A virus-like particle-based bivalent PCSK9 vaccine lowers LDL-cholesterol levels in non-human primates

Abstract

Elevated low-density lipoprotein cholesterol (LDL-C) is an important risk factor in the development of atherosclerotic cardiovascular disease (ASCVD). Inhibitors of proprotein convertase subtilisin/kexin type 9 (PCSK9), a negative regulator of LDL-C metabolism, have emerged as promising approaches for reducing elevated LDL-C levels. Here, we evaluated the cholesterol-lowering efficacy of virus-like particle (VLP) based vaccines that target epitopes found within the LDL receptor (LDL-R) binding domain of PCSK9. In both mice and non-human primates, a bivalent VLP vaccine targeting two distinct epitopes on PCSK9 elicited strong and durable antibody responses and lowered cholesterol levels. In macaques, a VLP vaccine targeting a single PCSK9 epitope was only effective at lowering LDL-C levels in combination with statins, whereas immunization with the bivalent vaccine lowered LDL-C without requiring statin co-administration. These data highlight the efficacy of an alternative, vaccine-based approach for lowering LDL-C.

Fig. 1. The structure of human PCSK9 in complex with LDL-R.

LDL-R is shown in blue and the prodomain and catalytic domain of PCSK9 are shown in salmon and gray, respectively. Targeted PCSK9 epitopes are shown in red (amino acids 153-163) and cyan (amino acids 207-223). The amino acid sequences of these epitopes from human, rhesus macaque, and mouse PCSK9 are shown; residues that differ from the human sequence are shown in magenta. Structures are generated using the RCSB Protein Data Bank, structure ID: 3M0C.

Fig. 2. Immunogenicity and cholesterol-lowering activity of PCSK9 VLPs in LDLR^+/− mice.

LDLR^+/− mice were immunized three times at three-week intervals with 5 µg of wild-type Qß VLPs, mPCSK9_153-163 VLPs, mPCSK9_207-223 VLPs, or a bivalent mPCSK9 vaccine consisting of a mixture of 5 µg mPCSK9_153-163 VLPs and 5 µg mPCSK9_207-223 VLPs. Blood plasma were obtained three weeks following the final immunizations and IgG antibody titers against (a, b) mPCSK9 peptides and (c) full-length recombinant mPCSK9 were determined by ELISA. The effects of vaccination on total cholesterol levels were determined by comparing the serum cholesterol level in plasma following the vaccination series with levels in plasma obtained prior to immunization (d, e). All error bars shown represent standard error of the mean (SEM). Experimental groups were compared statistically by unpaired two-tailed t test. ns not significant.

Fig. 3. Plasma PCSK9 levels in vaccinated LDLR^+/− mice.

Mice were immunized three times and plasma were collected two weeks following the final immunization. a Plasma PCSK9 levels in mice immunized with individual PCSK9 VLP vaccines or control, wild-type Qß VLPs. b Plasma PCSK9 levels in mice immunized with a bivalent PCSK9 VLP vaccine or control, wild-type Qß VLPs. Experimental groups were compared statistically by unpaired two-tailed t test.

Fig. 4. Liver LDL-R expression in immunized mice.

LDLR^+/− mice (n = 11) were immunized three times with a bivalent mPCSK9 VLP vaccine or, as a control, wild-type Qß VLPs. Three to four weeks following the final immunization, mice were euthanized and livers were collected. Liver LDL-R protein expression was measured by Western blot (shown in Supplementary Fig. 1), quantified using ImageJ, and then normalized by comparing the expression of LDL-R to β-actin (as a loading control). Each data point represents an individual mouse (n = 11) and means for each group are shown using a black line. Experimental groups were compared statistically by unpaired two-tailed t test.

Fig. 5. Durability of anti-PCSK9 antibody responses.

LDLR^+/− mice (n = 6; 3 males and 3 females) were immunized three times (at weeks 0, 3, and 6) with a bivalent mPCSK9 VLP vaccine and anti-PCSK9 antibody titers were measured by ELISA for over 1-year post-immunization. Data represents geometric mean end-point dilution titers, error bars show the geometric standard deviation (SD). Note: because of a medical issue (that was unrelated to vaccination) we were unable to obtain sera from one of the mice after week 35.

Fig. 6. Anti-PCSK9 antibody responses in immunized rhesus macaques.

a The study timeline. Groups of macaques (n = 7–8) were immunized four times (at days 0, 28, 56, and 161, denoted by red arrows) with rhPCSK9_207-223 VLP, a bivalent rhPCSK9 VLP vaccine, or, as a control, wild-type Qß VLPs. Plasma was obtained after three immunizations (at day 77) and b anti-PCSK9_207-223, c anti-PCSK9_153-163, and d anti-full length hPCSK9 IgG levels were measured by ELISA. Data represents mean ELISA values for each group of macaques, error bars show the SEM.

Fig. 7. Plasma LDL-C levels in immunized rhesus macaques.

Plasma LDL-C levels at different days post-prime in macaques immunized with bivalent rhPCSK9 VLPs (red), rhPCSK9_207-223 VLPs (blue) or, as a control, wild-type Qß VLPs (green). Macaques were treated daily with simvastatin from day 77–105 (represented by black bars below the x-axis). Each data point represents an individual animal, bars show mean values for each timepoint, and error bars represent SEM. Significance was determined by one-tailed t test.

Fig. 8. Circulating PCSK9 levels in immunized rhesus macaques.

Plasma was collected three weeks following the third immunization (day 77, a) or following a fourth immunization at necropsy (day 190, b). Circulating PCSK9 levels were determined by ELISA. Each data point represents an individual animal, lines represent mean values for each group of macaques, and error bars show SEM. Significance was determined by two-tailed t test.

Fig. 9. Boosting increases anti-PCSK9 antibody levels, lowers LDL-C, and increases expression of LDL-R in the liver.

a Anti-PCSK9 IgG levels were measured by ELISA prior to boosting (at day 161) and two weeks following the boost (day 175). b Plasma LDL-C levels were measured at day 175 and at necropsy and were compared to baseline (at d161, prior to prime). Error bars represent SEM. c LDL-R protein levels were measured by Western blot (shown in Supplementary Figure 5), quantified using ImageJ, and then normalized by comparing the expression of LDL-R to β-actin (as a loading control). Each data point represents an individual macaque and means for each group are shown using a black line. Experimental groups were compared statistically by one-tailed t test.