Effects of Six APOA5 Variants, Identified in Patients With Severe Hypertriglyceridemia, on In Vitro Lipoprotein Lipase Activity and Receptor Binding
Objective— The purpose of this study was to identify rare APOA5 variants in 130 severe hypertriglyceridemic patients by sequencing, and to test their functionality, since no patient recall was possible.
Methods and Results— We studied the impact in vitro on LPL activity and receptor binding of 3 novel heterozygous variants, apoAV-E255G, -G271C, and -H321L, together with the previously reported -G185C, -Q139X, -Q148X, and a novel construct -Δ139 to 147. Using VLDL as a TG-source, compared to wild type, apoAV-G255, -L321 and -C185 showed reduced LPL activation (−25% [P=0.005], −36% [P<0.0001], and −23% [P=0.02]), respectively). ApoAV-C271, -X139, -X148, and Δ139 to 147 had little affect on LPL activity, but apoAV-X139, -X148, and -C271 showed no binding to LDL-family receptors, LR8 or LRP1. Although the G271C proband carried no LPL and APOC2 mutations, the H321L carrier was heterozygous for LPL P207L. The E255G carrier was homozygous for LPL W86G, yet only experienced severe hypertriglyceridemia when pregnant.
Conclusion— The in vitro determined function of these apoAV variants only partly explains the high TG levels seen in carriers. Their occurrence in the homozygous state, coinheritance of LPL variants or common APOA5 TG-raising variant in trans, appears to be essential for their phenotypic expression.
Premature truncations of APOA5, Q139X, Q148X, and the IVS3+3g>c are associated with severe hypertriglyceridemia (HyperTG),1–3 behaving as phenocopies of lipoprotein lipase (LPL) deficiency. These rare APOA5 variants do not always lead to a deficiency in circulating plasma apoAV, and carriers present with a range of apoAV levels (reviewed in4). Kao et al5 identified a common polymorphism G185C in a Taiwanese study which occurs at a minor allele frequency of 0.04 in controls but at a 6.3-fold higher frequency of 0.27 in hypertriglyceridemic patients (P<0.001).
ApoAV is present on chylomicrons, VLDL, and HDL, but not on IDL or LDL, suggesting that VLDL-containing apoAV is cleared before the lipolytic cascade.6 One function of apoAV is to activate LPL, and Apoa5 knockout mice have 4-fold higher TG levels than wild type.7–9 Whereas LPL transgenic mice can rescue Apoa5 knockout mice from HyperTG, this is not entirely reciprocal8 suggesting that the effect of apoAV on plasma TG is dependent on heparin-sulfate proteoglycan (HSPGs)-bound LPL.7,8 In addition, apoAV-dependent TG catabolism acts by enhancing receptor-mediated endocytosis via members of the LDL-receptor (LDLR) family.10,11
We have identified 3 novel APOA5 missense variants (E255G, G271C, H321L) in patients with TG levels >10 mmol/L. Although LPL and APOC2 variants had been excluded for the G271C carrier, the coding exons of LPL and APOC2 were sequenced in the 2 other probands. Because family studies were not possible, the APOA5 variants were expressed in vitro together with G185C, Q139X, Q148X.1,2,5 In addition we designed a deletion construct Δ139 to 147, to study whether the deleted region could explain the difference in phenotypes between Q139X and Q148X. To fully characterize these mutant proteins their effects on LPL activation and receptor binding were undertaken.
See supplemental methods (available online at http://atvb.ahajournals. org) for full details.
In total 130 patients with TG levels >10 mmol/L were recruited into the study; seven patients from the UK, 28 patients from the Netherlands with LPL and APOC2 mutations excluded, and 95 patients from the Czech Republic.
Resequencing of APOA5
Sequence of the coding exons of APOA5 was performed in 4 fragments.
