	1.5	N	—	Y	N	na	5.1	89	32	230	167	na	na	N
4.4	R482W	F	25	22	1.8	N	—	Y	N	N	5.8	150	37	296	246	na	na	N
5.1	R482W	F	29	31	1.7	Y	30	Y	Yc	1	8.1	146	33	282	236	Y	N	N
5.2	R482W	F	39	25	1.5	Y	26	Y	N	1	9.7	105	37	159	174	N	Y	N
5.3	R482W	F	35	24	1.7	Y	36	Y	N	1	9.7	60	42	200	142	Y	Y	Yh
6.1	R482W	F	46	25	na	na	na	Y	N	N	5.6	122	37	247	209	na	na	N

Abbreviations: BMI, body mass index; F, female; HbA_1c, glycated hemoglobin A_1c; HDL, high-density lipoprotein; LDL, low-density lipoprotein; M, male; N, not present; na, not available; PT, patient codes; TGs, triglycerides; Y, yes, present;

Age at diagnosis, ^aat age 32 years; ^bat age 32 years; ^cat age 48 years; ^dportal vein thrombosis; ^ehepatitis C cirrhosis; ^fmild dilatation of ascending aorta; ^gearly coronary artery disease, ^hfirst-degree atrioventricular block.

Follow-Up Characteristics of 90 Previously Reported Patients

Detailed clinical and biochemical characteristics of all our patients have been previously reported (13, 20, 21, 92, 110-118). Follow-up data for these patients are presented in Tables 2 and 3. Ninety (67 women, 74%) patients from 42 different families with a median duration follow-up of 12 (15) years (range, 1-32 years) were included. The median age at the cross-sectional evaluation point was 41 (33) years (range, 1-77 years). Ninety-eight percent (n = 89) of patients presented with FPLD2 phenotypes, while one had generalized lipodystrophy without progeroid features (n = 1, T10I). Sixteen patients had gastroesophageal reflux, 11 had bowel problems, and 13 had myopathy. R482W was the most common variant (n = 37, 41%) in our cohort, while the other variants identified were as follows: R482Q (n = 36, 40%), R482L (n = 3, 3%), R349W (n = 2, 2%), c.1488 + 5G > C, intron 8 (n = 1, 1%), D300N (n = 1, 1%), G535Q (n = 1, 1%), R349W (n = 1, 1%), R541P (n = 1, 1%), R582C (n = 1, 1%), R582H (n = 1, 1%), R582L (n = 1, 1%), R584H (n = 1, 1%), R60G (n = 1, 1%), R62G (n = 1, 1%), and T10I (n = 1, 1%). The median BMI level was 26 (8) (range, 16-41). Sixty-four (72%) patients had diabetes with a median HbA_1c value of 7.2% (3%). Abnormal cardiac examination, nephropathy, and retinopathy were diagnosed in 29 (32%), 29 (32%), and 7 patients (7%), respectively. Seventy patients (83%) had dyslipidemia, 93% (n = 65) of whom had increased TG concentrations. The median levels of cholesterol, TGs, HDL, and LDL were 184 (69), 255 (317), 38 (16), and 100 (47) mg/dL, respectively. Thirteen patients (14%) carrying R482Q (n = 4), R482W (n = 4), D300N (n = 1), c.1488 + 5G > C (n = 1), R584H (n = 1), R62G (n = 1), and T10I (n = 1) variants had at least one episode of pancreatitis.

Table 2.

Clinical characteristics of our previously reported patients (13, 21, 92, 110, 111, 113–115, 117)

Pt	Variant	M/F	Age, y	BMI	Diabetes	Diabetes age, y	Dys lipidemia	Pancreatitis	Hepato steatosis	HbA1c, %	LDL, mg/dL	HDL, mg/dL	TGs, mg/dL	Cholesterol, mg/dL	Retinopathy	Nephro pathy	Cardiac abnormality
UM1	R482Q	F	41	33	N	—	Y	N	N	7.3	84	41	255	176	N	N	na
UM2	R482Q	F	42	31	N	—	Y	N	N	9.0	144	49	249	243	N	N	na
UM3	R482Q	F	33	23	Y	33	Y	N	Y	5.8	79	69	253	198	N	N	N
UM4	R482Q	F	47	na	Y	47	Y	N	Y	6.8	48	45	364	166	N	N	N
UM5	R482Q	F	62	34	Y	60	Y	N	Y	6.6	46	45	280	147	N	N	Yk
UM6	R482L	F	28	29	Y	20	N	N	Y	5.2	110	45	137	183	Y	Y	Yl
UM7	R482Q	F	38	27	Y	47	Y	N	Y	5.4	82	38	662	218	N	N	N
UM8	R482Q	M	16	31	N	—	N	N	N	5.0	89	58	104	168	N	N	na
UM9	R482Q	M	15	23	N	—	N	N	N	6.6	73	48	120	145	N	N	na
UM10	R482L	F	20	30	N	—	Y	N	N	11.1	56	37	262	145	N	N	na
UM11	c.1488 + 5G > C	F	18	22	Y	12	Y	Ya	Y	10.8	67	39	868	191	N	Y	Ym
UM12	R482W	F	20	26	Y	19	Y	Yb	N	11.9	152	32	359	256	N	N	N
UM13	R482W	F	55	20	Y	20	Y	Yc	N	5.2	153	34	355	187	Y	N	Yn
UM14	R482W	M	44	38	Y	na	Y	N	Y	7.7	125	28	326	196	N	N	na
UM15	R482Q	F	23	23	N	—	N	N	N	7.2	84	52	40	148	N	N	N
UM16	R482Q	F	63	38	Y	64	Y	N	Y	na	127	35	469	262	Y	N	Yn
UM17	R582H	F	47	26	Y	47	Y	N	Y	11.9	47	30	398	157	N	N	na
UM18	R62G	F	55	29	Y	38	Y	Yd	Y	7.2	77	19	150	125	N		Y
UM19	R482Q	M	58	27	Y	45	Y	N	Y	8.7	63	30	333	160	N	N	N
UM20	R482Q	F	58	27	Y	53	Y	N	Y	9.7	89	26	249	142	N	Y	Yn
UM21	R482Q	M	48	39	Y	46	Y	Ye	Y	na	na	32	1454	347	N	Y	Yl
UM22	R482Q	F	44	23	Y	23	N	N	Y	5.2	120	24	150	174	N	N	Yn
UM23	R482Q	M	76	na	N	—	na	N	N	5.2	57	29	118	124	N	N	Yn
UM24	R482Q	M	19	na	N	—	N	N	N	na	108	49	85	174	N	N	na
UM25	R482Q	F	32	32	N	—	Y	Yf	Y	5.6	121	43	308	226	N	N	N
UM26	R482W	F	45	41	Y	25	Y	N	Y	5.6	95	51	244	195	N	N	N
UM27	R482W	F	19	28	N	—	Y	N	N	7.3	73	36	na	143	N	N	na
UM28	R582L	F	44	28	N	—	N	N	Y	na	116	42	116	181	N	N	N
UM29	R60G	F	26	19	Y	26	Y	N	Y	6.7	84	35	402	159	N	Y	N
UM30	D300N	F	34	na	Y	44	Y	Yg	Y	5.3	93	38	126	156	N	N	Yl
UM31	R482Q	F	43	37	Y	26	Y	Y	Y	na	na	26	3461	236	N	Y	N
UM32	R482Q	F	29	18	N	—	Y	N	Y	8.5	226	39	72	262	N	N	na
UM33	R482Q	F	21	na	Y	30	Y	Yh	Y	na	na	na	484	248	N	N	N
UM34	R482Q	F	30	22	Y	42	Y	N	Y	6.8	112	31	3602	376	Y	N	na
UM35	R482Q	F	59	na	Y	36	na	N	N	7.6	111	41	372	226	N	N	N
UM36	R482L	F	40	20	Y	55	N	N	Y	8.4	81	58	105	160	N	N	Yk
UM37	R482W	F	60	26	Y	31	Y	N	Y	6.9	84	30	346	182	N	N	N
UM38	R482W	M	42	38	Y	43	Y	N	Y	6.4	117	38	195	193	N	N	Yl
UM39	R482W	M	43	29	Y	50	Y	N	N	5.9	29	36	363	137	N	N	Yl
UM40	R453W	F	54	22	N	—	Y	N	N	13.3	113	52	196	204	N	N	N
UM41	R541P	F	41	25	Y	57	Y	N	Y	5.8	130	45	192	213	N	Y	Yn
UM42	R584H	F	18	30	Y	25	Y	Yi	Y	na	91	39	434	237	N	N	Yn
UM43	R482Q	F	37	23	Y	29	Y	N	Y	8.2	145	36	211	223	na	N	Yn
UM44	R349W	F	37	26	Y	20	Y	N	Y	6.6	174	44	236	271	N	Y	Yn
UM45	R349W	M	26	17	Y	20	N	N	Y	11.9	107	37	89	162	N	N	Yo
UM46	R582C	F	9	24	N	—	N	N	Y	5.5	93	39	113	155	N	N	N
UM47	T10I	F	4	16	Y	6	Y	Yj	Y	8.0	150	17	4727	670	N	Y	Yn
UM48	G535Q	F	48	23	Y	47	Y	N	Y	9.3	na	47	912	303	N	Y	Yl
UM49	R482Q	F	70	na	Y	79	na	N	N	7.5	na	na	na	na	N	N	na
UM50	R482Q	M	77	26	Y	77	na	N	Y	7.5	na	na	na	na	Y	Y	na
UM51	R482Q	M	46	26	na	na	na	N	na	na	na	na	na	na	na		na
UM52	R482Q	M	71	26	Y	na	na	N	na	7.5	na	na	na	na	Y	Y	na
UM53	R482Q	F	47	22	N	—	Y	N	Y	5.3	127	41	212	284	N	N	N

