Mtfp1 ablation enhances mitochondrial respiration and protects against hepatic steatosis

Hepatic steatosis is the result of an imbalance between nutrient delivery and metabolism in the liver. It is the first hallmark of Non-alcoholic fatty liver disease (NAFLD) and is characterized by the accumulation of excess lipids in the liver that can drive liver failure, inflammation, and cancer. Mitochondria control the fate and function of cells and compelling evidence implicates these multifunctional organelles in the appearance and progression of liver dysfunction, although it remains to be elucidated which specific mitochondrial functions are actually causally linked to NAFLD. Here, we identified Mitochondrial Fission Process 1 protein (MTFP1) as a key regulator of mitochondrial and metabolic activity in the liver. Deletion of Mtfp1 in hepatocytes is physiologically benign in mice yet leads to the upregulation of oxidative phosphorylation (OXPHOS) activity and mitochondrial respiration, independently of mitochondrial biogenesis. Consequently, hepatocyte-specific knockout mice are protected against high fat diet-induced hepatic steatosis and metabolic dysregulation. Additionally, we find that deletion of Mtfp1 in liver mitochondria inhibits mitochondrial permeability transition pore opening in hepatocytes, conferring protection against apoptotic liver damage in vivo and ex vivo. Our work uncovers novel functions of MTFP1 in the liver, positioning this gene as an unexpected regulator of OXPHOS and a therapeutic candidate for NAFLD.


Introduction
Non-alcoholic fatty liver disease (NAFLD) is the most common chronic liver disease in both oxygen consumption rates (JO 2 ) and mitochondrial membrane potential (ΔΨ) changes with Rhodamine 123 (RH-123) 22 . JO 2 and RH-123 were recorded from mitochondria incubated with 127 respiratory substrates promoting the delivery of electrons to complex I (state 2; pyruvate, 128 glutamate, and malate (PGM) or complex II (state 2; succinate and rotenone), and also in the 129 presence of palmitoyl-carnitine plus malate then in the phosphorylating (state 3: ADP), non-130 phosphorylating (state 4: oligomycin (OLGM) to inhibit ATP synthase) ( Fig. 2A). Notably, and in 131 contrast to MTFP1-deficient cardiac mitochondria 15 , respiration in LMKO liver mitochondria was 132 significantly increased in phosphorylating (state 3) conditions in the presence of any of the 133 respiratory substrates we tested. Pyruvate, glutamate, and malate led to a 49% increase in state 3 134 respiration and succinate and rotenone led to a 57% increase in respiration (Fig. 2B). Interestingly,

135
we observed a 200% increase in state 3 respiration in the presence of palmitoyl carnitine, a fatty 136 acid ester derivative, pointing to an increased efficiency of fatty-acid derived energy metabolism 137 caused by hepatocyte-specific deletion of Mtfp1.

138
Additionally, LMKO liver mitochondria showed a higher respiratory control ratio (RCR) in 139 the presence palmitoyl carnitine plus malate (Fig 2C). Despite a marked increase in state 3 140 respiration, we did not observe a genotype-specific difference in mitochondrial membrane potential 141 ( Fig. 2D), which initially surprised us since increased oxygen consumption rates are typically 142 accompanied by reduction in membrane potential due to dissipation of the protonmotive force via 143 complex V (to synthesize ATP). The most parsimonious explanation for this result is that MTFP1 144 ablation promotes a commensurate increase in the activities of both cytochrome c oxidase 145 (complex IV) and the ATP synthase (complex V). Indeed, when we measured the specific activities 146 of complex IV (Fig. 2E) and complex V (Fig. 2F) in separate assays, we found a ~20% increase in 147 LMKO mitochondria relative to control littermate controls, suggesting a similar contribution of both 148 complexes to increase respiration while maintaining mitochondrial membrane potential. We further 149 confirmed our findings by measuring oxygen consumption rates in mitoplasts supplied with either 150 NADH, Cyt c, succinate and rotenone to drive electron transport via Complex II or NADH, Cyt c, 151 succinate and malonate to drive electron transport via Complex I (Fig. 2G). In both assays, oxygen 152 consumption was elevated in LMKO liver mitochondria demonstrating that Complex IV activity is 153 intrinsically augmented upon MTFP1 ablation independently of the protonmotive force. Together, components of the OXPHOS machinery or the induction of a general mitochondrial biogenesis 156 response. To differentiate between these possibilities, we assessed mitochondrial mass using Interestingly, we found a modest increase in mitochondrial elongation in LMKO mitochondria (Fig.