Expression and Purification of Recombinant ApoAV
For expression WT apoAV and all mutant proteins in pET20b+ vector were transformed into BL21(DES) cells (Novagen, UK) and cultured following standard conditions. Recombinant apoAV was purified under denaturing conditions on a Ni-NTA-His Bind resin column (Novagen, UK).
Ligand Binding Experiments
Details using either LR8 or LRP1 are described in detail elsewhere.10,11
Identification of Novel APOA5 Variants in Patients With Severe HyperTG
Our aim was to identify novel variants in APOA5 and assess their frequency in patients with severe HyperTG. Sequencing the APOA5 coding region in 130 individuals with TG levels >10 mmol/L identified 4 nonsynonymous variants: E255G (c.764 A>G), G271C (c.821 G>T), H321L (c.962 A>T), and A315V (c.824 C>T), all in exon 3. A315V was present in 3 HyperTG Czech individuals and occurs at polymorphic frequencies in a Czech cohort12 and was not studied further. Two intronic point variants (IVS1+87 G>A and IVS2+45 T>C) were predicted by splice recognition algorithms (http://www.fruitfly.org/seq_tools/splice.html) to not affect splicing and were not investigated further. The frequency of these 3 novel variants was 2.3%, and that of the common SNPs S19W (rs3135506) and −1131T>C (rs662799) was ≈3.5 and, ≈2.4 fold higher, respectively, in the Czech and Dutch HyperTG patients compared to reported normolipidemic controls from those countries.13,14
E255G (c.764A>G) was identified in a 32-year-old female Asian UK patient with gestational Type I hyperlipoproteinemia. She was also heterozygous for APOA5 S19W. Coding exons for both the LPL and APOC2 genes were sequenced and revealed that the proband was homozygous for LPL W86G (with a normal APOC2 sequence). This variant had been previously reported in a pediatric Type I hyperlipoproteinemic patient,15 so it was surprising that it was only under the stress of her 2 pregnancies that her plasma TGs rose to ≈75 mmol/L and she developed cutaneous xanthomas. Before and after her pregnancies she was able to maintain TGs levels between 2 and 10 mmol/L with a low-fat diet alone (Table). APOA5-G255 was not detected in 260 healthy white controls and in the absence of Asian controls, we tested for its presence in an Asian cohort (n=508) with type 2 diabetes (T2D).16 Ten patients were heterozygous for E255G (carrier frequency 1.7%), one of whom also carried the APOA5 −1131T>C rare allele. Fasting lipid levels were available on 9 E255G carriers and their mean TG levels were 1.64±0.48 mmol/L, which did not differ significantly from the sample mean (2.44±1.7 mmol/L, P=0.17 (supplemental Figure I, available online at http://atvb.ahajournals.org).
G271C (c.821G>T) was identified in the heterozygous state in a 43-year-old Dutch female with persistent HyperTG and bouts of pancreatitis. She abstained from alcohol and had a stable BMI of 28 kg/m2. Her pretreatment plasma TG was ≈20 mmol/L and when treated and she adhered to a fat-restricted diet, her TG levels dropped to 6.30 mmol/L (Table). She had no family history of cardiovascular disease, pancreatitis, or HyperTG. The patient was homozygous for −1131C. Plasma apoAV levels, as determined by ELISA,17 were 3762 ng/mL, 15-fold higher than mean levels in normolipidemic volunteers (257 ng/mL). On a nonreducing Western blot (supplemental Figure II) her plasma apoAV appeared mainly in a 120 kDa form, 3-fold larger than wild type (WT) apoAV (around 40 kDa). Her postheparin LPL activity and mass was 121% and 100%, respectively, compared to healthy controls. This variant was not found in 265 white healthy controls. LPL and APOC2 variants had been excluded in this proband.
