Participant-derived cell line transcriptomic analyses and mouse studies reveal a role for ZNF335 in plasma cholesterol statin response

Abstract

Background: Statins lower circulating low-density lipoprotein cholesterol (LDLC) levels and reduce cardiovascular disease risk. Though highly efficacious in general, there is considerable inter-individual variation in statin efficacy that remains largely unexplained.

Methods: To identify novel genes that may modulate statin-induced LDLC lowering, we used RNA-sequencing data from 426 control- and 2 μM simvastatin-treated lymphoblastoid cell lines (LCLs) derived from European and African American ancestry participants of the Cholesterol and Pharmacogenetics (CAP) 40 mg/day 6-week simvastatin clinical trial (ClinicalTrials.gov Identifier: NCT00451828). We correlated statin-induced changes in LCL gene expression with plasma LDLC statin response in the corresponding CAP participants. For the most correlated gene identified (ZNF335), we followed up in vivo by comparing plasma cholesterol levels, lipoprotein profiles, and lipid statin response between wild-type mice and carriers of a hypomorphic (partial loss of function) missense mutation in Zfp335 (the mouse homolog of ZNF335).

Results: The statin-induced expression changes of 147 human LCL genes were significantly correlated to the plasma LDLC statin responses of the corresponding CAP participants in vivo (FDR=5%). The two genes with the strongest correlations were zinc finger protein 335 (ZNF335 aka NIF-1, rho=0.237, FDR-adj p=0.0085) and CCR4-NOT transcription complex subunit 3 (CNOT3, rho=0.233, FDR-adj p=0.0085). Chow-fed mice carrying a hypomorphic missense (R1092W; aka bloto) mutation in Zfp335 had significantly lower non-HDL cholesterol levels than wild type C57BL/6J mice in a sex combined model (p=0.04). Furthermore, male (but not female) mice carrying the Zfp335^R1092W allele had significantly lower total and HDL cholesterol levels than wild-type mice. In a separate experiment, wild-type mice fed a control diet for 4 weeks and a matched simvastatin diet for an additional 4 weeks had significant statin-induced reductions in non-HDLC (-43±18% and -23±19% for males and females, respectively). Wild-type male (but not female) mice experienced significant reductions in plasma LDL particle concentrations, while male mice carrying Zfp335^R1092W allele(s) exhibited a significantly blunted LDL statin response.

Conclusions: Our in vitro and in vivo studies identified ZNF335 as a novel modulator of plasma cholesterol levels and statin response, suggesting that variation in ZNF335 activity could contribute to inter-individual differences in statin clinical efficacy.

Keywords: RNA-sequencing; Statin; cholesterol; gene expression; lipoprotein; lymphoblastoid cell lines; mice; zinc finger protein 335.

Figure 1.

Volcano plot illustrating gene expression statin response in 426 CAP LCLs. Approximate log₂ fold change is 2^(variance stabilized statin gene expression level – variance stabilized control gene expression level) calculated for each cell line and averaged across cell lines.

Figure 2.

Correlation of ZNF335 and CNOT3 LCL gene expression statin responses with in vivo plasma LDLC statin responses of the corresponding donors (A-B) and with each other (C).

Figure 3.

12-week old chow-fed mouse plasma (A) total, (B) HDL, and (C) non-HDL cholesterol levels split by sex and Zfp335 genotype. Male sample sizes were N=12, 15 (16 for TC), and 10, and female sample sizes were N=17, 15, and 5 (6 for TC) for wild type, heterozygotes, and homozygous Zfp335^R1092W, respectively. Values are mean ± SEM. *p<0.05 ***p<0.001

Figure 4.

Plasma non-HDL cholesterol levels in (A) male and (B) female mice before and during statin diet feeding and (C) statin-induced changes in non-HDL cholesterol split by sex and Zfp335 genotype. Male sample sizes were N=7, 15, and 4 and female sample sizes were N=12, 11, and 5 for wild type, heterozygotes, and homozygous Zfp335^R1092W, respectively. Values are mean ± SEM. *p<0.05

Figure 5.

Statin-induced changes in wild-type A) N=5 male and B) N=11 female mouse lipoprotein profiles measured by ion mobility. Mouse profiles were measured before and after 4 weeks of simvastatin-containing diet. The size intervals designating the major lipoprotein subclasses are based on those defined in humans (29).

Figure 6.

Statin-induced changes in male mouse LDL particle concentrations as measured by ion mobility. Sample sizes were N=5, 12, and 5 for wild type, heterozygotes, and homozygous Zfp335^R1092W, respectively. Values are mean ± SEM. Genotypes were compared using two-way ANOVA with Tukey’s multiple comparison test. **p<0.01