Abbreviations: BMI, body mass index; F, female; HbA_1c, glycated hemoglobin A_1c; HDL, high-density lipoprotein; LDL, low-density lipoprotein; M, male; N, not present; na, not available; PT, patient codes; TGs, triglycerides; Y, yes, present.

^a At age 18 years; ^bat age 19 years; ^cat age 45 years; ^dat age 46 years; ^eat age 10 years; ^fat age 43 years; ^gat age 26 years; ^hat age 38 years; ⁱat age 36 years; ^jat age 5 years; ^kdiastolic dysfunction; ^lventricular hypertrophy; ^mantrioventricular block; ⁿvalvular problems; ^oearly coronary artery disease.

Table 3.

Clinical characteristics of our previously reported patients (20, 110, 116, 127)

Pt	Variant	M/F	Age, y	BMI	Diabetes	Diabetes age, y	Dys lipidemia	Pancreatitis	Hepato steatosis	HbA1c, %	LDL, mg/dL	HDL, mg/dL	TGs, mg/dL	Cholesterol, mg/dL	Retinopathy	Nephro pathy	Cardiac abnormality
TR1	R482Q	F	69	26	Y	28	Y	N	Y	8.4	52	36	196	127	N	Y	Ya
TR2	R482Q	F	58	25	Y	38	Y	N	Y	7.8	130	44	221	210	N	N	N
TR3	R482Q	F	50	26	Y	42	Y	N	Y	6.3	158	35	1358	229	N	Y	N
TR4	R482Q	F	57	25	Y	45	Y	N	Y	6.0	84	44	285	185	N	N	N
TR5	R482Q	F	55	20	Y	40	Y	N	Y	8.0	131	46	197	217	N	N	N
TR6	R482Q	F	43	26	N	—	Y	N	Y	na	140	49	148	229	N	N	N
TR7	R482Q	M	58	29	Y	46	Y	N	Y	9.5	171	36	483	252	N	N	N
TR8	R482Q	F	32	19	N	—	N	N	N	5.4	144	42	129	212	N	N	N
TR9	R482W	M	55	na	N	—	Y	N	Y	6.0	54	27	553	164	N	N	N
TR10	R482W	F	38	20	Y	22	Y	Y	Y	9.8	126	33	3216	367	N	Y	N
TR11	R482W	M	65	25	Y	40	N	N	1	5.5	179	47	138	254	N	N	N
TR12	R482W	F	7	16	N		N	N	na	na	49	51	74	115	N	N	N
TR13	R482W	F	32	22	Y	22	Y	N	Y	10.2	61	21	799	175	N	Y	N
TR14	R482W	M	11	20	N	—	N	N	N	na	51	49	97	119	N	N	N
TR15	R482W	M	50	26	Y	42	Y	N	Y	6.9	97	39	232	182	N	N	N
TR16	R482W	M	47	30	Y	45	Y	N	Y	5.3	125	49	162	181	N	Y	N
TR17	R482W	F	39	20	Y	24	Y	N	Y	9.0	105	27	342	176	N	Y	Yb
TR18	R482W	F	46	22	Y	26	Y	N	Y	8.6	129	21	544	238	N	Y	Yc
TR19	R482W	F	31	21	N	—	Y	N	Y	4.3	61	28	485	107	N	N	N
TR20	R482W	F	34	20	Y	27	Y	N	Y	5.3	125	29	485	166	N	Y	N
TR21	R482W	F	58	25	Y	51	Y	N	Y	8.0	na	30	1080	317	N	Y	N
TR22	R482W	F	53	26	Y	40	Y	N	Y	11.4	87	47	293	193	N	N	N
TR23	R482W	F	41	24	N	—	Y	N	Y	5.4	110	30	208	182	N	N	N
TR24	R482W	F	41	25	N	—	Y	N	Y	5.0	97	36	268	187	N	N	N
TR25	R482W	F	17	21	Y	12	Y	Y	Y	12.2	126	21	2993	419	N	Y	N
TR26	R482W	M	37	25	N	—	N	N	Y	5.7	92	44	128	162	N	N	N
TR27	R482W	F	11	16	N	—	N	N	na	5.6	82	59	66	154	N	Y	N
TR28	R482W	F	18	23	Y	17	N	N	Y	5.2	68	57	na	147	N	Y	N
TR29	R482W	F	29	26	Y	23	Y	N	Y	5.4	110	32	450	190	N	N	N
TR30	R482W	F	23	28	Y	18	Y	N	Y	6.4	81	27	134	132	N	N	N
TR31	R482W	M	54	29	Y	49	Y	N	N	6.2	112	38	195	184	N	N	N
TR32	R482W	F	41	22	Y	21	Y	N	Y	11	104	22	803	226	Y	Y	Yd
TR33	R482W	F	70	na	Y	30	Y	N	Y	na	na	na	na	na	N	N	Ye
TR34	R482W	M	75	27	Y	45	Y	N	Y	8.7	104	21	422	176	N	Y	Yf
TR35	R482W	M	57	28	Y	46	Y	N	Y	7.1	118	47	136	171	N	Y	Yg
TR36	R482W	F	42	24	Y	25	Y	N	Y	na	na	na	na	na	N	Y	Yh
TR37	R482W	F	40	na	Y	28	Y	N	Y	na	na	na	na	na	N	N	N