H321L (c. 962 A>T) was identified in a 46-year-old Czech male who was also heterozygous for S19W. His mother had hypertension and T2D and his father suffered from severe dyslipidemia, but neither would participate in the study. In 1990 he was diagnosed with hyperlipidemia but was not treated until 1999 when diagnosed with T2D and grossly elevated TG levels of 63.4 mmol/L and suffered from acute pancreatitis. His pretreatment BMI and lipid levels are presented in the Table. H321L was not detected in 282 healthy Czech controls. The sequencing of coding exons of LPL and APOC2 identified that he was a carrier of the French Canadian LPL P270L mutation.18
Using Protein Prediction Methods to Understand the Potential Functional Domains of ApoAV
Molecular modeling was used to characterize the effect of these novel and previously reported variants on apoAV function. Results from mutation prediction algorithms, SIFT and PolyPhen, are shown in supplemental Table I. ApoAV is a 366 aminoacid apoprotein which is highly α-helical and free of β structure with a calculated molecular weight of 41.2 kDa. The 5 coiled-coil elements and 9 predicted Receptor Binding Domains (RBD19) represent protein–protein, protein–DNA, or protein–ion interaction domains (supplemental Table II). E255 is in an α-helix in a putative RBD. Glycine rotates easily and adds flexibility to the protein chain. Thus E>G is likely to create a conformational change with glycine inside the protein molecule, reducing the ability of apoAV to interact with HSPG-bound LPL.
Although WT apoAV has only a single cysteine residue at position 227, a G>C at either site 185 or 271 would allow the formation of a disulfide bond. Western blotting confirmed that these variant proteins form dimers and multimers, changing the tertiary structure (Figure 1). In addition, A171-R245 is a potential lipid-binding domain, including leucine residues at positions 162, 163, 172, and 173, and a tryptophan residue at position 170, which facilitate lipid interactions. Furthermore, residue 185 occurs within a RBD and may therefore have an effect on protein–protein interactions.
H321 lies outside any of the predicted coiled-coil or RBDs (supplemental Table II), but within a disordered domain (G201 to K335). Although hydrophilic histidine can stabilize the folded structures of proteins, hydrophobic leucine tends to be located inside the protein molecule. This change is likely to affect the interaction with LPL.
X139, X148, and Δ139 to 147 are all predicted to lie in a coiled-coil domain with a putative RBD at position 146 to 152 (supplemental Table II). The X148 and Δ139 to 147 maintain most of the coiled-coil domain, whereas X139 only retains 30%.
Expression of Recombinant APOAV Proteins
Recombinant WT apoAV, the 6 variants and the novel Δ139 to 147 variant were expressed in vitro with and without the C-terminal His6-tag sequence. This enabled us to establish that the C-terminal His6-tag sequence was not interfering in the LPL activity assay. The 2 truncated variants could not be detected with the apoAV polyclonal antibody directed against the C terminus of the protein. Expression of all proteins was confirmed by Western blot analysis under nonreducing conditions, using either the apoAV polyclonal antibody or a His6-tag antibody (Figure 1).
Effect of Recombinant ApoAV on Human VLDL Hydrolysis by HSPG-Bound LPL In Vitro
To mimic the activation of LPL at the vascular endothelium by the different apoAV proteins, the previously described assay which incorporates HSPG was used,8 and a concentration titration using WT apoAV confirmed that 5 μg/mL was within the assay linear range. The results are shown in Figure 2. Compared to the negative control (LPL-only) WT stimulated LPL activity by 35% (P=0.003), confirming apoAV as an activator/stabilizer of LPL. Compared to WT, apoAV-L321 had the largest impact on LPL activation (−36% [P<0.0001], next apoAV-G255 showed a −25% effect [P=0.005]. The two G>C substitutions showed different effects, with apoAV-C271 having a nonsignificant effect of −11% (P=0.24), whereas apoAV-C185 displayed a −23% on LPL activation (P=0.02). Of the two truncated proteins X139 showed 18% (P=0.04) lower LPL activity whereas X148 and Δ139 to 147 showed only borderline effects on LPL activity of −15% (P=0.09) and −16% (P=0.08), respectively.
