Postprandial metabolism of apolipoproteins B48, B100, C-III, and E in humans with APOC3 loss-of-function mutations

Background Apolipoprotein C-III (apoC-III) is a regulator of triglyceride (TG) metabolism, and due to its association with risk of cardiovascular disease, is an emergent target for pharmacological intervention. The impact of substantially lowering apoC-III on lipoprotein metabolism is not clear. Methods We investigated the kinetics of apolipoproteins B48 and B100 (apoB48 and apoB100) in chylomicrons, VLDL1, VLDL2, IDL, and LDL in patients heterozygous for a loss-of-function (LOF) mutation in the APOC3 gene. Studies were conducted in the postprandial state to provide a more comprehensive view of the influence of this protein on TG transport. Results Compared with non-LOF variant participants, a genetically determined decrease in apoC-III resulted in marked acceleration of lipolysis of TG-rich lipoproteins (TRLs), increased removal of VLDL remnants from the bloodstream, and substantial decrease in circulating levels of VLDL1, VLDL2, and IDL particles. Production rates for apoB48-containing chylomicrons and apoB100-containing VLDL1 and VLDL2 were not different between LOF carriers and noncarriers. Likewise, the rate of production of LDL was not affected by the lower apoC-III level, nor were the concentration and clearance rate of LDL-apoB100. Conclusion These findings indicate that apoC-III lowering will have a marked effect on TRL and remnant metabolism, with possibly significant consequences for cardiovascular disease prevention. Trial registration ClinicalTrials.gov NCT04209816 and NCT01445730. Funding Swedish Heart-Lung Foundation, Swedish Research Council, ALF grant from the Sahlgrenska University Hospital, Novo Nordisk Foundation, Sigrid Juselius Foundation, Helsinki University Hospital Government Research funds, Finnish Heart Foundation, and Finnish Diabetes Research Foundation.


INTRODUCTION
Apolipoprotein C-III (apoC-III) is a small protein (comprising 70 amino acid residues) found on triglyceride rich lipoproteins (TRLs) -chylomicrons and VLDL-that appears to be an important regulator of their intravascular metabolism (1)(2)(3). It is synthesised in the liver and intestine, and in the circulation transfers freely between lipoprotein particles, mainly TRL and HDL but it is also present on LDL (1,4). Genetic studies provide strong evidence linking variation in the APOC3 gene that encodes apoC-III with altered plasma triglyceride (TG) levels and risk of atherosclerosis (1)(2)(3)(5)(6)(7)(8)(9).
The existence of a strong, positive association between plasma apoC-III and TG levels has been appreciated for many years and raised the possibility that apoC-III was a key determinant of TRL metabolism. A mechanistic basis for this relationship emerged when it was discovered that apoC-III acted as an inhibitor of lipoprotein lipase (LpL), the main enzyme responsible for TG hydrolysis in both chylomicrons and VLDL (10)(11)(12)(13)(14)(15)(16). Indeed, the rate of TRL lipolysis was linked to the ratio of apoC-II (another small apolipoprotein that is an activator of LpL) to apoC-III on the particle surface (1)(2)(3). In the process of TRL lipolysis, cholesterol-enriched remnants are generated, and recent reports have identified these remnant particles as independent risk factors for atherosclerotic cardiovascular disease (ASCVD) (17)(18)(19)(20)(21)(22). According to current concepts, a slower lipolysis rate, due for example to higher apoC-III, favours remnant formation. There is evidence also, mainly in animal models, that a further action of apoC-III is to slow the rate of remnant removal by the liver, thereby prolonging the residence time of these particles in the circulation (1,4,13,18,23). The plasma level of apoC-III is determined by its rates of synthesis and catabolism, and human metabolic studies have shown that it is the apoC-III production rate that is a key determinant of plasma TG (24)(25)(26)(27). ApoC-III may also have an intrahepatic function in that increased secretion of apoC-III has been reported to stimulate VLDL production in cell models (28).
Rare loss-of-function (LOF) variants in APOC3 have been identified and shown to confer low TG levels (1,(5)(6)(7)(8)(9). One, the rs76353203 variant introduces a stop-codon in the signal peptide (R19X) preventing synthesis of the apoC-III protein (29). Interestingly, 5% of the Lancaster Amish are heterozygous carriers for this null mutation (29). Another, the rs138326449 variant is a splice variant (IVS2+1G-A) of APOC3 intron 2 (30). As compared with noncarriers of APOC3 mutations, heterozygotes for these two mutations have mean reductions of 44% in non-fasting TG levels (7). The impact of the R19X mutation has been examined in a kinetic study in heterozygotes for the null variant and their unaffected siblings (31). It was found that in the fasting state the genetically determined low apoC-III concentration led to accelerated delipidation rates for VLDL and more rapid conversion to LDL.