175
To gain insights into the mechanisms responsible for the increased specific activities, we 176 assessed the relative complex abundance by grouping mitochondrial proteins quantified by 177 proteomics in the liver of NCD-fed LMKO mice according to the macromolecular complexes to 178 which they belong (Fig. 2H, Supplemental Dataset 2). These data revealed a significant increase 179 in Complex V subunits, which could be confirmed by quantitative SDS-PAGE (Fig. 2I) and Blue-180 native polyacrylamide gel electrophoresis (BN-PAGE) analyses (Fig. 2J). BN-PAGE analysis of the steady-state levels of OXPHOS complexes in LMKO liver mitochondria revealed an increase in activity (Fig. 2F). Taken together, our data suggest that improved assembly and/or maintenance of Complex V along with an increased activity of Complex IV is responsible for the enhanced 185 respiration observed in LMKO liver mitochondria.

186
To gain insights into the relationship between MTFP1 and macromolecular complex 187 assembly in the inner mitochondrial membrane (IMM), we sought to assess the interactome of 188 MTFP1 in the liver. We generated a hepatocyte-specific transgenic mouse model enabling the 189 expression of FLAG-MTFP1 from the Rosa26 locus (Fig. 3A). We verified that the Hepatocyte FLAG-190 MTFP1 mice, expressed FLAG-MTFP1 correctly in the IMM by protease protection assay (Fig. 3B).

191
We then subjected liver mitochondria to co-immunoprecipitation (co-IP) and mass spectrometry

205
which was absent in LMKO mice, and that co-migrated with SLC25A4 (ANT1) (Fig. 3E), as we 206 recently described in the heart 15 . However, we found little overlap with the cardiac interactome of 207 MTFP1 15 (7 out of 113 proteins), implying that there may be tissue-specific physical interactions 208 and complex assembly associated with MTFP1 beyond ANT1 that regulate respiration. Altogether,

209
these data indicate that post-transcriptional modulation of mitochondrial gene expression and IMM 210 complex abundance and activities are induced in a specific manner by the deletion of MTFP1 in 211 hepatocytes, leading to enhanced mitochondrial respiration.

212
Given enhanced stimulated respiratory activity of LMKO liver mitochondria, we next S4C).

242
Given the protection against hepatocyte and liver damage observed in vivo and in vitro 243 upon the deletion of Mtfp1, we asked whether the protection against HFD-induced metabolic 244 dysregulation of the liver could be explained by differential sensitivity to cell death. However,

245
TUNEL assays performed on histological sections from HFD-fed control and LMKO mice revealed 246 an absence hepatocyte apoptosis (Fig. 4I), which is consistent with previous findings that HFD 247 causes limited liver cell death 25 . Altogether, our data demonstrate that Mtfp1 deletion in 248 hepatocytes confers metabolic resistance to hepatic steatosis in vivo (Fig. 4J) in a manner that is 249 independent of apoptotic resistance.

252
Having demonstrated that LMKO mice are protected against diet-induced steatosis, we

258
suggests that LMKO mice metabolize nutrient-derived lipids more efficiently than control 259 littermates, which can very likely be attributed to intrinsic differences in the livers of these mice.

260
Food and water intake were not altered between control and LMKO mice nor were distance 261 measurement on HFD ( Fig. S5A-C), pointing to a specific effect of MTFP1 on energy expenditure 262 by the liver that is revealed under HFD feeding. Metabolic dysregulation caused by HFD feeding is 263 known to increase de novo glucose synthesis by the liver of rodents 27 and so we sought to assess 264 gluconeogenesis by performing intraperitoneal pyruvate tolerance tests (IP-PTT). We observed a 265 1.5-fold increase blood glucose levels in HFD-fed control mice relative to NCD littermates, which 266 was reduced by 14% in the HFD-fed LMKO group (Fig. 5B), indicating that Mtfp1 deletion in 267 hepatocytes protects the liver against diet-induced dysregulation of gluconeogenesis. Consistent 268 with these findings, intraperitoneal glucose tolerance tests (IP-GTT) revealed LMKO mice to be To exclude that improved glucose tolerance in HFD-fed LMKO mice was the consequence of 273 altered insulin resistance, rather than improved gluconeogenesis, we performed intraperitoneal 274 insulin tolerance tests (IP-ITT). We observed no genotype-specific differences in insulin sensitivity 275 ( Fig. 5D) on either HFD or NCD, indicating that improved glucose tolerance in LMKO mice is not 276 caused by increased systemic insulin sensitivity. Concordantly, we did not observe differences in 277 basal insulin levels between control and LMKO mice fed a NCD (Fig S5E). While LMKO mice 278 gained less weight upon HFD feeding (Fig. 4C), we observed no genotype-specific differences in

