A human APOC3 missense variant and monoclonal antibody accelerate apoC-III clearance and lower triglyceride-rich lipoprotein levels

Abstract

Recent large-scale genetic sequencing efforts have identified rare coding variants in genes in the triglyceride-rich lipoprotein (TRL) clearance pathway that are protective against coronary heart disease (CHD), independently of LDL cholesterol (LDL-C) levels. Insight into the mechanisms of protection of these variants may facilitate the development of new therapies for lowering TRL levels. The gene APOC3 encodes apoC-III, a critical inhibitor of triglyceride (TG) lipolysis and remnant TRL clearance. Here we report a detailed interrogation of the mechanism of TRL lowering by the APOC3 Ala43Thr (A43T) variant, the only missense (rather than protein-truncating) variant in APOC3 reported to be TG lowering and protective against CHD. We found that both human APOC3 A43T heterozygotes and mice expressing human APOC3 A43T display markedly reduced circulating apoC-III levels. In mice, this reduction is due to impaired binding of A43T apoC-III to lipoproteins and accelerated renal catabolism of free apoC-III. Moreover, the reduced content of apoC-III in TRLs resulted in accelerated clearance of circulating TRLs. On the basis of this protective mechanism, we developed a monoclonal antibody targeting lipoprotein-bound human apoC-III that promotes circulating apoC-III clearance in mice expressing human APOC3 and enhances TRL catabolism in vivo. These data reveal the molecular mechanism by which a missense variant in APOC3 causes reduced circulating TG levels and, hence, protects from CHD. This protective mechanism has the potential to be exploited as a new therapeutic approach to reduce apoC-III levels and circulating TRL burden.

Figure 1

Human APOC3 A43T carriers exhibit lower apoC-III levels than non-carriers. (a) APOC3 A43T carriers were identified from exome-wide genotyping in the HHDL and Penn Medicine BioBank cohorts. The significance of the difference in A43T carrier frequency between the two cohorts was assessed by a Fisher's exact test. (b) TG concentration in overnight-fasted plasma of A43T carriers versus age-, sex-, and ancestry-matched controls (non-carriers) from the two cohorts. (c) Total apoC-III concentration in fasting plasma from A43T carriers and non-carrier controls. (d) A43T apoC-III concentrations in plasma samples of non-carriers and A43T carriers. (e) WT apoC-III concentrations in plasma samples of non-carriers and A43T carriers. (f) The mutant:WT apoC-III ratio of non-carriers and A43T carriers was compared to an expected ratio of 1:1 for no imbalance by a one-sample t test with an expected mean of 1.0. For b and c, n = 19 for A43T carriers and 76 for matched non-carriers. For d–f, n = 19 for A43T carriers and 21 for matched non-carriers. All measurements in b–f were replicated twice in the same plasma samples. All measurements are shown as mean ± s.d., and each data point depicts a single measure from an individual human participant plasma sample. **P < 0.01, ***P < 0.001, ****P < 0.0001, Student's unpaired two-sided t test. For f, ****P < 0.0001, one-sample t test.

