Role of Mcpip1 in obesity-induced hepatic steatosis as determined by myeloid and liver-specific conditional knockouts.

Monocyte chemoattractant protein-induced protein 1 (MCPIP1, alias Regnase 1) is a negative regulator of inflammation, acting through cleavage of transcripts coding for proinflammatory cytokines and by inhibition of NFκB activity. Moreover, it was demonstrated that MCPIP1 regulates lipid metabolism both in adipose tissue and in hepatocytes. In this study, we investigated the effects of tissue-specific Mcpip1 deletion on the regulation of hepatic metabolism and development of nonalcoholic fatty liver disease (NAFLD). We used control Mcpip1fl/fl mice and animals with deletion of Mcpip1 in myeloid leukocytes (Mcpip1fl/fl LysMCre ) and in hepatocytes (Mcpip1fl/fl AlbCre ), which were fed chow or a high-fat diet (HFD) for 12 weeks. Mcpip1fl/fl LysMCre mice fed a chow diet were characterized by a significantly reduced hepatic expression of genes regulating lipid and glucose metabolism, which subsequently resulted in low plasma glucose level and dyslipidemia. These animals also displayed systemic inflammation, demonstrated by increased concentrations of cytokines in the plasma and high Tnfa, Il6, IL1b mRNA levels in the liver and brown adipose tissue (BAT). Proinflammatory leukocyte infiltration into BAT, together with low expression of Ucp1 and Ppargc1a, resulted in hypothermia of 22-week-old Mcpip1fl/fl LysMCre mice. On the other hand, there were no significant changes in phenotype in Mcpip1fl/fl AlbCre mice. Although we detected a reduced hepatic expression of genes regulating glucose metabolism and β-oxidation in these mice, they remained asymptomatic. Upon feeding with a HFD, Mcpip1fl/fl LysMCre mice did not develop obesity, glucose intolerance, nor hepatic steatosis, but were characterized by low plasma glucose level and dyslipidemia, along with proinflammatory phenotype. Mcpip1fl/fl AlbCre animals, following a HFD, became hypercholesterolemic, but accumulated lipids in the liver at the same level as Mcpip1fl/fl mice, and no changes in the level of soluble factors tested in the plasma were detected. We have demonstrated that Mcpip1 protein plays an important role in the liver homeostasis. Depletion of Mcpip1 in myeloid leukocytes, followed by systemic inflammation, has a more pronounced effect on controlling liver metabolism and homeostasis than the depletion of Mcpip1 in hepatocytes.

alanine aminotransferase

aspartate aminotransferase

body mass index

extracellular matrix

glucose tolerance test

high‐density lipoprotein

high‐fat diet

lactate dehydrogenase

low‐density lipoprotein

nonalcoholic fatty liver disease

nonalcoholic steatohepatitis

Introduction

Monocyte chemoattractant protein‐induced protein 1 (MCPIP1, alias Regnase 1), encoded by the ZC3H12A gene, is a RNase that degrades mRNAs, pre‐miRNAs, and viral RNAs [1]. MCPIP1 functions as a negative regulator of inflammation, by direct cleaving of transcripts coding for proinflammatory cytokines like IL‐1β, IL‐6, IL‐8, and IL‐12 [2, 3, 4, 5], and by indirect inhibition of NFκB activity [6]. Apart from immunomodulatory effects, MCPIP1 activity was shown to regulate many biological processes, such as cell differentiation, angiogenesis, proliferation, and adipogenesis [7, 8, 9]. Adipogenesis of 3T3‐L1 cells is inhibited via the degradation of C/EBPβ and selected pre‐miRNAs by MCPIP1 [9, 10]. Additionally, the MCPIP1 level in human adipose tissue inversely correlates with patients' BMI [11]. Despite the results mentioned above, the functions of Mcpip1 in liver metabolism and nonalcoholic fatty liver disease (NAFLD) progression remain obscure.

Previous studies have shown that Mcpip1‐deficient mice spontaneously develop systemic inflammatory response, leading to splenomegaly, lymphadenopathy, hyperimmunoglobulinemia, and ultimately death within 12 weeks [3, 12]. Similarly, a hematopoietic deficiency of Mcpip1 resulted in severe systemic and multiorgan inflammation [13]. It was also shown that overexpression of Mcpip1 reduces liver injury in septic mice, by inhibiting inflammatory reaction in macrophages [14]. Additionally, Mcpip1 ameliorates liver damage, reduces inflammation, and promotes tissue regeneration in the hepatic ischemia/reperfusion injury model [15].

Nonalcoholic fatty liver disease describes a wide range of liver conditions affecting people who drink little to no alcohol. It is a chronic and progressive disease, characterized in the first stages by an excessive accumulation of triglycerides in hepatocytes [16]. Untreated NAFLD can progress from simple steatosis, to more severe nonalcoholic steatohepatitis (NASH), which is additionally characterized by liver inflammation and fibrosis [17]. Ultimately, NASH can further progress to cirrhosis and/or hepatocellular carcinoma, which over time might require liver transplantation [18].

In present study, we aimed to examine the contribution of Mcpip1 to liver metabolism and the pathogenesis of NAFLD. Specifically, we used Cre‐LoxP technology to delete Zc3h12a gene encoding Mcpip1 in myeloid leukocytes (Mcpip1fl/flLysMCre) and in hepatocytes (Mcpip1fl/flAlbCre). Myeloid cells are represented in the liver mostly by Kupffer cells that make up 80% of whole‐body macrophages, which is the largest population of tissue macrophages in solid organs [19]. Crosstalk between Kupffer cells and hepatocytes, mediated for example by IL‐1β, was shown to promote hepatic triglyceride storage through IL‐1β‐dependent suppression of PPARα activity and subsequent inhibition of β‐oxidation [20]. Also, blood monocytes infiltrating the liver during NAFLD are involved in its progression through the production of a vast array of proinflammatory mediators [21]. In our study, by comparing Mcpip1fl/flLysMCre and Mcpip1fl/flAlbCre animal models, we were able to analyze the tissue‐specific effects of Mcpip1 on liver metabolism and NAFLD development. Our results provide comprehensive data on the role of the Mcpip1 expression in the myeloid leukocytes and hepatocytes in the regulation of metabolism and sterile inflammation.

Results

Deletion of Mcpip1 in myeloid cells, but not in hepatocytes, results in dyslipidemia

Control Mcpip1fl/fl animals used in this study were born in both Mcpip1fl/flLysMCre and Mcpip1f/flAlbCre mice colonies. In the first analysis, we tested whether there is a difference between these animals. As compared in Fig. S1, there were no changes in mass, blood biochemistry, gene expression, nor in plasma cytokine concentrations in both Mcpip1fl/fl control cohorts (Fig. S1). Thus, in the next figures, all data are presented with one control group. Mcpip1fl/flLysMCre animals were already characterized by Li and coworkers, who demonstrated diminished expression of Zc3h12a encoding Mcpip1 in macrophages and spleens, but not in livers in comparison with Mcpip1fl/fl littermates [22]. On the other hand, Zc3h12a expression was lower in the livers, but not in VAT of Mcpip1fl/flAlbCre mice (Fig. S2).

