Journal of Clinical Lipidology xxx (xxxx) xxx

Contents lists available at ScienceDirect

Journal of Clinical Lipidology

journal homepage: www.elsevier.com/locate/jacl

Identification and functional analysis of novel homozygous LMF1 variants

in severe hypertriglyceridemia
Candy Bedoya, PhD a , Rishi Thomas, DO a , Anna Bjarvin, BSc a , Wilbur Ji, DO a ,
Hanien Samara, BSc a , Jody Tai, DO a , Laurie Green, BSc b, Philip H. Frost, MD b,
Mary J. Malloy, MD b, Clive R. Pullinger, PhD b, John P. Kane, MD, PhD b, Miklós Péterfy, PhD a,*
a
Department of Biomedical Sciences, Western University of Health Sciences, Pomona, CA, USA
b
Cardiovascular Research Institute, University of California San Francisco, San Francisco, CA, USA

A R T I C L E I N F O A B S T R A C T

Keywords: Background: The genetic basis of hypertriglyceridemia (HTG) is complex and includes variants in Lipase Matu-
Hypertriglyceridemia ration Factor 1 (LMF1), an endoplasmic reticulum (ER)-chaperone involved in the post-translational activation of
Lipase maturation factor 1 lipoprotein lipase (LPL).
Familial chylomicronemia syndrome
Objective: The objective of this study was to identify and functionally characterize biallelic LMF1 variants in
Multifactorial chylomicronemia syndrome
patients with HTG.
Methods: Genomic DNA sequencing was used to identify biallelic LMF1 variants in HTG patients without dele-
terious variants in LPL, apolipoprotein C-II (APOC2), glycosylphosphatidylinositol-anchored high-density lipoprotein
binding protein 1 (GPIHBP1) or apolipoprotein A-V (APOA5). LMF1 variants were functionally evaluated by in silico
analyses and assessing their impact on LPL activity, LMF1 protein expression and specific activity in transiently
transfected HEK293 cells.
Results: We identified four homozygous LMF1 variants in patients with severe HTG: two novel rare variants (p.
Asn147Lys and p.Pro246Arg) and two low-frequency variants (p.Arg354Trp and p.Arg364Gln) previously re-
ported at heterozygosity. We demonstrate that all four variants reduce the secretion of enzymatically active LPL
by impairing the specific activity of LMF1, whereas p.Asn147Lys also diminishes LMF1 protein expression.
Conclusion: This study extends the role of LMF1 as a genetic determinant in severe HTG and demonstrates that
rare and low-frequency LMF1 variants can underlie this condition through distinct molecular mechanisms. The
clinical phenotype of patients affected by partial loss of LMF1 function is consistent with Multifactorial Chylo-
micronemia Syndrome (MCS) and suggests that secondary factors and additional genetic determinants contribute
to HTG in these subjects.

Introduction metabolic diseases (obesity, type 2 diabetes), excessive alcohol con-

sumption, hypothyroidism, pregnancy and certain medications.2 Clin-
Severe hypertriglyceridemia (HTG), defined as plasma TG concen- ical features of MCS include abdominal pain, hepatosplenomegaly and
trations exceeding 885 mg/dL (10 mmol/L), is a relatively common (~1 atherosclerotic cardiovascular disease. Patients with MCS generally
in 600) form of dyslipidemia associated with impaired LPL-mediated respond well to lifestyle modifications that mitigate secondary factors
catabolism of TG-rich lipoproteins including chylomicrons and very and therapeutic interventions. In contrast, severe HTG in a small subset
low-density lipoproteins (VLDL).1 In most patients, severe HTG is the (~1 %) of patients is associated with familial chylomicronemia syn-
result of multifactorial chylomicronemia syndrome (MCS), a complex drome (FCS), a monogenic disorder that is influenced by secondary
disease involving contributions from genetic and secondary (i.e. factors to a much lesser extent than MCS. FCS is characterized by clinical
non-genetic) factors. The genetic component of MCS is characterized by presentation early in life, abdominal pain, hepatosplenomegaly, erup-
polygenic determinants, which provide predisposition to dyslipidemia. tive xanthomas, acute pancreatitis and a more modest response to life-
Secondary factors required to precipitate the condition include style changes or therapy.

* Corresponding author at: Western University of Health Sciences, Department of Biomedical Sciences, 309 E Second St., Pomona, CA 91766, USA.
E-mail address: [email protected] (M. Péterfy).

https://doi.org/10.1016/j.jacl.2024.10.004
Received 25 March 2024; Accepted 13 October 2024
Available online 19 October 2024
1933-2874/© 2024 National Lipid Association. Published by Elsevier Inc. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/4.0/).

Please cite this article as: Candy Bedoya et al., Journal of Clinical Lipidology, https://doi.org/10.1016/j.jacl.2024.10.004
C. Bedoya et al. Journal of Clinical Lipidology xxx (xxxx) xxx