Figure 2

Mice expressing APOC3 A43T have reduced TRL and circulating apoC-III levels. (a) Hepatic AAV vector levels, as assessed by qRT–PCR for the rabbit β-globin poly(A) sequence, from 25 mg of liver tissue from mice treated with WT or A43T APOC3 AAV. (b) Hepatic APOC3 mRNA levels (normalized to those of actin) in mice treated with WT or A43T APOC3 AAV. (c) Fasting plasma TG concentrations in Apoc3-knockout mice treated with Null, WT APOC3, or A43T APOC3 AAV at the indicated time points after AAV injection. (d) Fasting plasma TG concentrations in human-APOB-transgenic/Apobec1-knockout mice treated with the indicated AAVs. (e–g) Fasting plasma TG concentrations (e), plasma HDL-C concentrations (f), and plasma non-HDL-C concentrations (g) in WT mice treated with the indicated AAVs and co-treated with AAV encoding human CETP. (h,i) TG (h) and cholesterol (i) concentrations in FPLC-separated plasma fractions from day 28 plasma from the mice in e–g. Lipoprotein fractions are indicated above the fraction numbers. (j) Postprandial TG concentrations in Apoc3-knockout mice treated with the indicated AAVs and co-treated with AAV encoding human CETP, following olive oil gavage (OFTT, oral fat tolerance test). (k) Plasma [³H]TRL-TG FCR in WT mice treated with the indicated AAVs, 2 h after intravenous administration of [³H]triolein-labeled human TRLs. (l) Fasting plasma apoC-III concentrations in mice from e–g. (m) Immunoblots for apoC-I II using total protein from liver lysates of WT mice, 28 d after AAV administration. β-actin was used as a loading control. Cropped immunoblots are shown; corresponding uncropped blots are shown in Supplementary Figure 8. (n) Hepatic apoC-III secretion in WT mice treated with the indicated AAVs, 35 d after AAV administration and after treatment with [³⁵S]methionine tracer. ApoC-III secretion is defined as [³⁵S]methionine radioactivity in apoC-III bands isolated from protein electrophoresis, normalized to [³⁵S]methionine radioactivity in total TCA-precipitable protein from 2 μl of plasma. (o) ApoC-III secretion rates as measured by the slope of the curves in n. In c, data are shown from n = 5 Null mice, n = 7 WT mice, and n = 7 A43T mice. In a, b, d–g, and j–l, data are shown from n = 6 mice in each group. In n and o, data are shown from n = 5 mice from each group. Data show results from one representative experiment, and all experiments were repeated once in independent respective cohorts of mice. For data in a, b, k, and o, box length spans the 25th to 75th percentile range of the data points, with the middle line indicating the median and whiskers indicating the minimum and maximum values for the given data set. All other measurements show mean ± s.e.m. All data points represent measures from individual mice from a single experiment, and data in all panels were replicated in two independent experiments. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001, two-way ANOVA, WT versus A43T group; N.S., not significant.

Figure 3

The A43T substitution promotes circulating apoC-III clearance and renal uptake by perturbing apoC-III binding to lipoproteins. (a) Plasma [¹²⁵I]TC-modified WT or A43T apoC-III clearance in human-APOB-transgenic/Apobec1-knockout mice over the course of 24 h. Mice treated with WT APOC3 AAV were administered [¹²⁵I]TC-modified WT apoC-III, and those treated with A43T APOC3 AAV were administered [¹²⁵I]TC-modified A43T apoC-III. Normalized ¹²⁵I activity relative to plasma activity at 1 min is shown. (b) FCR of the plasma [¹²⁵I]TC-modified apoC-III shown in a. (c) Hepatic [¹²⁵I]TC activity in 30 mg of tissue for the mice in a. Activity was normalized to activity at 1 min. (d) Renal [¹²⁵I]TC activity in 30 mg of tissue from the mice in a. (e) ¹²⁵I activity in FPLC fractions of pooled plasma from each experimental group described in a, at 1 min. Activity is expressed as the fraction of total activity in plasma before FPLC separation. (f) ¹²⁵I activity in FPLC fractions after incubation of ¹²⁵I-labeled WT or A43T apoC-III (1 μg) with human plasma (200 μl) for 1 h at 37 °C. Data refer to a representative experiment and were replicated three times in independent experiments. (g) Percentage of total plasma ¹²⁵I activity in VLDL fractions versus unbound protein fractions after incubation of ¹²⁵I-labeled WT or A43T apoC-III (1 μg) with isolated human VLDL (100 μg of protein) for 1 h at 37 °C. Points represent the percentage of ¹²⁵I activity in fractions from one representative experiment of three experimental samples. (h) Percentage of total ¹²⁵I activity in HDL fractions versus unbound protein fractions after incubation of ¹²⁵I-labeled WT versus A43T apoC-III (1 μg) with isolated human HDL (200 μg of protein) for 1 h at 37 °C. Points indicate the ¹²⁵I activity in fractions from one representative experiment of three experimental samples. (i) Dissociation constant (K_d) from measurement of association and dissociation rate constants for binding of WT or A43T apoC-III to dimyristoylphosphatidylcholine surfaces by surface plasmon resonance. Points indicate observed K_d from a representative experiment of three replicate experimental samples. For a–d, n = 6 mice per group. Data in e show n = 1 pooled sample for each group of n = 6 mice in a–d. For f–i, results show the mean of three technical replicate experiments for each panel. Data were replicated by an independent repeat experiment. For data in b–d, box length spans the 25th to 75th percentile range of the data points, with the middle line indicating the median and whiskers indicating the minimum and maximum values for the given data set. All other measurements show mean ± s.e.m. where appropriate. *P < 0.05, **P < 0.01, Student's unpaired two-tailed t test; ****P < 0.0001, two-way ANOVA (a); ****P < 0.0001, Student's unpaired t test (b).