Starting from the 7th week of the experiment, Mcpip1fl/flLysMCre mice were characterized by reduced body weight, although they consumed an equal amount of food compared to Mcpip1fl/fl mice (Fig. 1A,B, Fig. S3A). At the age of 22 weeks, Mcpip1fl/flLysMCre exhibited hepatomegaly (0.053 ± 0.001 vs 0.037 ± 0.001 g), splenomegaly (0.029 ± 0.001 vs 0.004 ± 0.0004 g), and significantly lower cholesterol (1.6 ± 0.1 vs 2.4 ± 0.05 mmol·L−1) and HDL (0.68 ± 0.02 vs 1.52 ± 0.03 mmol·L−1) levels with concomitant increased concentration of LDL (0.39 ± 0.03 vs 0.25 ± 0.01 mmol·L−1), when compared to control Mcpip1fl/fl mice (Fig. S3B,C, Fig. 1C–E). We did not observe differences in the concentration of triglycerides, nor activity of ALT, AST, and LDH in Mcpip1fl/flLysMCre animals compared with Mcpip1fl/fl mice (Fig. 1F,G, Fig. S3D,E). Low plasma glucose level in Mcpip1fl/flLysMCre mice was demonstrated by glucose tolerance test (Fig. 1H,I). Additionally, all measured analytes did not differ between control and Mcpip1fl/flAlbCre mice; however, Mcpip1 depletion in hepatocytes resulted in hepatomegaly (0.037 ± 0.001 vs 0.046 ± 0.003 g) (Fig. 1C–I, Fig. S3B–E).

Fig. 1

Mcpip1fl/flLysMCre mice are characterized by reduced body weight, dyslipidemia, and low plasma glucose level. (A) Mcpip1fl/fl, Mcpip1fl/flLysMCre, and Mcpip1fl/flAlbCre mice were analyzed at the age of 10 and 22 weeks. At 10 weeks of age, body temperature, weight, locomotor activity, and brown adipose tissue were analyzed. After feeding for 12 weeks with chow diet, we additionally performed glucose tolerance test, plasma, liver, and visceral adipose tissue analysis. (B) Body weight measurements of Mcpip1fl/fl, Mcpip1fl/flLysMCre, and Mcpip1fl/flAlbCre mice fed chow diet. Plasma analysis of 22‐week‐old mice: (C) cholesterol; (D) HDL; (E) LDL; (F) triglycerides; (G) LDH levels. Glucose metabolism was tested after intraperitoneal injection of 2 g·kg−1 of glucose, by (H) glucose tolerance test, and by (I) calculations of the area under the curve (AUC). For Mcpip1fl/fl, n = 10 (B–G) and n = 8 (H, I); for Mcpip1fl/flLysMCre, n = 6 (B–G) and n = 5 (H, I); and for Mcpip1fl/flAlbCre, n = 8 (B–G) and n = 6 (H, I). The graphs show means ± SEM; *P < 0.05; **P < 0.01; ***P < 0.001.

Mcpip1 depletion leads to liver pathology

Histological analysis of livers collected from Mcpip1fl/flLysMCre and Mcpip1fl/flAlbCre mice showed pathology of the bile ducts, each of them with different nature. Livers of Mcpip1fl/fl mice did not show any pathological changes (Fig. 2A,B, left panel, proper bile ducts marked by asterisks, Fig. S4 left panel). The dominant pathology observed in Mcpip1fl/flLysMCre mice was located within large interstitial bile ducts and included hyperplasia and hypertrophy of cholangiocytes (Fig. 2A,B, middle panel, closed arrows, Fig. S4, middle panel). The hyalinization and ultimately cytoplasmic degranulation and cholangiocytes disintegration were also observed. Moreover, noted changes were accompanied by intense inflammatory infiltration (Fig. 2A, middle panel, open arrows, Fig. S4 middle panel) and local fibrosis (Fig. 2B, middle panel, arrowheads, Fig. S4, middle panel) without penetration into the liver parenchyma. Additionally, Mcpip1fl/flLysMCre mice were characterized by an increased extramedullary hematopoiesis in the liver located mainly in the walls of the sublobular veins, resulted in their significant thickening and fibrosis. The necrosis of single hepatocytes was found occasionally. In turn, histological analysis of livers collected from Mcpip1fl/flAlbCre mice revealed active and progressive proliferation of intrahepatic bile ducts located peripherally (Fig. 2A, right panel, open arrows) accompanied by extensive parenchymal inflammation and fibrosis (Fig. 2B, right panel, arrowheads), along with fibrosis in portal areas (Fig. S4 right panel, asterisk) reflecting most of the features of human primary biliary cholangitis (PBC) (presented and discussed in details in [23]).

Fig. 2

Mcpip1 depletion leads to liver pathology. Representative liver (A) hematoxylin and eosin (H&E) and (B) Picro Sirius Red (PSR) stainings. Asterisks point to proper bile ducts in control mice. Closed arrows point to dysplasia and proliferation of bile ducts with accompanying inflammation (open arrows) and collagen deposition (arrowheads) in Mcpip1fl/flLysMCre and Mcpip1fl/flAlbCre mice, respectively. The scale bar represents 200 µm.

Myeloid and hepatic Mcpip1 deletion alters expression of genes regulating glucose metabolism

In order to identify the molecular mechanism responsible for the low plasma glucose level and development of dyslipidemia in Mcpip1fl/flLysMCre animals, we analyzed a hepatic gene expression. The analysis revealed that Mcpip1fl/flLysMCre exhibits a significantly reduced expression of G6pc (regulating gluconeogenesis), Irs1, Irs2 (response to insulin), and Slc2a2 (glucose bidirectional transport) in livers (Fig. 3A). Changes in the livers of Mcpip1fl/flAlbCre mice were not as significant as in the case of Mcpip1fl/flLysMCre, we detected lower expression of Pck1, G6pc, and Irs1 in comparison with Mcpip1fl/fl mice (Fig. 3A).

Fig. 3

Expression of Mcpip1 is essential for proper liver metabolism. Expression of genes coding for proteins involved in (A) glucose metabolism; (B) lipid transport; (C) lipogenesis; (D) oxidation of fatty acids; (E) inflammation; and (F) fibrosis. For Mcpip1fl/fl, n = 8; for Mcpip1fl/flLysMCre, n = 6; and for Mcpip1fl/flAlbCre, n = 9. The graphs show means ± SEM; *P < 0.05; **P < 0.01; ***P < 0.001.

