Conditional deletion of CEACAM1 causes hepatic stellate cell activation

Objectives: Hepatic CEACAM1 expression declines with advanced hepatic fibrosis stage in patients with MASH. Global and hepatocyte-specific deletions of Ceacam1 impair insulin clearance to cause hepatic insulin resistance and steatosis. They also cause hepatic inflammation and fibrosis , a condition characterized by excessive collagen production from activated hepatic stellate cells (HSCs) . Given the positive effect of PPAR on CEACAM1 transcriptoin and on HSCs quiescence, the current studies investigated whether CEACAM1 loss from HSCs causes their activation.

We examined whether lentiviral shRNA-mediated CEACAM1 donwregulation (KD-LX2) activates cultured human LX2 stellate cells . We also generated LratCre+Cc1fl/fl mutants with conditional Ceacam1 deletion in HSCs and characterized their MASH phenotype. Media transfer experiments were employed to examine whether media from mutant human and murine HSCs activate their wild-type counterparts.

LratCre+Cc1fl/fl mutants displayed hepatic inflammation and fibrosis but without insulin resistance or hepatic steatosis. Their HSCs, like KD-LX2 cells, underwent myofibroblastic transformatio n and their media activated wild-type HDCs. This was inhibited by nicotinic acid treatment which stemmed the release of IL-6 and fatty acids, both of which activate the epidermal growth factor receptor (EGFR) tyrosine kinase. Gefitinib inhibition of EGFR and its downstream NF-B/IL-6/STAT3 inflammatory and MAPK-proliferation pathways also blunted HSCs activation in the absence of CEACAM1.

Loss of CEACAM1 in HSCs provoked their myofibroblastic transformation in the absence of insulin resistance and hepatic steatosis. This response is mediated by autocrine HSCs activation of the EGFR pathway that amplifies inflammation and proliferation.

Hepatic fibrosis, Inflammation, Hepatic steatosis, Stellate cell proliferation, Retinoic acid was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (whichthis version posted April 3, 2024. ; https://doi.org/10.1101/2024.04.02.586238doi: bioRxiv preprint 1. Introduction Metabolic-Associated Steatotic Liver Disease (MASLD), form erly termed non-alcoholic fatty liver disease, currently represents the most common cause of chronic liver disease worldwide [1]. MASLD spans a broad spectrum of metabolic disease with hepatic fibrosis defining its most aggressive form, Metabolic - Associated Steatohepatitis (MASH), together with inflammation, hepatocyte damage, and apoptosis. Hepatic fibrosis is on the rise and currently constitutes a leading etiology in patients with MASH, partly because of limited targeted therapy [2; 3] . This necessitates the need for further studies exploring its molecular and cellular basis. Histologically, hepatic fibrosis in patients with MASH is characterized by early lesions of perisinusoidal collagen deposition, followed by portal and eventually , bridging fibrosis [4]. It implicates the activation of hepatic stellate cells (HSCs) located in the Space of Disse between liver sinusoidal endothelial cells (LSECs) and hepatocytes . HSCs represent approximately 10% of resident liver cells . In healthy liver, PPAR activation maintains HSCs quiescent and containing large lipid droplets filled with vitamin A as retinyl esters (RE), triacylglycerols (TG) and cholesteryl esters (CE) [5]. Following transdifferentiation into proliferative, contractile, inflammatory myofibroblasts with enhanced extracellular matrix ( ECM) production, HSCs lose their retinoid content [6]. This is associated with reduced PPAR and reciprocal elevation in the level of PPAR [7; 8], which could be activated by all trans-retinoic acid [9] and PUFA [10] to increase HSCs proliferation via inducing the p38 and JNK MAPK pathways [8]. Further studies are needed to fully identify the factors that cause HSCs activation [11]. Virtually every liver cell contributes to HSCs activation, and they all express the Carcinoembryonic Antigen-related Cell Adhesion Molecule 1 (CEACAM1), with a dominant expression in hepatocytes where it promotes insulin clearance. Depletion of Ceacam1 gene globally [12] or exclusively in hepatocytes [13], causes chronic hyperinsulinemia, emanating chiefly from reduced insulin clearance, followed by hepatic insulin resistance, steatohepatitis and visceral obesity. It also provokes HSCs activation and a characteristic MASH-like fibrosis [14; 15]. Fed a high-fat diet, mice lacking CEACAM1 in hepatocytes develop advanced hepatocellular injury accompanied by chicken-wire fibrosis and apoptosis [14; 16]. Reciprocally, liver-specific was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (whichthis version posted April 3, 2024. ; https://doi.org/10.1101/2024.04.02.586238doi: bioRxiv preprint rescuing of CEACAM1 reverses metabolic dysregulation and hepatic fibrosis in global Cc1–/– null mice [14]. In contrast, CEACAM1 loss in endothelial cells promotes hepatic fibrosis, driven by increased production of endothelin1, without insulin resistance or hepatic steatosis [17]. Consistent with these data in genetically- modified mice, patients with MASH exhibit a progressive loss of CEACAM1 in liver [15] and particularly in LSECs [17] as the disease advances. CEACAM1 is also expressed in pericytes [18], including HSCs [19]. Herein we investigated whether its loss of CEACAM1 in HSCs induces their activation and sought to uncover underlying mechanisms. was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (whichthis version posted April 3, 2024. ; https://doi.org/10.1101/2024.04.02.586238doi: bioRxiv preprint 2. Materials and methods 2.1. Generation and metabolic phenotyping of LratCre+Cc1fl/fl mice As detailed in Supplemental data, Cc1loxp/loxp mice were crossed with LratCre transgenic mice expressing a Cre recombinase driven by mouse lecithin-retinol acyltransferase (Lrat) promoter [20]. Stellate cell-specific deletion of Ceacam1 in C57BL/6Jxhomozygotes ( LratCre+Cc1fl/fl) was confirmed by PCR reaction using gene-specific primers (Fig. S1). As littermate controls, this study used homozygotes of wild- type Ceacam1 allele with (LratCre+Cc1+/+) or without Cre (LratCre–Cc1+/+), and homozygotes of Ceacam1- floxed allele without Cre (LratCre–Cc1fl/fl) to rule out potential confounding effects of floxing and introducing Cre recombinase. Per institutiona lly approved protocols, a nimals were housed in a 12 -h dark -light cycle and fed standard chow ad libitum. Male mice were kept in cages with Alpha-dri bedding before undergoing metabolic phenotyping [intraperitoneal (IP) glucose and insulin tolerance tests -GTT and ITT, respectively]. Following recovery, mice were fasted for 18hrs, anesthetized with an IP injection of pentobarbital (1.1mg/kg BW), and their retro-orbital venous blood was drawn and tissues extracted for biochemical evaluation (Supplemental data). 2.2. Liver histology and immunohistochemical analysis As detailed in Supplemental data, fixed liver sections were stained with hematoxylin-eosin (H&E) or with 0.1% Sirius Red stain to evaluate hepatic fibrosis. Images were taken using Nikon Eclipse 90i Microscope and 10 randomly selected high power fields (20X) per sample were imaged with ImageJ (v1.53t) to quantify Sirius Red stain as %area [17]. For immunohistochemical (IHC) analysis, liver sections underwent antigen-retrieval, blocking with rabbit or mouse serum, stained overnight at 4°C with specific antibodies, blotted with species-specific biotinylated secondary antibodies before being hematoxylin-counterstained [17]. Images were taken using was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (whichthis version posted April 3, 2024. ; https://doi.org/10.1101/2024.04.02.586238doi: bioRxiv preprint Nikon Eclipse 90i Microscope and evaluated blindly to count positively stained cells in 5 fields/mouse at 40X magnification. 2.3. Isolating HSCs from human subjects Human HSCs were isolated , as described [21]. Briefly, dissociated parenchymal cells were suspended in 2mM EDTA buffer containing 5%FBS (Gibco) for 30min in the presence of anti-CD32 (Abcam, Cambridge, MA) and anti-CD45 (BD Biosciences, San Jose, CA). Quiescent HSCs (qHSCs) were sorted as CD32-CD45-UV+ cells using FASCAria (BD Biosciences). Activated HSC s (aHSCs) were obtained by plating qHSCs in DMEM (Gibco) supplemented with 20% FBS, 20ng/mL epidermal growth factor (EGF) and 10ng/mL fibroblast growth factor 2 (Peprotech, London, UK) , 100μM oleic acid , 100μM palmitic acid, and 5μM retinol (Sigma-Aldrich). RNA was isolated using RNeasy Micro Kit (Qiagen GmbH, Hilden, Germany). RNA samples were amplified using the Ovation Pico WTA system V2 (Tecan Genomics, San Carlos, CA). 2.4. Media transfer experiments in primary murine hepatic stellate cells Primary HSCs were isolated from ≥8-month-old control LratCre–Cc1fl/fl (recipient cells) and mutant mice LratCre+Cc1fl/fl (donor cells) [22]. Cells were cultured in 12-well-plates for 5 days. LratCre+Cc1fl/fl donor cells were washed twice and incubated in phenol red-free DMEM before treating with 500µM nicotinic acid (NA) (Sigma-Aldrich) or buffer alone for 24hrs. Media were collected, centrifuged at 380g for 3min to remove cell debris and the “conditioned media” were transferred to the twice-washed LratCre– Cc1fl/fl recipient HSCs. 24hrs later, cells were lysed for mRNA analysis (see below). In some experiments, 10µM Gefitinib (Sigma-Aldrich), an EGFR tyrosine kinase inhibitor [23], or dimethyl sulfoxide (DMSO- vehicle) were added to recipient cells for additional 24hrs before cell lysis. Media levels of free glycerol (Glycerol Assay Kit MAK117-1KT, Sigma-Aldrich), interleukin-6 (ELISA Kit, ab222503, Abcam) and TNF (ELISA Kit, ab100747, Abcam) were determined per manufacturer instructions [17]. was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (whichthis version posted April 3, 2024. ; https://doi.org/10.1101/2024.04.02.586238doi: bioRxiv preprint 2.5. Experiments with LX2 cells with stable downregulation of human CEACAM1 expression The immortalized human hepatic stellate LX2 cell line was infected with a human CEACAM1 shRNA lentiviral construct to establish a KD line with stable knockdown of hCEACAM1 and scramble control (Scr), as detailed in Supplemental data. KD-LX2 and Scr-LX2 cells were treated with DMSO, 5μM retinoic acid and/or 1μM Rosiglitazone (Sigma-Aldrich) for 24hrs before cell lysis and qRT-PCR analysis [16]. For lipid analysis, KD-LX2 and Scr-LX2 cells were seeded in 6-well-plates (4x104 cells/well) for 48hrs before being stained with Nile Red (Sigma-Aldrich) and evaluated with densitometry by ImageJ software to measure lipid content [24]. Media was collected to determine free glycerol levels using Glycerol Assay Kit (BioVision, Milpitas, CA) [24]. Media transfer from KD-LX2 to Scr-LX2 controls was performed as above and levels of free glycerol (MAK117-1KT, Sigma-Aldrich), interleukin-6 (ELISA-ab 178018, Abcam) and TNF (ELISA- ab181421, Abcam) were determined per manufacturer instructions. Cell growth was determined by MTT assay (Sigma-Aldrich) and absorbance read at 570nm in 96- well plates. Cell growth was calculated as percent of growth in the presence of effector minus basal growth divided by maximum growth in complete medium. 2.6. Immunoprecipitation and Western blot analysis As previously described [17], c ells were Triton-lysed and subjected to SDS -PAGE followed by Western blot analysis using antibodies as listed in Supplemental data. Proteins were detected by chemiluminescence, scanned and their density normalized against tubulin (Cell Signaling) or the total amount of proteins of the signaling molecule applied on parallel gels. For immunoprecipitation, 100µg of protein lysates were precleared with 20µl mixture of protein G and A sepharose beads (Invitrogen, Carlsbad, CA) at 40C for 2hrs. Proteins were immunoprecipitated from the precleared lysates by incubation with 2µg of their specific antibodies overnight at 40C, centrifuged and analyzed by SDS-PAGE and Western blot analysis. was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (whichthis version posted April 3, 2024. ; https://doi.org/10.1101/2024.04.02.586238doi: bioRxiv preprint 2.7. Quantitative real-time-PCR (qRT-PCR) Total RNA was isolated with PerfectPure RNA Tissue Kit (Fisher Scientific, Waltham, MA). cDNA was synthesized by iScript cDNA Synthesis Kit (Bio-Rad), using 1μg of total RNA and oligodT primers (Table S1). cDNA was evaluated with qRT-PCR (StepOne Plus, Applied Biosystems, Foster City, CA), and mRNA was normalized to GAPDH, unless otherwise mentioned. 