Figure 4

Anti-human-apoC-III monoclonal antibodies STT505 and STT5058 lower circulating apoC-III levels and promote TRL clearance. (a) Schematic of the experimental approach, in which the STT505 monoclonal antibody (mAb) or isotype control (ctrl) antibody was tested in C57BL/6 WT (B/6 WT) mice treated with WT APOC3 AAV for 3 weeks. (b) Plasma apoC-III levels over the course of 24 h following antibody administration. Values for the STT505 group are expressed as percentages of those for the control antibody group at the same time point. (c) Plasma apoC-III areas under the curve (AUCs) per mouse for each group in b. (d) Plasma apoB concentrations in mice from b over the course of 24 h. Values for the STT505 group are expressed as percentages of those for the control antibody group at the same time point. (e) Plasma apoC-III concentrations in mice after control or STT505 antibody administration and subsequent intragastric gavage of olive oil. (f) Plasma apoC-III AUCs per mouse for each group in e. (g) Postprandial plasma TG concentrations for the mice in e. (h) Postprandial TG elevation as measured by incremental AUC (i-AUC) per mouse for the groups in e. For i-AUCs, AUCs were calculated relative to a baseline defined as the mean plasma TG at time 0 for all mice in both the control and STT505 groups (72.74 mg/dl). (i) Schematic of the experimental approach, in which the STT5058 monoclonal antibody was tested in WT mice treated with WT APOC3 AAV for 3 weeks. (j,k) Plasma apoC-III (j) and apoB (k) levels over the course of 28 d following antibody administration. Values for the STT5058 group are expressed as percentages of those for the control antibody group at the same time point. (l) Clearance of [¹²⁵I]TC-modified apoC-III in WT mice that had been treated with APOC3 AAV 3 weeks before administration of STT5058 or control antibody (25 mg/kg subcutaneous dosing), followed 24 h later by intravenous administration of [¹²⁵I]TC-modified WT apoC-III. Clearance of radiolabeled apoC-III was measured over the course of 21 h. (m) FCR of [¹²⁵I]TC-modified apoC-III estimated from clearance curves in mice treated with STT5058 versus isotype control antibody. (n–p) Uptake of radiolabeled apoC-III in 20 mg of spleen (n), liver (o), or kidney (p) tissue 21 h after protein administration (values were normalized to plasma activity at 1 min for each mouse). (q) Proposed model of the contribution of TRL-associated apoC-III to TRL clearance. Top, WT apoC-III is bound to TRLs and is capable of inhibiting lipolysis and catabolism of circulating TRLs; a smaller pool of lipoprotein-free apoC-III may be cleared renally. Middle, A43T apoC-III has impaired binding to lipoproteins, augmenting apoC-III clearance by the kidney and promoting TRL lowering. Bottom, the STT505 and STT5058 monoclonal antibodies targeting apoC-III promote clearance of circulating apoC-III partially through an alternative splenic pathway, resulting in TRL lowering. For b and d, n = 7 mice per group. For c and f, n = 8 mice per group. For e, n = 9 mice in the control group and 10 mice in the STT505 group. Each experiment was replicated in an independent group of mice. For g and h, n = 10 mice in the control group and 9 mice in the STT505 group. For j and k, n = 7 mice for each group. Each experiment was replicated once in an independent group of mice. For l–p, n = 10 mice in each group. Each experiment in l–p was performed in one cohort of mice. All data show measures from individual mice. For data in c, f, h, and m–p, box length spans the 25th to 75th percentile range of the data points, with the middle line indicating the median and whiskers indicating the minimum and maximum values for the given data set. All other measurements show mean ± s.e.m. where appropriate. For b, d, e, and j–l, *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001, two-way ANOVA. For c, f–h, and m–p, *P < 0.05, **P < 0.01, ***P < 0.001, Student's unpaired two-tailed t test.