Serum CC-Chemokine Ligand 2 Is Associated with Visceral Adiposity but Not Fibrosis in Patients with Non-Alcoholic Fatty Liver Disease

Non-alcoholic fatty liver disease (NAFLD) is a chronic inflammatory disease caused by sustained injury to the liver parenchyma as a result of ectopic fat accumulation in the liver. As the liver manifestation of the metabolic syndrome, NAFLD affects an estimated 30% of people worldwide [1]. It is a complex systemic disease associated with a spectrum of inter-related pathologies, including obesity, type 2 diabetes, insulin resistance, and dyslipidemia [2]. Chronic liver injury and associated inflammation can lead to the development of non-alcoholic steatohepatitis (NASH), progressive fibrosis, and ultimately cirrhosis and hepatocellular carcinoma. Approximately, 20% of people with NAFLD will develop NASH and of those 25–40% develop progressive fibrosis placing them at high risk of liver-related complications (cirrhosis and hepatocellular carcinoma) [3]. Despite many clinical trials [4], there are currently no approved pharmacological treatments for NAFLD/NASH, and lifestyle modification is the main therapeutic approach.

Liver biopsy remains the gold standard for diagnosis of liver fibrosis, but it has health risks and diagnostic, logistic, and cost limitations. Given the prevalence of NAFLD in the community, there is a clear need for non-invasive biomarkers to identify patients at risk of progressive fibrosis and liver-related complications. Liver stiffness measurement (LSM, Fibroscan) is the most widely validated and accepted non-invasive method for fibrosis assessment, but it is still mainly limited to specialist centres as it requires costly equipment and specialist operators. Simple scores such as FIB4 and the NAFLD fibrosis score (NFS) based on routine serum biochemistry and clinical parameters have a high negative predictive value and are recommended to rule out advanced fibrosis, but a large proportion of patients present with indeterminate scores and require further investigation [5]. No single serum biomarker of fibrosis severity has yet been identified; however, several biomarker combinations are reported to have good diagnostic accuracy for predicting the presence of advanced fibrosis in NAFLD cohorts, including the commercially available ELF test, an algorithm based on circulating levels of hyaluronic acid, pro-collagen peptide 3, and tissue inhibitor of metalloproteinase 1 [6]. We and others have speculated that additional biomarker combinations, perhaps reflecting different aspects of the complex disease pathogenesis, could improve the diagnostic accuracy of these tests [7].

Activation of liver resident leukocytes in response to injury, and subsequent recruitment of additional inflammatory cells is a critical driver of inflammation and fibrogenesis in NASH, as in other etiologies of chronic liver disease [4]. Liver macrophages, and monocytes recruited in response to cell death and tissue injury [8], in particular, are key players in the initiation and perpetuation of fibrogenic injury and have been widely investigated as therapeutic targets [9]. For example, genetic or pharmacological inhibition of monocyte recruitment via the CC-chemokine ligand 2 (CCL2)/CCR2 chemokine axis prevented or reduced fibrosis in pre-clinical models of CLD [10, 11] and has also been shown to alleviate metabolic syndrome in obese animals [11‒16]. A dual CCR2/CCR5 antagonist, Cenicriviroc, showed promise in early clinical trials [17], but a phase 3 study was terminated early due to lack of efficacy (NCT03028740). Interestingly, the potential necessity of combinatorial therapeutic approaches in NASH has been raised, and a recent pre-clinical study demonstrated increased therapeutic efficacy of a CCR2/CCR5 antagonist and FGF21 compared to either drug alone [18]. Increased hepatic CCL2 mRNA expression and CCR2+ infiltrating myeloid cells have been demonstrated in patients with NASH compared to healthy controls [11, 19]. CCL2 expression was also reported to be significantly upregulated in peripheral blood mononuclear cells from patients with NASH compared to patients with NAFLD [20]. Several small cohort studies and a large meta-analysis reported elevated circulating CCL2 in patients with NASH and/or NAFLD compared to healthy controls [21‒24], whereas other studies report no association [25, 26]. With respect to liver fibrosis, one study of 75 obese patients with NAFLD reported no significant association between circulating CCL2 and fibrosis stage [25], whereas a separate study of 85 patients reported elevated CCL2 in patients with stage 3–4 fibrosis compared to stage 0–2 but no difference between NAFLD and NASH [18].

