Emerging Metabolic and Transcriptomic Signature of PNPLA3-Associated NASH

About 30% of the adult population in the USA is afflicted with NAFLD. Approximately 20%−25% of patients with NAFLD have the progressive subtype of NAFLD: NASH. The cardinal features of NASH on liver histology include steatosis, lobular inflammation, and ballooning with or without perisinusoidal fibrosis.⁽¹⁾ Currently, NASH is the second-leading indication for liver transplant in the USA, and is the fastest rising etiology leading to rising rates of HCC in the USA. The mechanistic understanding of why some patients with NAFLD have a nonprogressive course, while others progress to NASH leading to cirrhosis and HCC, is unclear. Several genome-wide association studies have consistently shown that a single nucleotide polymorphism in the patatin-like phospholipase domain-containing 3 (PNPLA3) gene, which changes isoleucine (I) to methionine (M) at residue 148 (rs738409 C>G polymorphism, PNPLA3^I148M), is one of the strongest genetic risk factors for NAFLD progression.⁽²⁾ PNPLA3^I148M is not only associated with greater risk of progressive steatohepatitis but also with cirrhosis and HCC.^(3,4) Although PNPLA3 function has been studied extensively, the molecular mechanisms of how the I148M variant drives hepatic fibrosis and carcinogenesis remain unclear.

To understand the mechanisms of PNPLA3^I148M in NASH pathogenesis, previous studies have generated transgenic and knock-in mice in C57BL/6J background that expressed either human or mouse PNPLA3^I148M variant.^(5,6) When fed with high sucrose or high fat diet, the mice expressing PNPLA3^I148M variant quickly progressed to steatosis as compared with wild-type (PNPLA3^WT) mice. However, the NASH and fibrosis phenotype was not apparent in these models. Therefore, although these studies provided key insights into the role of PNPLA3^I148M in steatosis development, how it modulates NASH and fibrosis progression remains to be further explored. In this issue of Hepatology, Banini et al. used a well-characterized diet-induced animal model of NAFLD (DIAMOND), which has been shown to sequentially develop steatosis, NASH, and fibrosis within 16 weeks on a high-fat (Western) diet with sugars in drinking water (WDSW).⁷ Mouse livers were transduced with adeno-associated virus–8 expressing human PNPLA3^WT, PNPLA3^I148M, and a control luciferase vector under thyroxin-binding globulin (TGB) promoter. The TGB promoter drives the transgene expression specifically in hepatocytes and spares the other cell types. When fed with WDSW, while all groups of mice gained similar body weight, the liver weight was significantly higher in PNPLA3^I148M mice. Serum triglycerides were also increased in PNPLA3^I148M+WDSW mice without affecting either glucose or insulin tolerance. As expected, PNPLA3^I148M+WDSW mice not only developed hepatic steatosis but also demonstrated features of accelerated NASH and fibrosis (within 8 weeks as compared with > 16 weeks for DIAMOND mice with WT PNPLA3). PNPLA3^I148M mice on standard chow diet and normal water, however, remained normal, reconfirming that PNPLA3^I148M alone is not sufficient to drive the pathology and rely on the environmental factors to manifest its effects.

Silencing of PNPLA3 using small interfering RNA (siRNA)–lipid nanoparticles prevented the development and progression of NASH and fibrosis in PNPLA3^I148M+WDSW mice. The mice were subjected to siRNA for 1 month after the mice were fed WDSW for 4 weeks (siRNA from 4 to 8 weeks). Although it is important that the siRNA prevented the disease progression, it would have been interesting to test its effect therapeutically, starting the siRNA at 8 weeks following WDSW.

Integrated analyses of RNA sequencing and metabolomics revealed that PNPLA3^I148M predominantly altered polyunsaturated fatty acid (PUFA), glutathione, and sphingolipid metabolism during NASH and fibrosis stages. Both PNPLA3^I148M and PNPLA3^WT led to depletion of n3 and n6 PUFAs (e.g., docosahexanoic acid), but the effect was more pronounced in PNPLA3^I148M mice. The loss of PUFAs was paralleled by an increased unsaturation of diacylglycerol and triacylglycerol, implying that the transfer of unsaturated fatty acids to other lipid molecules might be influenced by PNPLA3^I148M. De novo lipogenesis pathways, however, were not affected by PNPLA3^I148M variant.

Sphingolipid metabolism, particularly ceramide metabolism and signaling, was up-regulated in PNPLA3^I148M +WDSW mice. Although WDSW increased ceramide levels in all groups, mice expressing PNPLA3^I148M were affected more than others with increased gene expression of key enzymes involved in the ceramide biosynthetic pathway. PNPLA3 silencing decreased the same genes, indicating that the increased flux of ceramide biosynthetic pathway is a direct consequence of PNPLA3^I148M accumulation.