2.8. Statistical analysis Data were analyzed using one-way ANOVA analysis with Bonferroni correction or two-tailed Student- t-test using GraphPad Prism 6 software. Data were presented as means±SEM. P<0.05 was considered statistically significant. was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (whichthis version posted April 3, 2024. ; https://doi.org/10.1101/2024.04.02.586238doi: bioRxiv preprint 3. Results 3.1. CEACAM1 expression is induced by rosiglitazone and retinoic acid Rosiglitazone (Ro) elevated CEACAM1 mRNA levels by ~2-to-3–fold in human LX2 HSCs (Fig. 1A). This likely stems from the transcriptional activation of Ceacam1 promoter by the binding of liganded PPAR to the functional and well-conserved PPAR response element/retinoic acid receptor recognition site (PPRE/RXR) between nts–557 and –543 in Ceacam1 promoter [25]. A similar effect was exerted by retinoic acid (RA) either alone or combined with rosiglitazone (Fig. 1A). This resulted from the transcriptional activation by RA, as tested by a luciferase reporter assay in human HepG2 cells transfected with mouse Ceacam1 promoter [25] (Fig. 1B). As shown, Ro and RA induced the luciferase activity of PPREx3TK-Luc and PGL3-RARE-Luc (positive controls for Ro and RA, respectively) and of the mouse Ceacam1 promoter (–1100pLuc), individually or combined, without affecting that of the empty vector (PGL4.10) [Fig. 1B, *P<0.05 versus vehicle-treated (–)]. Mutating the PPRE/RXR sequence (–∆2), PPRE (–∆PPRE) or RXR (–∆RXR) abolished the positive effect of Ro and RA on Ceacam1 promoter activity. 3.2. Loss of CEACAM1 activates human LX2 stellate cells Activated primary human HSCs (aHSCs) exhibited lower (by >80%) CEACAM1 mRNA levels relative to quiescent cells (qHSCs) (Fig. 1C). To test whether CEACAM1 loss mediated HSC activation, we examined whether lentiviral shRNA-mediated repression of CEACAM1 by >90% (Fig. 2A.i and ii) could activate LX2-HSCs. Consistent with the loss of lipid content during HSC activation [26-28], knocking down CEACAM1 markedly reduced Nile red-stained fat-laden droplets relative to scrambled controls (Fig. 2B.i KD vs Scr in the graphical presentation of densitometry analysis). The lost cellular fat was recovered as free glycerol in the KD-LX2 culture media (Fig. 2B.ii). As predicted based on the known features of HSC activation, KD-LX2 cells exhibited reduced retinyl ester (RE) synthesis, as assessed by lower mRNA levels of enzymes catalyzing RE synthesis [lecithin-retinol acyltransferase (LRAT)] and lipolysis [lysosomal acid lipase-LAL (LIPA) [29]] (Fig. 2B.iii). Reciprocally, they manifested elevated PUFA-triacylglycerol was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (whichthis version posted April 3, 2024. ; https://doi.org/10.1101/2024.04.02.586238doi: bioRxiv preprint (PUFA-TG) synthesis [27], as suggested by their higher ratio of the mRNA of PUFA-specific fatty acid-CoA synthase 4/non-specific ACSL1 (ACSL4)/ACSL1 [30], and their two-to-threefold higher mRNA levels of lipogenic genes such as SREBP-1c and fatty acid synthase (FASN), DGAT1 (the last enzyme in TG synthesis) and ATGL (TG lipase). Activated HSCs undergo proliferation and resist apoptosis [26; 31]. Consistently, knocking down CEACAM1 markedly increased LX2 proliferation, as assessed by MTT assay (Fig. 2C.i). It also led to ~2- fold higher mRNA levels of -smooth muscle actin (-SMA or ACTA2) and COL11, markers of mesenchymal cell activation (myofibroblastic transformation) [26] (Fig. 2C.ii). The latter could result from increased activation of TGF canonical signaling pathway, as assessed by Western blot analysis of phosphorylated Smad2/3 (Fig. 2D). This demonstrated that CEACAM1 loss activated KD-LX2 cells. 3.3. Delineating the mechanism underlying LX2 activation by CEACAM1 deletion Following phosphorylation by epidermal growth factor (EGFR) and insulin (IR) receptors, CEACAM1 sequesters Shc and reduces its coupling to the receptors to suppress downstream Shc/MAPK- mediated cell growth and proliferation pathways [32; 33]. Consistently, insulin (100 nM) treatment for 5 min stimulated IR and MAPK phosphorylation in both groups of cells (Fig. S2B-C, + vs – insulin), and induced their proliferation, as assessed by MTT assay (Fig. S2D, + vs – insulin). In the absence of insulin, KD-LX2 cells manifested a higher basal phosphorylation of MAPK, but not IR (Fig. S2B-C), in parallel to higher cell growth relative to their Scr-LX2 controls (Fig. S2D). This suggests that an IR-independent pathway was implicated in KD-LX2 basal activation. Because lipolysis-derived FAs activate EGFR [33], which contributes to HSC activation [34], we then examined whether the FAs released from KD-LX2 cells [and captured as free glycerol in their media (Fig. 3A.i, black vs white bar)], could activate EGFR pathways to induce HSC myofibroblastic transformation. As Fig. 3B.i shows, incubating Src-LX2 in the conditioned media of KD-LX2 cells (Scr/Cond) markedly reduced CEACAM1 (CC1) mRNA levels (~75%) (grey vs white bar; P<0.05). This likely resulted from increased PPAR expression (Fig. 3B.ii) and its activation by the released FAs. was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (whichthis version posted April 3, 2024. ; https://doi.org/10.1101/2024.04.02.586238doi: bioRxiv preprint Consistently, blocking lipolysis by nicotinic acid (NA) normalized FAs content in KD-LX2 media (Fig. 3A.i, diagonally hatched vs white and vertically-hatched bars) and subsequently, restored CEACAM1 mRNA levels in Scr/Cond (Figs. 3B.i, horizontally hatched vs vertically hatched bar). Western blot analysis revealed higher EGFR phosphorylation in KD and Src/Cond cells in the absence of NA (Fig. 3C.i, – lanes 1 and 5 vs lane 3), but not in its presence (Fig. 3C.i, + vs – lanes/cell group). This demonstrated that EGFR was activated in response to FA-containing conditioned media. Consistent with reduced CEACAM1 level, CEACAM1/Shc binding was lower in Scr/Cond than Scr cells, as demonstrated by its repressed detection in the Shc immunopellet (Fig. 3C.ii, – lane 5 in Scr/Cond vs – lane 3 in Scr). This led to a reciprocal recovery of Shc in the EGFR immunopellet of Scr/Cond relative to Scr cells (Fig. 3C.iii, – lane 5 in Scr/Cond vs – lane 3 in Scr), and activation of downstream MAPK pathways (Fig. 3C.iv, – lane 5 in Scr/Cond vs – lane 3 in Scr) and NF-kB (Fig. 3C.v, – lane 5 in Scr/Cond vs – lane 3 in Scr). This induced cell proliferation, as assessed by elevated PCNA protein levels (Fig. 3C.vi, – lane 5 in Scr/Cond vs – lane 3 in Scr) and MTT assay (Fig. 3B.v, grey vs white bar). Additionally, PPAR1 mRNA levels were lowered in Scr/Cond-LX2 cells (Fig. 3B.iii, grey vs white bar), as expected during HSC activation and in contrast to the rise in PPAR levels [35], which was likely activated by the excess FAs produced in KD-LX2 and Src/Cond-LX2 cells (Table S2). Consistent with increased myofibroblastic transformation, ACTA2 mRNA levels were induced by ~two-to-threefold in KD and Scr/Cond LX2 cells (Fig. 3B.iv, grey and black vs white bar). Reversal of these processes in Scr/Cond cells by NA treatment (+ vs – lanes in Scr/Cond) further supported a role for FAs release from activated KD-LX2 in the autocrine activation of EGFR-Shc-MAPK to increase HSCs proliferation and activation. Interleukin-6 (IL-6), a transcriptional target of NF-B, was also elevated in KD media (Fig. 3Aii, black vs white bar). Consistent with the anti-inflammatory effect of NA [36] and its inhibition of IL-6 production [37], NA treatment reversed IL-6 level in KD media (Fig. 3Aii, + vs – lane) without affecting that of TNF (Fig. 3Aiii, + vs – lane). Because IL-6 transactivates EGFR [38], we then tested whether the rise in IL-6 contributed to EGFR basal activation in KD-LX2 and Scr/Cond-LX2 cells. To this aim, we carried out media transfer experiments in the absence and presence of Gefitinib, an EGFR tyrosine kinase inhibitor [23]. As Fig. 4A shows, Gefitinib inhibited EGFR phosphorylation in KD-LX2 and Scr/Cond-LX2 was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (whichthis version posted April 3, 2024. ; https://doi.org/10.1101/2024.04.02.586238doi: bioRxiv preprint cells (+ vs – lanes). In parallel, it reduced PPAR (Fig. 4B) and reciprocally induced PPAR1 mRNA levels in these cells (Fig. 4C, horizontally-hatched vs grey bar in Scr/Cond-LX2 and diagonally-hatched vs black bar in KD-LX2 cells) to stimulate their CEACAM1 mRNA levels (Fig. 4D). This was associated with the ability of Gefitinib to prevent Scr/Cond-LX2 activation, as demonstrated by reduction and normalization of ACTA2 mRNA levels (Fig. 4E) and their cell proliferation (Fig. 4F, + vs – lanes). 3.4. Activation of primary HSCs from Cc1–/– null mice In support of HSCs activation when their CEACAM1 is absent, primary HSCs from global Cc1–/– nulls exhibited higher mRNA levels of Ppar, Pcna, and Acta2 than HSCs from wild-type mice (Table S3). They also exhibited higher Srebp-1c and Fasn mRNA levels. Moreover, their media induced the mRNA levels of these genes in wild-type HSCs (Table S3). NA treatment normalized these parameters in HSCs from Cc1–/– and Cc1+/+/Cond cells (Table S3). This proposed that CEACAM1 loss in HSCs activated them and caused their myofibroblastic transformation to contribute to hepatic fibrosis in Cc1–/– nulls [14]. 3.5. LratCre+Cc1fl/fl mice with conditional deletion of Ceacam1 in HSCs are insulin sensitive Because Ceacam1 loss in endothelial cells and hepatocytes could also contribute to hepatic fibrosis in Cc1–/– nulls [14], we then assessed the effect of deleting Ceacam1 exclusively from HSCs on hepatic fibrosis. To this end, we generated LratCre+Cc1fl/fl mice with conditional deletion of Ceacam1 in HSCs, as demonstrated by their intact Ceacam1 expression in bone marrow macrophages, hepatocytes and endothelial cells (liver and heart) (Fig. 5A). These mice exhibited normal body weight, visceral fat mass, plasma NEFA and triacylglycerol levels (Table 1). LratCre+Cc1fl/fl mice exhibited normal hepatic insulin clearance (steady-state C-peptide/insulin molar ratio) and normo-insulinemia relative to their control counterparts (Table 1). They also showed normal tolerance to exogenous glucose and insulin (Fig. 5B and 5C, respectively), with normal fasting and fed blood glucose levels (Table 1). Consistent with normo- insulinemia, hepatic triacylglycerol levels were normal (Table 1) and H&E stain did not detect lipid droplet deposition in liver sections (Fig. 5D.d). Moreover, the mRNA levels of genes involved in fatty acid transport was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (whichthis version posted April 3, 2024. ; https://doi.org/10.1101/2024.04.02.586238doi: bioRxiv preprint (CD36 translocase) and lipogenesis [Srebp-1c, and fatty acid synthase (Fasn)] were normal (Table S4). Together, this demonstrated that conditional Ceacam1 deletion from HSCs did not cause insulin resistance or hepatic steatosis, consistent with intact expression of CEACAM1 in hepatocytes. 