The present study examined broader effects of apoC-III on TG metabolism by investigating in APOC3 LOF carriers in the post-prandial state. Our objective was to elucidate the regulatory role of apoC-III in this more physiologically relevant situation and thereby gain better insight into the potential for therapeutic interventions targeted at apoC-III to alter circulating levels of TRL and their remnants, and so potentially protect against cardiovascular disease (5)(6)(7)(8)(9). Our approach allows the derivation of production and clearance rates of intestinally derived apoB48-lipoproteins and liver-derived apoB100containing lipoproteins, integrating the dynamics of TRL metabolism in the fed state with continuous hepatic secretion of lipoproteins. The kinetics of apoC-III and apoE were also investigated in APOC3 LOF heterozygote carriers.

RESULTS
Key characteristics of the subject groups that participated in the kinetic study are shown in Table 1 and a broader range of cardiometabolic variables is provided in Supplementary Tables 1  (TG) was also significantly lower by 57%, as was the fasting apoB48 level (by 71%, P=0.025). In contrast, mean LDL-C and total plasma apoB were virtually the same in the two groups. There were no differences in other lipid-related variables such as HDL-C, apoAI, apoA5, apoE and Angptl3 (Table 1). No between group differences were seen in biomarkers of TG and fatty acid metabolism including free fatty acid levels, β-hydroxybutyrate (a marker of fatty acid oxidation), and the activities of lipoprotein lipase and hepatic lipase.
Triglycerides and phospholipids, specifically phosphatidylcholines, were subject to lipidomic analysis to ascertain if alterations in apoB metabolism altered their fatty acid composition. No systematic differences were detected between the APOC3 LOF carriers and non-carriers in the fatty acid species present in chylomicrons, VLDL1, VLDL2, IDL or LDL ( Supplementary Figures 2A and 2B).

ApoB100, apoB48 and triglyceride metabolism in APOC3 LOF heterozygotes
Tracers of deuterated leucine and glycerol were used to determine the kinetic properties of apoB100, apoB48 and triglyceride in these subjects. The total set of lipoprotein concentration and protein enrichment data used as input to the compartmental model is shown in Supplementary Figure 1. A good fit of the model to the observed data was obtained (Supplementary Figure 1) and production, transfer and clearance rates for lipoprotein classes were derived as set out in Table 2. To illustrate the impact of the APOC3 LOF variants on the flow of apoB100 down the VLDL to LDL delipidation cascade, mean enrichment curves for VLDL1, VLDL2, IDL and LDL apoB100 are presented in Figure   1. Here it can be seen that, compared to non-variant carriers, APOC3 LOF carriers exhibited a very rapid transfer of apoB100 (and hence lipoprotein particles since each particle has one integral apoB protein) from VLDL1 to VLDL2, from VLDL2 to IDL, and finally from IDL to LDL. The derived fractional transfer rates (FTR) between these fractions reflected this metabolic behaviour ( Table 2); FTRs were increased significantly in the APOC3 LOF carriers by about 3-fold compared to the nonvariant carriers. Similarly, the overall fractional clearance rates (FCRs) for VLDL1-, VLDL2-, and IDL-apoB100 were increased by approximately the same degree. Total clearance from a density interval (reflected in the FCR) is a combination of transfer to the next component in the delipidation chain (lipolysis) and direct clearance of the apoB-containing particle from the circulation (fractional direct clearance rate, FDCR). It was noteworthy that the FDCRs for VLDL1 apoB100 and for total VLDLapoB100 were significantly higher in the APOC3 LOF carriers compared to non-variant carriers ( Table 2). Further key findings were that the FCR for LDL-apoB100 did not differ between the two groups, and that the production rate for VLDL1 apoB100, direct production rate for VLDL2 apoB100, and total input of apoB into IDL plus LDL was comparable in APOC3 LOF carriers versus non-variant carriers ( Table 2, Supplementary Table 3). There was a nominally significant difference noted in direct LDL apoB100 synthesis (Supplementary Table 3).