The genetics of severe HTG is complex and reflects the contrasting reported here harbored deleterious coding variants in LPL, APOC2,
etiologies of FCS and MCS. FCS results from rare, highly penetrant GPIHBP1 or APOA5. As copy number variation (CNV) was not assessed,
biallelic (i.e. homozygous and compound heterozygous) loss-of-function we cannot exclude the possibility of CNV variation in canonical genes
coding-variants in canonical genes involved in plasma TG catabolism.1 contributing to the HTG phenotype. Single nucleotide variants (SNVs)
Over 95 % of FCS cases are due to defects in LPL, but variants in other leading to amino acid changes in LMF1 were checked against the dbSNP
genes regulating the lipolytic activity (APOC2, APOA5), database (https://www.ncbi.nlm.nih.gov/snp) and the Genome Aggre-
trans-endothelial transport (GPIHBP1) and biosynthesis (LMF1) of LPL gation Database (gnomAD) (https://gnomad.broadinstitute.org) to
have also been identified.3 The genetic etiology of MCS is heterogeneous identify rs numbers and assess global minor allele frequencies (MAFs).
and involves either the presence of rare heterozygous loss-of-function The potential impact of missense variants on LMF1 function was
coding-variants of variable penetrance in one of the canonical genes analyzed by the Protein Variation Effect Analyzer (PROVEAN; http
or the accumulation of multiple common single nucleotide variants ://provean.jcvi.org), Sorting Intolerant From Tolerant (SIFT; htt
(SNVs) of individually low penetrance.1 Given the contrasting genetic ps://sift.bii.a-star.edu.sg) and Polymorphism Phenotyping v2 (Poly-
architecture of FCS and MCS, genetic analysis plays an important role in Phen-2; http://genetics.bwh.harvard.edu/pph2) algorithms.
the differential diagnosis between these conditions.4
Lipase maturation factor 1 (LMF1) is an ER-bound chaperone Biochemical analyses
required for the post-translational maturation of newly synthesized LPL
as well as hepatic lipase (HL) polypeptides into catalytically active en- Plasma lipid concentrations and post-heparin lipase activities were
zymes.5 First described as the gene defective in a naturally occurring measured as described previously.6,18 Sequential ultracentrifugation
mouse model of severe HTG, LMF1 has since been implicated in human was used for density-based separation of lipoproteins, including chylo-
HTG through the identification of a handful of rare biallelic micron and very low-density lipoprotein (CM/VLDL, d < 1.006 g/mL),
loss-of-function variants in such patients.6 Best documented among low-density lipoprotein (LDL, 1.019 < d < 1.063 g/mL) and high-density
these are the p.Tyr439Ter, p.Trp464Ter, p.Gly172Arg, p.Arg53- lipoprotein (HDL, 1.063 < d < 1.21 g/mL).19
GlyfsTer5 and p.Ser137Leu alleles, which impair the lipase-maturation
activity of LMF1.6–9 Although functionally uncharacterized, the Functional analysis of LMF1 variants
recently reported homozygous p.Trp460Ter, p.Arg233Ter and p.
(Thr244_Gln299del) variants are also likely to be causative for HTG, as LMF1 variants were functionally analyzed using an in vitro assay
they result in major structural changes and may reduce the expression based on the reconstitution of LPL maturation in LMF1-deficient
and/or activity of the LMF1 protein.10–12 Additional missense LMF1 HEK293 cells, essentially as described previously (Fig. 1).9 Briefly,
variants (p.Ser14Trp, p.Pro86Leu, p.Pro248Ser, p.Thr395Ile, p. cells were co-transfected in quintuplicate with a mixture of plasmids
Arg451Trp) have also been uncovered in patients with HTG, but their expressing human LPL, wild-type (WT) or mutant LMF1 and firefly
clinical relevance remains undetermined in the absence of functional luciferase (FFluc) for normalization. For the quantitative assessment of
evaluation.13–15 LMF1 protein expression, LMF1 variants were expressed as N-terminal
Although monogenic LMF1 deficiency represents a very rare genetic fusion proteins with Gaussia luciferase (Gluc-LMF1). Twentyfour hours
cause of severe HTG, the identification of patients with this molecular after transfection, fresh heparin-containing media (10 U/ml) were
etiology has clinical relevance. First, it may allow refined phenotyping added to cells and harvested for fluorometric LPL activity assay (Cell
and the discovery of unique clinical features in FCS due to LMF1 defi- Biolabs, USA) 5 h later. FFluc and Gluc activities were measured in cell
ciency. Furthermore, the identification of LMF1-deficient patients in lysates using the Luc-Pair™ Duo-Luciferase HS Assay Kit (GeneCopoeia,
sufficient numbers may allow rigorous testing of therapeutic approaches USA). LMF1 activity was defined as LPL activity in medium normalized
of HTG in this group. Indeed, only 1 of 66 FCS patients harbored a ho- by FFluc activity in cell lysates and LMF1 protein expression was
mozygous rare variant in LMF1 in a clinical trial assessing the efficacy of calculated as FFluc-normalized Gluc activity (Fig. 1). LMF1 specific
APOC3 inhibition, a promising therapeutic approach in FCS.16 Finally, activity was calculated as the ratio of medium LPL activity and intra-
the identification of novel LMF1 variants may lead to insights into cellular LMF1 protein mass as determined by Gluc activity assays. In
structure-function relationship and a better understanding of the mo- each experiment, quintuplicate replicates have been averaged and the
lecular function of this protein. With these objectives in mind, the goal of results are presented as the mean ± SEM of three independent
the present study was to investigate the contribution of biallelic LMF1 experiments.
deficiency in a large cohort of dyslipidemic patients and extend the
LMF1 allelic spectrum responsible for severe HTG in the population. Statistical analyses