Mcpip1 depletion disturbs lipid metabolism in the livers of Mcpip1fl/flLysMCre and Mcpip1fl/flAlbCre mice

Besides regulating glucose metabolism, the liver is also a prominent regulator of lipid metabolism, due to the ability to perform lipoprotein synthesis and lipid storage [24]. Therefore, in the following set of experiments, we have evaluated the expression of key genes related to lipid transport, lipogenesis, and fatty acid oxidation. As shown in Fig. 3B–D, more significant changes in hepatic lipid metabolism were detected in Mcpip1fl/flLysMCre than in Mcpip1fl/flAlbCre mice (Fig. 3B–D). The expression of Cd36 was increased, while Mttp was downregulated in the livers of Mcpip1fl/flLysMCre mice (Fig. 3B). Among the three master regulators of lipogenesis (Srebf1, Cebpb, Pparg) [25, 26], we detected a lower expression of Srebf1 in samples from Mcpip1fl/flLysMCre mice, together with the downregulation of SREBP1c‐regulated genes, Acc1, Fasn, and Scd1 (Fig. 3C). Furthermore, the expression of Ppara was reduced in the hepatic tissue of Mcpip1fl/flLysMCre mice, together with its downstream targets Acox1 and Cpt1a, but not Acadl (Fig. 3D). In livers isolated from Mcpip1fl/flAlbCre animals, there were no differences in the expression of genes regulating lipid transport (Cd36, Fabp1, Mttp), but the amount of Acox1, Cpt1a, and Scd1 transcripts was significantly reduced in comparison with control Mcpip1fl/fl mice (Fig. 3B–D).

Deletion of Mcpip1 in myeloid cells, but not in hepatocytes, induces systemic inflammation

To further characterize the phenotype of both Mcpip1 knockout strains, we performed a Luminex Assay. Out of 45 analytes, 14 were not detected in all samples, whereas an expression of 20 was significantly changed between Mcpip1fl/flLysMCre and control mice (Table 1). The level of cytokines and its receptors (Tnf‐alpha, Il‐6 R alpha, Tnf RI, Tnf RII) was significantly upregulated in Mcpip1fl/flLysMCre in comparison with Mcpip1fl/fl mice. Cytokines that are chemotactic for immune cells, for example, Ccl8, Ccl12, Ccl19, were significantly induced in Mcpip1fl/flLysMCre animals. Concentrations of other inflammation markers, such as Syndecan‐1/CD138, Chi3‐L1, and p‐Selectin, were also increased in Mcpip1fl/flLysMCre animals compared with control mice (Table 1). In myeloid Mcpip1 knockouts, we detected a reduced amount of three proteins, namely Vegf‐R2, thrombospondin 4, and endoglin. Surprisingly, concentrations of all tested analytes were not changed between Mcpip1fl/fl and Mcpip1fl/flAlbCre mice, indicating a key immunomodulatory role of myeloid Mcpip1.

Table 1

Immune profile of plasma samples from 22‐week‐old mice determined by Luminex Assay. Out of 45 analytes, 14 has not been detected, namely IL‐2, IL‐1β, IL‐3, IL‐7, CCL4/MIP‐1 beta, CCL2/MCP1, RAGE, IL‐17e/IL‐25, IL‐17/IL‐17a, IL‐1a, IL‐4, IL‐33, CCL5/RANTES, FGF‐basic. For Mcpip1fl/fl n = 6, for Mcpip1fl/flLysMCre n = 7, for Mcpip1fl/flAlbCre n = 7. The table shows means ± SEM; nd, not detected.

Function	Analyte (pg·mL−1)	Mcpip1fl/fl	Mcpip1fl/flLysMCre	Mcpip1fl/flAlbCre
Interleukins	IL‐5	nd	0.047 ± 0.013*	nd
	IL‐6	nd	0.018 ± 0.006*	nd
	IL‐6 R alpha	14.99 ± 1.06	30.12 ± 2.41*	14.55 ± 0.76
	IL‐10	nd	0.030 ± 0.007*	nd
	IL‐12 p70	nd	0.140 ± 0.065	nd
	IL‐13	nd	0.252 ± 0.025*	nd
	IL‐16	0.463 ± 0.104	0.561 ± 0.174	0.434 ± 0.075
	IL‐27	nd	0.023 ± 0.008*	nd
Chemokines	CCL8/MCP‐2	127.16 ± 10.59	525.62 ± 60.80*	141.83 ± 11.88
	CCL7/MARC	0.063 ± 0.015	0.204 ± 0.015*	0.060 ± 0.019
	CCL12/MCP‐5	0.027 ± 0.011	0.280 ± 0.037*	0.046 ± 0.010
	CCL3/MIP‐1 alpha	0.970 ± 0.055	3.837 ± 1.172*	0.890 ± 0.051
	CCL20/MIP‐3 alpha	0.126 ± 0.040	0.097 ± 0.048	0.028 ± 0.028
	CCL19/MIP‐3 beta	0.006 ± 0.004	0.040 ± 0.008*	0.0003 ± 0.0002
	CCL11/Eotaxin	1.44 ± 0.08	1.52 ± 0.15	1.34 ± 0.08
	CXCL16	0.358 ± 0.027	0.203 ± 0.045	0.267 ± 0.053
Growth factors	GM‐CSF	0.0005 ± 0.0002	0.0009 ± 0.0004	0.0004 ± 0.0001
	M‐CSF	4.84 ± 1.03	2.55 ± 0.48	6.73 ± 1.68
	G‐CSF	0.125 ± 0.032	0.059 ± 0.014	0.097 ± 0.021
TNF	TNF‐alpha	2.94 ± 0.03	4.13 ± 0.23*	2.97 ± 0.09
	TNF RI	0.753 ± 0.054	1.160 ± 0.120*	0.759 ± 0.046
	TNF RII	3.04 ± 0.20	8.63 ± 1.03*	3.48 ± 0.49
Angiogenesis	VEGF	nd	0.0030 ± 0.0006*	nd
	VEGF R2	49.70 ± 2.11	29.77 ± 2.13*	48.09 ± 0.68
	Endoglin	1.88 ± 0.30	1.00 ± 0.14*	2.27 ± 0.23
Others	Thrombospondin‐4	705.99 ± 99.09	380.89 ± 46.32*	556.98 ± 31.23
	CHI3‐L1	29.40 ± 3.70	145.56 ± 21.41*	42.59 ± 9.01
	Icam‐1	18.06 ± 2.37	13.69 ± 2.36	17.88 ± 0.75
	p‐Selectin/CD‐62P	29.01 ± 4.49	46.23 ± 8.13	29.09 ± 4.93
	Syndecan‐1/CD138	4.58 ± 0.74	17.51 ± 5.14*	7.13 ± 0.69
	IFN gamma	0.03100 ± 0.00007	0.0320 ± 0.0002	0.03100 ± 0.00009

P < 0.05; **P < 0.01; ***P < 0.001.

Systemic inflammation in Mcpip1fl/flLysMCre mice was followed by a proinflammatory hepatic gene expression pattern. Expression of Il6, Il1b, as well as Tnfa, was significantly upregulated in the livers of Mcpip1fl/flLysMCre in comparison with Mcpip1fl/fl mice. Livers of Mcpip1fl/flLysMCre were characterized by increased infiltration of macrophages, T and B lymphocytes, reflected by upregulated expression of Cd68, Cd3e, and Cd19. However, there were no changes between Mcpip1fl/fl and Mcpip1fl/flAlbCre samples, beside increased expression of Cd3e (Fig. 3E). Moreover, we have found a significantly increased expression of profibrotic genes—Tgfb, Mmp3, and Mmp9—in the livers of Mcpip1fl/flLysMCre. On the other hand, Mcpip1fl/flAlbCre mice were characterized by slightly, but not significantly higher expression of these selected genes (Fig. 3F).