Effect of Recombinant ApoAV on Lipid Emulsion Hydrolysis by HSPG-Bound LPL In Vitro
To test whether the presence of apoAV in human plasma influenced the LPL hydrolysis of VLDL, we repeated the assay using a lipid emulsion as a TG source (Figure 3). Essentially the results were the same, however the different effects seen for the 2 G>C mutants in the VLDL assay were reversed.
Binding of Recombinant ApoAV to LDLR Relatives (LRs) by Ligand Blotting
Direct ligand blotting was carried out to compare mutant and WT protein binding to the chicken LDLR relative, LR8 based on Dichlberger et al.10 Binding to the 95-kDa LR8 from chicken follicle extract is shown in Figure 4. Binding of apoAV-L321, -E255, -C185 as well as Δ139 to 147 were comparable to WT binding, suggesting that the 9 deleted amino acids are not involved in receptor binding, which must be C-terminal of apoAV-Q148 because neither apoAV-X139 nor -X148 bound to LR8. ApoAV-C271 showed no binding to the LR8.
Binding of Recombinant ApoAV to LRP1 Using Surface Plasmon Resonance
Binding studies using surface plasmon resonance (SPR) to LRP1 confirmed these results, identifying apoAV -X139, -X148, and -C271 as nonbinders. Dissociation constants (KD) were calculated to determine overall affinity for this interaction. Compared to WT, apoAV-C185 displayed close to unchanged affinity for LRP1, KD=134×10−9 mol/L and KD=107×10−9 mol/L, respectively. ApoAV-G255, -L321 and Δ139 to 147 showed significant differences in affinity for LRP1 as given by their KD-values; KD_G255=348×10−9 mol/L; KD_L321=310×10−9 mol/L; KD_Δ139 to 147 l=224×10−9 mol/L. These 3 mutants also displayed a significant loss of number of binding sites compared to WT as shown by the lower response in Figure 5 and could represent conformational changes in the mutant proteins. KD could not be calculated for nonbinders.
We have identified 3 novel APOA5 missense variants, E255G, G271C, and H321L, in patients with severe HyperTG (>10 mmol/L at a frequency of ≈2.3%). However, in addition to these variants, the E255G proband was homozygous for the previously reported LPL W86G15 and the H321L carrier was heterozygous for LPL P207L18 suggesting that these variants in APOA5 are making a rather minor contribution to the disorder. We expressed these APOA5 variants together with those previously reported1,2,5 and carried out functional studies in vitro obtaining novel insights into the structure: function relationship of apoAV.
Wang et al resequenced coding exons in the APOA5, LPL, and APOC2 in 110 HyperTG patients. They identified no rare APOA5 variants, but 10% had LPL or APOC2 variants compared to 0.2% in normolipidemic controls.20 The carrier frequency of the S19W was 4.7-fold higher in HyperTG patients than controls.20 In our study the carrier frequency of rare APOA5 and LPL variants was 2.3% and 1.5%, respectively. Rare allele frequency differences for S19W and 1131T>C in HyperTG patients compared to same nationality controls were ≈3.5- and ≈2.4-fold higher, respectively.
The impact of E255G on TG levels in vivo is modest at best. The proband was also homozygous for LPL W86G, a variant known to cause Type I hyperlipoproteinemia in the homozygous state.15 Interestingly, the E255G proband only expressed severe HyperTG during her 2 pregnancies. Gestational HyperTG has been reported in patients homozygous or compound heterozygotes for LPL variants21,22 and is suggested to be related to increased VLDL secretion from the liver, especially in the third trimester of pregnancy, overwhelming the hydrolytic capability of the mutant LPL.22 ApoAV-G255 does show some reduced functionality and is less effective in activating LPL (as seen when VLDL or lipid emulsion were the TG-source), and this could exacerbate the HyperTG. However this variant occurred at a carrier frequency of 1.7% in a study of T2D Indian Asians and was not associated with TG levels that differed significantly from noncarriers. The change from a negatively charged glutamic acid to small neutral glycine could affect protein tertiary structure, α-helices formation or the putative RBD predicted at 251 to 257.