Materials and methods Data are presented as mean ± standard error of the mean (SEM).
Statistical analyses were performed with the SigmaPlot 11.0 software.
Study subjects ANOVA was used to compare groups and Holm-Sidak post-hoc tests
were employed to assess significant differences from the control group.
Hypertriglyceridemic subjects with fasting triglyceride levels above
400 mg/dL were identified in the Genomic Resource in Arteriosclerosis Results
and Metabolic Disease at the Cardiovascular Research Institute, Uni-
versity of California, San Francisco (UCSF). All subjects gave informed Case reports
consent and the study was approved by the UCSF Institutional Review
Board. To investigate the role of LMF1 in the etiology of HTG, we sequenced
all coding exons of the LMF1 gene in 386 patients (Caucasian 69 %,
Genomic DNA analysis and variant annotation Hispanic 12 %, East Asian 12 %, African American 2 %, South Asian 1 %,
Native American 1 %, mixed ethnicity 3 %) with fasting plasma TG
Genomic DNA was isolated as previously described.17 Exons and levels over 400 mg/dL. This effort led to the identification of the first
exon-intron boundaries in the LMF1, LPL, APOC2, GPIHBP1, APOE and homozygous loss-of-function LMF1 variant (p.Tyr439Ter) associated
APOA5 genes were PCR amplified (primer sequences available upon with severe HTG and has been described previously.6 Here, we report 5
request) and PCR products were sequenced using an automated DNA additional patients from the same cohort harboring novel homozygous
sequencer ABI 377XL (Applied Biosystems). None of the patients missense variants in LMF1 (Table 1).

2
C. Bedoya et al. Journal of Clinical Lipidology xxx (xxxx) xxx

Fig. 1. Schematic diagram of the functional assay of LMF1 lipase-maturation activity. LMF1-deficient cells are cotransfected with vectors expressing Gaussia
luciferase-LMF1 (Gluc-LMF1) fusion proteins, LPL and firefly luciferase (FFluc) for normalization. Two days after transfection, LPL and luciferase activities were
assessed in the culture media and cell lysates, respectively.

Table 1
Clinical characteristics of hypertriglyceridemic patients with homozygous mutations in LMF1.
Patient 1 Patient 2 Patient 3 Patient 4 Patient 5

Age at first presentation ​ 29 33 11 65 72

(year)
Sex ​ female male male male female
Ethnicity ​ Hispanic Hispanic Hispanic Caucasian Caucasian
BMI (kg/m2) ​ 25.8 21.9 24.2 29.1 na
Type 2 diabetes ​ yes no no yes yes
Cardiovascular disease ​ yes no no yes yes
Response to lifestyle/ ​ yes yes yes unknown unknown
therapy
APOE genotype ​ E4/E4 na E3/E3 E3/E3 E3/E3
Triglycerides: ​ ​ ​ ​ ​ ​
Total (mg/dL) 1711 2200 1350 1095 1717
CM/VLDL (mg/dL) 1540 na 1290 na na
LDL (mg/dL) 46 na 16 na na
HDL (mg/dL) 30 na 44 na na
Cholesterol: ​ ​ ​ ​ ​ ​
Total (mg/dL) 397 506 195 251 454
CM/VLDL (mg/dL) 338 na 147 na na
LDL (mg/dL) 25 na 33 na na
HDL (mg/dL) 14 na 15 na na
Post-heparin lipase ​ ​ ​ ​ ​ ​
activities:
LPL (mmol FFA/hour/ 0.84 [5.8 ± 0.3] na na na na
mL plasma) *
HL (mmol FFA/hour/ 1.92 [4.8 ± 0.8] na na na na
mL plasma) *
LMF1 variant(s) ​ p.Asn147Lys p.Asn147Lys p.Pro246Arg p.Arg354Trp (hom) p. p.Arg354Trp (het) p.
(hom) (hom) (hom) Arg364Gln (hom) Arg364Gln (hom)

BMI, body mass index; VLDL, very low-density lipoprotein; LDL, low-density lipoprotein; HDL, high-density lipoprotein; LPL, lipoprotein lipase; CM, chylomicron; HL,
hepatic lipase; LMF1, lipase; maturation factor 1; na, not available; hom, homozygote; het, heterozygote.
*
Lipase activities in brackets show normolipidemic controls (means ± SE, n = 31).

Patient 1 was a Hispanic female first seen at the UCSF Lipid Clinic at (26th percentile) and hypercholesterolemia (397 mg/dL) associated
age 29 because of elevated TG levels (500–1500 mg/dL). At the age of with elevated CM/VLDL- and low LDL-cholesterol. Commensurate with
50, her TG was 1711 mg/dL due to severely diminished post-heparin LPL variable adherence to low-fat diet and changes in body weight over the
activity (0.15th percentile). The patient also exhibited low HL activity years, her TG levels fluctuated between 600 and 1700 mg/dL. She