High‐fat diet does not induce obesity and hepatic steatosis in Mcpip1fl/flLysMCre mice

In agreement with literature data, High‐fat diet (HFD) feeding increased body mass and liver steatosis of Mcpip1fl/fl mice (Fig. 4A–C). However, starting from the sixth week of experiment, Mcpip1fl/flLysMCre mice were characterized by significantly decreased body weight that cannot be fully attributed to lower food consumption (Fig. 4A, Fig. S5A). Additionally, Mcpip1fl/flLysMCre mice, after 12 weeks of HFD, were characterized by very limited liver steatosis (3% vs 37% in Mcpip1fl/fl) (Fig. 4B and Fig. S6 middle panel, Fig. 4C), hepatomegaly (0.059 ± 0.002 vs 0.031 ± 0.001 g), splenomegaly (0.031 ± 0.003 vs 0.002 ± 0.0003 g), (Fig. S5B,C), and diffuse hematopoietic foci (Fig. 4B, middle panel, arrowheads). On the other hand, 12 weeks of HFD feeding to Mcpip1fl/flAlbCre mice resulted in hepatomegaly (0.045 ± 0.001 vs 0.031 ± 0.001 g), (Fig. S5B) without impact on body mass nor liver steatosis (Fig. 4A,B). No changes in Mcpip1fl/flAlbCre mice phenotype, characterized as in chow diet fed mice by very pronounced collagen deposition (Fig. 4B, right panel, open arrows) and inflammation (Fig. 4B, right panel, orange closed arrows) accompanying proliferation of intrahepatic bile ducts (Fig. 4B, right panel, black closed arrows), were observed after HFD as well. Changes in lipid profile measured in plasma of both Mcpip1 knockouts, after 12 weeks of HFD, resembled results obtained in mice fed chow food: We detected low cholesterol (2.19 ± 0.03 vs 3.53 ± 0.19 mmol·L−1) and HDL (0.80 ± 0.04 vs 1.79 ± 0.08 mmol·L−1) concentration only in Mcpip1fl/flLysMCre mice (Fig. 4D,E). Twelve weeks of HFD increased the LDL level in Mcpip1fl/fl in comparison with chow diet (0.44 ± 0.02 vs 0.25 ± 0.01mmol·L−1), blunting differences between these mice and myeloid knockouts (Fig. 4F). There were no changes in the LDL, triglycerides, ALT, and AST in Mcpip1fl/flLysMCre when compared to Mcpip1fl/fl mice (Fig. 4F,G, Fig. S5D,E). Only the LDH level was raised in the plasma of Mcpip1fl/flLysMCre mice (689.00 ± 48.42 vs 387.57 ± 52.26 U·L−1) (Fig. 4H). Mcpip1fl/flAlbCre mice were characterized by increased cholesterol level (4.20 ± 0.08 vs 3.53 ± 0.19 mmol·L−1), but all other analytes did not differ between these animals and Mcpip1fl/fl (Fig. 4D–H, Fig. S5D,E). Finally, feeding with HFD led to hyperglycemia and an impaired glucose tolerance detected in all three strains in comparison with mice fed a chow diet (Fig. 4I,J). However, based on GTT test and fasting glucose measurement, Mcpip1fl/flLysMCre animals were characterized by low plasma glucose level, in comparison with Mcpip1fl/fl counterparts (Fig. 4I,J).

Fig. 4

Mcpip1fl/flLysMCre mice do not develop obesity and hepatic steatosis upon feeding with high‐fat diet. Body weight measurements of Mcpip1fl/fl, Mcpip1fl/flLysMCre, and Mcpip1fl/flAlbCre mice fed high‐fat diet for 12 weeks. (B) Representative liver hematoxylin & eosin (H&E), Picro Sirius Red (PSR), and Oil Red O (ORO) stainings; arrowheads point to hematopoietic foci in Mcpip1fl/flLysMCre. Close arrows point to proliferation of intrahepatic bile ducts accompanying by pronounced inflammation and collagen deposition (open arrows) in Mcpip1fl/flAlbCre mice. The scale bar represents 200 µm. (C) Quantitative analysis of Oil Red O‐stained liver tissues; plasma analysis of 22‐week‐old mice: (D) cholesterol; (E) HDL; (F) LDL; (G) triglycerides; (H) LDH levels. Glucose metabolism was tested after intraperitoneal injection of 2 g·kg−1 of glucose by (I) glucose tolerance test and by (J) calculations of the AUC. For Mcpip1fl/fl, n = 7; for Mcpip1fl/flLysMCre, n = 6; and for Mcpip1fl/flAlbCre, n = 6. The graphs show means ± SEM; *P < 0.05; **P < 0.01; ***P < 0.001.

The profile of cytokines, chemokines, and growth factors in the plasma of HFD‐fed mice showed similar tendencies to animals fed chow food (Table S3). Concentrations of proinflammatory mediators (interleukins, chemokines, soluble receptors, and ECM proteins) were significantly increased in Mcpip1fl/flLysMCre mice compared with their Mcpip1fl/fl counterparts (Table S3). Again, there were no differences between Mcpip1fl/flAlbCre and Mcpip1fl/fl mice. Interestingly, HFD feeding of Mcpip1fl/flLysMCre mice further elevated the amount of MIP‐1 alpha, syndecan‐1, and CHI3‐L1, in comparison with animals on chow food (Table 1, Table S3).

Mcpip1 deficiency changes hepatic mRNA expression profile in HFD‐fed mice

The expression of key gluconeogenesis regulators (Pck1, G6pc) was not changed in myeloid knockouts, but impaired glucose metabolism in these mice is at least partially explained by a low level of Irs1 and Slc2a2 (Fig. 5A). Lipid metabolism was also altered in Mcpip1fl/flLysMCre animals fed for 12 weeks with HFD, as manifested by the reduced expression of Fabp1 and Mttp regulating lipid transport, and low levels of Srebf1, Acc1, Fasn, and Scd1 mRNAs that are involved in lipogenesis (Fig. 5B,C). Additionally, the expression of genes regulating beta‐oxidation (Ppara, Acox1, and Cpt1a) was significantly reduced in Mcpip1fl/flLysMCre mice (Fig. 5D). Additionally, there was a significant increase of proinflammatory and profibrotic transcripts in Mcpip1fl/flLysMCre mice in comparison with their Mcpip1fl/fl counterparts (Fig. 5E,F).

Fig. 5

Disturbed hepatic gene expression profile is linked to lean phenotype of HFD‐fed Mcpip1fl/flLysMCre mice. Expression of genes coding for proteins involved in (A) glucose metabolism; (B) lipid transport; (C) lipogenesis; (D) oxidation of fatty acids; (E) inflammation; and (F) fibrosis. All results were calculated in comparison with appropriate gene in Mcpip1fl/fl mice fed chow diet (ΔΔCt method). Dotted lines represent results obtained from Mcpip1fl/fl mice fed chow diet. For Mcpip1fl/fl, n = 7; for Mcpip1fl/flLysMCre, n = 8; and for Mcpip1fl/flAlbCre, n = 9. The graphs show means ± SEM; *P < 0.05; **P < 0.01; ***P < 0.001.