Triglyceride kinetics in VLDL fractions revealed similar between-group differences as was seen for apoB100 ( Table 2). The FTR for VLDL1 to VLDL2 conversion, the fractional direct removal rate of VLDL1 triglyceride (VLDL1-TG FDCR) and the overall FCRs for VLDL1 and VLDL2 were 2 to 3fold higher in APOC3 LOF carriers. Again, production rates for TG in VLDL1 and VLDL2 were similar in the two groups.
Metabolism of apoB48-containing particles in the chylomicron and VLDL density ranges was altered in the APOC3 LOF carriers compared to non-variant carriers (Figure 2, Table 2, Supplementary   Figure 1). There was a remarkable reduction in post-prandial lipaemia (area-under-curve, AUC) in response to the standard fat-meal in the APOC3 LOF carriers compared to non-variant carriers (Table 1, Figure 2) that was evident for both plasma TG (AUC reduced by 55%) and plasma apoB48 (AUC reduced by 76%). In line with the changes seen for VLDL apoB100-containing particles, overall clearance rates for chylomicrons (CM-apoB48 FCR), and apoB48-containing VLDL1 and VLDL2 were 2-to 3-fold higher in the APOC3 LOF carriers relative to non-variant carriers. These differences were nominally significant with the exception of the VLDL1-apoB48 FCR and the TG-apoB48 VLDL1 FCR which showed a trend of approximately the same magnitude. Production rates of apoB48-containing particles either in the basal (fasted) state or post-prandially following the test meal were not different between the two groups. LOF carriers compared to non-variant carriers. Production rates for intestinally derived apoB48-containing chylomicrons after a meal, and for apoB100-containing VLDL1 and VLDL2 released from the liver appear not to be affected in the APOC3 LOF carriers. The substantially and significantly lower circulating pool sizes for VLDL1-apoB100 (74%), VLDL2-apoB100 (69%) and IDL-apoB100 (83%) in the APOC3 LOF carriers (Supplementary Table 3) are the result of enhanced clearance both down the delipidation chain -lipolysis rates are several-fold higher -and by direct catabolism of VLDL apoB100-containing particles. Higher lipolysis and overall clearance rates are also seen for apoB48-containing chylomicrons (although the difference in direct chylomicron clearance did not reach significance). Estimation of the total transit time for an apoB100-containing VLDL1 particle to become LDL was 4.3 hours in APOC3 LOF carriers versus 17.7 hours in non-variant carriers (total transit time was taken as the sum of the residence times for VLDL1, VLDL2 and IDL apoB100containing particles: residence time is the reciprocal of the FCR). Likewise, the mean residence time for a chylomicron particle (reciprocal of CM-apoB48 FCR in Table 2) is 7.7 minutes in APOC3 LOF carriers compared to 28 minutes in non-variant carriers.

ApoC-III and apoE metabolism in APOC3 LOF heterozygotes
Kinetic parameters were determined for apoC-III and apoE using enrichment data and individual compartmental models for each protein ( Table 3). In the APOC3 LOF carriers, the circulating pool size of apoC-III was reduced in line with the lower plasma concentration for this apolipoprotein. This difference compared to non-variant carriers was attributable to both a 57% lower production rate and a 72% higher tractional clearance rate for the protein. In contrast, there were no differences in the rates of production and clearance of apoE ( Table 3).

DISCUSSION
This detailed metabolic investigation in subjects with genetically determined lower apoC-III was designed to uncover the regulatory actions that this apolipoprotein exerts on triglyceride transport in humans. It was found that subjects heterozygous for APOC3 LOF variants exhibited marked and highly specific alterations in chylomicron and VLDL kinetics. First, the degree of post-prandial lipaemia following a standardised fat-rich meal was greatly reduced in APOC3 LOF carriers. Second, on average the overall clearance rates of VLDL1, VLDL2, IDL and chylomicrons were about 3-fold higher in those with APOC3 LOF compared to non-variant carriers matched for age and BMI. This resulted in substantial reductions in the circulating levels of these lipoprotein classes, and in the time taken for large VLDL particles to transit down the delipidation cascade (from VLDL1 to LDL). Third, a potentially significant finding from a mechanistic viewpoint was that the rate of direct clearance of VLDL (especially VLDL1) from the circulation was higher in the APOC3 LOF carriers. Fourth, it was noteworthy that having a lower apoC-III level was not associated with altered production rates for VLDL or chylomicrons. Nor, did it impact on the production or clearance rate of LDL which explains why total plasma apoB and LDL-C were the same in APOC3 LOF carriers as in non-variant carriers.