Abbreviations: BMI, body mass index; CoD, cause of death; F, female; HbA_1c, glycated hemoglobin A_1c; HDL, high-density lipoprotein; LDL, low-density lipoprotein; M, male; MI, myocardial infarction; N, not present; na, not available; PT, patient codes; TGs, triglycerides; Y, yes, present.

^a Coronary artery disease, catheter; ^bcardiomyopathy; ^ccoronary artery disease, catheter, MI; ^dcoronary artery disease, catheter, MI, died at age 41; ^eearly coronary disease, died at age 70 years, CoD: MI; ^fearly coronary disease, died at age 75 years, CoD: stroke; ^gearly coronary disease, died at age 58 years, CoD: coronary artery disease; ^hearly coronary disease, died at age 42 years, CoD: heart failure after MI.

R482W and R482Q Correlations of Our In-House Patients

Comparison of the clinical features of the 2 most common variants, R482W (n = 55) and R482Q (n = 36), in our in-house data demonstrated that patients carrying R482W were younger (age 31 [30] years vs 46 [26] years; P < .001), were of the same sex (females %, 75% vs 72%; P = .812), and had similar BMI levels (25 [6] vs 26 [8]; P = .08) in comparison to those harboring the R482Q variant.

The median levels of HbA_1c, cholesterol, TGs, and LDL were similar between the 2 groups (all values listed as R482W first followed by R482Q) (HbA_1c [6.3 (4) vs 7.2 (2) %, P = .785], cholesterol [182 (61) mg/dL vs 215 (75) mg/dL; P = .121], TGs [280 (183) mg/dL vs 251 (296) mg/dL; P = .543], and LDL [101 (55) mg/dL vs 111 (55) mg/dL; P = .334], and HDL [36 (15) vs 41 (11) mg/dL; P = .061]). Other than the ages at diagnosis of diabetes (28 [21] vs 44 [16] years; P = .002) and hepatic steatosis (34 [19] vs 45 [23] years; P = .012), which were earlier in patients carrying R482W variants, all accompanying clinical features that have been reported, including the history of pancreatitis attack, the presence of diabetes, fatty liver disease, dyslipidemia, hypertriglyceridemia, nephropathy, retinopathy, hypertension, abnormal cardiac examination, and the respective ages of diagnosis of the aforementioned complications, were similar between the patients carrying R482W and R482Q.

Genotype-Phenotype Correlations of 494 Patients

Combined analyses of our patients (n = 115) from 4 centers and all other cases (n = 379) curated from 95 studies in the literature were performed. Detailed clinical characteristics of all patients are presented in Supplementary Table S1 (128). The study included a total of 494 patients (females/males: 366/126) with a mean age of 33 (±17) years (median: 33; range, 1-77 years). The range of pathogenic variants was as follows: R482W (n = 190), R482Q (n = 93), R349W (n = 24), R527H (n = 20), T10I (n = 17), R62G (n = 9), S583L (n = 9), R60G (n = 8), R644C (n = 9), R582H (n = 9), T528M (n = 8), P4R (n = 7), T655FSX49 (n = 6), R582C (n = 5), L306V (n = 5), R28W (n = 4), R584H (n = 4), D47N (n = 4), N466D (n = 3), R471G (n = 3), R482L (n = 3), R527C (n = 3), c.1488 + 5G > C (n = 3), R133L (n = 3), A529V (n = 2), C591F (n = 2), D300N (n = 2), L59V (n = 2), L92F (n = 2), R439C (n = 2), R527P (n = 2), R571S (n = 2), S573L (n = 2), A529T (n = 1), C588R (n = 1), D136H (n = 1), D192V (n = 1), D230N (n = 1), D300H (n = 1), E145K (n = 1), E202K (n = 1), E262K (n = 1), G465D (n = 1), G535Q (n = 1), H506D (n = 1), K486E (n = 1), K486T (n = 1), K515E (n = 1), L387V (n = 1), L421P (n = 1), R399C (n = 1), R399H (n = 1), R453W (n = 1), R471C (n = 1), R541P (n = 1), R545H (n = 1), R582L (n = 1), P485R (n = 1), S395 (n = 1), and V440M (n = 1). Clinical characteristics of the common patients carrying the most common variants are presented in Table 4. Most patients were heterozygous for the changes (n = 464), with a minority presenting with either homozygous (n = 25) or compound heterozygous (n = 5) variants. Domain topology of the variants included in this analysis and associated phenotypes are presented in Fig. 2. BMI values of all variants included in this analysis are shown in Fig. 3. Circulating concentrations of leptin, adiponectin, and FGF-21 are presented in Fig. 4A–C.

Table 4.