3
C. Bedoya et al. Journal of Clinical Lipidology xxx (xxxx) xxx

developed type 2 diabetes at age 56, suffered a stroke at 67 and died 2 protein, respectively (Fig. 2). p.Arg354Trp and p.Arg364Gln are low-
years later. Genomic DNA sequencing identified a homozygous frequency variants (MAF 1–5 %) located in the second predicted cyto-
c.441C>G variant in LMF1 corresponding to the p.Asn147Lys missense plasmic loop (Loop D) of LMF1 (Fig. 2) and have previously been re-
change. Cascade screening revealed that the proband’s brother (Patient ported at heterozygosity in patients with severe HTG.8,20,21
2) was also homozygous for p.Asn147Lys and suffered from severe HTG To assess the potential functional impact of the variants identified,
(2200 mg/dL) and hypercholesterolemia (506 mg/dL). Hyperlipidemia we evaluated the evolutionary conservation of affected amino acids
was ameliorated (TG = 406 mg/dL, TC = 211 mg/dL) after a month on a across orthologs of human LMF1. Residues p.Asn147, p.Pro246 and p.
very low-fat diet and plasma lipids normalized (TG = 115 mg/dL, TC = Arg364 are invariant from human to sea urchin suggesting that amino
106 mg/dL) on a combination therapy including fibrate (160 mg/d), acid changes at these positions are likely to affect function, whereas p.
niacin (Niacin) (1 g/d), ezetimibe (110 mg/d) and statin (40 mg/d) Arg354 is highly variant across taxa (Fig. 2). Furthermore, while p.
drugs. Asn147Lys, p.Pro246Trp and p.Arg364Gln are predicted to be func-
Patient 3 of Hispanic ancestry presented at the Lipid Clinic with a tionally deleterious by multiple algorithms, the analysis of p.Arg354Trp
chief complaint of recurrent abdominal pain and BMI of 24.1 kg/m2 at provided mixed results (Table 2).
the age of 11 years. A very high-fat diet was noted and physical exam-
ination was positive for hepatosplenomegaly, but negative for xantho- Functional analysis of LMF1 variants
mas. Plasma lipid analysis revealed severe HTG (1350 mg/dL) and
elevated cholesterol (195 mg/dL) associated with CM/VLDL. To investigate whether the identified variants affect the lipase-
Sequencing of LMF1 identified a homozygous c.737C>G variant corre- chaperone activity of LMF1, we evaluated the ability of mutant pro-
sponding to the p.Pro246Arg missense variant. The patient was advised teins to reconstitute LPL maturation in a LMF1-deficient cell line. In this
to follow a low-fat diet, after which his TG normalized and abdominal assay, LPL activity secreted into the cell culture media serves as a proxy
pain did not recur. He was lost to follow-up at the age of 12 years. for intracellular LMF1 activity in transfected cells.22 Indeed, the previ-
Patient 4 was a 65-year-old overweight Caucasian male, who pre- ously characterized p.Tyr439Ter variant completely abolished LPL ac-
sented at the Lipid Clinic with a BMI of 29.1 kg/m2, type 2 diabetes and a tivity (<1 % of wild-type) in the media confirming that this variant is a
history of severe six-vessel coronary artery disease (CAD), myocardial null-mutation (Fig. 3a).6 Of the novel variants, p.Asn147Lys markedly
infarction (MI) and bypass surgery. He had severe HTG (1095 mg/dL) impacted the secretion of active LPL (11 % of wild-type), whereas p.
and elevated plasma cholesterol (251 mg/dL). DNA sequencing identi- Pro246Arg, p.Arg354Trp and p.Arg364Gln resulted in significant, but
fied two homozygous LMF1 variants (c.1060C>T and c.1091G>A) less severe reductions in LPL activity (39 %, 78 % and 67 % of wild-type,
corresponding to the p.Arg354Trp and p.Arg364Gln missense changes in respectively) (Fig. 3a).
the LMF1 protein. The observed reductions in media LPL activity could be due to
Patient 5, a Caucasian female first seen at the age of 72, presented reduced LMF1 expression, impaired lipase-maturation function of LMF1
with a history of severe three-vessel CAD, MI at age 56 and was diag- or a combination of the two. To investigate the mechanism underlying
nosed with type 2 diabetes at age 68. Her plasma TG ranged from 555 to diminished LPL activities, we first determined whether LMF1 protein
1717 mg/dL and DNA sequencing revealed the presence of heterozygous expression was impacted by the four variants. For a quantitative analysis
c.1060C > T/p.Arg354Trp and homozygous c1091G>A/p.Arg364Gln of LMF1 protein mass, we expressed Gluc-LMF1 fusion proteins with the
LMF1 variants. variants and measured Gluc activity in cell lysates. As shown in Fig. 3b,
Other than overweight in Patients 1 and 4, type 2 diabetes in Patients p.Asn147Lys significantly diminished LMF1 expression (35 % of wild-
4 and 5 (Table 1) and high dietary fat intake in Patient 3, no additional type), whereas p.Pro246Arg, p.Arg354Trp, p.Arg364Gln and p.
secondary causes of HTG including elevated alcohol consumption, Tyr439Ter had no effect.
metabolic syndrome, hypothyroidism, renal disease or TG-elevating To assess the impact of variants on the specific activity of LMF1, we
medications could be identified in any of the patients.2 Furthermore, determined the ratio of LPL activity in the culture media and intracel-
none of the patients’ medical records indicated lactescent plasma or a lular Gluc activity (i.e. LMF1 protein mass) in each assay. This analysis
history of acute pancreatitis. revealed that all four variants significantly impaired LMF1 specific ac-
tivity (Fig. 3c). Taken together, our results indicate that the variants
In silico characterization of LMF1 variants analyzed in this study impair LPL secretion by two distinct mechanisms:
all variants diminish the lipase maturation function of LMF1, whereas p.
Our genetic analyses in HTG patients identified four homozygous Asn147Lys also reduces the expression of mutant LMF1 protein.
missense LMF1 variants (Table 2). p.Asn147Lys and p.Pro246Arg are
novel rare variants (MAF < 0.1 %) located in the second predicted
transmembrane region (TM2) and ER-facing loop (Loop C) of the LMF1

Table 2
Homozygous LMF1 variants associated with severe hypertriglyceridemia.
p.Asn147Lys p.Pro246Arg p.Arg354Trp p.Arg364Gln