High‐fat diet feeding for 12‐week blunted differences between Mcpip1fl/flAlbCre and Mcpip1fl/fl mice. Out of 26 tested genes, only the level of Pck1 was decreased in comparison with controls (Fig. 5A–F). Overall changes in the hepatic gene expression pattern in HFD‐fed mice—Mcpip1fl/flLysMCre and Mcpip1fl/flAlbCre knockouts vs Mcpip1fl/fl—were very similar to mice fed chow food (Figs 3A–F and 5A–F). Thus, our results indicate that the deletion of Mcpip1 protein in myeloid leukocytes has a more severe impact on metabolic regulatory function than hepatic Mcpip1 deficiency. Additionally, Mcpip1 acts independently from HFD‐induced obesity effects on liver metabolic regulation.

Myeloid Mcpip1 deficiency changes production of adipokines by white adipose tissue

In order to analyze the lean phenotype of Mcpip1fl/flLysMCre mice (Fig. 6A), we evaluated visceral white adipose tissue deposits and performed indirect calorimetry measurements at 10‐ and 22‐week time points. Body mass was unchanged (Fig. S7A) and visceral fat (VAT) was characterized by slightly decreased total mass in 10‐week‐old Mcpip1fl/flLysMCre animals in comparison with controls (Fig. S7B). In 22‐week‐old Mcpip1fl/flLysMCre mice, VAT mass was significantly reduced (0.001 ± 0.0003 vs 0.010 ± 0.001 g) with concomitant altered gene expression profile (Fig. 6B–F). Out of four tested adipokines (leptin, resistin, adiponectin, fatty acid‐binding protein 4), the expression of Lep was downregulated by 83.1%, while Retn was increased by 130% in VAT of Mcpip1fl/flLysMCre mice (Fig. 6C). Transcripts of genes involved in glucose metabolism (Insr, Slc2a4) were not changed, but the mRNA level of Cpt1a was significantly reduced, and Pnpla2 (encoding adipose triglyceride lipase) was enhanced in VAT collected from Mcpip1fl/flLysMCre animals (Fig. 6D,E). The mRNA levels of proinflammatory cytokines (Il6, Il1b, Tnfa), monocytes, and lymphocytes B markers (Cd68, Cd19) were not changed, but myeloid deletion of Mcpip1 led to enhanced T lymphocyte infiltration into VAT (increase of Cde3 level by 295%, Fig. 6F). Again, there were no differences between Mcpip1fl/flAlbCre and Mcpip1fl/fl mice (data not shown). VAT of Mcpip1fl/flLysMCre mice fed with HFD was also characterized by low Lep and high Retn mRNA levels in comparison with controls (Fig. S8A). Additionally, myeloid deletion of Mcpip1 led to a reduced amount of Insr mRNA and lipolysis‐related genes (Cpt1a, Lpl, Lipe) with concomitant increase in Il1b and Cd3e transcripts (Fig. S8B–D).

Fig. 6

Myeloid deletion of Mcpip1 impairs functions of white adipose tissue. (A) Mass of 22‐week‐old Mcpip1fl/fl and Mcpip1fl/flLysMCre mice and (B) visceral adipose tissue/body ratio; expression of genes in VAT from 22‐week‐old mice: (B) adipokines; (C) genes regulating glucose metabolism; (D) lipolysis; and (E) gens related to inflammation. Mean + SEM, for Mcpipfl/fl, n = 6 (A, B) and n = 4 (C–F); for Mcpipfl/flLysMCre, n = 3 *P < 0.05, **P < 0.01.

At 10 weeks of age, indirect calorimetry analyzing energy expenditure showed no differences between Mcpip1fl/flLysMCre and control littermates (Fig. S9). Twenty‐two‐week‐old Mcpip1fl/flLysMCre mice were characterized by diminished oxygen consumption, reduced carbon dioxide production resulting in significant drop of energy expenditure. Thus, increased energy burning cannot explain phenotype of mice with myeloid depletion of Mcpip1.

Myeloid Mcpip1 deletion causes hypothermia and impairs locomotor activity

In the next set of experiments, we analyzed brown adipose tissue (BAT) collected from 10‐ and 22‐week‐old Mcpip1fl/fl and Mcpip1fl/flLysMCre mice. At 10 weeks, there was no difference in BAT mass (0.003 ± 0.0003 vs 0.003 ± 0.0005 g), but its weight was significantly reduced in 22‐week‐old Mcpip1fl/flLysMCre knockouts (0.002 ± 0.0001 vs 0.003 ± 0.0002 g, Fig. 7A). Although the expression of thermoregulatory genes, namely uncoupling protein 1 (Ucp1) and peroxisome proliferator‐activated receptor gamma coactivator 1‐alpha (Ppargc1a), was not changed in young mice, there was a downregulation of both transcripts at 22 weeks by 95.7% and 54.2%, respectively (Fig. 7B). Similarly, myeloid deletion of Mcpip1 led to enhanced monocytes and T lymphocytes infiltration into BAT (139% increase of Cd68, and 644% of Cde3 levels) and increased Tnfa mRNA level by 624% in 22‐week‐old, but not in 10‐week‐old mice (Fig. 7C,D). Apart from diminished expressions of Ucp1 and Ppargc1a, 22‐week‐old Mcpip1fl/flLysMCre knockouts were characterized by hypothermia (35.5 °C vs 37.6 °C in controls) (Fig. 7E). Finally, it is well known that reduced locomotor activity is a major contributor to diet‐induced obesity in mice [27]. That is why, in the last set of experiments, we investigated both average velocity and total traveled distance by 10‐ and 22‐week‐old Mcpip1fl/flLysMCre and Mcpip1fl/flAlbCre knockouts. Locomotor activity of all strains at 10 weeks was the same (Fig. 7F,G). However, 22‐week‐old Mcpip1fl/flLysMCre mice were characterized by significantly reduced locomotor activity. Although the average velocity of Mcpip1fl/flLysMCre was not changed, their total traveled distance was reduced by 70% in comparison with Mcpip1fl/fl mice (Fig. 7F,G).

Fig. 7

Myeloid Mcpip1 deletion impairs locomotor activity and causes hypothermia. (A) Mass of brown adipose tissue; expression of genes involved in (B) thermoregulation and (C) leukocyte infiltration; (D) inflammation; (E) body temperature; (F) average velocity and (G) total distance traveled. For Mcpip1fl/fl, n = 5 (A–D) and n = 10 (E–G); for Mcpip1fl/flLysMCre, n = 3 (A–D) and n = 5 (E–G); and for Mcpip1fl/flAlbCre, n = 6. The graphs show means ± SEM; *P < 0.05; **P < 0.01; ***P < 0.001.