Clinical characteristics of the patients harboring the most common variants. The results are presented as median (interquartile ranges) or as number of patients (percentages)

	All variants	R482W	R482Q	R349W	R527H	T10I	R62G	S583L	R582H	T528M	R60G	R644C	P4R	R582C
	n = 494	n = 190 (39%)	n = 93 (19%)	n = 24 (5%)	n = 20 (4%)	n = 17 (3%)	n = 9 (2%)	n = 9 (2%)	n = 9 (2%)	n = 8 (2%)	n = 8 (2%)	n = 8 (2%)	n = 7 (1%)	n = 5 (1%)
Common presenting phenotype		FPLD2	FPLD2	aFPLD	MAD	GLDS	FPLD2	FPLD2	aFPLD	aFPLD2	FPLD2	aFPLD2	APS	aFPLD
Sex (female %)	366/492 (74%)	134/190 (71%)	82/93 (88%)	15/24 (63%)	13/20 (65%)	8/17 (44%)	5/9 (55%)	6/9 (66%)	6/9 (86%)	7/8 (88%)	5/8 (63%)	4/8 (50%)	7/7 (100%)	4/5 (80%)
Age at diagnosis, y	33 (25) n = 471	31 (26) n = 185	44 (25) n = 93	30 (20) n = 24	15 (21) n = 20	14 (6) n = 17	38.5 (12) n = 8	40 (17) n = 9	33 (15) n = 7	44 (23) n = 8	26 (13) n = 8	39 (12) n = 7	25 (17) n = 7	22 (43) n = 5
BMI	23 (6) n = 347	24 (5) n = 134	25 (4) n = 78	19(3) n = 12	18 (9) n = 16	15 (2) n = 15	NA n = 1	27 (7) n = 8	26 (4) n = 3	24 (7) n = 7	NA n = 1	26 (6) n = 5	18(2) n = 4	22 (5) n = 4
Presence of additional features
Diabetes mellitus	241/395 (61%)	96/153 (63%)	42/64 (66%)	11/20 (55%)	1/7 (14%)	13/17 (72%)	6/9 (67%)	4/9 (44%)	4/7 (57%)	4/8 (50%)	4/8 (50%)	6/8 (75%)	1/7 (14%)	3/7 (43%)
Age at diagnosis, y	30 (22) n = 176	27 (19) n = 61	40 (17) n = 31	22.5(12.3) n = 8	NA	12 (6) n = 11	38 (24) n = 5	NA	32 (23) n = 4	44 n = 3	42(18) n = 4	44 (15) n = 6	NA	23 n = 3
Hypertriglyceridemia	300/392 (77%)	125/157 (80%)	56/73 (76%)	15/20 (75%)	4/15 (27%)	17/17 (100%)	3/3 (100%)	7/8 (88%)	3/3 (100%)	8/8 (100%)	2/2 (100%)	6/8 (75%)	2/6 (33%)	3/4 (75%)
Dyslipidemia	355/428 (83%)	142/165 (86%)	66/82 (80%)	17/24 (75%)	7/15 (47%)	17/17 (100%)	5/5 (100%)	7/7 (100%)	5/5 (100%)	8/8 (100%)	2/2 (100%)	7/8 (88%)	2/7 (29%)	4/5 (80%)
Age at diagnosis, y	26 (20) n = 111	28 (22) n = 48	33 (16) n = 20	20 (17) n = 8	NA n = 1	13.5 (6) n = 12	28 n = 2	NA	33 n = 3	32 n = 2	NA	NA n = 1	25.5 n = 2	12 n = 3
Pancreatitisp	39/494 (8%)	11/190 (6%)	6/92 (6.5%)	1/24 (4%)	—	5/17 (23.5%)	2/9 (22%)	—	1/7 (14%)	—	1/8 (13%)	1/8 (13%)	—	—
Age at diagnosis, y	20 (18) n = 23	26 (25) n = 4	24 (24) n = 6	NA n = 1	—	6 n = 3	NA n = 1	—	NA	—	NA	NA n = 1	—	—
Hepatic steatosis	149/172 (87%)	32/38 (84%)	28/38 (73%)	6/6 (100%)	NA	16/16 (100%)	1/1 (100%)	NA	1/1 (100%)	1/1 (100%)	1/1 (100%)	NA	1/1 (100%)	1/1 (100%)
Age at diagnosis, y	32 (25) n = 101	35 (21) n = 41	40 (21) n = 33	24.5 (11) n = 4	NA	15 n = 3	NA	NA	NA n = 1	NA n = 1	NA n = 1	NA	NA n = 1	NA n = 1

Abbreviations: aFPLD, atypical form of familial partial lipodystrophy; APS, atypical progeroid syndrome; BMI, body mass index; FPLD2, typical form of familial partial lipodystrophy, type 2; GLDS, generalized lipodystrophy syndrome; MAD, mandibuloacral dysplasia; NA, not applicable.

Figure 2.

Domain topology of the variants in LMNA and the association of the most common clinical phenotypes. FPLD2, familial partial lipodystrophy type 2; FSGS, focal segmental glomerulosclerosis.

Figure 3.

Body mass index (BMI) values of the patients harboring different LMNA variants.

Figure 4.

Concentrations of leptin, adiponectin, and fibroblast growth factor-21 (FGF-21).

R482W vs R482Q Variants

Comparison of available data of the clinical features of the 2 most common variants, R482W and R482Q, demonstrated that patients carrying R482W were predominantly female (134/190 vs 82/93, respectively; P < .001) and were younger (n = 185, 31 [26] years vs n = 93, 44 [23] years, respectively; P < .001) and had lower BMI levels (n = 134, 24 [5] vs n = 78, 25 [4], respectively; P = .037) in comparison to those harboring the R482Q variant at the cross-sectional evaluation point.

The median levels of HbA_1c, cholesterol, TGs, LDL, and HDL were similar between the 2 groups (all values listed as R482W first followed by R482Q) (HbA1c [n = 56, 6.8 (4) vs n = 37, 7.2 (3); P = .997], cholesterol [n = 92, 193 (70) mg/dL vs n = 65, 198 (67) mg/dL; P = .907], TGs [n = 107, 293 (290) mg/dL vs n = 66, 248 (259) mg/dL; P = .141], LDL [n = 88, 105 (54) mg/dL vs n = 57, 111 (51) mg/dL; P = .606], and HDL [n = 93, 37 (18) vs n = 63, 41 (13); P = .065 mg/dL]). Other than age at diagnosis of diabetes (n = 61, 27 [18] vs n = 31, 40 [17] years; P < .001) and hepatic steatosis (n = 41, 35 [21] vs 40 [21] years; P = .029), which were earlier in patients carrying R482W variants, all accompanying clinical features that have been reported, including the history of pancreatitis attack, the presence of diabetes, fatty liver disease, dyslipidemia, hypertriglyceridemia, nephropathy, retinopathy, hypertension, abnormal cardiac examination, and the respective ages of diagnosis of the aforementioned complications, were similar between the patients carrying R482W and R482Q.