The results in this group of unrelated subjects (five of whom were heterozygous for a previously unstudied APOC3 LOF variant) are in line with, and build on, the metabolic studies conducted in the Amish population where carriers of a null mutation in APOC3 (R19X) were compared to unaffected siblings. In the investigation by Reyes-Soffer et al (31), the mutation in APOC3 was associated with approximately 50% lower plasma apoC-III levels, 35% lower plasma TG and a 36% lower total VLDL-apoB. There was no difference in VLDL production rates between the APOC3 LOF carriers and their unaffected siblings, and the lower VLDL (and TG) level was attributed to an increased conversion rate to IDL and LDL. While a trend to increased direct clearance of VLDL-apoB was seen in this earlier study, it was not significant. We were able to establish that lack of apoC-III did substantially affect the delipidation rates of both chylomicrons and VLDL, and provided evidence for a significant effect on apoB100-containing VLDL direct clearance. The observation that our APOC3 LOF carriers had significantly lower apoC-III production rates establishes the non-significant trend seen in the Amish study (31). It is likely that decreased production of apoC-III is a consequence of the genetic variation in these two reports, and that the significantly increased FCR for this apolipoprotein seen in our study is due to the more rapid clearance of TRL particles containing this protein on their surface rather than a direct result of the LOF mutation. The data from these highly informative studies in genetically defined subjects, combined with metabolic investigations across subject groups with a wide range of TG levels that show a strong, inverse association between high plasma apoC-III levels and lower VLDL and chylomicron clearance rates, lead to the conclusion that this apolipoprotein is a major regulator of TRL lipolysis and a suitable target for intervention (1,4,13,18,23). Examination of these rare individuals with LOF variants not only reduces the potential for confounding factors to influence interpretation of the kinetic findings but also demonstrates vividly the large differences in TRL metabolism that accompany APOC3 LOF even in the heterozygous state. The observation that the clearance rate of LDL-apoB100 (and hence LDL particles) was unaffected by geneticallydetermined low apoC-III levels in both the present and previous investigations (31) indicates that apoC-III has no impact on LDL receptor activity, and provides a possible explanation as to why apoE kinetics did not differ between APOC3 LOF carriers and non-carriers (apoE and apoB100 are the major ligands for the LDL receptor) (17).
Given the substantial impact that reduced apoC-III levels have on the metabolism of intestinallyderived chylomicrons and liver-derived VLDL, the question arises as to what physiological role this protein has in people with normal or optimal TG levels (17). One possibility is that apoC-III has a counter-regulatory action which moderates the degree of TRL lipolysis in a tissue bed, for example, cardiac or skeletal muscle. In our evolutionary past, access to fat-rich foods, especially animal-based, was likely to be episodic and required the expenditure of energy over a prolonged period (to meet the challenge of hunting prey). Lipoprotein lipase appears to be a high capacity, very efficient enzyme for removing TG from the core of TRL particles and potentially an inhibitor present on TRL may operate to ensure delivery to multiple tissue sites rather than have the full TG load in a chylomicron or VLDL1 particle removed at 'first-pass'. APOC3 LOF variants are rare in populations, and this may be due to the requirement to have, when fat in the diet is scarce (as it has been throughout much of human history), a prolonged circulation time for TRL -of the order of near 30 minutes for chylomicrons and about 12 hours for VLDL compared to the <8 minutes and 3 hours respectively seen when apoC-III is very low.
The major limitation of this investigation relates principally to the rarity of APOC3 LOF variant carriers in the population. We had to screen large numbers of genetically defined individuals to identify our subjects while the previous study relied on an isolated population cohort where rare variants may be enriched (31). It follows, therefore, that the number of LOF carriers is small and we have to accept a sex imbalance relative to the non-carrier group. We believe this did not significantly impact of the interpretation of the kinetic results since the differences between the groups were so marked.
Genetic studies have provided convincing evidence that APOC3 LOF even in the heterozygous state is associated with reduced risk of ASCVD (5)(6)(7)(8)(9)38). This protective effect may be attributable to TGlowering consequences of having a low apoC-III, or according to some studies apoC-III may have a more direct atherogenic action (1,(39)(40)(41). The observation in studies of apoB metabolism in APOC3 LOF and other investigations (7,8) that low levels of apoC-III are not linked to reduction in total plasma apoB or LDL-C prompts consideration of the role of this apolipoprotein in determining the abundance of remnant lipoproteins in the circulation. These partially-lipolyzed TRL are cholesterolrich and are thought to have an atherogenic potential that is equal to or exceeds that of LDL (17)(18)(19)(20)(21).