3.6. Increased inflammation in LratCre+Cc1fl/fl livers H&E stai ning indicated diffused mononuclear inflammatory foci in the liver parenchyma of LratCre+Cc1fl/fl mutants without ballooning or altered hepatocellular architecture starting at 10 months of age [Fig. 5D.d (and graph) vs Fig. S3A.d from 8-month-old mice]. Immunohistochemical (IHC) analysis revealed increased macrophage recruitment (CD68) and activation (M ac2) [Fig. 6A.i-ii (and graphs), panels d vs a-c, respectively ]. It also showed elevated immunostained myeloperoxidase (MPO) levels [Fig. 6A.iii (and graph), panel d vs a-c], which together with increased mRNA of MPO and elastase (Table S4), demonstrated an increase in neutrophil accumulation in the liver parenchyma of mutant livers. In addition to MPO, a granulocyte-specific transcription factor (STAT3) was activated (phosphorylated) at 10 months (Fig. 6B), but not at 8 months of age (Fig. S3B) . This likely resulted from the ~2-to-3-fold concomitant rise in hepatic IL-6 production (Fig. 6C vs Fig. S3C) and in the plasma levels of this pro-inflammatory cytokine (Fig. 6D). Consistent with IL-6 as a transcriptional target of NF-B, the p65(NF-B) subunit was basally activated (phosphorylated) in the livers of 10-month-old (Fig. 6B), but not 8-month-old (Fig. S3B) LratCre+Cc1fl/fl mutants. This likely resulted from reduced Shc sequestration in the absence of Ceacam1 and the reciprocal increase in its coupling to EGFR [33]. In addition to IL-6, activated p65(NF-B) could induce Mcp-1/Ccl2 transcription [39], as shown in Fig. 6C, to recruit monocytes/macrophages into active inflammatory foci in mutant livers. Together with IL-6, Ccl2 could induce CD11b+ macrophage pool (Fig. 6C) and its differentiation toward the M2 type [40; 41], which is partly mediated by elevated levels of IL- 4/IL-13 type 2 cytokines (Fig. 6C). Together with no increase in the mRNA levels of Th1-derived cytokine IFN (Fig. 6C) or in plasma TNF levels (Fig. 6D), this points to the mounting of an M2 response in mutant livers, mediated partly by sustained IL-6/STAT3 phosphorylation [42]. In contrast to macrophages, IHC was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (whichthis version posted April 3, 2024. ; https://doi.org/10.1101/2024.04.02.586238doi: bioRxiv preprint (Fig. 6A.iv-v) and qRT-PCR (Table S4) analyses revealed no significant increase in pro-inflammatory CD4+T and CD8+T lymphocytes. Moreover, there was no increase in the anti-inflammatory Treg immunostain (Foxp3) (Fig. 6A.vi), or in hepatic IL-10 expression (Table S4). Thus, liver injury in LratCre+Cc1fl/fl mice was associated with a Th2 response marked by elevated IL-4/IL-13 secretion by hepatic lymphocytes that could activate infiltrated myeloid cells (macrophages and neutrophils) to induce their M2 genes expression. 3.7. Spontaneous fibrosis in LratCre+Cc1fl/fl livers Because activated hepatic macrophages could initiate and maintain the myofibroblastic transformation of HSCs [41], we then tested whether LratCre+Cc1fl/fl mice developed hepatic fibrosis. Based on Sirius Red staining, LratCre+Cc1fl/fl, but not their controls, develop ed an extensive interstitial chicken - wire pattern of collagen deposition starting at 10 months of age (Fig. 7A.d vs a-c, and vs Fig. S4A.d at 8 months). Consistently, the mRNA levels of pro-fibrogenic genes ( Acta2, Col11, Col31, and Tgf) were induced in the livers of 10 -month-old (Fig. 7B.i), but not 8 -month-old (Fig. S4B) mutants. Hepatic fibrosis could be mediated by the activation of the canonical TGF –SMAD2/3 profibrogenic pathway , as demonstrated by SMAD2 phosphorylation (Fig. 7C) with no change in the expression of its inhibitor, Smad7 (Fig. 7B). Activated HSCs modulate the extracellular matrix (ECM) composition, mediated by MAPK, NF-B and TGF-SMAD2/3 pathways. This involves the regulation of the expression of the matrix metalloproteinases (MMPs) and the tissue inhibitor of metalloproteinases (TIMPs) that are implicated in the production as well as the resolution of excess collagen and other ECM components. Consistently, 10- month-old LratCre+Cc1fl/fl livers displayed higher mRNA (Fig. 7B.i) and protein levels (Fig. 7C) of MMP9, MMP13 and TIMP1 relative to controls. They also exhibited a ~2-fold increase in the mRNA levels of hepatic Mmp 2, Timp 2 and Timp 3 (Fig. 7B.i). Whereas MMP9, TIMP1 and TIMP2 are pro-fibrogenic, MMP2 and TIMP3 block fibroblastic activation and increase collagen clearance [43]. MMP13 can promote was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (whichthis version posted April 3, 2024. ; https://doi.org/10.1101/2024.04.02.586238doi: bioRxiv preprint collagen production as well as its clearance [43]. This demonstrated that the loss of CEACAM1 in HSCs regulated ECM formation by a complex of MMPs and TIMPs favoring hepatic fibrosis. Coupled with oxidative stress, TGF  signaling could cause hepatocellular injury [44]. Consistently, mutant livers manifested higher mRNA levels of genes implicated in oxidative stress (Nox1 and Nox4) (Table S4) and hepatocytes injury ( Txn, Nqo, Nrf1 and Hgf) ( Fig. 7B.ii ). This could drive liver dysfunction, as determined by higher plasma alanine transaminase (ALT) and aspartate aminotransferase (AST) content in 10-month-old but not 8-month-old mutants as compared to control mice (Fig. 7D). 3.8. Conditioned media from LratCre+Cc1fl/fl HSCs activates wild-type HSCs via an EGFR -mediated mechanism As above, NA treatment blocked the release of FAs as glycerol (Fig. 8A.i, + vs – lane) and IL-6 (Fig. 8A.ii, + vs – lane) into the media of LratCre+Cc1fl/fl HSCs (KO). Thus, we next examined whether media from KO-HSCs (KO-Med) could activate wild -type (WT) HSCs and whether this could be blocked by NA treatment. As Fig. 8B.i shows, incubating WT -HSCs with KO -Med (WT/KO-Med) repressed Ceacam1 expression by ~65% relative to WT-HSCs incubated in regular culture media (Reg -Med) (– lanes, grey vs white bar ). This likely result ed from increased Ppar and reduced Ppar1 expression (Fig. 8B.ii-iii, respectively, – lanes, grey vs white bars). This reciprocal change in Ppar and Ppar1 expression, together with higher expression of Acta2 (Fig. 8B.iv) and Pcna (Fig. 8B.v) in WT/KO -Med than WT/Reg -Med cells (grey vs white bars) demonstrated a higher myofibroblastic activation and proliferation of WT/KO-Med than WT/Reg-Med cells. Furthermore, NA treatment reversed these changes in Ppar Ppar1, Acta2 and Pcna mRNA levels in parallel to restoring Ceacam1 expression in WT/KO-Med (Fig. 8B.i-vi, + vs – bars, horizontal vs vertical bars). Gefitinib had a similar effect on the mRNA of these genes in WT/KO-Med (Fig. 8B.vii-x, + vs – bars, horizontal vs vertical bars). PPAR are activated by the FAs that are released from LratCre+Cc1fl/fl HSCs. As Table S5 shows, these KO-HSCs exhibited reduced RE synthesis with a reciprocal increase in PUFA -TG, as assessed by was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (whichthis version posted April 3, 2024. ; https://doi.org/10.1101/2024.04.02.586238doi: bioRxiv preprint the 2-to-4–fold reduction in the mRNA levels of Lrat and Lal/Lipa, with the reciprocal ~12-fold increase in Acsl4/Acsl1 and the ~4-to-6–fold increase in the mRNA levels of Dgat1 and Atgl (Table S5). It is likely that the increase in FAs release activated PPAR to reduce Ceacam1 expression in WT/KO-Med. This would lower Shc sequestration and elevate its reciprocal coupling with EGFR to activate downstream pro- fibrogenic and proliferation pathways (increased Acta2 and Pcna, respectively). Reversal of these changes in lipid metabolism in WT/KO-Med by Gefitinib further demonstrated that EGFR activation mediated the myofibroblastic transformation of HSCs by Ceacam1 loss. was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (whichthis version posted April 3, 2024. ; https://doi.org/10.1101/2024.04.02.586238doi: bioRxiv preprint 4. Discussion The current study demonstrated that CEACAM1’s expression in cultured human LX2-HSCs is supported by autocrine PPAR and retinoic acid transcriptional upregulation, and that activation of primary human HSCs significantly repressed CEACAM1 expression. On the other hand, loss of CEACAM1 in LX2 and primary murine HSCs activated them. This was manifested by reduced PPAR1 and retinoic acid levels with reciprocal elevation in PPAR and PUFA-TG content, respectively. Because CEACAM1 inhibits FASN activity under normo-insulinemic conditions [45], suppressing Ceacam1 transcription by PPAR [25] (and by the loss of PPAR), likely mediated the increase in TG synthesis in mutant HSCs. In light of the anti-lipogenic and anti-fibrogenic effect of FASN inhibitors [46], the current data propose a key role for HSCs’ CEACAM1 in preventing hepatic fibrosis. CEACAM1 expression is highest in hepatocytes. Its deletion in these cells impaired hepatic insulin clearance to cause hyperinsulinemia-driven insulin resistance, de novo lipogenesis and inflammation [15]. It also caused hepatic fibrosis, whereas hepatocytes-specific rescuing of CEACAM1 reversed steatosis and fibrosis in parallel to restoring insulin sensitivity in Cc1–/– null mice. This points to a key role for hyperinsulinemia-driven steatosis in hepatic fibrosis caused by CEACAM1 loss in hepatocytes [31]. In contrast, Ceacam1 deletion from endothelial cells caused hepatic fibrosis in the absence of insulin resistance and hepatic steatosis [17]. The phenotype was driven by hyperactivation of the vascular endothelial growth factor (VEGFR)/NF-B pathway and increased synthesis of endothelin1 and of its pro- fibrogenic signals via its receptor A in HSCs. Inflammation in this mutant preceded hepatic fibrosis and implicated macrophage activation in addition to mounting a Th1 response by T lymphocytes. Like its deletion from endothelial cells, specific deletion of Ceacam1 from HSCs caused hepatic fibrosis in the absence of insulin resistance and hepatic steatosis. However, it occured concurrently to inflammation and was mediated by activation of EGFR by PUFA and IL-6, a transcriptional target of NF- kB. Sustained activation of the IL-6/STAT3 pathway could mediate the mounting of a Th2/M2 response in was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (whichthis version posted April 3, 2024. ; https://doi.org/10.1101/2024.04.02.586238doi: bioRxiv preprint LratCre+Cc1fl/fl livers, as in Stat1 nulls that exhibited activation of the M2 macrophage pool without a significant increase in pro-inflammatory T lymphocytes [47]. In addition to EGFR/NF-B pathway, the EGFR/MAPK proliferative pathway was also activated in KD-LX2 HSCs devoid of CEACAM1. This resulted from the increased coupling of Shc to EGFR when its reciprocal sequestration by CEACAM1 was absent [33], as with respect to VEGFR [48] and the insulin receptor [32]. Thus, activation of NF-B and the MAPK pathways downstream of these growth factor receptors constitutes a unifying mechanism underlying hepatic fibrosis when their shared substrate, CEACAM1, is lost. This agrees with the reported PPAR-driven HSCs proliferation and hepatic fibrosis via activation of the P38-JNK MAPK pathway in LX2 and murine HSCs [8]. EGFR is implicated in HSCs activation [34; 49] as demonstrated by the reversal of hepatic fibrogenesis, hepatocyte proliferation and liver injury in experimental models of hepatic fibrosis by inhibitors of EGFR tyrosine kinase activity [50; 51]. Yet, inhibiting EGFR phosphorylation to curb hepatic fibrosis has not gained traction at the clinical setting. Instead, targeting inflammation and lipogenesis constitutes the main current therapeutic approach, particularly at the early stages of the disease [2; 3]. This includes the use of a combinational therapy of PPAR agonists and incretins to retard/attenuate hepatic fibrosis in patients with MASLD/MASH [52]. It is likely that the effectiveness of these drugs is mediated, at least partly, by the transcriptional activation of CEACAM1 [16], which would in turn, counter inflammation in immune cells [53] and lipogenesis in hepatocytes (by inhibiting FASN and limiting chronic hyperinsulinemia). We have previously shown that hepatic CEACAM1 expression is progressively reduced with advancing fibrosis stages in patients with MASH [15]. Moreover, single cell RNA-sequencing showed lower CEACAM1 expression in hepatocytes and LSECs of patients with fibrosis/cirrhosis [17]. The current study demonstrated that activation of primary human HSCs repressed CEACAM1 expression and that deleting CEACAM1 from immortalized human LX2 activated them. The observations in LratCre+Cc1fl/fl mutants further emphasized the regulation of hepatic fibrosis by CEACAM1’s loss in HSCs, independently of its paracrine role in hepatocytes and endothelial cells. The underlying mechanisms converge at the level was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (whichthis version posted April 3, 2024. ; https://doi.org/10.1101/2024.04.02.586238doi: bioRxiv preprint of NF-B inflammation and MAPK