Sexual Dimorphism

Among all patients, a higher percentage of female patients were reported to have diabetes (79% vs 21%; P = .012), hypertriglyceridemia (78% vs 22%, P = .019), and pancreatitis (90% vs 10%; P = .020). The distribution of dyslipidemia, hypertension, pancreatitis, hepatic steatosis, nephropathy, and abnormal cardiac examination were similar between the female and male patients. The median ages at lipodystrophy diagnosis, dyslipidemia, pancreatitis, and hepatic steatosis were similar between female and male patients. The levels of BMI, the concentrations of cholesterol, TGs, HDL, LDL, and percentage of HbA_1c were also similar between female and male patients.

Dyslipidemia

Dyslipidemia was the earliest biochemical evidence described in 354 (83%) of the 427 patients (277 females, 78%), diagnosed at a median age of 26 (20) years (range, 7-68 years). The median age and BMI of patients with dyslipidemia at the cross-sectional evaluation point was 35 (24) years (range, 1-75 years) and 24 (6) (range, 11-41). The 4 most frequent variants were R482W (n = 142), R482Q (N = 66), T10I (n = 17), and R349W (n = 19).

Hypertriglyceridemia was the most frequent cause of dyslipidemia (n = 291, 91%). TG concentrations corresponding to all the variants included in this analysis are presented in Fig. 5. Of all patients whose levels of lipids had been reported numerically, increased concentrations of TGs, HDL, cholesterol, LDL were identified in 77%, 77%, 55%, and 71% of all patients, respectively.

Figure 5.

Triglyceride concentrations and the association of all variants included in these analyses.

The concomitant diagnosis of dyslipidemia and hepatic steatosis was reported in 89% (n = 109/123), with diabetes present in 70% (n = 212/303), abnormal cardiac examination present in 22% (n = 77/354), pancreatitis present in 10% (n = 36/354), nephropathy present in 8% (n = 29/354), and retinopathy present in 3% (n = 11/354) of these cases. Those with dyslipidemia had 8.4 times the odds (95% CI, 1.1-62.3) ([37/317]/[1/72]) of developing acute pancreatitis compared to those without dyslipidemia, while those with hypertriglyceridemia had 12.0 times the odds (95% CI, 1.6-89.0) ([35/265]/[1/91]) compared to those without. Circulating TGs were negatively correlated with age at diabetes onset (Spearman rho [r_s] = −0.312; P = .001) and HDL levels ([r_s] = −0.557; P < .001), and positively correlated with HbA_1c ([r_s] = 0.310; P < .001) and cholesterol levels ([r_s] = 0.402; P < .001). There were no other statistically significant correlations between TG levels, age at lipodystrophy diagnosis, BMI values, and LDL concentrations. The patients with hypertriglyceridemia had higher BMI levels (23 [6] vs 22 [7]; P = .003) and lower HbA_1c concentrations (8 [3] vs 6 [1]%; P = .004) in comparison to those without.

Diabetes

Diabetes had been reported in 241 (61%) of 395 patients (191 females, 79%) at a median age onset of 30 (21) years (range, 6-79 years). Patients with diabetes had a median age of 39 (24) years (range, 4-77 years) with a BMI level of 23 (6) (range, 12–41) at the cross-sectional evaluation point. The median age at onset of dyslipidemia was significantly earlier than that of diabetes (n = 108, 26 [21] vs n = 179, 30 [22] years; P = .020). The most common variants identified among patients with diabetes were R482W (n = 96), R482Q (n = 42), T10I (n = 13), R349W (n = 11), R644C (n = 7), T655fsX49 (n = 6), R582H (n = 6), R62G (n = 5), T528M (n = 4), S583L (n = 4), and R60G (n = 4).

The concomitant presence of diabetes with dyslipidemia was reported in 95% (n = 213/224), hypertriglyceridemia in 91% (n = 181/198), hepatic steatosis in 93% (n = 92/99), nephropathy in 27% (n = 64/241), abnormal cardiac examination in 34% (n = 83/241), pancreatitis in 13% (n = 32/241), and retinopathy in 2% (n = 12/241) of all cases. The median ages of diagnosis of dyslipidemia, pancreatitis, and hepatic steatosis were 27.5 (19), 23 (20), and 34 (25) years, which had been reported in 70, 20, and 73 patients, respectively. Patients with diabetes were diagnosed at older ages (39 [24] vs 25 [23] years; P < .001) with higher BMI levels (23 [6] vs 22 [7]; P = .046) in comparison to those without diabetes. Patients with diabetes had significantly higher levels of HbA_1c (7.6 [3] vs 5.5 [1]; P < .001), TGs (350 [421] vs 209 [204] mg/dL; P < .001), cholesterol (195 [73] vs 178 [56] mg/dL; P = .008), and lower concentrations of HDL (36 [13] vs 41 [10] mg/dL; P < .001) in comparison to those without diabetes, while the levels of LDL were similar between the 2 groups. Patients who were reported to have diabetes had 3.2 times the odds (95% CI, 1.4-7.5) ([32/209]/[7/147]) of having pancreatitis compared to those who did not have diabetes.

Pancreatitis

Among 39 patients (female: 35, 90%) with an episode of acute pancreatitis history, the median age at acute pancreatitis diagnosis was 20 (17) years (range, 5-48 years). The distribution of all the patients with a history of pancreatitis included in this analysis is shown in Fig. 6. Patients had a median age of 29 (18) years (range, 4-55 years) and BMI of 24 (9) (range, 11–39) at the cross-sectional evaluation point. Patients with a history of acute pancreatitis harbored R482W (n = 11), R482Q (n = 6), T10I (n = 4), c.1488 + 5G > C (n = 3), D47N (n = 2), D300N (n = 2), R582H (n = 2), R62G (n = 2), D136H (n = 1), G465D (n = 1), L306V (n = 1), R349W (n = 1), R584H (n = 1), R60G (n = 1), and R644C (n = 1) variants (see Table 1). Ten (26%) patients were reported to have multiple pancreatitis ( > 1) episodes. Evaluations of patients with a history of acute pancreatitis revealed that dyslipidemia was present in 97% (n = 37/39), hypertriglyceridemia in 90% (n = 35/39), hepatic steatosis in 90% (n = 23/25), diabetes in 83% (n = 32/39), abnormal cardiac examination in 46% (n = 11/39), nephropathy in 28% (n = 11/39), and retinopathy in 10% (n = 4/39) of all patients.

Figure 6.

The distribution of all patients with a history of acute pancreatitis. The percentages shown demonstrate the occurrence of pancreatitis for each variant.