ApoAV-C271 and -C185
The patient heterozygous for G271C had a 15-fold higher plasma apoAV level compared to the normal plasma pool. The estimated molecular weight of her plasma apoAV-G271C was 3-fold higher (120 kDa) than WT, and the unprecedented apoAV levels suggest that this is a function of polymerization of the apoAV. The Western blot shows that the majority of this protein exists in the multimeric form. In vitro LPL activation by apoAV-C271 was not significantly different from the WT but apoAV-C271 did not bind to the receptors, and this is likely to represent the basis for the HyperTG. Another contributing factor to the HyperTG is likely to be the homozygosity for the rare allele of the APOA5 −1131T>C SNP (Table) which itself is associated with 40% TG-raising effect in the homozygous state.23
G185C was identified at a polymorphic frequency of 4.2% in healthy Taiwanese controls, but at significantly higher frequency (≈27%) in Taiwanese HyperTG patients (P<0.001),5 but is absent in whites.13 Compared to G185 homozygotes (mean TG 1.06 mmol/L), heterozygous G185C individuals had a modest 15% higher TG (1.22 mmol/L), whereas the CC185 individuals had TG levels of 21.0 mmol/L, with >10 fold higher risk of HyperTG.5 Thus C185 appears only to present with severe TG-raising effects in the homozygous state. In vitro, apoAV-C185 forms multimers, but the majority of this protein, in contrast to -C271, appears to be in the monomeric form (illustrated in Figures 1 and 4⇑a). However, sequence analysis predicts that, in contrast to apoAV-C271, LPL activation would be significantly impaired by apoAV-C185, because of the relative position of the mutated glycine in the protein. G185C showed normal binding to receptors LR8 and LRP1, confirming that residue 185 is more likely to be important in LPL activation than receptor binding. The differences between G185C and G271C in binding to LRs could be explained by the quarternary structure of the protein. We speculate that apoAV-C185 favors binding to the receptor over multimerization to itself or to C227, the only cysteine residue present in WT apoAV, whereas apoAV-C271 multimerizes more effectively, which could sterically blocks accessing LR8.
The Czech patient, heterozygous for both H321L and APOA5 S19W, was also heterozygous for the LPL variant P207L. He presented with severe HyperTG and T2D. Of all the recombinant apoAV variants tested in vitro, -L321 showed the most reduced LPL activation in the VLDL and lipid emulsion assays, and clearly showed an effect on LRP1 binding in the more sensitive SPR system. LPL P207L has been well characterized in vitro24 and in vivo.18 In a study of 34 P207L carriers, in the presence of a APOC3 TG-raising SNP, mean TG were 10.31 mmol/L compared to 5.58 mmol/L in noncarriers of this SNP.18 We propose that the presence of APOA5 H321L and S19W compound the effect of P207L, resulting in pretreatment TG levels reaching 63 mmol/L, considerably higher than that reported by Garenc et al.18
ApoAV-X139, -X148, and Δ139 to 147
The structure–function studies of these 2 reported truncation variants reflect the disparity in their observed LPL activity in vivo. The Q139X proband had only 21% residual LPL activity (comparable to homozygous LPL deficiency)1 whereas the Q148X homozygous proband had 60% residual LPL activity.2 The deletion Δ139 to 147 construct was created to test whether these 9 amino acids were of particular importance in LPL activation. In the VLDL assay X139 showed 18% lower LPL activity than WT, whereas X148 and Δ139 to 147 had little effect. In the lipid emulsion assay, both X139 and X148 showed significantly reduced LPL activation (10% and 9%, respectively). However, in vivo, in the heterozygous state, apoAV from the WT allele is likely to be present in plasma and thus the VLDL experiment should represent a more realistic assay. This supports the in vivo observation that the penetrance of these variants is low1,2 and common APOA5 SNPs, as well as obesity and age, are required for expression of HyperTG. It seems unlikely that residues 139 to 147 are crucial for LPL activation. When tested against LRP1 and LR8, neither truncation showed binding, suggesting that the receptor-binding domain is located C-terminal to residue 148.