proliferation pathways downstream of EGFR in HSCs (and in hepatocytes at the basal state) and of VEGFR in endothelial cells. Consistent with elevated serum IL-6 levels in patients with advanced hepatic fibrosis [54], loss of Ceacam1 in HSCs caused an elevation in plasma IL-6 levels concurrently wth hepatic fibrosis in LratCre+Cc1fl/fl mutants. Together, this proposes that inducing CEACAM1 expression could become an effective therapeutic approach to curb fibrosis, not only in early stages of the disease, but also at a later stage. In summary, the current report provides an in vivo demonstration of a novel mechanistic link between a distinct CEACAM1/EGFR/NF-B signaling module in murine HSCs and hepatic fibrosis, an advanced component of MASH. This was supported by studies in human LX2 HSCs demonstrating an autocrine regulation of hepatic fibrosis by the loss of CEACAM1 in stellate cells. Further analysis is required to translate our observations to MASH pathogenesis in humans. was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (whichthis version posted April 3, 2024. ; https://doi.org/10.1101/2024.04.02.586238doi: bioRxiv preprint Credit authorship Contribution statement H.T.M., H.E.G., S.A., S.G.L., S.V., H.L.S., S.Z., R.A., and G.D.B. researched data. H.T.M., H.E.G., and S.A.planned and organized experiments, collected and analyzed data, prepared the illustration and drafted the manuscript. S.L.F., R.F.S., L.A.vG, and T.D.H. Jr discussed data and edited the manuscript. S.M.N. conceived and oversaw the work, including its study design and data analysis, leading scientific discussions and reviewing/editing the manuscript. Declaration of competing interest None declared. Acknowledgments S.M.N. is partly supported by the Osteopathic Heritage Foundation J.J.Kopchick Eminent Research Chair. Financial support This work was supported by NIH grants: R01-HL112248, R01-DK054254, R01-DK124126 (to S.M.N), R01-DK128289 (to S.L.F). S.V. is supported by FWO 1243121N, and L.A.vG. by FWO G071922N. Data availability statement All data relevant to the study are included in the article/online supplemental information was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (whichthis version posted April 3, 2024. ; https://doi.org/10.1101/2024.04.02.586238doi: bioRxiv preprint

[1] Younossi, Z.M., 2019. Non-alcoholic fatty liver disease - A global public health perspective. J. Hepatol. 70(3):531-544. [2] Harrison, S.A., Allen, A.M., Dubourg, J., Noureddin, M., Alkhouri, N., 2023. Challenges and opportunities in NASH drug development. Nat. Med. 29(3):562-573. [3] Wang, S., Friedman, S.L., 2023. Found in translation-Fibrosis in metabolic dysfunction-associated steatohepatitis (MASH). Sci. Transl. Med. 15(716):eadi0759. [4] Brunt, E.M., Kleiner, D.E., Carpenter, D.H., Rinella, M., Harrison, S.A., Loomba, R., et al., 2021. NAFLD: Reporting Histologic Findings in Clinical Practice. Hepatology 73(5):2028-2038. [5] Friedman, S.L., 2008. Mechanisms of hepatic fibrogenesis. Gastroenterology 134(6):1655-1669. [6] Jophlin, L.L., Koutalos, Y., Chen, C., Shah, V., Rockey, D.C., 2018. Hepatic stellate cells retain retinoid-laden lipid droplets after cellular transdifferentiation into activated myofibroblasts. Am. J. Physiol. Gastrointest. Liver Physiol. 315(5):G713-G721. [7] Hellemans, K., Michalik, L., Dittie, A., Knorr, A., Rombouts, K., De Jong, J., et al., 2003. Peroxisome proliferator-activated receptor-beta signaling contributes to enhanced proliferation of hepatic stellate cells. Gastroenterology 124(1):184-201. [8] Kostadinova, R., Montagner, A., Gouranton, E., Fleury, S., Guillou, H., Dombrowicz, D., et al., 2012. GW501516-activated PPARbeta/delta promotes liver fibrosis via p38-JNK MAPK-induced hepatic stellate cell proliferation. Cell Biosci. 2(1):34. [9] Berry, D.C., Noy, N., 2009. All-trans-retinoic acid represses obesity and insulin resistance by activating both peroxisome proliferation-activated receptor beta/delta and retinoic acid receptor. Mol. Cell. Biol. 29(12):3286-3296. was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (whichthis version posted April 3, 2024. ; https://doi.org/10.1101/2024.04.02.586238doi: bioRxiv preprint [10] Forman, B.M., Chen, J., Evans, R.M., 1997. Hypolipidemic drugs, polyunsaturated fatty acids, and eicosanoids are ligands for peroxisome proliferator-activated receptors alpha and delta. Proc. Natl. Acad. Sci. U. S. A. 94(9):4312-4317. [11] Friedman, S.L., 2024. Hepatic Fibrosis and Cancer: The Silent Threats of Metabolic Syndrome. Diabetes Metab. J. [12] DeAngelis, A.M., Heinrich, G., Dai, T., Bowman, T.A., Patel, P.R., Lee, S.J., et al., 2008. Carcinoembryonic antigen-related cell adhesion molecule 1: a link between insulin and lipid metabolism. Diabetes 57(9):2296-2303. [13] Ghadieh, H.E., Russo, L., Muturi, H.T., Ghanem, S.S., Manaserh, I.H., Noh, H.L., et al., 2019. Hyperinsulinemia drives hepatic insulin resistance in male mice with liver-specific Ceacam1 deletion independently of lipolysis. Metabolism 93:33-43. [14] Helal, R.A., Russo, L., Ghadieh, H.E., Muturi, H.T., Asalla, S., Lee, A.D., et al., 2021. Regulation of hepatic fibrosis by carcinoembryonic antigen-related cell adhesion molecule 1. Metabolism 121:154801. [15] Zaidi, S., Asalla, S., Muturi, H.T., Russo, L., Abdolahipour, R., Belew, G.D., et al., 2024. Loss of CEACAM1 in hepatocytes causes hepatic fibrosis Eur. J. Clin. Invest. in press. [16] Ghadieh, H.E., Muturi, H.T., Russo, L., Marino, C.C., Ghanem, S.S., Khuder, S.S., et al., 2018. Exenatide induces carcinoembryonic antigen-related cell adhesion molecule 1 expression to prevent hepatic steatosis. Hepatol. Commun. 2(1):35-47. [17] Muturi, H.T., Ghadieh, H.E., Abdolahipour, R., Stankus, H.L., Z., Y., Liu, J.K., et al., 2023. Loss of CEACAM1 in endothelial cells causes hepatic fibrosis. Metabolism 144:155562. [18] Rueckschloss, U., Kuerten, S., Ergun, S., 2016. The role of CEA-related cell adhesion molecule-1 (CEACAM1) in vascular homeostasis. Histochem. Cell Biol. 146(6):657-671. [19] El Taghdouini, A., Sorensen, A.L., Reiner, A.H., Coll, M., Verhulst, S., Mannaerts, I., et al., 2015. Genome-wide analysis of DNA methylation and gene expression patterns in purified, uncultured human liver cells and activated hepatic stellate cells. Oncotarget 6(29):26729-26745. was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (whichthis version posted April 3, 2024. ; https://doi.org/10.1101/2024.04.02.586238doi: bioRxiv preprint [20] Mederacke, I., Hsu, C.C., Troeger, J.S., Huebener, P., Mu, X., Dapito, D.H., et al., 2013. Fate tracing reveals hepatic stellate cells as dominant contributors to liver fibrosis independent of its aetiology. Nat. Commun. 4:2823. [21] El Taghdouini, A., Najimi, M., Sancho-Bru, P., Sokal, E., van Grunsven, L.A., 2015. In vitro reversion of activated primary human hepatic stellate cells. Fibrogenesis Tissue Repair 8:14. [22] Mederacke, I., Dapito, D.H., Affo, S., Uchinami, H., Schwabe, R.F., 2015. High-yield and high- purity isolation of hepatic stellate cells from normal and fibrotic mouse livers. Nat. Protoc. 10(2):305-315. [23] Schiffer, E., Housset, C., Cacheux, W., Wendum, D., Desbois-Mouthon, C., Rey, C., et al., 2005. Gefitinib, an EGFR inhibitor, prevents hepatocellular carcinoma development in the rat liver with cirrhosis. Hepatology 41(2):307-314. [24] Hinds, T.D., Jr., Stechschulte, L.A., Cash, H.A., Whisler, D., Banerjee, A., Yong, W., et al., 2011. Protein phosphatase 5 mediates lipid metabolism through reciprocal control of glucocorticoid receptor and peroxisome proliferator-activated receptor-gamma (PPARgamma). J. Biol. Chem. 286(50):42911-42922. [25] Ramakrishnan, S.K., Khuder, S.S., Al-Share, Q.Y., Russo, L., Abdallah, S.L., Patel, P.R., et al., 2016. PPARalpha (peroxisome proliferator-activated receptor alpha) activation reduces hepatic CEACAM1 protein expression to regulate fatty acid oxidation during fasting-refeeding transition. J. Biol. Chem. 291(15):8121-8129. [26] Tsuchida, T., Friedman, S.L., 2017. Mechanisms of hepatic stellate cell activation. Nat. Rev. Gastroenterol. Hepatol. 14(7):397-411. [27] Testerink, N., Ajat, M., Houweling, M., Brouwers, J.F., Pully, V.V., van Manen, H.J., et al., 2012. Replacement of retinyl esters by polyunsaturated triacylglycerol species in lipid droplets of hepatic stellate cells during activation. PLoS One 7(4):e34945. [28] Sauvant, P., Cansell, M., Atgie, C., 2011. Vitamin A and lipid metabolism: relationship between hepatic stellate cells (HSCs) and adipocytes. J. Physiol. Biochem. 67(3):487-496. was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (whichthis version posted April 3, 2024. ; https://doi.org/10.1101/2024.04.02.586238doi: bioRxiv preprint [29] Tuohetahuntila, M., Molenaar, M.R., Spee, B., Brouwers, J.F., Wubbolts, R., Houweling, M., et al., 2017. Lysosome-mediated degradation of a distinct pool of lipid droplets during hepatic stellate cell activation. J. Biol. Chem. 292(30):12436-12448. [30] Tuohetahuntila, M., Spee, B., Kruitwagen, H.S., Wubbolts, R., Brouwers, J.F., van de Lest, C.H., et al., 2015. Role of long-chain acyl-CoA synthetase 4 in formation of polyunsaturated lipid species in hepatic stellate cells. Biochim Biophys Acta 1851(2):220-230. [31] Wobser, H., Dorn, C., Weiss, T.S., Amann, T., Bollheimer, C., Buttner, R., et al., 2009. Lipid accumulation in hepatocytes induces fibrogenic activation of hepatic stellate cells. Cell Res 19(8):996- 1005. [32] Poy, M.N., Ruch, R.J., Fernstrom, M.A., Okabayashi, Y., Najjar, S.M., 2002. Shc and CEACAM1 interact to regulate the mitogenic action of insulin. J. Biol. Chem. 277(2):1076-1084. [33] Abou-Rjaily, G.A., Lee, S.J., May, D., Al-Share, Q.Y., Deangelis, A.M., Ruch, R.J., et al., 2004. CEACAM1 modulates epidermal growth factor receptor--mediated cell proliferation. J. Clin. Invest. 114(7):944-952. [34] Bhushan, B., Michalopoulos, G.K., 2020. Role of epidermal growth factor receptor in liver injury and lipid metabolism: Emerging new roles for an old receptor. Chem. Biol. Interact. 324:109090. [35] She, H., Xiong, S., Hazra, S., Tsukamoto, H., 2005. Adipogenic transcriptional regulation of hepatic stellate cells. J. Biol. Chem. 280(6):4959-4967. [36] Ganji, S.H., Kashyap, M.L., Kamanna, V.S., 2015. Niacin inhibits fat accumulation, oxidative stress, and inflammatory cytokine IL-8 in cultured hepatocytes: Impact on non-alcoholic fatty liver disease. Metabolism 64(9):982-990. [37] Zhou, E., Li, Y., Yao, M., Wei, Z., Fu, Y., Yang, Z., 2014. Niacin attenuates the production of pro- inflammatory cytokines in LPS-induced mouse alveolar macrophages by HCA2 dependent mechanisms. Int. Immunopharmacol. 23(1):121-126. was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (whichthis version posted April 3, 2024. ; https://doi.org/10.1101/2024.04.02.586238doi: bioRxiv preprint [38] Wang, Y., van Boxel-Dezaire, A.H., Cheon, H., Yang, J., Stark, G.R., 2013. STAT3 activation in response to IL-6 is prolonged by the binding of IL-6 receptor to EGF receptor. Proc. Natl. Acad. Sci. U. S. A. 110(42):16975-16980. [39] Ueda, A., Ishigatsubo, Y., Okubo, T., Yoshimura, T., 1997. Transcriptional regulation of the human monocyte chemoattractant protein-1 gene. Cooperation of two NF-kappaB sites and NF-kappaB/Rel subunit specificity. J. Biol. Chem. 272(49):31092-31099. [40] Roca, H., Varsos, Z.S., Sud, S., Craig, M.J., Ying, C., Pienta, K.J., 2009. CCL2 and interleukin-6 promote survival of human CD11b+ peripheral blood mononuclear cells and induce M2-type macrophage polarization. J. Biol. Chem. 284(49):34342-34354. [41] Tacke, F., 2017. Targeting hepatic macrophages to treat liver diseases. J. Hepatol. 66(6):1300- 1312. [42] Quero, L., Tiaden, A.N., Hanser, E., Roux, J., Laski, A., Hall, J., et al., 2019. miR-221-3p Drives the Shift of M2-Macrophages to a Pro-Inflammatory Function by Suppressing JAK3/STAT3 Activation. Front. Immunol. 10:3087. [43] Giannandrea, M., Parks, W.C., 2014. Diverse functions of matrix metalloproteinases during fibrosis. Dis. Model Mech. 7(2):193-203. [44] Sun, X., Harris, E.N., 2020. New aspects of hepatic endothelial cells in physiology and nonalcoholic fatty liver disease. Am. J. Physiol. Cell. Physiol. 