Nucleotide variant ​ c.441C > G c.737C > G c.1060C > T c.1091G > A
rsID ​ rs182685983 rs1307363051 rs143076454 rs35168378
MAF (gnomAD): ​ ​ ​ ​ ​
Overall 2.2e-5 (6/278,182) nd (0/197,034) 1.3e-2 (3545/265,172) 2.5e-2 (6759/272,046)
Latino 1.1e-4 (4/35,230) nd (0/28,346) 1.0e-2 (342/34,026) 1.6e-2 (550/34,992)
European nd (0/126,946) nd (0/86,456) 1.8e-2 (2206/120,438) 2.9e-2 (3646/124,880)
ACMG classification ​ VUS VUS benign benign
ClinVar classification ​ VUS nr benign benign
In silico predictions: ​ ​ ​ ​ ​
PROVEAN (score) deleterious (− 4.48) deleterious (− 8.46) deleterious (− 2.76) deleterious (− 2.65)
SIFT (score) damaging (0.003) damaging (0.000) tolerated (0.085) tolerated (0.080)
PolyPhen-2 (score) probably damaging (0.992) probably damaging (1) benign (0.002) probably damaging (0.977)

MAF, minor allele frequency; gnomAD, genome Aggregation Database; nd, not detected; nr, not reported; PROVEAN, Protein Variation Effect Analyzer; SIFT, Sorting
Intolerant From Tolerant; PolyPhen-2, Polymorphism Phenotyping v2; VUS, variant of unknown significance.

4
C. Bedoya et al. Journal of Clinical Lipidology xxx (xxxx) xxx

Fig. 2. Location and evolutionary conservation of LMF1 variants. The schematic illustrates the predicted membrane topology of LMF1 and the location of
p.Asn147Lys, p.Pro246Arg, p.Arg354Trp and p.Arg364Gln variants (red circles). The heavy line represents the DUF1222 domain. The evolutionary sequence
conservation of TM2, Loop C and Loop D including the 4 variants are shown. Amino acids identical to the human sequence (top) are represented by dots. Fully
conserved residues across 42 orthologous LMF1 sequences, only a representative subset of which is shown, appear in bold. TM, transmembrane domain.

Discussion ancestry (n = 199, TG > 885 mg/dL) have been found to carry rare
biallelic LMF1 variants in a recent study, consistent with our results.24
Previous genetic studies revealed that FCS is responsible for only a In the present study, we identified two novel homozygous rare var-
small fraction (~1 %) of severe HTG and biallelic LMF1 variants iants (p.Asn147Lys and p.Pro246Arg) and two homozygous low-
represent a rare cause even within this subset of patients.23 Indeed, no frequency LMF1 variants (p.Arg354Trp and p.Arg364Gln) that have
such variants have been observed in large cohorts of moderate-to-severe been previously reported at heterozygosity in patients with severe
(n = 413, TG > 300 mg/dL)20 or severe (n = 563, TG > 885 mg/dL)23 HTG.8,21,23 Both patients homozygous for the p.Arg364Gln variant in
HTG, and a single homozygous rare LMF1 variant was identified in a our study also harbored p.Arg354Trp consistent with linkage disequi-
third report (n = 385, TG > 885 mg/dL).8 Thus, the frequency of ho- librium between these two variants, as previously observed.8 To explore
mozygous LMF1 variants observed in the present study (~1 %) is sub- the potential causality of the identified LMF1 variants in HTG, we per-
stantially higher than previously reported and likely explained by the formed functional analyses and assessed their impact on quantitative (i.
ethnic composition of our study population. While previous studies e. expression level) and qualitative (i.e. specific activity) aspects of the
focused exclusively on populations of European descent, the present LMF1 protein.
cohort included patients with Hispanic ancestry (12 %), a group with The analysis of p.Asn147Lys variant revealed a marked reduction in
increased overall burden of rare variants in canonical genes of TG LMF1 protein expression and specific activity. While the underlying
metabolism. Indeed, nearly 2 % of severe HTG patients of Hispanic molecular mechanism has not been investigated, we hypothesize that