Other Clinical Characteristics and Mortality

Evidence of hypertension was reported in 21% (n = 105) of patients. The median age of diagnosis of hypertension was 24 (17) years (range, 10-71 years). Among patients who had been evaluated with liver ultrasound or biopsy, 87% (n = 149/172) were found to have hepatic steatosis. The median age at diagnosis was 33 (25) years (n = 105; range, 6-65 years). Abnormal cardiac examination was reported in 26% (n = 130) of all patients. Early coronary heart disease (n = 30), valvular abnormalities (n = 30), ventricular hypertrophy (right or left) (n = 27), conduction (n = 24), and dilated cardiomyopathy (n = 23) were among the most common cardiac problems. Whereas 32% (n = 41) of cases with cardiac defects were found to have variants in exon 8, the majority of cases (68%) had defects affecting regions outside exon 8 (P < .001). Defects outside exon 8 were most frequently found in exon 1 (38%; n = 34/52), exon 6 (25%; n = 22/25), and exon 11 (11%; n = 10/24). Evidence of nephropathy and retinopathy were reported in 15% (n = 76) and 3% (n = 15) of all patients. While mild proteinuria as a complication of diabetes was the most common evidence of nephropathy reported in most studies, there were also reports on cases with chronic renal disease (n = 5) and focal segmental glomerulosclerosis (n = 9).

Skeletal abnormalities, including clavicular hypoplasia, scoliosis, acro-osteolysis, mandibular hypoplasia, short stature, joint contractures, micrognathia, evident shoulders, thin clavicles, vertebral sclerosis, and dental crowding were mostly reported among patients carrying the R527H variant. Other variants presenting with skeletal abnormalities were as follows: T10I (eg, joint contractures, spinal deformity, mandibular hypoplasia), R60G (eg, contractures), K486E (eg, mandibular hypoplasia), R527C (eg, short stature, joint contractures), R571S (eg, short stature, clinodactyly, joint contractures), A529V (eg, short statures, dysmorphic extremities, mandibular hypoplasia, acro-osteolysis), and A529T (eg, short stature, clavicle aplasia).

Based on our analysis, personalized treatment strategies mostly focusing on treating the complications of lipodystrophies, including diabetes, pancreatitis, hypertriglyceridemia, hypertension, and cardiac failure, were followed. Twenty-five percent (n = 122) of all patients were reported to be treated with combination treatments such as insulin, metformin, fibrates, statins, angiotensin-converting enzyme inhibitors, diuretics, and pioglitazone. Treatment with metreleptin was also well tolerated and resulted in improvements of the complications in several cohorts (21, 23, 42, 104, 107, 112, 117)

Cardiovascular disease was the leading cause of mortality occurring at a median age of 43 (20) years (n = 16; range, 7-75 years). The following causes of death were reported: paroxysmal arrhythmia (n = 1), myocardial infarction (n = 5), heart failure (n = 4), graft failure from heart transplant (n = 1), hemorrhagic stroke (n = 2), bacterial endocarditis (n = 1), sepsis due to foot infection (n = 1), cerebral hemorrhage after hemodialysis (n = 1), and other (n = 1).

Discussion

This study reports the distribution of various clinical and metabolic characteristics, including dyslipidemia, diabetes, pancreatitis, hypertension, cardiac pathology, nephropathy, and hepatic steatosis, in the largest number of patients with LMNA-related lipodystrophy syndromes to date, using clinical cases reported in more than 100 studies in addition to our cohort of 115 patients, some of whom we had been following for more than 20 years. The results confirm that dyslipidemia precedes the onset of diabetes. As each rare variant may present with different clinical features, this study demonstrates that the most common variants, R482W and R482Q, do not exhibit distinct metabolic characteristics.

Consistent with earlier publications, we observed disparities among clinical phenotypes, with certain variants exhibiting a milder phenotype and others having a more aggressive clinical course (4, 27). Several authors have tried to derive specific phenotypic associations with respect to lamin domain topology. As an example, the pathogenic variants affecting only the structure of exon 11 (protein domains 566-656), which is not transcribed in lamin C, as opposed to exon 8, had initially been held accountable for the milder phenotypes; however, the patients with homozygosity in exon 11 or variants, such as R582H, have blurred this hypothesis (27, 61, 65, 84, 102). Further, those pathogenic variants located in the amino-terminal (5′-end of the gene) to the nuclear localization signal (NLS, 417-422) (exon 7; protein domains 386-466) of lamin A had been initially classified as class 1 laminopathies, while those carboxy-terminal (3′-end of the gene) to the NLS were clustered as class 2 laminopathies (129). However, the presumption that the class 1 laminopathies would affect the cardiac, skeletal muscle, and neurological system, whereas class 2 would cause partial lipodystrophy or progeria syndromes, has been proven otherwise through analyses of the R28W, D230N, R399C, and R471G variants (31, 35, 40, 41). Our study showed that patients with the T10I variant, who typically presented with generalized lipodystrophy, exhibited the most severe complications. All patients with T10I were diagnosed with dyslipidemia, with the greatest circulating TGs compared to other variants. Patients carrying the R582H variant, which is related to mandibuloacral dysplasia, and the P4R variant, which is associated with atypical progeroid syndromes, showed a milder phenotype than those carrying the R482Q, R482W, R349W, and R62G variants, which most commonly present with the FPLD2 phenotype. The most common variants in our study, R482Q and R482W, were associated with similar metabolic characteristics and complications, despite the fact that those with the R482W variant were younger and had statistically lower BMI levels at the time they were evaluated. Given that we had more cases of the R482W variant, it is possible that the efforts to expand family studies to investigate younger family members may have skewed the data collected. To state that R482W is representative of a more severe clinical phenotype would require age-adjusted studies looking at hard end points such as mortality, and organ-specific morbidity (diabetes complications, liver failure, cardiovascular disease, etc). The classification of these patients with respect to different variants may serve as a map for those interested in the field to better understand the associations and develop individualized approaches for all the distinctions in these overlapping phenotypes.

Our study showed that the median ages of diagnosis of lipodystrophy and several metabolic complications were comparable between females and males with LMNA-related lipodystrophy syndromes. This finding should be interpreted carefully, as it does not necessarily contradict that females with FPLD2 are diagnosed at an earlier age than males, since our study included many types of lipodystrophy syndromes including those with progeroid forms (4). Although this finding does not cover the whole lifespan, similar ages at onset of complications for both sexes support the notion that age correlates with phenotype (2).

Dyslipidemia and diabetes, the 2 most common metabolic abnormalities in patients with lipodystrophy syndromes, were seen frequently among all LMNA variants (3). While previous studies have reported a prevalence of dyslipidemia ranging from 59% to 100% in patients with FPLD2 (4, 37, 130, 131), our study showed that 83% of all patients had dyslipidemia, which was also comparable in the most common variants, R482W and R482Q, which were 86% and 80%, respectively. It was still interesting to note the totality of the distribution width in each variant, which likely represents contributions not only from the LMNA variants themselves, but also from additional genetic, epigenetic, and environmental factors.