318(6):C1200-C1213. [45] Najjar, S.M., Yang, Y., Fernstrom, M.A., Lee, S.J., Deangelis, A.M., Rjaily, G.A., et al., 2005. Insulin acutely decreases hepatic fatty acid synthase activity. Cell Metab. 2(1):43-53. [46] O'Farrell, M., Duke, G., Crowley, R., Buckley, D., Martins, E.B., Bhattacharya, D., et al., 2022. FASN inhibition targets multiple drivers of NASH by reducing steatosis, inflammation and fibrosis in preclinical models. Sci. Rep. 12(1):15661. was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (whichthis version posted April 3, 2024. ; https://doi.org/10.1101/2024.04.02.586238doi: bioRxiv preprint [47] Li, J., Bessho, K., Shivakumar, P., Mourya, R., Mohanty, S.K., Dos Santos, J.L., et al., 2011. Th2 signals induce epithelial injury in mice and are compatible with the biliary atresia phenotype. J. Clin. Invest. 121(11):4244-4256. [48] Abu Helal, R., Muturi, H.T., Lee, A.D., Li, W., Ghadieh, H.E., Najjar, S.M., 2022. Aortic Fibrosis in Insulin-Sensitive Mice with Endothelial Cell-Specific Deletion of Ceacam1 Gene. Int. J. Mol. Sci. 23(8). [49] Wan, S., Liu, X., Sun, R., Liu, H., Jiang, J., Wu, B., 2024. Activated hepatic stellate cell-derived Bmp-1 induces liver fibrosis via mediating hepatocyte epithelial-mesenchymal transition. Cell Death Dis. 15(1):41. [50] Fuchs, B.C., Hoshida, Y., Fujii, T., Wei, L., Yamada, S., Lauwers, G.Y., et al., 2014. Epidermal growth factor receptor inhibition attenuates liver fibrosis and development of hepatocellular carcinoma. Hepatology 59(4):1577-1590. [51] Wang, D., Xu, H., Fan, L., Ruan, W., Song, Q., Diao, H., et al., 2023. Hyperphosphorylation of EGFR/ERK signaling facilitates long-term arsenite-induced hepatocytes epithelial-mesenchymal transition and liver fibrosis in sprague-dawley rats. Ecotoxicol. Environ. Saf. 249:114386. [52] Wong, V.W.S., Zelber-Sagi, S., Cusi, K., Carrieri, P., Wright, E., Crespo, J., et al., 2022. Management of NAFLD in primary care settings. Liver int. 42(11):2377-2389. [53] Horst, A.K., Najjar, S.M., Wagener, C., Tiegs, G., 2018. CEACAM1 in Liver Injury, Metabolic and Immune Regulation. Int. J. Mol. Sci. 19(10). [54] Gunes, A., Schmitt, C., Bilodeau, L., Huet, C., Belblidia, A., Baldwin, C., et al., 2023. IL-6 Trans- Signaling Is Increased in Diabetes, Impacted by Glucolipotoxicity, and Associated With Liver Stiffness and Fibrosis in Fatty Liver Disease. Diabetes 72(12):1820-1834. was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (whichthis version posted April 3, 2024. ; https://doi.org/10.1101/2024.04.02.586238doi: bioRxiv preprint Table 1 Plasma and tissue biochemistry in mice at 10 months of age. LratCre–Cc1+/+ LratCre+Cc1+/+ LratCre–Cc1fl/fl LratCre+Cc1fl/fl Body weight (g) 29.4 ± 2.2 26.7 ± 1.2 28.3 ± 1.7 28.5 ± 1.1 % WAT/BW 1.9 ± 0.9 1.6 ± 0.5 1.2 ± 0.4 1.2 ± 0.3 NEFA (mEq/l) 0.5 ± 0.1 0.6 ± 0.1 0.4 ± 0.1 0.5 ± 0.1 TG (mg/dl) 48.4 ± 8.9 50.4 ± 6.5 42.8 ± 13.2 43.4 ± 7.0 Insulin (pM) 80.5 ± 10.7 74.7 ± 4.7 73.7 ± 2.1 78.5 ± 7.0 C-peptide (pM) 198.6 ± 36.4 193.1 ± 36.7 177.9 ± 22.6 219.4 ± 45.6 C/I molar ratio 2.3 ± 0.3 2.6 ± 0.3 2.4 ± 0.3 2.9 ± 0.7 Fast Glucose (mg/dl) 67. ± 9. 60. ± 6. 71. ± 6. 64. ± 5. Fed Glucose (mg/dl) 116. ± 7. 118. ± 4. 112. ± 6. 116. ± 9. Hepatic TG (µg/mg) 58.7 ± 8.7 65.5 ± 9.2 59.1 ± 9.9 58.7 ± 15.2 Blood was drawn from m ale mice (10 months of age, n ≥5/genotype) at 2100 h to assess fed glucoselevels. Following a recovery period of 3 days, mice were fasted overnight before blood was drawn and tissues were excised at 1100h. Except of blood glucose levels, other values refer to plasma levels, unless otherwise mentioned. Hepatic TG is measured as µg/mg protein. Values are expressed as mean ± SEM. BW: Body weight; WAT: White adipose tissue; %WAT/BW: visceral obesity; C/I: Steady -state C-peptide/Insulin molar ratio as a measure of hepatic insulin clearance; NEFA: Non-esterified fatty acid; TG: Triacylglycerol. was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (whichthis version posted April 3, 2024. ; https://doi.org/10.1101/2024.04.02.586238doi: bioRxiv preprint Figure legends Fig. 1. Regulation of CEACAM1 transcription. (A) immortalized human hepatic stellate cells LX2 cells were treated with DMSO ( –) (white bars), 1 μM Rosiglitazone (Ro) ( grey bars), 5 μM Retinoic Acid (RA) (black bars), and Ro plus RA (hatched bars) for 24 hrs before being subjected to qRT-PCR analysis of CEACAM1 (CC1) mRNA levels. Data are expressed as mean ± SEM; * P<0.05 vs vehicle (–). (B) to analyze the transcriptional regulation of Ceacam1 promoter activity in HepG2 , wild-type (nts −1100) mouse Ceacam1 promoter and block mutants (small letters) of the PPRE/RXRα site (nts –557 to –543) (–∆2); RXRα (–548 and –543) (–∆RXRα); or PPRE (nts –557 and -551) (–∆PPRE) were subcloned into pGL4.10 promoterless plasmid. Luciferase activity was measured in triplicate in response to DMSO ( –, white), rosiglitazone (Ro, grey), retinoic acid (RA, checkered) or rosiglitazone plus retinoic acid (Ro/RA, black). PPREx3-TK-luc and PGL3-RARE-Luc were used as positive controls for PPRE and RXR, respectively. PGL4.10 empty vector was used as a negative control. Luciferase light units are expressed as mean ± SEM in relative light uni ts. *P<0.05 treatment versus vehicle. (C) CEACAM1 mRNA was evaluated in spontaneously activated (aHSCs) (black bars) and quiescent (qHSCs) (white bars) primary human HSCs . Data are expressed as mean ± SEM; * P<0.05 vs qHSCs. was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (whichthis version posted April 3, 2024. ; https://doi.org/10.1101/2024.04.02.586238doi: bioRxiv preprint Fig. 2. Activation of LX2 human hepatic stellate cells by CEACAM1 deletion. (A) LX2 cells were subjected to shRNA-mediated knockdown of CEACAM1 (KD) and (i) analyzed by qRT-PCR in triplicate to assess the decrease in CEACAM1 mRNA in KD-LX2 (Black bars) versus Scr-LX2 scrambled control cells (white bars). mRNA was normalized to GAPDH mRNA and data represented as mean ± SEM; * P<0.05 vs Scr- LX2; (ii) CEACAM1 protein levels were assessed by immunoblotting (Ib) the upper half of the membrane with -CEACAM1 (-CC1) antibody and the lower half with -tubulin to normalize per loaded proteins. (B) to examine lipid metabolism, (i) cells were grown in at least 3 plates/stable line, stained with Nile Red to depict fat (yellow) droplets. Fat stains were evaluated by densitometry and presented graphically as mean ± SEM; * P<0.05 vs Scr-LX2; (ii) free glycerol level was assayed in the media of the stained cells as a measure of lipolysis. Experiments were done in triplicate. Data are expressed as mean ± SEM; * P<0.05 vs Scr-LX2; (iii) qRT-PCR mRNA analysis of genes implicated in lipid metabolism was performed in triplicate. Values are expressed as mean ± SEM. *P< 0.05 vs Scr-LX2. (C) to assess LX2 activation, (i) KD-LX2 and Scr-LX2 cells were subjected to MTT assay in triplicate. Data represent mean ± SEM; *P<0.05 vs Scr-LX2; (ii) qRT-PCR analysis was performed in triplicate to assess ACTA2 and COL11 mRNA levels as markers of fibrogenic activity. Data represent mean ± SEM; * P<0.05 vs Scr-LX2. (D) to examine TGF signaling, cell lysates were immunoblotted with -phosphoSmad2 or -phosphoSmad3 antibody (-pSmad) followed by re-immunoblotting (re-Ib) with -Smad 2 or -Smad 3 antibodies, respectively, for normalization. was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (whichthis version posted April 3, 2024. ; https://doi.org/10.1101/2024.04.02.586238doi: bioRxiv preprint Fig. 3. Scr-LX2 activation by conditioned media from KD -LX2 cells . Src -LX2 and KD-LX2 cells were incubated with nicotinic acid ( NA) (+) or with vehicle (–) for 24 hr s before (A) media were collected and assayed for glycerol (captured lipolysis-derived fatty acids) (i), IL-6 (ii) and TNF (iii) levels. Data represent mean ± SEM; *P<0.05, KD-LX2 vs Scr-LX2 cells/treatment type; †p<0.05, NA-treated vs untreated/cell line. (B) media of KD-LX2 cells (conditioned media) were transferred to pre-washed Scr group (Scr/Cond) before cells were harvested for qRT-PCR analysis of the mRNA of CC1, PPAR, PPAR1 and ACTA2 (Fig. 3B.i- iv) and cell growth by MTT assay in triplicate and repeated twice (Fig. 2B.v). *P<0.05 untreated KD and Scr/Cond vs untreated Scr-LX2 cells; †P<0.05 NA-treated vs untreated/cell line; §P<0.05 untreated Scr/Cond vs untreated KD, and ¶P<0.05 NA-treated KD vs NA-treated Scr and NA-treated Scr/Cond cells. The latter indicates that although NA treatment decreased the mRNA of PPAR and ACTA2 as well as cell growth, it did not completely restore their values to those in Scr-LX2 controls, as it did in Scr/Cond cells. This is likely due to persistent absence of CEACAM1 (and low PPAR1) with sustained TNF (iii) levels in these donor KD-LX2 cells. (C) Western blot analysis of EGFR signaling in cells described above: liver l ysates were subjected to immunoblotting (Ib) with antibodies against (i) phospho-EGFR (α-pEGFR), (iv) phospho-MAPK (α-pMAPK), (v) phospho - NF-kB ( α-pNF-kB), and (vi) α-PCNA and in parallel gels, with their specific antibodies for normalization. (ii) some lysates were subjected to immunoprecipitation (Ip) with Shc antibody followed by immunoblotting (Ib) with antibodies against CEACAM1 ( α-CC1) and Shc (α-Shc). (iii) Lysates were subjected to immunoprecipitation (Ip) with α-pEGFR antibody followed by immunoblotting (Ib) with α- Shc and α-EGFR antibodies. Gels represent two separate experiments. The apparent molecular mass (kDa) is indicated at the right hand-side of each gel. was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (whichthis version posted April 3, 2024. ; https://doi.org/10.1101/2024.04.02.586238doi: bioRxiv preprint Fig. 4. EGFR-mediated myofibroblastic transformation of KD-LX2 cells. Src -LX2 and KD-LX2 cells were incubated with Genitinib (10µM) or vehicle for 24 hr s. The conditioned media was collected from both Genitinib-treated and untreated KD-LX2 cells, and the supernatant was collected and transferred to Src-LX2 and incubated for 24 hrs (Scr/Cond-LX2 cells). (A) lysates were analyzed by immunoblotting (Ib) with α- pEGFR and α-EGFR for normalization to assess EGFR activation. Gel is representative of 2 different experiments performed on 2 different sets of mice. (B-E) lysates were analyzed by qRT-PCR analysis in triplicate to assess mRNA of PPAR, PPAR1, CC1, and ACTA2. (F) cells were grown in 96 well plates to examine cell growth in triplicate by MTT assay (repeated twice). Values are expressed as mean ± SEM; *P<0.05 untreated KD and Scr/Cond vs untreated Scr-LX2 cells; †P<0.05 NA-treated vs untreated/cell line; §P<0.05 untreated Scr/Cond vs untreated KD, and ¶P<0.05 NA-treated KD vs NA-treated Scr and NA-treated Scr/Cond cells. was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (whichthis version posted April 3, 2024. ; https://doi.org/10.1101/2024.04.02.586238doi: bioRxiv preprint Fig. 5. Metabolic phenotyping of LratCre+Cc1fl/fl mice. (A) Primary cells were isolated from male mutants and their littermate controls at 2-4 months of age except for HSCs which were isolated from mice at 8 months of age (n=2-5/genotype). Ceacam1 mRNA levels were analyzed by qRT-PCR in triplicate and normalized to 18s. Values are expressed as mean ± SEM. (B-C) 10-month-old male mice (n≥ 7–8/genotype) were injected intraperitoneally with insulin or glucose to assess glucose disposal in response to insulin (B) and glucose (C). Values were expressed as mean ± SEM. (D) Livers were removed from 10-month-old Lrat+Cc1fl/fl male mice and their 3 littermate controls (n=4 –5/genotype), sectioned and stained with H&E staining to identify foci of inflammatory cell infiltrates in mutants (panel d) and their littermate controls (panels a-c). Values are expressed as mean ± SEM in the accompanying inflammatory foci quantification graph. *P<0.05 mutants vs the 3 littermate controls. was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (whichthis version posted April 3, 2024. ; https://doi.org/10.1101/2024.04.02.586238doi: bioRxiv preprint Fig. 6. Increased inflammation in LratCre+Cc1fl/fl livers. (A) livers were removed from 10-month-old LratCre+Cc1fl/fl mutants and their three controls (n=4-5 mice/genotype) and