Discussion

We have previously demonstrated that hepatic steatosis accompanying diet‐induced obesity diminishes the amount of Mcpip1 protein in murine primary hepatocytes [28]. Similarly, the Mcpip1 protein levels were reduced in murine and human adipose tissue collected from obese individuals [11, 28]. Additionally, Mcpip1 was shown to regulate lipid and glucose metabolism in hepatic and adipose tissue. Mechanistically, in hepatocytes Mcpip1 induces both the expression and activity of hepatic peroxisome proliferator‐activated receptor gamma via the TXNIP/PGC‐1α pathway [28]. On the other hand, in 3T3‐L1 cells, the downregulation of MCPIP1 at early stages of differentiation promotes the conversion of preadipocytes to adipocytes, resulting in an accumulation of fat. Conversely, an overexpression of MCPIP1 in 3T3‐L1 cells impaired adipogenesis by the direct cleavage of C/EBPβ mRNA and selected pre‐miRNAs [9, 10]. However, the metabolic role of Mcpip1 was not analyzed in tissue‐specific knockout animals.

In the present study, we compared how deletion of Mcpip1 in the myeloid leukocytes or in hepatocytes changes metabolism in response to chow or high‐fat diet. We demonstrated a lack of metabolic manifestation in Mcpip1fl/flAlbCre mice fed chow or HFD food. Although Mcpip1 knockout in hepatocytes reduces the expression of genes involved in glucose metabolism (Pck1, G6pc, Irs1) in liver, plasma glucose concentration was not changed when compared to Mcpip1fl/fl counterparts. Similarly, the lack of Mcpip1 in hepatocytes did not influence lipid metabolism in livers of mice fed chow or HFD. Notably, our observations are in accordance with another study, demonstrating that the deletion of Mcpip1 in Mcpip1fl/flAlbCre mice is asymptomatic, in terms of blood morphology, cytokines, and chemokines detected in plasma. Instead, it leads to a massive hyperplasia of cholangiocytes and the development of primary biliary cholangitis [23]. The important role of Mcpip1 in regulation of hepatic homeostasis was recently demonstrated by Sun and coworkers. By using mice with Zc3h12a deletion and overexpression in liver cells, the authors proved that Mcpip1 functions to ameliorate liver damage, reduce inflammation, prevent cell death, and promote regeneration after ischemia/reperfusion (I/R) injury [15]. The protective function of Mcpip1 was mediated by increasing hypoxia‐inducible factor 1α expression [15]. Additionally, Mcpip1 regulates the outcome of liver I/R through modifying macrophage polarization. Mcpip1 knockdown shifted the balance of M1/M2 status toward M1 via the sensitization of the NF‐κB p65 pathway leading to deteriorated disease [29].

In contrast to Mcpip1fl/flAlbCre mice, the deletion of Mcpip1 in myeloid leukocytes was strongly manifested. Mcpip1fl/flLysMCre animals at 22 weeks of age were characterized by dyslipidemia, low plasma glucose level, disturbed liver homeostasis, and systemic proinflammatory phenotype, together with lethargy and hypothermia. Our observations are in line with the first characterization of Mcpip1fl/flLysMCre mice done by Li and coworkers, who described that after 5 months of age, mice failed to thrive and developed cachexia, splenomegaly, lymphadenopathy, and multiorgan inflammation [22]. Late‐onset inflammatory syndrome described in Mcpip1fl/flLysMCre mice by Dobosz and coworkers was also observed in our diet‐induced obesity model [30]. Twenty‐two‐week‐old myeloid‐specific Mcpip knockout mice developed splenomegaly, significant systemic inflammation manifested by increased concentrations of proinflammatory cytokines and chemokines, compared with Mcpip1fl/fl controls. In the livers of Mcpip1fl/flLysMCre mice, we also observed leukocyte infiltration, bile duct pathology, and extramedullary hematopoiesis (EMH). The pathologic EMH may be caused by one of many hematological diseases, bone marrow irradiation, or anemia [31, 32]. It can be also seen as part of the response to infection or systemic inflammation [31, 32], detected in fact in both total and myeloid‐specific Mcpip1 knockouts [3, 12, 22]. Additionally, Kidoya and coworkers have shown that the post‐transcriptional regulation mediated by Mcpip1 is essential for hematopoietic stem and progenitor cell homeostasis. Mcpip1 regulates self‐renewal of hematopoietic stem and progenitor cells through modulating the stability of Gata2 and Tal1 mRNA, and its dysfunction leads to the rapid onset of abnormal hematopoiesis [33].

Similarly to our results, total Mcpip1 knockout mice displayed a decrease in serum levels of HDL‐cholesterol and total triglycerides, but an increase in LDL/VLDL‐cholesterol levels when compared to wild‐type mice [34]. As demonstrated here, this phenotype cannot be attributed to Mcpip1 depletion in hepatocytes, since dyslipidemia was detected only in Mcpip1fl/flLysMCre, but not in Mcpip1fl/flAlbCre mice. According to Pin et al., an increased level of LDL is dependent on HDL level reduction, due to the fact that LDL maturation requires the acquirement of cholesterol esters from HDL particles. If this process is disturbed, low‐density, cholesterol‐poor LDL particles lose their affinity to LDL receptors and remain in circulation [35]. High level of LDL‐cholesterol in Mcpip1−/− mice was also explained by the reduced expression of LDL‐receptor in hepatocytes [34]. In addition, it was shown that high resistin level—observed also in VAT of myeloid Mcpip1 knockouts—induces degradation of LDL—receptor via PCSK9 protease [36].

In our opinion, changes in lipid and glucose metabolisms in Mcpip1fl/flLysMCre mice are secondary to the inflammatory syndrome, rather than a direct product of myeloid leukocytes Mcpip1 activity on the control of hepatic metabolism. In fact, low plasma glucose level and dyslipidemia are common features of murine sterile inflammation and cachexia models [35, 37, 38, 39, 40]. The low plasma glucose level in Mcpip1fl/flLysMCre mice might be explained by the altered expression of genes encoding proteins involved in gluconeogenesis (G6pc), insulin response (Irs1, Irs2), and glucose bidirectional transport (Slc2a2). Additionally, a raised level of Tnfα in the plasma of these mice could account for decreased gluconeogenesis rates [41]. Twenty‐two‐week‐old Mcpip1fl/flLysMCre mice were also characterized by a decreased mass of visceral adipose tissue and altered profile of adipokine expression. A lower mass of VAT in Mcpip1fl/flLysMCre animals correlated with reduced leptin expression that was earlier shown to be released from white adipose tissue in proportion to the size of fat depots [42]. Leptin, by inhibiting food intake, increases energy expenditure via the inhibition of lipogenesis and stimulation of lipolysis functions as a negative feedback signal in the regulation of energy balance. A low level of leptin may also explain the reduced expression of genes involved in beta‐oxidation like Ppara or Cpt1a [42]. A reduced serum leptin level was also observed under nonobese physiological conditions in mice with depleted tissue macrophages. Macrophage depletion led to systemic inflammation induced by neutrophil infiltration and reduced body fat mass due to reduced energy intake [43].

Immune cell infiltration, such as macrophages and T cells, into white adipose tissue is an established phenomenon related to the pathophysiology of this tissue, especially in relation to obesity and diabetes [44]. Additionally, recent transcriptomic studies of BAT from HFD‐induced obese mice demonstrated enhanced expression of gene markers for leukocytes, monocytes, and macrophages in these mice [45, 46, 47]. In BAT of 22‐week‐old Mcpip1fl/flLysMCre mice, we detected high levels of Cd68 and Cd3e with concomitant reduction of genes regulating thermogenesis (Ucp1, Ppargc1a). It had already been shown that proinflammatory M1 macrophages infiltrating BAT impair its ability to appropriately respond to thermogenic stimuli, by the secretion of TNFα and IL‐1β that suppress the induction of Ucp1 gene expression [48, 49]. Since Mcpip1 inhibits the classic activation of M1 macrophages and promotes M2 polarization, we postulate that the hypothermia of Mcpip1fl/flLysMCre mice might be at least partially explained by an increased infiltration of M1 macrophages into BAT and a high level of Tnfa [29, 50].