Of the naturally-occurring variants only X148,2 C185,5 and the recently identified Q97X25 have been found in the homozygous state, associated with severe HyperTG. Whereas E255G and H321L probands coinherited LPL variants in the homozygous or heterozygous state, respectively, all the other rare APOA5 variants were coinherited with S19W in the heterozygous state or, in the case of G271C, with −1131T>C in the homozygous state. The family studies of X139 and X148 carriers1,2 clearly show that these rare APOA5 variants are only of clinical importance when coinherited with an additional APOA5 TG-raising SNP, primarily S19W. The case of Q97X adds strength to this hypothesis, because the obligate heterozygous parents and heterozygous sibling are all normoTG and homozygous for the common alleles of both S19W and −1131T>C.
A major limitation of this study was that because of the inability to recall these patients and their families, we could not perform cosegregation analysis nor measure postheparin LPL activity. Only in the case of G271C carrier did we have plasma apoAV measures. However, these are the first in vitro studies of apoAV variants. It now seems clear that the impact of rare variants on TG levels is only evident in the homozygous state, or with coinheritance of LPL variants, or in the presence of common APOA5 variants. Thus, missense or truncation variants in APOA5 might occur more commonly, but without these compounding effects they are unlikely to be associated with HyperTG.
We thank Prof Jorgen Gliemann, Aarhus University, Aarhus, Denmark for LRP1 and receptor-associated protein (RAP).
Sources of Funding
B.D., J.A.C., W.P., R.W., S.E.H., and P.J.T. are supported by the BHF (RG2005/014, and PG 04/110/17827). A.D. and W.J.S. were supported by the grant “GOLDII-Genomics of Lipid-Associated Disorders-II”, part “GEN-AU Genome Research in Austria,” and is funded by the Austrian Ministry for Education, Research, and Culture, and the Austrian Research Foundation (FWF P20218). S.K.N. and G.O. were supported by Swedish Medical Research Council. M.M. was supported by the “Deutsche Forschungsgemeinschaft and LiDia (Lipids and Diabetes).”
B.D. and W.W.Z. contributed equally to this study.
Original received April 18, 2008; final version accepted July 8, 2008.
Marcais C, Verges B, Charriere S, Pruneta V, Merlin M, Billon S, Perrot L, Drai J, Sassolas A, Pennacchio LA, Fruchart-Najib J, Fruchart JC, Durlach V, Moulin P. Apoa5 Q139X truncation predisposes to late-onset hyperchylomicronemia due to lipoprotein lipase impairment. J Clin Invest. 2005; 115: 2862–2869.
Priore Oliva C, Pisciotta L, Li VG, Sambataro MP, Cantafora A, Bellocchio A, Catapano A, Tarugi P, Bertolini S, Calandra S. Inherited apolipoprotein A-V deficiency in severe hypertriglyceridemia. Arterioscler Thromb Vasc Biol. 2005; 25: 411–417.
Kao JT, Wen HC, Chien KL, Hsu HC, Lin SW. A novel genetic variant in the apolipoprotein A5 gene is associated with hypertriglyceridemia. Hum Mol Genet. 2003; 12: 2533–2539.
O'Brien PJ, Alborn WE, Sloan JH, Ulmer M, Boodhoo A, Knierman MD, Schultze AE, Konrad RJ. The novel apolipoprotein A5 is present in human serum, is associated with VLDL, HDL, and chylomicrons, and circulates at very low concentrations compared with other apolipoproteins. Clin Chem. 2005; 51: 351–359.