Consistent with previous studies indicating that the lipoprotein abnormalities initially contribute to the development of diabetes, we also found that the age at dyslipidemia onset was earlier than that of diabetes (15, 29, 132). Hypertriglyceridemia, which we have previously demonstrated to have a positive correlation with liver lipid content and a negative association with leg fat mass (13), was the most common lipid abnormality among patients with dyslipidemia carrying different variants. While the consensus in 2018 by Grundy et al (133) recommends adding therapeutic modalities at specified cutoff levels for each lipid subfraction, it is important to note that normal mean levels of different lipid subfractions have been reported to be much lower, depending strictly on a variety of factors, including age, race, ethnicity, and secular trends. It is also worth mentioning that Aberra et al (134) have reported that TG concentrations even in the reference ranges may be associated with increased cardiovascular risks. Similarly, despite the recommendations on initiating TG-lowering treatment at levels greater than or equal to 500 mg/dL, mild-to-moderate hypertriglyceridemia (>177 mg/dL) has also been shown to be associated with high risk of acute pancreatitis in a prospective cohort study by Pedersen et al (135). The strong association of high cholesterol, LDL, and TGs, and low HDL levels with the development of atherosclerosis and cardiovascular disease, the presence of cardiac complications, and the risk of acute pancreatitis in lipodystrophy syndromes highlight the importance of strict lipid control at the earliest possible point in the clinical care of patients afflicted with these diseases (124).

Diabetes, the second most prevalent characteristic in our study, has been reported to range in prevalence from 28% to 77% among patients with FPLD2 (4, 130, 131, 136). According to our findings, the prevalence of diabetes was 61% in all patients with LMNA-related lipodystrophy syndromes, as well as among those harboring the most common R482W variant. It is interesting to note the discrepancies in cause-and-effect relationships among patients of similar ages who have different metabolic derangements. Although there was a positive correlation between TGs and HbA_1c levels, those with diabetes had even higher TG concentrations. This result may be anticipated, since both are independent risk factors for the development of the other (137). However, along with the earlier onset of dyslipidemia, this distinction may also suggest that hypertriglyceridemia is a risk factor for the development of diabetes, and it gets worse once diabetes has been established (132).

Severe atherosclerotic cardiovascular disease and acute pancreatitis have been reported to be the 2 most common causes of mortality and morbidity in patients with lipodystrophy syndromes (138, 139). Comparable to a previous study (140) reporting the overall prevalence of pancreatitis as 4% (140) among a cohort of 74 patients predominantly carrying R482W variants (n = 51, 69%), we found that the overall prevalence of pancreatitis was 8%, while the patients with R482W and R482Q variants had a lower prevalence of 6%. The difference is presumably due to the pathogenic variants associated with the generalized type, which we also included in our study for comparative purposes, which are associated with more severe complications. Although the age of onset of acute pancreatitis was found to be younger in comparison to dyslipidemia, it is well known that dyslipidemia precedes the onset of acute pancreatitis in lipodystrophy syndromes. This contradictory finding has presumably arisen due to the symptomatic nature of acute pancreatitis, and possible data skewness of higher number of patients with dyslipidemia. However, one patient had normal lipid levels despite pancreatitis, highlighting that acute pancreatitis may in some instances precede dyslipidemia, and since some biological lipid abnormalities may arise without biochemical evidence of dyslipidemia (135). Further, pancreatic acinar cell–specific LMNA-deletion in mice has shown that these animals develop histological abnormalities comparable to chronic pancreatitis due to changes in p53 signaling, cell cycle, and basal transcription factors, which collectively lead to cell-autonomous endoplasmic reticulum stress with loss of acinar cells of the pancreas (141). In line with previous reports (13, 55, 140) demonstrating an increased risk of pancreatitis with hypertriglyceridemia in FPLD2, we also found that the risk of acute pancreatitis increased in the presence of diabetes and dyslipidemia. The patients with hypertriglyceridemia had 12.0 times the odds of having acute pancreatitis in comparison to those who did not, while those with diabetes had 3.2 times the odds.

Cardiovascular events were the most prevalent cause associated with mortality. We have previously shown that 78% of all patients with lipodystrophy may have evidence of a severe cardiac pathology (110). In line with the literature, this study showed that cardiac defects due to variants in LMNA, which is associated with poor prognosis and higher rates of sudden cardiac death, were more common in patients carrying variants outside exon 8 (18). As several prediction models underpin the changes in the protein structure of lamins with respect to the subdomains and hotspot mutations, the exact underlying mechanisms still need further research (18, 142).

Although the abnormal cardiac observations reported among the patients carrying the R482W variant outnumbered the cases seen among all other variants, its prevalence was similar to that of R482Q, the second most frequent variant. Hence, this higher number could just indicate a possible reporting bias, since R482W is the most frequent variant overall. Although the prevalence of cardiovascular abnormalities in our study was not particularly high, significant valvular, conduction, or morphological abnormalities were found in most patients who were able to undergo a comprehensive cardiac examination. These findings support the need for routine cardiac monitoring, in accordance with our previous recommendations (110).

Despite serious metabolic complications, most patients with LMNA variants had similarly low or normal BMI values in comparison to the healthy population, though the distribution did include some individuals with BMI levels as high as in the 40s. We have previously demonstrated this particular discordance between BMI and metabolic dysfunction, which in our view may be related to the distinct distribution of adiposity observed in lipodystrophy syndromes (127). Loss of lower-extremity adipose tissue, which is very sensitive to insulin, contributes to excess free fatty acids in circulation, which may then lead to the deterioration of cardiovascular and metabolic health (14, 143, 144). Further, the loss of brown adipose tissue characteristics is inversely correlated with BMI (145). Brown adipose tissue is predominantly found in the neck and upper body, resonating the distribution pattern of preserved depots in FPLD2, and thus may contribute to metabolic dysfunction. We believe that the limited storage capacity conferred by reduced overall adipose tissue and the altered characteristics of the residual adipose depots are deserving of additional studies.

It is also important to note that the adipose tissue deficiency and perturbations in circulating factors coming out the adipose tissue per se cause systemic metabolic effects. These are well-established observations given the overlapping clinical phenotypes in all forms of lipodystrophy, classical studies of fat transplantation in rodent models, and studies of leptin therapy in patients as well as animal models (21, 115, 146). While it is known that leptin levels are quite low in patients with the LMNA-related lipodystrophy from previous reports and the data compiled here, the fact that leptin therapy is not universally effective in this subgroup suggests that molecular mechanisms leading to the adipocyte loss in some depots and alterations in the preserved depots may be additional important contributors to overall pathophysiology (13, 30, 49, 112, 113, 147, 148). Other circulating factors that have been evaluated here are adiponectin and FGF-21. Interestingly, adiponectin levels were quite varied according to genotype and were not universally low as previously implicated in singular reports (104, 149-151). On the other hand, FGF21 levels were not low and surprisingly elevated in some genotypes, suggesting a potential compensatory role for this hormone (152, 153).