Additional pathway analyses revealed that PNPLA3^I148M expression leads to excessive endoplasmic reticulum (ER) stress along with oxidative stress as indicated by increased p-JNK, Atf4, and Chop levels. Both ER and oxidative stress are known to promote NASH and fibrosis. However, it is not clear whether the ER and oxidative stress are due to a direct consequence of PNPLA3^I148M expression and accumulation. It will be interesting to investigate how compounds that ameliorate oxidative and ER stress (e.g., antioxidants, tauroursodeoxycholic acid) affect NASH and fibrosis in PNPLA3^I148M+WDSW mice.

Because NASH is an inflammatory disease, it is expected that multiple inflammatory pathways will be up-regulated in PNPLA3^I148M+WDSW mice. Indeed, both innate and adaptive arms of the immune system were activated in these mice, hallmarked by up-regulation of TNFα, IL-6, signal transducer and activator of transcription (STAT) 1, RIG-1 (retinoic acid inducible gene-1), and MHC-II (major histocompatibility complex II) pathway. Gut–liver axis plays a pivotal role in hepatic inflammation and NASH progression. Microbial dysbiosis, intestinal barrier disruption, and translocation of microbial products to the liver leads to activation of toll-like receptors and hepatic inflammation.⁽⁸⁾ Lipotoxicity-induced hepatocyte death also leads to macrophage activation and liver inflammation, but this phase needs longer time to ensue as compared with microbial dysbiosis and gut leakiness, which can occur relatively rapidly. Whether PNPLA3^I148M+WDSW mice had altered gut micro-biome compared with the controls on the same diet, and the status of intestinal permeability, remain to be investigated. However, activation of the Jak-STAT3 pathway in the liver appears to be one of the key pathways that is regulated by the PNPLA3^I148M variant. Several cytokines and growth factors such as that of the IL-6 family (IL-6, IL-11, oncostatin, and LIF), IL-22, and EGF, along with ceramides, can activate STAT3. While silencing of PNPLA3^I148M with siRNA suppressed STAT3 activation, how this is achieved will be an interesting avenue to explore in future studies.

HSCs are the key cell types that secrete collagen when activated and are responsible for progressive liver fibrosis. To explore how PNPLA3^I148M-variant expression in hepatocytes influences HSCs, the authors used HepG2 cells, which are known to carry the PNPLA3^I148M variant. The addition of high glucose–containing media (4.5 g/L) to HepG2 cells induced the expression of PNPLA3^I148M, and when the conditioned media from the HepG2 cells were added to a human HSC cell line (LX2), it activated collagen transcript levels. Interestingly, when HepG2 cells were pretreated with a STAT3 inhibitor, the conditioned media failed to activate profibrotic genes in LX2 cells. Although interesting, a few points needs to be addressed: (1) how high glucose concentration in the conditioned media might have affected HSC activation; (2) inclusion of controls to account for the carryover of STAT3 inhibitors in the conditioned media, and whether the STAT3 inhibitors directly suppressed the LX2 fibrogenic gene signature; (3) validation of these experiments with primary hepatocytes isolated from the PNPLA3^I148M+WDSW mice and HSCs from WT mice; and (4) whether the presence of PNPLA3^I148M variant in HSC has a cell-intrinsic role in its activation needs further elucidation.⁽⁹⁾

In summary, the findings from this study open up multiple avenues to further explore and shed light on key aspects of metabolomic and transcriptomic changes that are imparted by hepatocyte-specific expression and silencing of human PNPLA3^I148M in a mouse model of NASH and fibrosis, which had not been studied before. There is emerging consensus regarding the pathogenic role of PNPLA3 in fibrosis progression in NASH. Silencing PNPLA3 by siRNA-based therapies are currently being explored as potential treatment in NASH-related fibrosis.⁽¹⁰⁾ Improved understanding of metabolomics and transcriptomic signature of PNPLA3-associated NASH will open up new targets for therapies as well as novel biomarkers of target engagement and early prediction of treatment response in NASH.

Abbreviations:

ER

endoplasmic reticulum

I

isoleucine

LX2

human HSC cell line

M

methionine

PNPLA3

patatin-like phospholipase domain-containing 3

PUFA

polyunsaturated fatty acid

siRNA

small interfering RNA

STAT

signal transducer and activator of transcription

WDSW

Western diet with sugars in drinking water

WT

wild-type