Lookene A, Beckstead JA, Nilsson S, Olivecrona G, Ryan RO. Apolipoprotein A-V-heparin interactions: implications for plasma lipoprotein metabolism. J Biol Chem. 2005; 280: 25383–25387.
Merkel M, Loeffler B, Kluger M, Fabig N, Geppert G, Pennacchio LA, Laatsch A, Heeren J. Apolipoprotein AV accelerates plasma hydrolysis of triglyceride-rich lipoproteins by interaction with proteoglycan-bound lipoprotein lipase. J Biol Chem. 2005; 280: 21553–21560.
Schaap FG, Rensen PC, Voshol PJ, Vrins C, van der Vliet HN, Chamuleau RA, Havekes LM, Groen AK, van Dijk KW. ApoAV reduces plasma triglycerides by inhibiting very low density lipoprotein-triglyceride (VLDL-TG) production and stimulating lipoprotein lipase-mediated VLDL-TG hydrolysis. J Biol Chem. 2004; 279: 27941–27947.
Dichlberger A, Cogburn LA, Nimpf J, Schneider WJ. Avian apolipoprotein A-V binds to LDL receptor gene family members. J Lipid Res. 2007; 48: 1451–1456.
Mailly F, Palmen J, Muller DP, Gibbs T, Lloyd J, Brunzell J, Durrington P, Mitropoulos K, Betteridge J, Watts G, Lithell H, Angelico F, Humphries SE, Talmud PJ. Familial lipoprotein lipase (LPL) deficiency: a catalogue of LPL gene mutations identified in 20 patients from the UK, Sweden, and Italy. Hum Mutat. 1997; 10: 465–473.
Ireland H, Konstantoulas CJ, Cooper JA, Hawe E, Humphries SE, Mather H, Goodall AH, Hogwood J, Juhan-Vague I, Yudkin JS, di Minno G, Margaglione M, Hamsten A, Miller GJ, Bauer KA, Kim YT, Stearns-Kurosawa DJ, Kurosawa S. EPCR Ser219Gly: elevated sEPCR, prothrombin F1+2, risk for coronary heart disease, and increased sEPCR shedding in vitro. Atherosclerosis. 2005; 183: 283–292.
Schaap FG, Nierman MC, Berbee JF, Hattori H, Talmud PJ, Vaessen SF, Rensen PC, Chamuleau RA, Kuivenhoven JA, Groen AK. Evidence for a complex relationship between apoA-V and apoC-III in patients with severe hypertriglyceridemia. J Lipid Res. 2006; 47: 2333–2339.
Wang J, Cao H, Ban MR, Kennedy BA, Zhu S, Anand S, Yusuf S, Pollex RL, Hegele RA. Resequencing genomic DNA of patients with severe hypertriglyceridemia (MIM 144650). Arterioscler Thromb Vasc Biol. 2007; 27: 2450–2455.
Murugasu CG, Armstrong G, Creedon G, Cavanna JS, Galton DJ, Tomkin GH. Acute hypertriglyceridaemic pancreatitis in a pregnant Indian: a new lipoprotein lipase gene mutation. J R Soc Med. 1998; 91: 205–207.
Talmud PJ, Hawe E, Martin S, Olivier M, Miller GJ, Rubin EM, Pennacchio LA, Humphries SE. Relative contribution of variation within the APOC3/A4/A5 gene cluster in determining plasma triglycerides. Hum Mol Genet. 2002; 11: 3039–3046.
Peterson J, Ayyobi AF, Ma Y, Henderson H, Reina M, Deeb SS, Santamarina-Fojo S, Hayden MR, Brunzell JD. Structural and functional consequences of missense mutations in exon 5 of the lipoprotein lipase gene. J Lipid Res. 2002; 43: 398–406.