Understanding how specific LMNA gene variants alter or impair lamin A/C function is critical to investigate mechanisms of lipodystrophy syndromes and to reveal therapeutic targets. Regarding mechanisms leading to the diverse pathologies observed in different laminopathy syndromes, several theories exist. Specific cell types may be more sensitive to perturbed lamin A/C function, which may explain why cells like adipocytes, cardiomyocytes, and myocytes are predominantly affected by laminopathies. The nuclear lamina maintains the shape of the nuclear membrane, is associated with chromatin, and interacts with other intracellular intermediate filament and transmembrane nuclear proteins to coordinate movement and shape change in response to external and internal stimuli. Due to the association of lamin A/C with chromatin, lamin A/C plays a key role in nuclear location and accessibility of transcribed genes, some of which may exert control over the differentiation and maintenance of different cell types. Lamin A/C also maintains shape and stability for the nucleus while undergoing cell division and the overall cell structure under plastic stressors, like mechanical stress. Our laboratories are currently investigating the mechanisms by which adipocyte loss occurs in different LMNA-related lipodystrophy phenotypes. We have demonstrated that LMNA is essential to the maintenance of mature adipocytes by creating an adipocyte-specific LMNA knockout mouse with Cre recombinase activity driven by the adiponectin promoter (154). These mice recapitulate many of the features observed in FPLD2 clinically, including adipose tissue loss coincident with puberty, dyslipidemia, hepatic steatosis, and insulin resistance. Ongoing work in the laboratory includes development of tamoxifen-inducible conditional LMNA knockouts, which allows us to examine the earliest molecular events after lamin A/C deletion that lead to adipocyte loss. We are also actively working on the creation of a topographic map of lamin-associated domains in chromatin, which may reveal key genes involved in adipocyte maintenance. In addition, 6 different human variants have been recreated in mice through knockin-mutant murine models. Finally, we are examining adipose tissue biopsies of affected teenagers before and after adipose loss at single-cell nuclei resolution to correlate the changes we observe in the conditional knockout models. We believe that these approaches will allow us to determine the precise etiology of adipose loss in patients and rodent models and venture into precision therapeutics.

Study Limitations

Our study was limited by the retrospective nature of our data and the potential reporting bias of the previous 100 published reports. Our analyses could include only those cases with detailed clinical characteristics. We did not include patient cohorts that lacked the genotype-phenotype characteristics per individual. Although the presence of clinical features related to disease was noted, the ages at which certain clinical features were diagnosed were not included in most of the studies either. For this report, we did not focus on clinical characteristics such as pancreatic exocrine function or morphology, smooth muscle function, neurocognitive characteristics, mood, pain perception, and other mental diseases or reproductive function. However, we are keenly aware that such characteristics need to be investigated in future studies either because there is a link to lamin A/C function specifically at a tissue level or abnormalities have been reported in a limited fashion in case series. Finally, it is important to note the differences between different countries’ access to health care when interpreting the results. Even regarding our own cohort, the parameters such as the ages at diagnosis or onset of certain complications completely depend on the information collected from the patients. Hence, we believe that although it is important to describe the comparisons, strong statements on genotype-phenotype correlations should be avoided.

Conclusion

This study identifies genotype-phenotype relationships in the largest cohort of LMNA-related lipodystrophy syndrome patients to date. This study will serve as a comprehensive directory for obtaining a deeper understanding of the clinical characteristics and underlying mechanisms of this rare disorder. Our report helps to quantify prevalence of rare complications associated with different LMNA variants. Dyslipidemia was the most common and earliest metabolic dysfunction, affecting 83% of all patients, leading to the earliest symptomatic and significant clinical manifestation, acute pancreatitis, which was diagnosed at a median age of 20 years. The patients with hypertriglyceridemia had 12.0 times, while those with diabetes had 3.2 times, the odds of having acute pancreatitis in comparison to those who did not. Even though patients carrying different variants had varying features, most of the clinical and metabolic features between the patients harboring the most common variants, R482W and R482Q, were indistinguishable for the variables we interrogated for this analysis.

Abbreviations

BMI

body mass index

FGF-21

fibroblast growth factor-21

FMR

ratio of percentage fat mass of trunk to percentage fat mass of legs

FPLD2

familial partial lipodystrophy type 2

HbA_1c

glycated hemoglobin A_1c

HDL

high-density lipoprotein

LDL

low-density lipoprotein

LMNA

lamin A/C gene

MI

myocardial infarction

TGs

triglycerides

Funding

This work was supported by the National Institutes of Health (National Institute of Diabetes and Digestive and Kidney Diseases grant number 2R56DK125513-04).

Author Contributions

E.A.O. conceived the idea for this work and funded team members. She also designed the studies in which cases are collected and has collected cases and their data. She assisted with the initial draft and reviewed all drafts including the final one. E.A.O. assumes full responsibility for the integrity of the data. O.B. collected data, performed statistical analyses, interpreted the data, and wrote the initial draft. M.C.F.F., N.R.G., M.C.G., D.G., A.D.G., I.Y.S., and B.A. participated in patient care and data collection. J.N.M., R.L.S., K.S.H., B.O., K.D., B.A., and O.A.M. edited and revised the manuscript critically. All authors read and approved the final manuscript.

Disclosures

E.A.O. has intellectual property related to the use of metreleptin and related compounds in patients with all forms of lipodystrophy. She has provided consulting services to Ionis and Akcea in the past 3 years. She is actively providing consulting services to Amryt Pharmaceuticals, Regeneron Pharmaceuticals Inc, and Third Rock Ventures. She has received either grant or clinical trial support from Amryt Pharmaceuticals and Ionis related to the lipodystrophy space. She has previously received writing support or educational grants from Amryt Pharmaceuticals and Regeneron Pharmaceuticals, Inc, as well as Akcea Therapeutics. Her current clinical trial portfolio also includes support from Rhythm Pharmaceuticals, GI Dynamics, Novo Nordisk, Fractyl (current), and Gemphire Therapeutics (in the last 3 years). She is on the scientific advisory board of Rejuvenate Bio and has ongoing grant support. O.A.M. has received grant support from Regeneron Pharmaceuticals, Inc, CombiGene AB, and Rejuvenate Bio. J.N.M. and R.S. have received grant support from Regeneron Pharmaceuticals Inc, Rejuvenate Bio, and CombiGene. B.A. receives consulting fees from Amryt Pharmaceuticals and Regeneron Pharmaceuticals, Inc. The remaining authors have nothing to disclose with respect to this manuscript.

Data Availability

Original data generated and analyzed during this study are included in the data repositories listed in “References.”