subjected to (A) immunohistochemical (IHC) analysis with: (i) CD68 to assess macrophage recruitment, (ii) Mac-2 to examine macrophage activation, (iii) MPO to evaluate neutrophil accumulation, CD4 (iv) and CD8 (v) to immunostain T cells and (vi) Foxp3 to determine the anti-inflammatory Treg pool. Representative images taken at 50 μm magnification are shown with insets at 20 μm. Values are expressed as mean ± SEM in the accompanying quantification graph. *P<0.05 mutants (d) vs the 3 littermate controls (a-c). (B) liver lysates were subjected to immunoblotting (Ib) with antibodies against the phosphorylated p65 subunit of NF-B (α- pNF-kB), and α-pStat3. To normalize against added proteins, gels were analyzed by SDS gel electrophoresis in parallel and proteins immunoblotted with specific antibodies. Representative gels include 2 different mice/genotype. (C) liver lysates (n=6/each genotype) were analyzed in duplicate by qRT-PCR using gene-specific primers and normalized to Gapdh. Values are expressed as mean ± SEM. *P<0.05 vs all three controls. (D) Male mice (8 and 10 months of age, n≥6/genotype/age group) were fasted overnight before blood was drawn at 1100 in the next morning and their plasma IL-6 and TNF levels were analyzed. Values are expressed as mean ± SEM. *P<0.05 vs all three controls. was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (whichthis version posted April 3, 2024. ; https://doi.org/10.1101/2024.04.02.586238doi: bioRxiv preprint Fig. 7. Spontaneous hepatic fibrosis in LratCre+Cc1fl/fl mice. Livers were removed from 10-month-old Lrat+Cc1fl/fl male mice and their 3 littermate controls (n=4–5/genotype). (A) Sirius red staining revealed increased deposition of interstitial chicken-wire pattern of collagen fibers in mutants (panel d) versus their littermate controls (panels a-c). Values are expressed as mean ± SEM in the accompanying quantification graph. *P<0.05 mutants vs the 3 littermate controls. (B) liver lysates (n=6/each genotype) were analyzed in duplicate by qRT-PCR using gene-specific primers and normalized to Gapdh to assess mRNA levels of genes involved in inflammation (i) and in hepatocytes injury (ii). Values are expressed as mean ± SEM. *P<0.05 vs all three controls. (C) Western Blot analysis of liver lysates from LratCre+Cc1fl/fl male mice (lanes 7-8) and their LratCre–Cc1+/+ (lanes 1–2), LratCre+Cc1+/+ (lanes 3-4) and LratCre–Cc1fl/fl (lanes 5– 6) controls. Phosphorylated Smad2 (α-pSmad2) normalized against α-Smad2. The protein levels of α- MMP9, α-MMP13 and α-Timp1 were normalized against α-Tubulin. Gels represent two different mice/genotype. The apparent molecular mass (kDa) is indicated at the right hand-side of each gel. (D) Male mice (8 and 10 months of age, n≥6/genotype/age group) were fasted overnight before blood was drawn at 1100 in the next morning and their plasma ALT and AST levels were analyzed. Values are expressed as mean ± SEM. *P<0.05 vs all three controls. was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (whichthis version posted April 3, 2024. ; https://doi.org/10.1101/2024.04.02.586238doi: bioRxiv preprint Fig. 8. EGFR-mediated a ctivation of LratCre+Cc1fl/fl HSCs. Primary HSCs were isolated from ≥8 LratCre+Cc1fl/fl mice (KO) and a combination of wild-type ( LratCre–Cc1+/+, LratCre+Cc1+/+ and LratCre– Cc1fl/fl) mice (WT). They were treated with (+) or without (–) NA and the media were collected and combined. (A) free glycerol (i), IL-6 (ii) and TNF (iii) levels were assayed in the media. Values are mean ± SEM. *P<0.05 KO (–) vs WT (–); †P<0.05 NA-treated vs untreated/mouse group. (B) the conditioned KO-Med was transferred to WT HSCs (WT/KO-Med) for 24 hrs, while a parallel set of WT-HSCs was incubated in regular culture media (Reg-Med) and the mRNA levels were analyzed by qRT-PCR analysis. In some experiments, conditioned media without NA were transferred to WT -HSCs and the cells were treated with or without Gefinitib (vi-x). Cells were harvested for qRT-PCR analysis in triplicate of the mRNA levels of Ceacam1 (i, vi), Ppar (ii, vii), Ppar (iii, viii), Acta2 (iv, ix) and Pcna (v, x). Values are expressed as mean ± SEM; *P<0.05 untreated K O and WT/KO-Med vs WT/Reg-Med; †P<0.05 treated vs untreated/cell line; §P<0.05 untreated WT/KO-Med vs untreated K O, and ¶P<0.05 treated K O vs treated WT/Reg-Med and treated WT/KO-Med cells. was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (whichthis version posted April 3, 2024. ; https://doi.org/10.1101/2024.04.02.586238doi: bioRxiv preprint Supplemental Materials 1. Supplemental material and methods 1.1. Luciferase assay As described [1; 2], human hepatocellular carcinoma cells (HepG2) were cultured overnight to reach ∼60%‐70% confluence. Transfection was performed with promoter constructs and the pRL‐ thymidine kinase Renilla luciferase promoter (Promega) using Lipofectamine 2000 (Invitrogen). Empty pGL4.10 vector was used as the negative control, pGL3-RARE-Luc (plasmid 13458; Addgene Cambridge, MA) and PPREx3 -TK-Luc (Plasmid 1015; Addgene) were used as positive controls for PPRE and RXR, respectively. 24 hrs post-transfection, cells were serum-starved before being treated for 24 hrs with dimethyl sulfoxide (DMSO), 5 μM Retinoic Acid (Sigma-Aldrich Saint Louis, MO) and/or 1 μM Rosiglitazone (Sigma -Aldrich Saint Louis, MO) . Luciferase activity was assessed using the Dual-Luciferase Reporter Assay System (Promega). 1.2. Generation of human CEACAM1 shRNA lentiviral construct and treatment The immortalized human hepatic stellate LX2 cells [initially characterized by the Friedman laboratory [3]] were cultured in Dulbecco's Modified Eagle Medium (DMEM) High Glucose media (Gibco), supplemented with 2% FBS, 1% penicillin/streptomycin and 1% L -Glutamine. The free Web -based tool (http://www.genelink.com/sirna/shRNAi.asp) was used to design a putative siRNA against human CEACAM1 (hCEACAM1) and to desig n oligonucleotides that encode a corresponding small hairpin RNA (shRNA). Origene (Rockville, MD) constructed the shRNA plasmid with oligonucleotides: TGAATCCATGCCATTCAATGTTGCAGAGG and the homologous 120 sequence. The hCEACAM1 shRNA construct was co -transfected together with vectors expressing gag -pol, REV and VSV -G into 293FT cells was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (whichthis version posted April 3, 2024. ; https://doi.org/10.1101/2024.04.02.586238doi: bioRxiv preprint (Invitrogen) to generate a third-generation lentiviral construct. Transfection was achieved by Lipofectamine 2000 (Invitrogen) using 100 ng total DNA per cm 2 of the growth plate or well. The supernatants were harvested and the cell debris was removed by centrifugation at 2000g. The supernatant was used to infect human LX2 stellate cells after addition of 5 ng/ml polybrene (Sigma -Aldrich) and to establish ShCC1 line (KD) with stable downregulation of hCEACAM1 and scramble shRNA control (Scr). After 72 hrs, cells were selected by puromycin resistance (Gibco). For insulin treatment, KD-LX2 and Scr -LX2 were serum -starved with phenol -free DMEM (Gibco) supplemented with 25 mM HEPES and 0.1%BSA for 16 hrs before being stimulated with or without 100 nM Insulin (Sigma-Aldrich) for 5 min before they were collected for Western blotting. In some experiments, KD-LX2 and Scr-LX2 cells were stimulated with and without 100 nM Insulin (Sigma-Aldrich) for 24 hours and subjected to MTT assay (Sigma-Aldrich) and absorbance read at 570 nm in 96-well plates. Cell growth was calculated as percent of growth in the presence of insulin minus basal growth divided by maximum growth in complete medium [4]. 1.3. Generation of LratCre+Cc1fl/fl mice As in the conditional T cell-specific null mouse [5], the targeting Ceacam1 construct inserted a loxP- neo cassette in intron 6 and a loxP fragment in intron 9, deleting a sequence that encodes the cytoplasmic domain that is required for CEACAM1 phosphorylation by growth factor receptors [6; 7]. Cc1loxp/loxp mice were crossed with LratCre transgenic mice expressing a Cre recombinase directed by mouse lecithin-retinol acyltransferase [phosphatidylcholine -retinol-O-acyltransferase] (Lrat) promoter [8]. Heterozygotes were backcrossed >6x with C57BL/6J mice (Jackson laboratory). Stellate cell -specific deletion of Ceacam1 in homozygotes (LratCre+Cc1fl/fl or Lrat+Cc1fl/fl), was confirmed by PCR using gene-specific primers (Fig. S1). As controls, we used homozygotes with wild -type Ceacam1 allele with ( LratCre+Cc1+/+; Lrat+Cc1+/+) or without Cre ( LratCre–Cc1+/+; Lrat–Cc1+/+), and homozygotes with Ceacam1-floxed allele, without Cre (LratCre–Cc1fl/fl; Lrat–Cc1+/+). Genotyping of offspring was done by PCR analysis of ear DNA: FLOX A + FLOX B primers were used to detect the 383 bp wild -type allele (Cc1+/+), FLOX A + FLOX C detected the was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (whichthis version posted April 3, 2024. ; https://doi.org/10.1101/2024.04.02.586238doi: bioRxiv preprint 488 bp knockout-type allele (Cc1fl/fl). The LratCre reaction detected a 343 bp band in the LratCre+Cc1+/+ and LratCre+Cc1fl/fl. 1.4. Glucose and insulin tolerance tests As routinely done [9], male mice were kept in cages with Alpha -dri bedding (Shepherd Specialty Papers) and fasted for 6 hrs before being injected intraperitoneally with either 1.5g/kg BW dextrose solution (GTT or 0.75units/kg BW human regular insulin (Novo Nordisk, Princeton, NJ) (ITT) and their tail blood glucose was measured at 0-120 min (GTT) or 0-180 post-injection (ITT). 1.5. Biochemical parameters Following their move to Alpha-dri bedding (Shepherd Specialty Papers), mice were fasted overnight for 18 hrs, anesthetized with an IP injection of pentobarbital (1.1mg/kg BW) at 1100h the next morning and their venous blood drawn into heparinized micro -hematocrit capillary tubes (Catalog number 22 -362566, Fisherbrand, Waltham, MA). Plasma was processed and tissues excised to determine hepatic triacylglycerol as previously described [10]. Plasma was analyzed for insulin (80 -INSMSU-E01 ELISA kit; Alpco, Salem, NH), C-peptide (80-CPTMS-E01 ELISA kit; Alpco), non-esterified fatty acids (NEFA-C enzymatic colorimetric assay; Wako, Richmond, VA), Tumor necrosis factor-alpha (TNFα ELISA Kit, ab100747, Abcam), Interleukin- 6 (IL-6 ELISA Kit, ab222503, Abcam), alanine transaminase ( ALT ELISA KIT, 700260; Cayman Chemical Company, Ann Arbor, MI), and aspartate aminotransferase (AST ELISA, 701640; Cayman Chemical Company). 1.6. Liver histology and immunohistochemical analysis As routinely done, liver sections were fixed in 10% neutral buffered formalin (Sigma-Aldrich, Milwaukee, WI), paraffin-embedded, cut, mounted and stained with hematoxylin-eosin (H&E) to assess steatosis and lobular inflammation. Fibrosis was assessed by staining deparaffinized and rehydrated was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (whichthis version posted April 3, 2024. ; https://doi.org/10.1101/2024.04.02.586238doi: bioRxiv preprint slides with 0.1% Sirius Red stain (Direct Red 80, Sigma-Aldrich, St Louis, MO) in saturated aqueous solution of picric acid (Sigma-Aldrich), washed in 0.5% acetic acid (Sigma-Aldrich), hydrated in different alcohol percentages, and cleared in Xylene (Sigma-Aldrich). Images were taken using Nikon Eclipse 90i Microscope (Nikon, Melville, NY) as described [11]. 10 randomly selected high power fields (20x) per sample were imaged and quantified with ImageJ (v1.53t) to determine the percentage of Sirius red (SR) stain area within the whole area of imaged hepatic tissue. Briefly, each image was RGB-stacked, subjected to standardized thresholding of individual channel to isolated Sirius Red stain and then quantified as % area. Image quantifications were averaged and the mean within each experimental group was plotted in the accompanying graph. For immunohistochemical (IHC) analysis, FFPE sections were deparaffinized, dewaxed in xylene, rehydrated before quenching endogenous peroxidases with 3% H2O2 buffer. Antigen-retrieval was then performed by incubating slides in 10mM sodium citrate buffer (pH 6.0) or 10 mM Tris-EDTA buffer (pH 9.0) in a microwave. Blocking was done using 1% rabbit or mouse serum (Vector labs, Burlingame, CA) for 1 hr at room temperature. Slides were stained overnight at 4°C in a humidifier chamber with α-CD68 (1:150, polyclonal, Abcam, Cambridge, MA), α-Mac2 (1:250 monoclonal, Abcam), α-MPO (1:1000 monoclonal, Abcam),, α-CD4 (1:1000 monoclonal, Abcam), α-CD8 (1:2000 monoclonal, Abcam), and α-Foxp3 (1:100 Invitrogen, Waltham, MA). Slides were then incubated with appropriate species-specific biotinylated ImmPRESS HRP horse anti-mouse or anti-rabbit secondary antibody (Vector Labs, Burlingame, CA) for 30min before being treated with 3,3’-Diaminobenzidine (DAB, Vectastain kit-Vector labs, Newark, CA) and counterstain with Mayer’s hematoxylin. Images were taken using Nikon Eclipse 90i Microscope (Nikon, Melville, NY), as previously described [11]. Sections were evaluated blindly, and positively stained cells were counted in 5 fields/mouse at 40X magnification. 