Finally, Mcpip1fl/flLysMCre mice did not develop hepatic steatosis, despite being fed a HFD. This kind of obesity‐resistant phenotype was already described in methyl‐CpG–binding domain 2 knockout mice (Mbd2) and in animals bearing a C305F mutation in the mouse double minute 2 homolog (Mdm2) [51, 52]. Mbd2 deficiency in mice fed HFD led to reduced epididymal adipose tissue mass, improved glucose tolerance, hyperlipidemia, and attenuated HFD‐induced hepatosteatosis, features also present in Mcpip1fl/flLysMCre animals. However, our model differs from Mbd2 mice in respect to energy expenditure, adipose inflammation, and lipolysis [51]. Also, Liu and coworkers described that Mdm2C mice demonstrated decreased fat accumulation along with improved glucose tolerance and increased energy expenditure after prolonged HFD feeding [52]. In contrast to the models mentioned above, myeloid deletion of Mcpip1 led to diminished oxygen consumption and reduced carbon dioxide production, resulting in a significant drop of energy expenditure. Thus, increased energy burning cannot explain the lean phenotype of Mcpip1fl/flLysMCre mice, but perhaps fecal energy loss could explain this phenomenon. Additionally, mice analyzed at 22 weeks of age did not have enhanced lipolysis nor reduced inflammation, such as Mbd2 or Mdm2 animals. A future study, utilizing 4‐week‐old animals fed with HFD, might help to understand this phenomenon, because young Mcpip1fl/flLysMCre animals do not have a proinflammatory phenotype that might influence obesity development. According to results published by Li and coworkers in 2‐month‐old myeloid Mcpip1 knockouts and Mcpip1fl/fl controls, there were no differences in serum concentrations of proinflammatory cytokines and chemokines. Additionally, minimal inflammation in the lungs and livers of Mcpip1fl/flLysMCre mice was detected at 3 months of age, whereas inflammatory cell infiltration was significant at 6 months [22]. Diminished intestinal food assimilation might be an alternative explanation of resistance to high‐fat diet‐induced obesity in Mcpip1fl/flLysMCre mice. Although we do not have experimental data to support this view, this kind of molecular mechanism has already been described. Recently, the genetic inactivation of the Golgi‐resident protein GRASP55 in mice was demonstrated to reduce whole‐body fat mass via impaired intestinal fat absorption, causing resistance to high‐fat diet‐induced body weight gain [53].

In conclusion, we have demonstrated that the depletion of Mcpip1 in myeloid leukocytes results in systemic inflammation, dyslipidemia, low plasma glucose level, and disturbed liver lipid metabolism protecting from HFD‐induced obesity. On the other hand, Mcpip1fl/flAlbCre mice did not exhibit signs of systemic inflammation, nor disturbed glucose and lipid metabolisms.

Materials and methods

Animals, diets, and genotyping

To obtain myeloid or hepatocyte‐specific knockout of Mcpip1 protein, we used Mcpip1fl/fl mice with two LoxP sites flanking the two sides of exon 3 of the Zc3h12a gene encoding Mcpip1 [22]. Next, these mice were crossed with myeloid cells (LysMCreTg/Wt, Jackson Laboratory; Bar Harbor, ME, USA; 004781) or hepatocytes expressing Cre transgenic mice (AlbCreTg/Wt, Jackson Laboratory; Bar Harbor, Maine, United States of America; 018961). In the second round of crosses, Mcpip1fl/fl females were crossed with Mcpip1fl/fl LysMCreTg/Wt males to obtain both control Mcpip1fl/fl LysMCreWt/Wt animals (designed as Mcpip1fl/fl) and Mcpip1fl/fl LysMCreTg/Wt knockouts (designed as Mcpip1fl/flLysMCre). Similarly, Mcpip1fl/fl females were crossed with Mcpip1fl/flAlbCreTg/Wt males to obtain controls (Mcpip1fl/fl) and knockouts (Mcpip1fl/flAlbCre). In all studies, C57BL/6N male mice were used that were fed ad libitum for 12 weeks with a chow diet (ZooLab; AIN93G) or HFD, (60% kcal from fat) (ZooLab; DP‐1E‐60S), starting from the age of 10 weeks. Animals were housed under SPF conditions in ventilated cages in a temperature‐controlled environment with a 14/10‐h light/dark cycle. To measure the amount of consumed food, the mice were housed in metabolic cages (Tecniplast; Buguggiate, Italy; 3600M021). For genotyping, DNA was extracted from tail tissue with a KAPA Mouse Genotyping Kit (KAPA Biosystems; Woburn, MA, USA; KK7302) according to the manufacturer's instructions. Sequences of primers (Sigma; Saint Louis, MO, USA) used for genotyping are listed in Table S1. All animal procedures were conducted in accordance with the Guide for the Care and Use of Laboratory Animals (Directive 2010/63/EU of the European Parliament) and carried out under a license from the 2nd Local Institutional Animal Care and Use Committee in Kraków (study no. 106/2017 and 360/2020).

Blood and tissue samples

Mice were fasted for 4 h, before their blood was collected from the orbital sinus after isoflurane anesthesia into EDTA (MP Biomedicals; Santa Ana, CA, USA; 195173) containing tubes, and then, the animals were euthanized. Plasma was prepared by centrifugation (1000  , 10 min) and stored at −80 °C. Following euthanasia by cervical dislocation, organs were isolated and prepared for histological analysis, or snap‐frozen in liquid nitrogen and stored at −80 °C.

Plasma analysis

Murine plasma was used for the subsequent analysis: cholesterol, HDL‐cholesterol (HDL), LDL‐cholesterol (LDL), glucose, triglycerides, alanine aminotransferase (ALT), aspartate aminotransferase (AST), and lactate dehydrogenase (LDH). These parameters were measured using an automatic biochemical analyzer, Pentra 400 (Horiba; Kioto, Japan), by an enzymatic photometric method according to the vendor's instructions.

Glucose tolerance test

A glucose tolerance test (GTT) was performed on mice aged 22 weeks. Mice were fasted for 4 h, after which they were intraperitoneally injected with saturated glucose (Sigma; G7021) solution (2 g·kg−1 body weight). Blood was collected from the tail vein before (0 min) and at 15, 30, 45, 60, 120 min after glucose application. Glucose concentrations were measured by an Accu‐Check glucometer (Roche Diagnostics; Basel, Switzerland). The area under the curve (AUC) of blood glucose concentration was calculated geometrically by applying a trapezoidal method [54].