5
C. Bedoya et al. Journal of Clinical Lipidology xxx (xxxx) xxx

this variant impairs the stability and increases the turnover rate of
LMF1. The p.Asn147Lys missense variant changes an evolutionarily
highly conserved residue within the second transmembrane domain
(TM2) of LMF1 (Fig. 2). We speculate that the replacement of an un-
charged amino acid for a positively charged one within the membrane-
spanning helix may affect its insertion into the lipid bilayer resulting in
topological changes that impair the activity and trigger the degradation
of the protein. Consistent with our results, a previously characterized
HTG-associated variant (p.Ser137Leu) in TM2 also abolished LMF1
expression and function.9
While p.Pro246Arg, p.Arg354Trp and p.Arg364Gln had no detect-
able effect on protein expression, all three variants reduced the specific
activity of LMF1 highlighting the role of corresponding protein domains
in lipase maturation. The p.Pro246Arg variant affects a residue that is
invariant from human to invertebrates (Fig. 2) and resulted in a major
reduction in LMF1 specific activity (35 % of wild-type). As p.Pro246Arg
is located in Loop C, a soluble domain of LMF1 facing the ER lumen, this
variant may affect the interaction with LPL.25 Relative to p.Asn147Lys
and p.Pro246Arg, the p.Arg354Trp and p.Arg364Gln variants had
modest effects on LMF1 function with 81 % and 68 % of wild-type
specific activity, respectively. Both p.Arg354Trp and p.Arg364Gln are
located in Loop D, which is predicted to face the cytoplasm (Fig. 2) and
may affect the interaction of LMF1 with other proteins involved in lipase
maturation.26,27 Further studies will be required to explore the molec-
ular mechanisms responsible for the functional impact of these variants.
Our results on p.Arg354Trp and p.Arg364Gln contrast with previous
studies reporting either no functional impact of these variants or <50 %
activity for p.Arg354Trp.8,21 These discrepancies are likely due to dif-
ferences in the assays used to assess LMF1 activity. The activity assay
employed in the present study represents multiple improvements over
previous ones. Importantly, the measurement of LPL activity and LMF1
protein expression in the same transfected cell population allows us to
quantitatively assess LMF1 specific activity for the first time. Further-
more, our use of LMF1 reconstitution in LMF1-deficient cells instead of
overexpression in LMF1-competent cells, as performed previously,
eliminates the potential confounding effects of stoichiometric imbalance
and compensatory changes associated with overexpression approaches.
Beyond the elevation of plasma TG, severe HTG is also associated
with a characteristic dyslipidemic profile that includes hypercholester-
olemia in conjunction with low LDL- and HDL-cholesterol levels.28–30
Indeed, with the exception of a pediatric patient (Patient 3), the subjects
in our study exhibited hypercholesterolemia and LDL/HDL-cholesterol
levels below the 1st population percentiles.31 Lipoprotein analysis
revealed that 75–85 % of total cholesterol was CM/VLDL-associated
indicating that hypercholesterolemia results from the accumulation of
these lipoproteins likely due to defective lipolysis. Consistent with this
mechanism, post-heparin LPL activity was severely diminished (14 % of
controls) in Patient 1. Another notable observation in this patient was
low (40 % of controls) post-heparin HL activity, a likely reflection of the
fact that the posttranslational maturation of this lipase is also dependent
on LMF1, although to a lesser degree than that of LPL.6,32 Indeed,
combined lipase deficiency (i.e. reduced activity of both LPL and HL) is a
consistent observation in patients with LMF1 loss-of-function variants
and Lmf1-deficient mouse models.6,7,14,33,34 Complete HL deficiency is
associated with several-fold increased TG content in LDL and HDL par-
ticles and modestly elevated LDL-TG and HDL-TG levels have been
observed in patients with heterozygous mutations in the HL gene.35–38
Consistent with HL-deficiency, Patients 1 and 3 presented with LDL and
HDL cholesterol:TG molar ratios that are 3–10-fold lower than normal
population averages.39,40 Although we cannot exclude the possibility
that these changes are secondary to severe HTG in our patients, elevated
LDL- and HDL-TG levels due to partial HL deficiency may be a dis-
Fig. 3. Functional analysis of LMF1 variants. (a) LPL activity secreted into the tinguishing feature of HTG associated with LMF1 mutations. Further
culture medium, (b) LMF1 protein expression, and (c) LMF1 specific activity
studies involving larger patient populations with biallelic LMF1 rare
were assessed in transfected LMF1-deficient cells.
variants will be required to investigate this possibility.