1.7. Isolation of primary murine cells Primary hepatocytes, bone marrow -derived macrophages, and endothelial cells from livers and hearts were isolated from ketamine/xylazine -anesthetized 2 -to-3–month-old male mice, as previously was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (whichthis version posted April 3, 2024. ; https://doi.org/10.1101/2024.04.02.586238doi: bioRxiv preprint described [11]. Briefly, hepatocytes [12] were maintained in DMEM (Gibco Lab, Gaithersburg, MD) -10% FBS, 1% glutamine, and 1% penicillin/streptomycin (Gibco). Bone marrow -derived macrophages [11] were obtained by flashing femurs and tibias with sterile Roswell Park Memorial Institute (RPMI) -1640 medium from Gibco supplemented with 10% horse serum, 1% L-glutamine, 1% Penicilin/Streptomycin and 10ng/ml Recombinant Mouse factor macrophage colony -stimulating factor [M -CSF] R&D Systems, Mi nneapolis, MN). Primary endothelial cells [13; 14] isolated from livers and hearts were maintained in DMEM -F12 (Gibco) supplemented with 20% FBS, 1% glutamine, and 1% penicillin/streptomycin (Gibco), 100 ug/ml Heparin (Sigma -Aldrich) and 100ug/ml Endothelial Cell Growth supplement [ECGS) (EMD Millipore Corporation Burlington MA). Primary mouse hepatic stellate cells (HSCs) were isolated from ≥8 -month-old mice, as described [15]. Mice were anesthetized and using laparotomy, liver and the inferior vena cava (IVC) were exposed and liver were perfused with in situ with EGTA solution (8,000 mg/l NaCl, 400 mg/l KCl, 88.17 mg/l NaH2PO4.H2O, 120.45 mg/l, NaHPO4 2,380 mg/l HEPES, 350 mg/l NaHCO3, 190 mg/l EGTA, and 900 mg/l Glucose) for 1-2 min, at a rate of 5ml/min. This was followed by perfusion with the Pronase solution for 5 min at 5ml/min. Pronase solution was prepared by dissolving 14 mg/mouse of pronase (Sigma -Aldrich) in 35 ml of enzyme buffer solution (8,000 mg/l NaCl, 400 mg/l KCl, 88.17 mg/l NaH2PO4, 120.45 mg/l, NaHPO4 2,380 mg/l HEPES, 350 mg/l NaHCO3, and 560 CaCl2.2H2O). The livers were then perfused for 7 min with collagenase solution composed of 3.7 U/mouse of Collagenase D (Roche, Pleasanton, CA) dissolved in 40 ml enzyme buffer solution. After perfusion, the partially digested livers were excised, placed into a sterile Petri dish containing Pronase/Collagenase solution prepa red by dissolving 25 mg/mouse of pronase and 4.4 U/mouse to 50 ml enzyme buffer solution and minced under cell culture hood. 1% DNase was added (Roche, Pleasanton, CA) before further digesting the liver at 400C for 20 min and the digest passed through a 70m cell strainer (Sigma-Aldrich) to remove undigested materials. The cell suspension was centrifuged at 580g for 10 min at 40. The pellet was then washed using 4°C Gey’s buffered salt solution B (GBSS/B) (8,000 mg/l NaCl, 370 mg/l KCl, 210 mg/l MgCl 2 .6H2O, 70 mg/l MgSO 4 7H2O, 75 mg/l Na 2HPO4  2H2O, 30 mg/l KH2PO4, 991 mg/l glucose, 227 mg/l NaHCO 3, and 225 mg/l CaCl 2 H2O, supplemented with DNase I by centrifugation at 580g at 40C for 10 min. HSCs were purified from the rest of cells including hepatocytes by was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (whichthis version posted April 3, 2024. ; https://doi.org/10.1101/2024.04.02.586238doi: bioRxiv preprint density gradient-mediated separation. Cells were resuspended in 32 ml Gey’s buffered salt solution (GBSS) before 16 ml Nycodenz solution [4.94 g of Nycodenz (Accurate Chemicals Westbury, NY) in 15 ml GBSS/B without NaCl] was added. Gently overlaid cells-Nycondenz suspension with 1.5 ml GBSS using 3 ml syringe with a 26 -gauge needle. Centrifuge the cell suspension at 1.380g at 4 0C for 17 min without brake. The interphase containing enriched HSC between the GBSS and Nycondenz layer was removed and washed using GBSS by centrifugation at 580g, 40C for 10 min. The cell pellet containing HSCs was resuspended in prewarmed culture media DMEM (Gibco, Gaithersburg, MD) supplemented with 10% fetal bovine serum (FBS), 0.04% Gentamycin, and 1% Antibiotic/antimyocotic (Gibco), placed in a humidified tissue culture incubator (37°C, 5%CO2). 1.8. Western blot analysis Proteins were extracted from tissues or cells using lysis buffer containing 150 mM NaCl -50 mM Hepes (pH 7.6), 0.02% sodium azide,1% Triton X-100, 50 mM NaF, PMSF, Na3VO4 and proteinase Inhibitor Cocktail Tablet (Roche Diagnostics Penzberg, Germany). Protein concentration was determined by Bio-Rad Protein Assay (Bio-Rad Laboratories, Hercules, CA) and 10-30 µg proteins were denatured by boiling in sodium dodecyl sulfate (SDS) buffer for 3 min and resolved by 7% SDS -PAGE, transferred onto nitrocellulose membrane (Bio -Rad Laboratories). After blocking with 3% -5% dry milk or bovine serum albumin (BSA) dissol ved in TBST (Tris -Buffered Saline, pH 7.2, 0.1% Tween 20), the membrane was incubated at 4 0C overnight with polyclonal antibodies (1:1000): phospho - Akt Ser 473, Akt, phospho-p44/42 MAPK (Thr 202/Tyr204), p44/42 MAPK, phospho -SMAD2Ser 465/467, SMAD2, phospho -SMAD3Ser 423/425, SMAD3, phospho-EGFR1 (Y1173), EGFR, Shc, phospho-NF-κB p65 (Ser536), NF-κB, phospho-Stat3 (Tyr705) Stat3, MMP9, MMP13 and Timp1 (Cell signaling, Danvers, MA), phospho -Insulin receptor beta (pIR β) (phospho- Y1361), Insulin receptor beta (C18C4) (Abcam), PCNA (Santa Cruz, Dallas, TX) , phospho-tyrosine (4G10) (EDM Milipore, Billerica, MA). The following custom-made rabbit polyclonal antibodies were used: Ab 3759 against the mouse CEACAM1 extracellular domain, and anti-human CEACAM1 clone 18/20, a gift from the late Dr. Bernhard Singer (University Hospital Essen, Germany). Next day the membrane was washed three times with TBST and incubated with horseradish peroxidase conjugated secondary antibodies antibody (GE was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (whichthis version posted April 3, 2024. ; https://doi.org/10.1101/2024.04.02.586238doi: bioRxiv preprint Healthcare Life Sciences, Amersham, Sunnyvale, CA, USA), for one hour at room temperature. For normalization, membranes were reprobed with monoclonal antibodies against tubulin (Cell signaling). Proteins were detected by chemiluminescence (Thermo Scientific, Rockford, IL). Name Company Catalog Number Primary Antibodies for Western blot Akt Cell Signaling Technology 9272 phospho-Akt (Ser473) Cell Signaling Technology 4051 EGFR1 Cell Signaling Technology 4267 phospho-EGFR1 (Y1173) Cell Signaling Technology 4407 Insulin receptor beta Abcam ab69508 phospho-Insulin receptor beta (pIRβ) (phospho-Y1361) Abcam Ab303492 p44/42 MAPK Cell Signaling Technology 9102 phospho-p44/42 MAPK (Thr202/Tyr204) Cell Signaling Technology 9101 Phosphotyrosine (4G10) EDM Millipore 05-321 MMP2 Cell Signaling Technology 40994 MMP9 Cell Signaling Technology 13667 MMP13 Cell Signaling Technology 69926 NF-κB p65 Cell Signaling Technology 8242 Phospho-NF-κB p65 (Ser536) Cell Signaling Technology 3031 PCNA Santa Cruz, Dallas Sc-56 Shc Cell Signaling Technology 50832 Smad2 Cell Signaling Technology 5339 phospho-Smad2 (Ser465/467) Cell Signaling Technology 3108 Smad3 Cell Signaling Technology 9523 phospho-Smad3 (Ser423/425) Cell Signaling Technology 9520 Stat3 Cell Signaling Technology 1264 phospho-Stat3 (Tyr705) Cell Signaling Technology 9145 TIMP1 Cell Signaling Technology 8946 Tubulin Cell Signaling Technology 2148 Secondary Antibodies for Western blot Donkey anti-rabbit IgG antibody (HRP) GE Healthcare Life Sciences NA934V Sheep anti-mouse IgG (HRP) GE Healthcare Life Sciences NXA931V IHC Antibodies-Mouse CD4 Abcam ab183685 CD8 Abcam ab209778 CD68 Abcam ab125212 Foxp3 Invitrogen 700914 MPO Abcam ab208670 Mac2 Abcam ab76245 was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (whichthis version posted April 3, 2024. ; https://doi.org/10.1101/2024.04.02.586238doi: bioRxiv preprint

[1] Ramakrishnan, S.K., Khuder, S.S., Al-Share, Q.Y., Russo, L., Abdallah, S.L., Patel, P.R., et al., 2016. PPARalpha (peroxisome proliferator-activated receptor alpha) activation reduces hepatic CEACAM1 protein expression to regulate fatty acid oxidation during fasting-refeeding transition. J. Biol. Chem. 291(15):8121-8129. [2] Ghadieh, H.E., Muturi, H.T., Russo, L., Marino, C.C., Ghanem, S.S., Khuder, S.S., et al., 2018. Exenatide induces carcinoembryonic antigen-related cell adhesion molecule 1 expression to prevent hepatic steatosis. Hepatol. Commun. 2(1):35-47. [3] Xu, L., Hui, A.Y., Albanis, E., Arthur, M.J., O'Byrne, S.M., Blaner, W.S., et al., 2005. Human hepatic stellate cell lines, LX-1 and LX-2: new tools for analysis of hepatic fibrosis. Gut 54(1):142-151. [4] Liu, J., Ramakrishnan, S.K., Khuder, S.S., Kaw, M.K., Muturi, H.T., Lester, S.G., et al., 2015. High- calorie diet exacerbates prostate neoplasia in mice with haploinsufficiency of Pten tumor suppressor gene. Mol. Metab. 4(3):186-198. [5] Nagaishi, T., Pao, L., Lin, S.H., Iijima, H., Kaser, A., Qiao, S.W., et al., 2006. SHP1 phosphatase- dependent T cell inhibition by CEACAM1 adhesion molecule isoforms. Immunity 25(5):769-781. [6] Najjar, S.M., Philippe, N., Suzuki, Y., Ignacio, G.A., Formisano, P., Accili, D., et al., 1995. Insulin- stimulated phosphorylation of recombinant pp120/HA4, an endogenous substrate of the insulin receptor tyrosine kinase. Biochemistry 34:9341-9349. [7] Abou-Rjaily, G.A., Lee, S.J., May, D., Al-Share, Q.Y., Deangelis, A.M., Ruch, R.J., et al., 2004. CEACAM1 modulates epidermal growth factor receptor--mediated cell proliferation. J. Clin. Invest. 114(7):944-952. was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (whichthis version posted April 3, 2024. ; https://doi.org/10.1101/2024.04.02.586238doi: bioRxiv preprint [8] Mederacke, I., Hsu, C.C., Troeger, J.S., Huebener, P., Mu, X., Dapito, D.H., et al., 2013. Fate tracing reveals hepatic stellate cells as dominant contributors to liver fibrosis independent of its aetiology. Nat. Commun. 4:2823. [9] Helal, R.A., Russo, L., Ghadieh, H.E., Muturi, H.T., Asalla, S., Lee, A.D., et al., 2021. Regulation of hepatic fibrosis by carcinoembryonic antigen-related cell adhesion molecule 1. Metabolism 121:154801. [10] Russo, L., Ghadieh, H.E., Ghanem, S.S., Al-Share, Q.Y., Smiley, Z.N., Gatto-Weis, C., et al., 2016. Role for hepatic CEACAM1 in regulating fatty acid metabolism along the adipocyte-hepatocyte axis. J. Lipid Res. 57(12):2163-2175. [11] Ghadieh, H.E., Abu Helal, R., Muturi, H.T., Issa, D.D., Russo, L., Abdallah, S.L., et al., 2020. Loss of Hepatic Carcinoembryonic Antigen-Related Cell Adhesion Molecule 1 Links Nonalcoholic Steatohepatitis to Atherosclerosis. Hepatol. Commun. 4(11):1591-1609. [12] Poy, M.N., Yang, Y., Rezaei, K., Fernstrom, M.A., Lee, A.D., Kido, Y., et al., 2002. CEACAM1 regulates insulin clearance in liver. Nat. Genet. 30(3):270-276. [13] Muturi, H.T., Ghadieh, H.E., Abdolahipour, R., Stankus, H.L., Z., Y., Liu, J.K., et al., 2023. Loss of CEACAM1 in endothelial cells causes hepatic fibrosis. Metabolism 144:155562. [14] Muturi, H.T., Khuder, S.S., Ghadieh, H.E., Esakov, E.L., Noh, H., Kang, H., et al., 2021. Insulin Sensitivity Is Retained in Mice with Endothelial Loss of Carcinoembryonic Antigen Cell Adhesion Molecule 1. Cells 10(8). [15] Mederacke, I., Dapito, D.H., Affo, S., Uchinami, H., Schwabe, R.F., 2015. High-yield and high-purity isolation of hepatic stellate cells from normal and fibrotic mouse livers. Nat. Protoc. 