In the present study, we did observe substantially lower concentrations of VLDL1, VLDL2 and IDL (likely associated with decreased levels of apoB100-containing remnants) and also plasma apoB48 (a marker of chylomicron remnant concentration) in the group of APOC3 LOF carriers. Whether this reduced abundance of remnants is sufficient to account fully for the decrease in ASCVD risk seen in APOC3 LOF carriers is yet to be determined. Finally, it is worthwhile comparing the findings of studies in subjects with APOC3 LOF variants associated with reduced apoC-III production to the results of clinical trials of agents designed to suppress apoC-III synthesis. Olezarsen, an anti-sense oligonucleotide drug that interferes with APOC3 mRNA translation, was found at the highest dose to reduce apoC-III by 74%, plasma TG by 60%, plasma apoB by 10% but LDL-C was largely unaffected (32). The accord between the natural occurring mutations and pharmacological intervention is striking. Previous in vitro observations have suggested a role for apoC-III in the regulation of VLDL assembly and secretion (1,2,28). However, the lack of effect of APOC3 LOF variants on chylomicron apoB48 and VLDL apoB100 production rates indicates that this finding may not translate to the human situation. The APOC3 LOF variants would likely affect intracellular (intrahepatic) apoC-III content as well as the circulating concentration of the apoprotein.
In conclusion, in a group of unrelated subjects studied in the physiological setting of post-prandial lipaemia we found that APOC3 LOF variants were associated with dramatic and highly specific changes in triglyceride metabolism. Lipolysis rates were increased several-fold and plasma concentrations of TRL and IDL decreased as a consequence. There was no impact on production rates of apoB48-containing lipoproteins or apoB100-containing lipoproteins which is an important observation indicating that use of lipid-lowering agents directed at apoC-III may not impair TG export from liver and intestine. The observation that reducing apoC-III levels is associated with a rapid transit of TRL particles down the delipidation pathway suggests that the rate remnant formation will be diminished and further, the finding that direct clearance of VLDL is higher when apoC-III is low indicates that remnant removal may be enhanced. The net effect of genetically determined lower apoC-III on ASCVD risk is well documented; however, how this is linked to the substantial perturbations in TRL metabolism given that LDL levels are not altered is yet to be defined.

METHODS
SubjectsThis study was performed in 12 subjects (white of Finnish ancestry): 6 (3 men and 3 postmenopausal women) were heterozygotes for APOC3 LOF mutations (5 for rs138326449 and 1 for rs7653203) and 6 'control' subjects (all men) who were non-carriers of any known APOC3 LOF variant and matched for age and body mass index (BMI) with the carriers. The six APOC3 variant carriers were identified initially from previously genotyped and whole exome sequenced participants in the THL Biobank in Finland (https://thl.fi/en/web/thl-biobank/for-researchers/sample-collections), and their genotypes were subject to confirmatory analysis. The non-carrier subjects were selected from previous kinetic study cohorts (42)(43)(44). General inclusion criteria were age 20-75 years, BMI <35 kg/m 2 and non-smoking status. Exclusion criteria were a history of any cardiovascular or severe disease, any condition affecting lipid levels, abnormalities in thyroid or kidney function or haematological abnormalities. None of the subjects used any medication or hormones known to influence lipid metabolism.
Metabolic study protocolSubjects were admitted to the clinical research unit at the Helsinki University Central Hospital after a 12 h fast at 8.00 am. To examine the kinetics of apolipoprotein B-48, B100, apo C-III, apo E we used a tracer of deuterated leucine (5,5,5-D3 Euriso-Top, d3-leucine) at a dose of 7 mg/kg body weight. To examine the kinetics of triglyceride we used a tracer of deuterated glycerol (D1,1,2,3,3, D5 Euriso-Top, d5-glycerol) at a dose of 500 mg (44,45). Two hours after tracer administration a standard fat-rich meal (927 kcal) comprising bread, butter, cheese, ham, boiled egg, fresh red pepper, low-fat (1%) milk, orange juice and tea or coffee (63 g carbohydrate, 69 g fat and 40 g protein) was consumed within 10 minutes. Blood samples were drawn before tracer injection and at frequent intervals thereafter until 10 h post-administration when a dinner was served.