6
C. Bedoya et al. Journal of Clinical Lipidology xxx (xxxx) xxx

Severe HTG is a manifestation of two distinct genetic etiologies Bjarvin: Investigation. Wilbur Ji: Writing – review & editing, Investi-
including monogenic FCS and, much more commonly, MCS, a polygenic gation. Hanien Samara: Writing – review & editing, Investigation. Jody
form of the condition characterized by additional contributions from Tai: Writing – review & editing, Investigation. Laurie Green: Writing –
secondary (i.e. non-genetic) factors.1 Distinguishing clinical character- review & editing, Investigation. Philip H. Frost: Writing – review &
istics of FCS vs MCS include a 4–10-fold higher prevalence of acute editing, Investigation. Mary J. Malloy: Writing – review & editing,
pancreatitis in FCS and higher prevalence of insulin Investigation. Clive R. Pullinger: Writing – review & editing, Investi-
resistance-associated metabolic conditions (i.e. elevated BMI, type 2 gation, Funding acquisition, Data curation. John P. Kane: Writing –
diabetes, cardiovascular disease) in MCS.29,30,41 Furthermore, MCS pa- review & editing, Investigation, Funding acquisition, Data curation.
tients generally respond well to the management of secondary factors of Miklós Péterfy: Writing – review & editing, Writing – original draft,
HTG and pharmacological treatment, whereas FCS patients respond less Visualization, Validation, Supervision, Software, Resources, Project
well to these interventions.2 In the current study, none of the patients administration, Methodology, Investigation, Funding acquisition,
had a history of acute pancreatitis, whereas secondary causes of HTG Formal analysis, Data curation, Conceptualization.
could be identified in Patients 1, 4 and 5 (overweight and/or diabetes)
and Patient 3 (high dietary fat intake), and cardiovascular disease was
documented in Patients 1, 4 and 5 (Table 1). Importantly, all patients Conflict of Interest
responded well to therapeutic interventions and lifestyle changes. Taken
together, the clinical characteristics suggest MCS as the underlying ge- None.
netic etiology in all five patients in the present study.
Genetic analysis of canonical genes involved in TG catabolism (i.e. Acknowledgements
LPL, APOC2, APOA5, LMF1 and GPIHBP1) is an important tool to
discriminate between FCS and MCS.1,42 Accordingly, the presence of The authors would like to acknowledge and thank the patients
rare biallelic loss-of-function variants defines FCS, whereas rare het- involved in this study. This work was supported by the National In-
erozygous loss-of-function variants, or their absence, in canonical genes stitutes of Health (R01HL127155 and R15HL154071 to M.P.), the Jo-
indicate MCS. In the present study, we identified patients with rare seph Drown Foundation, the Campini Foundation, and by gifts from
homozygous loss-of-function variants in LMF1, which seemingly fulfills Peter Read, Harold Dittmer, Susan Boeing and Donald Yellon.
the genetic criteria for FCS. However, our results also demonstrated that
in contrast to the previously characterized p.Tyr439Ter null-mutation, References
the four variants identified here only partially impair LMF1 function.
The p.Arg354Trp and p.Arg364Gln variants exhibit 20–30 % reduced 1. Dron JS, Hegele RA. Genetics of hypertriglyceridemia. Front Endocrinol (Lausanne).
2020;11:455.
LMF1 activity, which is insufficient to cause severe HTG. Indeed, het- 2. Laufs U, Parhofer KG, Ginsberg HN, Hegele RA. Clinical review on triglycerides. Eur
erozygous Lmf1 knock-out mice and heterozygous carriers of function- Heart J. 2020;41:99–109c.
ally confirmed null-alleles in humans, such as p.Trp464Ter and p. 3. Lewis GF, Xiao C, Hegele RA. Hypertriglyceridemia in the genomic era: a new
paradigm. Endocr Rev. 2015;36:131–147.
Arg53Glyfs*5, exhibit normal plasma TG levels.8,9,34 Thus, additional 4. Paquette M, Bernard S. The evolving story of multifactorial chylomicronemia
genetic and/or non-genetic factors are likely required for the phenotypic syndrome. Front Cardiovasc Med. 2022;9, 886266.
expression of p.Arg354Trp and p.Arg364Gln variants, which is consis- 5. Peterfy M. Lipase maturation factor 1: a lipase chaperone involved in lipid
metabolism. Biochim Biophys Acta. 2012;1821:790–794.
tent with a diagnosis of MCS. The most severe variant characterized in 6. Peterfy M, Ben-Zeev O, Mao HZ, et al. Mutations in LMF1 cause combined lipase
the present study, p.Asn147Lys, retained only 11 % LMF1 activity in deficiency and severe hypertriglyceridemia. Nat Genet. 2007;39:1483–1487.
vitro and 14 % post-heparin LPL activity in vivo, whereas p.Pro246Arg 7. Cefalu AB, Noto D, Arpi ML, et al. Novel LMF1 nonsense mutation in a patient with
severe hypertriglyceridemia. J Clin Endocrinol Metab. 2009;94:4584–4590.
exhibited 35 % residual LMF1 activity. Despite the relatively large ef-
8. Serveaux Dancer M, Di Filippo M, Marmontel O, et al. New rare genetic variants of
fects of p.Asn147Lys and p.Pro246Arg, the clinical characteristics of LMF1 gene identified in severe hypertriglyceridemia. J Clin Lipidol. 2018;12:
patients harboring these variants are not consistent with FCS. Thus, our 1244–1252.
9. Peterfy M, Bedoya C, Giacobbe C, et al. Characterization of two novel pathogenic
results demonstrate that even ~90 % reduction in LMF1 activity is
variants at compound heterozygous status in lipase maturation factor 1 gene causing
insufficient to cause severe HTG without the contribution of secondary severe hypertriglyceridemia. J Clin Lipidol. 2018;12:1253–1259.
factors and manifests in MCS. 10. Cefalu AB, Spina R, Noto D, et al. Identification of a novel LMF1 nonsense mutation
In conclusion, the present report extends the allelic spectrum of responsible for severe hypertriglyceridemia by targeted next-generation sequencing.
J Clin Lipidol. 2017;11:272–281. e278.
LMF1 in severe HTG and provides novel insights into the structure- 11. Tanaka M, Takase S, Ishiura H, Yamauchi T, Okazaki S, Okazaki H. A novel
function relationship of LMF1 protein. Our study also illustrates the homozygous nonsense variant of LMF1 in pregnancy-induced hypertriglyceridemia
importance of functional analyses of rare variants in canonical genes and with acute pancreatitis. J Clin Lipidol. 2023;17:327–331.
12. Serveaux Dancer M, Marmontel O, Wozny AS, et al. Involvement of a homozygous
the consideration of their effect size before making a genetic diagnosis of exon 6 deletion of LMF1 gene in intermittent severe hypertriglyceridemia. J Clin
FCS or MCS. Lipidol. 2020;14:756–761.
13. Hegele RA, Berberich AJ, Ban MR, et al. Clinical and biochemical features of
different molecular etiologies of familial chylomicronemia. J Clin Lipidol. 2018;12:
Ethical statement 920–927. e924.
14. Liu Y, Xu J, Tao W, Yu R, Zhang X. A Compound Heterozygous Mutation of Lipase
This research was approved by the Institutional Review Board of the Maturation Factor 1 is Responsible for Hypertriglyceridemia of a Patient.
J Atheroscler Thromb. 2019;26:136–144.
University of California, San Francisco. All subjects gave informed 15. Heidemann BE, Bemelmans RHH, Marais AD, Visseren FLJ, Koopal C. Clinical
consent and all procedures were performed in compliance with relevant heterogeneity in monogenic chylomicronaemia. BMJ Case Rep. 2022:15.
laws and institutional guidelines. 16. Witztum JL, Gaudet D, Freedman SD, et al. Volanesorsen and Triglyceride Levels in
Familial Chylomicronemia Syndrome. N Engl J Med. 2019;381:531–542.
17. Pullinger CR, Stock EO, Movsesyan I, et al. Identification and metabolic profiling of
Use of AI and AI-assisted technologies statement patients with lysosomal acid lipase deficiency. J Clin Lipidol. 2015;9:716–726. e711.
18. Stock EO, Ferrara CT, O’Connor PM, et al. Levels of prebeta-1 high-density
lipoprotein are elevated in 3 phenotypes of dyslipidemia. J Clin Lipidol. 2018;12:
AI has not been used in the preparation of this manuscript.
99–109.
19. Havel RJ, Eder HA, Bragdon JH. The distribution and chemical composition of
CRediT authorship contribution statement ultracentrifugally separated lipoproteins in human serum. J Clin Invest. 1955;34:
1345–1353.
20. Johansen CT, Wang J, McIntyre AD, et al. Excess of rare variants in non-genome-
Candy Bedoya: Writing – review & editing, Methodology, Investi- wide association study candidate genes in patients with hypertriglyceridemia. Circ
gation. Rishi Thomas: Writing – review & editing, Investigation. Anna Cardiovasc Genet. 2012;5:66–72.