Histological analysis

Livers were fixed in 4% paraformaldehyde formalin solution (Chempur; Piekary Śląskie, Poland; 114321730) and divided into two parts. One piece was submerged in a 30% sucrose solution overnight for cryoprotection and then frozen in OCT medium (Tissue‐Tek; Sakura, Tokyo, Japan; 4583) at −80 °C. Frozen sections were stained using an Oil Red O (ORO) method to analyze fat accumulation and imaged under 100× magnification. Ten images of each section were randomly collected and analyzed by the columbus v.2.4.2 software (Perkin Elmer; Waltham, Massachusetts, USA) with an algorithm adapted for Oil Red O‐stained sections. The second piece was processed using standard paraffin procedures, and 5‐μm paraffin tissue sections were stained with hematoxylin and eosin (H&E) for general histology and Picro Sirius Red (PSR) for collagen deposition and then visualized using a standard light microscope (Olympus BX51; Olympus Corporation, Tokyo, Japan).

Luminex assay

Magnetic Luminex Assay (R&D Systems; Minneapolis, MN, USA; LXSAMSM) was performed on plasma samples, according to the manufacturer's instructions. All samples were analyzed in duplicate on a MAGPIXs (Luminex, Austin, TX, USA). Median fluorescence intensity (MFI) data, obtained using a weighted 5‐parameter logistic curve‐fitting method, was used to calculate analyte concentrations.

RNA isolation and RT‐PCR

Total RNA was isolated from liver tissue using a modified guanidinium isothiocyanate method [55]. Briefly, fragments of frozen liver tissue were lysed in fenozol (A&A Biotechnology; Gdansk, Poland; 203‐50), and after addition of chloroform (POCH; Gliwice, Poland; 234431116), the aqueous phase containing RNA was collected into new Eppendorf tubes. Next, RNA was precipitated, centrifuged, washed in ethanol (POCH; Gliwice, Poland; 396420113), and finally dissolved in nuclease‐free water (Sigma; W4502). The concentration of RNA was assessed by a NanoDrop 1000 Spectrophotometer (Thermo Fisher Scientific; Waltham, MA, USA). Reverse transcription was performed using 1 µg of total RNA, oligo(dT) primer (Sigma) and M‐MLV reverse transcriptase (Promega; Madison, WI, USA; M170B). Real‐time PCR was carried out using Sybr Green Master Mix (A&A Biotechnology; 2008‐100A) and QuantStudio Real‐Time PCR System (Applied Biosystems; Waltham, MA, USA). Gene expression was normalized to elongation factor‐2 (EF2), after which the relative level of transcripts was quantified by a 2−ΔΔCT method. Sequences of primers (Genomed/Sigma) and annealing temperatures are listed in Table S2. The primers were designed to cross an exon‐exon boundary to allow the specific amplification of cDNA targets.

Measurement of body temperature

Mice body temperature was measured by rectal probing (physiological monitoring unit THM150; Visualsonics, Toronto, Canada). The probe was inserted for 2 cm into the rectum and held in position for 10 s before temperature was determined. Measurements were conducted for five days at exactly the same time of day.

Open‐field test

All mice were handled for three consecutive days prior to the start of behavioral measurements. On the day of the behavioral experiment, the mice were brought in their home cages to the testing room and left for 1 h to acclimate. General locomotor activity and movement velocity were monitored in an open field in arena cages (40 × 40 × 38 cm) made from transparent plexiglass, equipped with an array of infrared photo beams and software for automatic tracking of mouse ambulation (Activity Monitor 7, Med Associates Inc., St. Albans, VT, USA). The open‐field arena was illuminated with white light at a level of 150 lux. The duration of each behavioral session was 30 min for each animal. After each trial, the cages were wiped with 70% ethanol and left to dry. To eliminate the effect of circadian rhythm on locomotor activity, all mice were subjected to the open‐field test between 6 and 9 pm.

Indirect calorimetry measurements

Indirect calorimetry was performed to assess energy expenditure in wild‐type (Mcpipfl/fl) and Mcpipfl/fl LysMCre mice at 10 and 22 weeks of age. Animals were placed into individual cages—Columbus Instruments Comprehensive Lab Animal Monitoring System chambers (Oxymax/CLAMS; Columbus Instruments, Columbus, OH, USA). VCO2 and VO2 were measured after 24‐h acclimatization period and calculated based on the input and output rates of O2 consumption and CO2 production. The respiratory exchange ratio (RER = VCO2/VO2) and heat production [kcal·min−1 = (3.815 × 1.232 × RER) × (VO2)] were calculated using software provided by the manufacturer (Columbus Instruments, Columbus, OH, USA). Data were normalized to body weight.

Statistical analysis

Results are expressed as mean ± SEM. The exact number of experiments or animals used is indicated in the figure legends. One‐way ANOVA with Tukey's post hoc test was applied for comparison of multiple groups. To analyze the effects of two variable factors, a two‐way ANOVA was used. The P values are marked with an asterisks in the charts (*P < 0.05; **P < 0.01; ***P < 0.001).

Conflict of interest

The authors declare no conflict of interest.

Author contributions

NP contributed to conceptualization, methodology, investigation, writing—original draft, and visualization; DZ contributed to investigation and writing—review and editing; JoKo contributed to resources and writing—review and editing; EK contributed to investigation and writing—review and editing; KW‐L contributed to investigation; AJ contributed to investigation; MF contributed to resources; JJ contributed to conceptualization and writing—review and editing; JeKo contributed to conceptualization, methodology, writing—original draft, supervision, project administration, and funding acquisition.

Ethics approval

All animal experiments were approved by the 2nd Local Institutional Animal Care and Use Committee in Kraków (study no. 106/2017). All institutional and national guidelines for the care and use of laboratory animals were followed.

Peer Review

The peer review history for this article is available at https://publons.com/publon/10.1111/febs.16040.

Fig. S1. Comparison of Mcpip1fl/fl controls from Mcpip1fl/flLysMCre and Mcpip1fl/flAlbCre strains.

Fig. S2. Expression of Zc3h12a encoding Mcpip1 in the liver and white adipose tissue.

Fig. S3. Daily food intake, ALT and AST activity and liver mass of Mcpip1fl/fl, Mcpip1fl/flLysMCre and Mcpip1fl/flAlbCre mice fed chow diet.

Fig. S4. Representative histology images of liver from mice fed chow diet.

Fig. S5. Daily food intake, ALT and AST activity and liver mass of Mcpip1fl/fl, Mcpip1fl/flLysMCre and Mcpip1fl/flAlbCre mice fed HFD.

Fig. S6. Representative histology images of liver from mice fed HFD.

Fig. S7. Comparison of 10‐week‐old Mcpip1fl/fl and Mcpip1fl/flLysMCre mice.

Fig. S8. Analysis of gene expression profile in visceral adipose tissue from HFD‐fed Mcpip1fl/fl and Mcpip1fl/flLysMCre mice.

Fig. S9. Indirect calorimetry measurements in Mcpip1fl/fl and Mcpip1fl/flLysMCre mice.

Table S1. Sequences of primers used for genotyping.

Table S2. Sequences of primers used for real‐time PCR.

Table S3. Immune profile of plasma samples from 22‐week‐old mice fed with HFD for 12 weeks, determined by Luminex Assay.

Click here for additional data file.