The subjects remained physically inactive and only water was drunk (ad libitum) during the study.
The subjects returned the following morning again in the fasting state to give blood at 24 h post tracer administration. In addition, the subjects came back fasting at 8.00 on days 2, 3, 4 and 7 to allow measurements of D3-leucine enrichments of apoB100 in IDL and LDL.
Tracer enrichment in apolipoproteins and triglycerides, multicompartmental modelling and parameter estimationThe protocol for kinetic investigation and the compartmental model structure for apoB48/apoB100/triglyceride has been described in detail previously (27,(44)(45)(46). The experimental protocol and modelling for apoC-III and apoE have also been reported (27,46).
Modelling and parameter estimation were performed using SAAMII (47).

Lipoprotein isolation and Biochemical analysesChylomicrons, VLDL1 (Sf 60-400), and VLDL2
(Sf 20-60) were isolated from blood samples by density gradient centrifugation (48) and IDL and LDL were prepared using sequential fixed density centrifugation (46). Concentrations of triglycerides and cholesterol in total plasma and lipoprotein fractions were analysed using the Konelab 60i analyser (Thermo Fisher Scientific, Finland). Other laboratory tests were performed using standard methods.
Total plasma and postprandial apoB-48 levels were measured by ELISA (Shibayagi, Shibukawa, Japan). Plasma levels of apoC-III were determined with an immunoturbidimetry-based method Lipidomic analysisLipids were extracted from the lipoprotein fractions using the BUME method (49). The lipid extracts were diluted in chloroform:methanol [1:2] with 5 mM ammonium acetate and the lipids were quantified by direct infusion (shotgun) analysis on a QTRAP 5500 mass spectrometer (Sciex, Concord, Canada) equipped with a robotic nanoflow ion source, TriVersa NanoMate (Advion BioSciences, Ithaca, NJ). The analysis of triglycerides was performed in positive ion mode by neutral loss detection of 10 common acyl fragments formed during collision induced dissociation according to previous work (50). The phosphatidylcholines were also detected in positive mode using precursor ion scanning of m/z 184.1 as described previously (51). Glyceryl-d5-hexadecanoate (CDN isotopes, Quebec, Canada) and diheptadecanoyl (17:0/17:0) phosphatidylcholine were added during the extraction and used for quantification.
Lipase MeasuresA bolus injection of 75 IU/kg heparin was injected intravenously after an overnight fast to release lipase in the circulation. An immunochemical method was used to measure selectively lipoprotein and hepatic lipase activities in post-heparin plasma isolated from blood samples taken 10 minutes after the heparin bolus (52).

Determination of intra-abdominal fat depotsSubcutaneous abdominal and intra-abdominal fat
content were determined using a 1.5-T whole-body device (53). Subjects were advised to fast for 4 hours before imaging.
Statistical analysesAll statistical analyses were performed using R (version 4.0.2). P-values for baseline characteristics in Table 1 were calculated using a t-test for continuous variables and a chisquare test for categorical variables (as per standard settings in package "tableone"). P-values for kinetic parameters (Table 2 and Table 3) and lipidomics ( Supplementary Figures 2A and 2B) were calculated using the Mann-Whitney U test using the "wilcox.test()" function in R.
Study approvalThe study design was approved by the Ethics committee of Helsinki University Central Hospital, Helsinki, Finland (Clinical Trials NCT04209816 and NCT01445730) The study was performed in accordance with the Declaration of Helsinki and the European Medicines Agency Note for guidance on Good Clinical practice. All study participants gave a written informed consent form before any study procedures were initiated.

AUTHOR CONTRIBUTIONS
The authors contributed to the present work as follows:        Data are mean values for each group of subjects. Production rates are given in mg/d(ay), fractional transfer rates and fractional direct clearance rates in pools/d. The plasma pool of each apoB100-containing lipoprotein class (VLDL1, VLDL2, IDL and LDL) is given in mg within the appropriate circle. ApoB48 and apoB100 kinetic rate constants are in upright text; kinetic rate constants for TG are in italics. The mean transit time was calculated as the sum of the residence times for VLDL1, VLDL2 and IDL apoB-containing particles. Residence time is the reciprocal of the overall fractional clearance rate e.g. for VLDL1 apoB100 in non-variant carriers the FCR is 13.1 pools/24hours ( Table 2) which gives a residence time of 24/13.1 = 1.83 hours. Asterisks indicate significant differences between APOC3 LOF variants vs. non-variant carriers.