7
C. Bedoya et al. Journal of Clinical Lipidology xxx (xxxx) xxx

21. Surendran RP, Visser ME, Heemelaar S, et al. Mutations in LPL, APOC2, APOA5, 33. Cao C, Liu Y, Liu L, Wang X. Identification of a compound heterozygous LMF1
GPIHBP1 and LMF1 in patients with severe hypertriglyceridaemia. J Intern Med. variants in a patient with severe hypertriglyceridemia - case report and literature
2012;272:185–196. review. J Atheroscler Thromb. 2024;31:1106–1111.
22. Yin F, Doolittle MH, Peterfy M. A quantitative assay measuring the function of lipase 34. Ehrhardt N, Bedoya C, Peterfy M. Embryonic viability, lipase deficiency,
maturation factor 1. J Lipid Res. 2009;50:2265–2269. hypertriglyceridemia and neonatal lethality in a novel LMF1-deficient mouse model.
23. Dron JS, Wang J, Cao H, et al. Severe hypertriglyceridemia is primarily polygenic. Nutr Metab (Lond). 2014;11:37.
J Clin Lipidol. 2019;13:80–88. 35. Hegele RA, Little JA, Vezina C, et al. Hepatic lipase deficiency. Clinical,
24. Gill PK, Dron JS, Dilliott AA, et al. Ancestry-specific profiles of genetic determinants biochemical, and molecular genetic characteristics. Arterioscler Thromb. 1993;13:
of severe hypertriglyceridemia. J Clin Lipidol. 2021;15:88–96. 720–728.
25. Doolittle MH, Ehrhardt N, Peterfy M. Lipase maturation factor 1: structure and role 36. Ruel IL, Lamarche B, Mauger JF, et al. Effect of fenofibrate on plasma lipoprotein
in lipase folding and assembly. Curr Opin Lipidol. 2010;21:198–203. composition and kinetics in patients with complete hepatic lipase deficiency.
26. Sha H, Sun S, Francisco AB, et al. The ER-associated degradation adaptor protein Arterioscler Thromb Vasc Biol. 2005;25:2600–2607.
Sel1L regulates LPL secretion and lipid metabolism. Cell Metab. 2014;20:458–470. 37. Knudsen P, Antikainen M, Uusi-Oukari M, et al. Heterozygous hepatic lipase
27. Roberts BS, Babilonia-Rosa MA, Broadwell LJ, Wu MJ, Neher SB. Lipase maturation deficiency, due to two missense mutations R186H and L334F, in the HL gene.
factor 1 affects redox homeostasis in the endoplasmic reticulum. EMBO J. 2018;37. Atherosclerosis. 1997;128:165–174.
28. Blom DJ, O’Dea L, Digenio A, et al. Characterizing familial chylomicronemia 38. Tahvanainen E, Syvanne M, Frick MH, et al. Association of variation in hepatic
syndrome: baseline data of the APPROACH study. J Clin Lipidol. 2018;12: lipase activity with promoter variation in the hepatic lipase gene. The LOCAT Study
1234–1243. e1235. Invsestigators. J Clin Invest. 1998;101:956–960.
29. O’Dea LSL, MacDougall J, Alexander VJ, et al. Differentiating familial 39. Balling M, Afzal S, Davey Smith G, et al. Elevated LDL triglycerides and
chylomicronemia syndrome from multifactorial severe hypertriglyceridemia by atherosclerotic risk. J Am Coll Cardiol. 2023;81:136–152.
clinical profiles. J Endocr Soc. 2019;3:2397–2410. 40. Holmes MV, Millwood IY, Kartsonaki C, et al. Lipids, lipoproteins, and metabolites
30. Paquette M, Bernard S, Hegele RA, Baass A. Chylomicronemia: differences between and risk of myocardial infarction and stroke. J Am Coll Cardiol. 2018;71:620–632.
familial chylomicronemia syndrome and multifactorial chylomicronemia. 41. Belhassen M, Van Ganse E, Nolin M, et al. 10-year comparative follow-up of familial
Atherosclerosis. 2019;283:137–142. versus multifactorial chylomicronemia syndromes. J Clin Endocrinol Metab. 2021;
31. Balder JW, de Vries JK, Nolte IM, Lansberg PJ, Kuivenhoven JA, Kamphuisen PW. 106:e1332–e1342.
Lipid and lipoprotein reference values from 133,450 Dutch Lifelines participants: 42. Hegele RA, Dron JS. 2019 george lyman duff memorial lecture: three decades of
age- and gender-specific baseline lipid values and percentiles. J Clin Lipidol. 2017; examining DNA in patients with dyslipidemia. Arterioscler Thromb Vasc Biol. 2020;
11:1055–1064. e1056. 40:1970–1981.
32. Boedeker JC, Doolittle MH, White AL. Differential effect of combined lipase
deficiency (cld/cld) on human hepatic lipase and lipoprotein lipase secretion. J Lipid
Res. 2001;42:1858–1864.