10(2):305-315. was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (whichthis version posted April 3, 2024. ; https://doi.org/10.1101/2024.04.02.586238doi: bioRxiv preprint Table S1 Primer sequences of mouse (m) and human (h) genes used in qRT-PCR. Primer Forward Sequence (5'-3') Reverse Sequence (5'-3') mAcsl1 TCTTTCTTGCCTCTCGCC GTCGTCCATAAGCAGCCT mAcsl4 GCCAGTGTGAACGTATCC GTGAAGAGTATCCAATCCTACAG mAtgl TGACCATCTGCCTTCCAGA TGTAGGCGCAAGACA mα-Sma CGTGGCTATTCCTTCGTTAC TGCCAGGAGACTCCATCC mCcl2 AGGTCCCTGTCATGCTTCTG GGGATCATCTTGCTGGTGAA mCd4 TCACCTGGAAGTTCTCTGACC GGAATCAAAACGATCAAACTGCG mCd4 CTCTGGCTGGTCTTCAGTATGA TCTTTGCCGTATGGTTGGTTT mCd11b TACGTAATTGGGGTGGGAA GTGCCCTCAATTGCAAAGAT mCd36 TCTTGGCTACAGCAAGGCGACATA AGCTATGCAGCATGGAACATGACG mCd68 CCTCGCCCTAGTCCAAGGTC CGATTCGGATTTGAATTTGGGCT mCeacam1 AATCTGCCCCTGGCGCTTGGAGCC AAATCGCACAGTCGCCTGAGTACG mCol1α1 TAGGCCATTGTGTATGCAGC ACATGTTCAGCTTTGTGGACC mCol3α1 CATACCTGGTACCGGTGGTC CACCGACTTCACCCTTTGGA mCpt1α AAGGCAGAAGAGTGGGCTTTCACT ACCTTGGCTGCGGTAAGACTATGT mDesmin AAGATGGCCTTGGATGTGGA GTTGATCCTGCTCTCCTCGC mElastase GTTGGGCACAAACAGACC GCAAACTCAGCCACAGG mFasn ACTGTGAGAAGCATGTCCCTGGAA AAGCAACCTCCACTCCTCTGCTTA mF4/80 CAAGGAGGACAGAGTTTATCGTG CTTTGGCTATGGGCTTTCAGTC mFoxp3 CCCAGGAAAGACAGCAACCTT TTTCACAACCAGGCCACTTG mIfnγ ATGAACGCTACACACTGCATC CCATCCTTTTGCCAGTTCCTC mGapdh CCAGGTTGTCTCCTGCGACT ATACCAGGAAATGAGCTTGACAAAGT mHgf CTTCTCCTTGGCCTTGAATG CCTGACACCACTTGGGAGTA mIl-4 AGGTCACAGGAGAAGGGACGCC TGCGAAGCACCTTGGAAGCCC mIl-6 GGCCTTCCCTACTTCACCAG ATTTCCACGATTTCCCAGAG mIl-10 CACAAAGCAGCCTTGCAGAA AGAGCAGGCAGCATAGCAGTG mIl-13 TGTTTCGCCACGGCCCCTTC TGCTCAAGCTGCTGCCTGCC mMmp2 CAACGGTCGGGAATACAGCAG CCAGGAAAGTGAAGGGGAAGA mLal/Lipa GCAAAGGTCCCAGACCAGTT TCATCAAAACTGAAGGCCCAGA was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (whichthis version posted April 3, 2024. ; https://doi.org/10.1101/2024.04.02.586238doi: bioRxiv preprint mLrat CAGATATGGCTCTCGGATCAG GACAATAGATGCTAATCCCAAGAC mMmp9 CGTGTCTGGAGATTCGACTTGA TGGAAGATGTCGTGTGAGTTCC mMmp13 CCTTCTGGTCTTCTGGCACAC GGCTGGGTCGTCACACTTCTCTGG mTimp1 GCATGGACATTTATTCTCCACTGT TCTCTAGGAGCCCGATCTG mTimp2 GCCAAAGCAGTGAGCGAGAAG GGGGAGGAGATGTAGCAAGGG mTimp3 CACGGAAGCCTCTGAAAGTC CCCAAAATTGGAGAGCATGT mNox1 GGATCCATGGCCTGGGTGGGAT GGATGCCTGCAACTCCCCTTATGG mNox4 TCCAAGCTCATTTCCCACAG CGGAGTTCCATTACATCAGAGG mNqo1 TATCCTTCCGAGTCATCTCTAGCA TCTGCAGCTTCCAGCTTCTTG mNrf1 AGCACGGAGTGACCCAAAC TGTACGTGGCTACATGGACCT mPparγ1 AGATCATCTACACGATGCTGGCCT ATAAAGTCACCAAAGGGCTTCCGC mPparβ/δ TGCTGGTATCGGCTCAATAA TCCTGCCACTTGCTCACTAC mPcna CACGTATATGCCGAGACCTTAGC CTCCACTTGCAGAAAACTTCACC mSmad7 GTTGCTGTGAATCTTACGGG ATCTGGACAGCCTGCA mSrebp-1c GGAGCCATGGATTGCACATT GCTTCCAGAGAGGAGGCCA mTgfβ GTGGAAATCAACGGGATCAG ACTTCCAACCCAGGTCCTTC mTnfα ACGGCATGGATCTCAAAGAC CGGACTCCGCAAAGTC mTxn GCCAAAATGGTGAAGCTGAT TGATCATTTTGCAAGGTCCA m18s TTCGAACGTCTGCCCTATCAA ATGGTAGGCACGGCGACT hα-SMA CGTGGCTATTCCTTCGTTAC TGCCAGCAGACTCCATCC hATGL ACCAGCATCCAGTTCAACCT ATCCCTGCTTGCACATCTCT hACSL4 CACGAATATCTTCTGTGATTTCAA ACGCGGTTCCTTTTTGC hACSL1 GCTTGCATTGTCCTGTGTTG GGAGTGGGCTGCAGTGAC hCEACAM1 TCTACCCTGAACTTTGAAGCCCA TGAGAGACTTGAAATACATCAGCAC TG hCD36 ATGTAACCCAGGACGCTGAG GTCGCAGTGACTTTCCCAAT hCol1a1 GATTCCCTGGACCTAAAGGTGC AGCCTCTCCATCTTTGCCAGCA hCPT-1α TCCAGTTGGCTTATCGTGGTG CTAACGAGGGGTCGATCTTGG hDGAT1 AGTTGCTGAAGCCACTGTCA CAACAAGGACGGAGACGC hFASN AACTCCAAGGCAACAGTCACCAT CAGCTGCTCCACGAACTCAA hGAPDH ATCCATGACAACTTTGTTATCGTG ATGACCTGGCCCACAGCCTT hLAL/LIPA CTAGAATCTGCCAGCAAGCC TGTGCCTTAACCGAATTCCT was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (whichthis version posted April 3, 2024. ; https://doi.org/10.1101/2024.04.02.586238doi: bioRxiv preprint hLRAT CCAAGACTGCTGAAGCAAGA TACTGCAGATATGGCACCCC hPPARβ/δ AAGAGCTTGGAGCTCGGC TGAAAGCGTGTCCGTGATGA hPPARγ1 CGTGGCCGCAGATTTGAA CTTCCATTACGGAGAGATCCAC hSREBP-1c CAGCATAGGGTGGGTCAAAT GAGCCGTGCGATCTGGA was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (whichthis version posted April 3, 2024. ; https://doi.org/10.1101/2024.04.02.586238doi: bioRxiv preprint Table S2 The effect of nicotinic acid on lipid metabolism of HSCs from KD-LX2 transferred into Scr-LX2 HSCs. KD KD +NA Scr Scr +NA Scr/Cond Scr/Cond +NA hLRAT 1.1±0.1* 1.7±0.2¶† 5.1 ± 0.3 5.2 ± 0.4 1.5 ± 0.1* 4.8 ± 0.3† hLAL(LIPA) 1.0±0.1* 1.1±0.1¶ 3.8 ± 0.3 4.1 ± 0.2 1.6 ± 0.1*§ 3.8 ± 0.2† hACSL1 1.1±0.1* 1.7±0.1¶† 3.0 ± 0.2 3.0 ± 0.1 1.2 ± 0.3* 3.2 ± 0.1† hACSL4 4.8 ± 0.3* 4.2±0.2¶† 1.1 ± 0.3 0.8 ± 0.1 3.3 ± 0.1*§ 1.1 ± 0.1† hACSL4/1 4.5 ± 0.4* 2.6±0.1¶† 0.4 ± 0.1 0.3 ± 0.1 2.8 ± 0.5*§ 0.3 ± 0.1† hSREBP-1c 1.8 ± 0.1* 1.1±0.1¶† 0.4 ± 0.1 0.5 ± 0.1 1.1 ± 0.2*§ 0.6 ± 0.1† hFASN 5.2 ± 0.6* 3.1±0.2¶† 0.7 ± 0.1 0.8 ± 0.1 3.4 ± 0.2*§ 0.8 ± 0.1† hDGAT1 3.4 ± 0.2* 3.0±0.1¶ 0.9 ± 0.2 1.0 ± 0.1 2.9 ± 0.2* 1.1 ± 0.3† hATGLl 4.3 ± 0.1* 3.2±0.3¶† 1.4 ± 0.2 1.1 ± 0.3 3.1 ± 0.5*§ 1.3 ± 0.2† Cells were treated with nicotinic acid (NA) or with buffer alone before the media were collected (conditioned) and transferred to pre-washed Scr-LX2 cells (Scr/Cond). qRT-PCR was conducted in triplicate to assess mRNA levels normalized to GAPDH. Values are expressed as mean ± SEM. *p<0.05 untreated KD and Scr/Cond vs untreated Scr-LX2 cells. †p<0.05 NA-treated vs untreated/cell line. §p<0.05 untreated Scr/Cond vs untreated KD. ¶p<0.05 NA-treated KD vs NA-treated Scr and NA-treated Scr/Cond cells. was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (whichthis version posted April 3, 2024. ; https://doi.org/10.1101/2024.04.02.586238doi: bioRxiv preprint Table S3 qRT-PCR analysis in primary cells incubated in conditioned media from Cc1–/– HSCs. Cc1–/– Cc1–/– +NA Cc1+/+ Cc1+/+ +NA Cc1+/+/Cond Cc1+/+/Cond +NA Ceacam1 negligible negligible 1.12±0.01 1.05±0.02 0.53±0.02*§ 1.08±0.03† Pparβ/δ 2.12±0.01* 1.45±0.03¶† 1.11±0.01 1.05±0.02 1.94±0.03* 1.05±0.03† Pcna 2.27±0.02* 1.43±0.03¶† 1.07±0.04 1.02±0.02 2.03±0.03* 1.06±0.03† -Sma 2.29±0.02* 1.44±0.03¶† 1.12±0.01 1.05±0.02 2.13±0.04* 1.08±0.03† Srebp-1c 1.28±0.03* 0.58±0.02¶† 0.34±0.02 0.34±0.02 0.83±0.02* 0.36±0.02† Fasn 2.38±0.01* 1.64±0.02¶† 1.25±0.03 1.26±0.03 2.01±0.03* 1.40±0.06† Primary HSCs were isolated from ≥8 -month-old mice (n=7/genotype), pre-incubated with (+) or without ( –) nicotinic acid (NA) before Cc1–/– media were collected and transferred into Cc1+/+ control HSCs (Cc1+/+/cond). qRT-PCR analysis was carried out in triplicate, normalized against Gapdh and presented as mean ± SEM. *p<0.05 untreated HSCs from Cc1–/– and Cc1+/+/Cond vs untreated Cc1+/+cells. †p<0.05 NA-treated vs untreated/cell line. §p<0.05 untreated Cc1+/+/Cond vs untreated Cc1–/–. ¶p<0.05 NA-treated Cc1–/– vs NA-treated Cc1+/+ and NA-treated Cc1+/+/Cond cells. was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (whichthis version posted April 3, 2024. ; https://doi.org/10.1101/2024.04.02.586238doi: bioRxiv preprint Table S4 mRNA levels of genes in liver lysates. Lrat–Cc1+/+ Lrat+Cc1+/+ Lrat–Cc1fl/fl Lrat+Cc1fl/fl Lipid metabolism Srebp1c 1.03 ± 0.03 1.01 ± 0.14 1.04 ± 0.09 1.14 ± 0.08 Fasn 2.92 ± 0.57 3.00 ± 0.74 3.42 ± 0.38 3.74 ± 0.30 Cd36 1.92 ± 0.20 1.95 ± 0.56 2.25 ± 0.26 2.21 ± 0.21 Inflammation F4/80 2.10 ± 0.37 2.78 ± 0.35 3.63 ± 0.51 12.81 ± 1.75* Cd68 1.93 ± 0.11 2.29 ± 0.37 3.27 ± 0.43 10.02 ± 1.80* Mpo 1.62 ± 0.18 1.77 ± 0.20 2.71 ± 0.11 4.61 ± 0.42* Elastase 0.66 ± 0.07 0.69 ± 0.17 0.75 ± 0.19 2.52 ± 0.70* Cd4 1.18 ± 0.08 1.25 ± 0.09 1.53 ± 0.20 1.52 ± 0.17 Cd8 0.84 ± 0.06 0.95 ± 0.06 1.00 ± 0.07 1.03 ± 0.09 Il-10 1.00 ± 0.28 0.96 ± 0.12 1.10 ± 0.29 0.90 ± 0.15 Oxidative stress Nox1 0.99 ± 0.08 1.28 ± 0.12 1.33 ± 0.11 2.16 ± 0.17* Nox4 1.00 ± 0.07 0.89 ± 0.10 1.05 ± 0.10 1.53 ± 0.19* Liver lysates from male mice (10 months of age, n=6/each genotype) were analyzed in duplicate by qRT- PCR using gene-specific primers and normalized to Gapdh. Values are expressed as mean ± SEM. *P<0.05 vs all three controls. was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (whichthis version posted April 3, 2024. ; https://doi.org/10.1101/2024.04.02.586238doi: bioRxiv preprint Table S5 The effect of EGFR inhibitor, Gefitinib, on lipid metabolism of primary HSCs from LratCre+Cc1fl/fl mice. Lrat+Cc1fl/fl Lrat+Cc1fl/fl Gef WT WT + Gef WT/ Cond WT/ Cond+Gef mLrat 0.8 ± 0.1* 1.1 ± 0.1 3.7 ± 0.1 3.7 ± 0.0 1.3 ± 0.1* 3.4 ± 0.2† mLal(Lipa) 0.7 ± 0.0* 1.5 ± 0.0¶† 2.1 ± 0.1 2.1 ± 0.1 0.8 ± 0.1* 1.8 ± 0.1† mAcsl1 0.5 ± 0.1* 1.1 ± 0.1¶† 2.6 ± 0.1 2.5 ± 0.1 0.9 ± 0.1*§ 2.0 ± 0.1† mAcsl4 4.0 ± 0.2* 3.0 ± 0.2¶† 1.1 ± 0.1 1.1 ± 0.0 2.4 ± 0.1*§ 1.2± 0.0† mAcsl4/Acsl1 6.7 ± 1.9* 3.8 ± 1.0¶† 0.4 ± 0.2 0.4 ± 0.3 2.5 ± 0.2*§ 1.6 ± 0.4† mDgat1 2.9 ± 0.1* 1.6 ± 0.1¶† 0.5 ± 0.1 0.4 ± 0.1 1.2 ± 0.1*§ 0.6 ± 0.1† mAtgl 4.1 ± 0.2* 2.8 ± 0.2¶† 1.1 ± 0.1 1.0 ± 0.2 2.3 ± 0.1*§ 1.0 ± 0.0† HSCs were isolated from LratCre+Cc1fl/fl and a combination of wild -type (WT) mice ( LratCre–Cc1+/+, LratCre+Cc1+/+ and LratCre–Cc1fl/fl). Some WT cells were incubated with conditioned media from LratCre+Cc1fl/fl HSCs in the presence of Gefitinib (Gef) or its absence. qRT-PCR analysis was carried out in triplicate, normalized against Gapdh and presented as mean ± SEM. *p<0.05 untreated HSCs from LratCre+Cc1fl/fl and WT/Cond vs untreated WT cells. †p<0.05 treated vs untreated/cell line. §p<0.05 untreated WT/Cond vs untreated LratCre+Cc1fl/fl. ¶p<0.05 treated LratCre+Cc1fl/fl vs NA-treated WT and treated WT/Cond cells. was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (whichthis version posted April 3, 2024. ; https://doi.org/10.1101/2024.04.02.586238doi: bioRxiv preprint Fig. S1. Genotyping of LratCre+Cc1fl/fl mutants. PCR amplification was performed on ear DNA using primers for the FloxA/FloxB pair detected the 382bp wild -type allele ( Cc1+/+) and the FloxA/FloxC pair detected the 488bp null when the primer sets of the Cre reaction were combined. The LratCre+ allele was detected by the 343bp Lrat promoter. Nucleotide sequences were listed in the inserted table. FP denotes forward primer and RP denotes reverse primer. was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (whichthis version posted April 3, 2024. ; https://doi.org/10.1101/2024.04.02.586238doi: bioRxiv preprint Fig. S2. CEACAM1 loss in LX2 cells induces cell growth. Serum-starved Src-LX2 and KD-LX2 cells were treated with insulin (100 nM) for 5 min, lysed and proteins analyzed by immunoblotting (Ib) : (A) the α-pTyr immunopellet (Ip) with α-pCEACAM1 (α-pCC1) antibody; (B) total lysates with α-pIRβ followed by re- immunoblotting (relb) with α-pIRβ to normalize for loading; (C) total lysates with α-pMAPK followed by re - immunoblotting (relb) with α-MAPK. (D) Cells were grown in 96 well plates and treated with insulin for 24 hrs before cell growth was assayed in triplicate by MTT assay. Values represent mean ± SEM; *p<0.05 KD- LX2 vs Scr-LX2 cells/treatment group; †p<0.05 insulin-treated vs untreated/cell line. was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (whichthis version posted April 3, 2024. ; https://doi.org/10.1101/2024.04.02.586238doi: bioRxiv preprint Fig. S3. Absence of inflammatory infiltration in 8 -month-old mice. Livers were removed from 8 -month-old LratCre+Cc1fl/fl male mice and their 3 littermate controls (n=4 –5/genotype) to carry out (A) H&E staining as in the legend of Fig. 5. (B) Western analysis to assess the phosphorylation of the p65 subunit of NF-B and Stat3 (activation), as in the legend of Fig. 6 and (C) qRT-PCR analysis of the mRNA levels of genes involved in inflammation, as in the legend of Fig. 6. was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (whichthis version posted April 3, 2024. ; https://doi.org/10.1101/2024.04.02.586238doi: bioRxiv preprint Fig. S4. Absence of hepatic fibrosis in 8 -month-old mice. Livers were removed from 8 -month-old LratCre+Cc1fl/fl male mice and their 3 littermate controls (n=4–5/genotype) to carry out (A) Sirius red staining of liver sections from mutant mice (panel d) and their littermate controls (panels a-c), as in the legend of Fig. 7. (B) qRT-PCR analysis of the mRNA levels of hepatic genes involved in fibrosis, as in the legend of Fig. 7. (C) Western analysis to assess Smad3 phosphorylation (activation), as in the legend of Fig. 7. was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (whichthis version posted April 3, 2024. ; https://doi.org/10.1101/2024.04.02.586238doi: bioRxiv preprint