Hit-and-run epigenome editing durably lowers cholesterol in mice

Main text

Epigenome editing functions like a “dimmer switch” to fine-tune target gene expression by changing the epigenetic marks without directly altering the primary DNA sequence, unlike genome editing technologies.¹ Permanent epigenome editing using programmable editors holds great promise for the treatment of human diseases. A recent study published in Nature demonstrated that transient delivery of epigenome editors achieved durable silencing of Pcsk9 and cholesterol reduction in mice.² This paves the way for the development of in vivo hit-and-run epigenetic silencing-based therapeutics.

Epigenome editors are engineered by fusing effector domains (EDs) derived from natural transcriptional factors (repressors or activators) with a programmable DNA-binding domain (DBD), such as catalytically deactivated Cas9 (dCas9),³ transcription activator-like effectors (TALEs)^4,5 or zinc-finger proteins (ZFPs).⁶ Among the different EDs, the Krüppel-associated box (KRAB) family of transcriptional repressors is the most popular for epigenetic silencing. However, KRAB-associated histone marks are labile unless continuously deposited by a chromatin-bound repressor. Several groups have taken advantage of key EDs from a repressive complex that permanently silences endogenous retroviruses throughout development and adult life, including KRAB, the catalytic domain of the de novo DNA-methyltransferase A (cdDNMT3A), and its inactive cofactor DNMT3-like (DNMT3L) to achieve prolonged target-gene repression.^7,8,9 Transient delivery of the triple-engineered transcriptional repressors (ETRs) led to efficient, durable, and specific epi-silencing of endogenous genes in cultured cells.

Using an engineered reporter cell line, the authors screened the most effective triple-ETR combination for each of the three programmable DBD platforms targeting the CpG island encompassing the promoter region of Pcsk9: dCas9, TALEs, and ZFPs. Pharmacodynamic studies revealed that ZFP-based ETRs were more potent than dCas9- and TALE-based architectures in silencing Pcsk9, which was selected for in vivo studies. This is an interesting observation warranting further investigation. Nevertheless, the small size and independence of short-lived guide RNAs (gRNAs) make ZFPs advantageous in terms of the delivery and stability of the epigenome editing complex.

The ZFP-ETRs showed a high specificity profile, as examined by transcriptomic and DNA methylation analyses of ETR-treated cells. RNA sequencing (RNA-seq) revealed that only eight additional genes (four downregulated and four upregulated) were differentially expressed after treatment with ZFP-ETRs targeting Pcsk9, at a lower magnitude than for Pcsk9. None of these genes were in the proximity of Pcsk9, and the 40 genes adjacent to Pcsk9 showed no significant changes, suggesting that epigenetic editing is highly target specific. This is further supported by the genome-wide analysis of CpG methylation, which revealed 18 other differentially methylated regions in addition to Pcsk9. The observed off-target perturbations are likely caused by the mismatched binding of the ZFP arrays used in ZFP-ETRs, as the triple dCas9-based ETR combination targeting Pcsk9 did not induce significant off-target transcriptional or methylation deregulation.

The modified mRNA encoding the triple ZFP-ETRs packaged in lipid nanoparticles (LNPs) was systemically delivered into C57BL/6 mice through intravenous administration to assess their in vivo performance. The ETR/LNP-treated mice showed a rapid and profound reduction in PCSK9, which was then stabilized at around 40% of vehicle treatment levels until 330 days, the last time point analyzed. The blood cholesterol levels were also reduced following the treatment. The stability of epigenetic silencing induced by ETR/LNPs was further verified by hepatectomy-induced liver regeneration, corroborating the impressive in vivo durability of epigenetic editing. However, it remains to be determined whether the long-term durability of epigenome editing is broadly generalizable across target genes. A recent preprint showed that this may not be the case, at least for angiopoietin-like 3 (ANGPTL3), another therapeutic target gene for hypercholesterolemia, and angiotensinogen (AGT), a therapeutic target for high blood pressure.¹⁰

To simplify the epi-silencing platform, the authors further converted the triple ZFP-ETRs into an all-in-one format (referred to as evolved ETR [EvoETR]) with the cdDNMTA3A and DNMT3L heterodimer at the N terminus of ZFP and KRAB at the C terminus, similar to previously described epigenetic editors.^8,9 The EvoETR was more efficient than the triple-ETR combination in terms of deposing DNA methylation at the Pcsk9 promoter. Moreover, the EvoETR was more specific compared with the triple-ETR combination owing to the use of only one ZFP instead of three different ones. The in vivo performance of EvoETR in mice was also impressive. A 75% reduction of circulating PCSK9 was recorded at 6 weeks following EvoETR/LNP delivery, comparable with the results achieved with conventional gene editing of Pcsk9.

These proof-of-principle results are poised to fuel the already burgeoning excitement surrounding therapeutic epigenome editing. Notably, a cohort of leading companies, including Chroma Medicine, Epic Bio, Modalis Therapeutics, Navega Therapeutics, Tune Therapeutics, and Omega Therapeutics, is racing to develop epigenome editing therapies for a diverse range of human diseases, spanning from cancer and muscular dystrophy to cardiovascular diseases, hepatitis B infection, and pain relief. Some of these entities have reported promising preclinical results in non-human primates.^11,12

In summary, the heritable nature of the epigenetic marks deposited by epigenome editing technology makes it highly attractive as a one-and-done therapy, a feature shared only with gene editing technologies. However, epigenome editing does not rely on the generation of harmful DNA breaks to inactivate therapeutic target genes, representing a major safety advantage over conventional gene editing, base editing, and more recent prime editing. Another important feature of epigenome editing differing from gene editing is that it is reversible by either pharmacological intervention or epigenome activators. This would allow for temporally controlled silencing of the targeted gene and the reversal of treatment-related adverse effects. The advancements in epigenome editing hold immense potential for reshaping the treatment landscape, heralding an exciting frontier in the field of gene therapy.