286
To determine whether hepatocyte deletion of Mtfp1 prevents steatosis in a cell-autonomous 287 manner, we established an assay in which we could mimic HFD-induced steatosis in primary 288 hepatocytes isolated from NCD-fed mice. Primary hepatocytes were plated and cultured in the 289 presence of Intralipid (IntLip): a complex lipid emulsion composed of linoleic, oleic, palmitic, and 290 stearic acids, which are the most abundant fatty acids found in HFD. We optimized our assay 291 conditions to ensure that limited damage, death, or differentiation under both treated and non-292 treated (NT) conditions was occurring. Indeed, IntLip feeding of primary hepatocytes isolated from 293 NCD-fed control mice led to a 1.6-fold increase in intracellular lipid accumulation after 24 hours, 294 which could be visualized by live-cell imaging with BODIPY fluorescence (Fig. 5H) and quantified 295 by high-throughput confocal imaging ( Figure 5I)      genes that were targeted in previous studies combined with the potential for metabolic cross-talk 349 between tissues in vivo has made it challenging to pinpoint how modulating mitochondrial activity 350 may be used to combat liver disease. Here, we discovered that Mitochondrial Fission Process 1 351 (MTFP1) plays an important role metabolic role in the liver of mice that is critical for NAFLD but not 352 under basal conditions. MTFP1 was first identified as a metabolically-regulated inner mitochondrial 353 membrane IMM) protein [16][17][18]33 initially implicated in mitochondrial fission and cell death resilience in a variety of cell lines [16][17][18]38 . In the liver, the discovery that MTFP1 protein expression is predictive of hepatocellular carcinoma survival and recurrence risk in humans 15 , prompted us to directly investigate its role in the liver.

407
The discovery that MTFP1 ablation in the liver is benign and even beneficial under stress 408 conditions is in stark contrast to our recent report that Mtfp1 deletion in cardiomyocytes causes which is significantly smaller than the fully assembled mito-ribosome and ATP synthase complexes with whose subunits it interacts and thus we posit that the impacts on mitochondrial function that by Ppif), both of which have been previously physically and functionally connected to mitochondrial 447 ATP synthesis 62-66 and the mPTP. Our data also support a role for MTFP1 as a regulator of mPTP 448 activity in the liver, as inactivation of hepatocytes slows mPTP opening response in liver 449 mitochondria and protects against liver cell death in primary hepatocytes and in mice (Fig. 6). This 450 pro-survival role of MTFP1 in the liver appears to be independent from the metabolic protection its 451 deletion confers against HFD-induced hepatic steatosis, since HFD-feeding does not cause cell 452 death (Fig. 4I).

453
The metabolic protection conferred to MTFP1-deficient livers that are already hyper-

549
Triglyceride assay: snap frozen liver tissue was used to assess hepatic triglyceride content by

577
The cytochrome c oxidase (COX) activity was measured in the presence of Ascorbate-Na (2mM;

757
Maximum peptide charge was set to 7 and 5 amino acids were required as minimum peptide 758 length. A false discovery rate of 1% was set up for both protein and peptide levels. The iBAQ 759 intensity was used to estimate the protein abundance within a sample.

760
Quantitative analysis was based on pairwise comparison of intensities. Values were log-761 transformed (log2). Reverse hits and potential contaminant were removed from the analysis.

762
Proteins with at least 2 peptides (including one unique peptide) were kept for further statistics.

766
They have therefore been set aside and considered as differentially abundant proteins.

798
For RT-qPCR, 1 µg of total RNA was converted into cDNA using the iScript Reverse Transcription Briefly, a catheter (22G feeding needle) was connected to a pump and inserted into the vena cava.  Table 4.

930
The graph represents the relative values of each complex ratio between LMKO and controls.