Our previous multiplex analysis of serum biomarkers of advanced fibrosis (n = 432) did not identify CCL2 but this cohort largely comprised patients with chronic hepatitis C infection [27]. In the current study, we investigated the relationship between circulating CCL2 and liver fibrosis, assessed by LSM, and other clinically relevant parameters in patients with NAFLD. Impaired intestinal barrier function and subsequent release of pro-inflammatory gut-derived microbial products that induce inflammatory cytokines and chemokines such as CCL2 are implicated in CLD pathogenesis, including NASH [28]. Therefore, we also measured serum fatty acid binding protein 2 (FABP2), an enterocyte-specific protein that has been suggested as a biomarker of mucosal damage and intestinal permeability in patients with CLD [29, 30] and diabetes [31].

This is a retrospective study of patients with NAFLD who attended a liver clinic at the Princess Alexandra Hospital between October 2015 and August 2017 for clinical assessment (n = 252) (source population previously described in [5]). Clinical assessment included anthropometric measurements, routine haematological, biochemical and serological tests, ELF test, LSM, and liver ultrasound. Transient elastography was performed using a FibroScan (Echosens, Paris, France) as previously described, with a cut-off value of ≥8.2 kPa for clinically significant fibrosis, for both standard M and XL probes [5]. Metabolic syndrome was defined per the International Diabetes Federation guidelines [32].

An aliquot of fasted serum collected at clinical assessment was stored at −80°C. Serum CCL2 and FABP2 levels were determined in 250 patients from the original study who had serum samples available, including 228 patients who had LSM meeting quality criteria, using enzyme-linked immunosorbent assays (ELISAs) (catalogue # DY279 and DY3078, respectively, RnD Systems, Minneapolis, USA) according to the manufacturer’s instructions. Both kits utilise a monoclonal mouse capture antibody and a biotinylated polyclonal goat antibody and follow an identical protocol. Sera were diluted 1:2 in the manufacturer’s recommended diluent (1% BSA in PBS, catalogue # DY995) for analysis. The mean of duplicate samples was taken for analysis. CCL2 was below the detection limit of the ELISA in 4 (1.6%) samples, which were imputed as the minimum detectable value. FABP was below the detection limit of the ELISA in 7 (2.7%) samples, which were imputed as the minimum detectable value. Serum MMP7 was previously measured in the same patient cohort [7] and further analysed in relation to CCL2 and FABP2 here. MMP7 was below or above the limit of detection in 2 and 1 samples, respectively, which were imputed as the minimum and maximum detectable values.

Patient characteristics were summarized with mean (standard deviation) or median (interquartile range, [IQR]) for normally distributed or skewed continuous variables, respectively, and frequency (%) for categorical variables. Characteristics of patients in different LSM categories (<8.2, 8.2+) were compared using the χ² test for categorical parameters and the Student’s t test or Mann-Whitney test for continuous variables. Association between CCL2 and clinically significant fibrosis, defined as LSM ≥8.2 kPa, was examined using a binary logistic regression model (n = 228 patients). Spearman’s correlation was used to examine monotonic relationships between continuous variables (pairwise deletion, i.e., all available data pairs for each variable were used) and visualized using a correlation matrix. A binary logistic regression model was used to investigate associations between CCL2 and categorical variables. All analyses were performed using Stata 17.0 (StataCorp. 2021. Stata Statistical Software: Release 17. College Station, TX: StataCorp LLC) and R statistical package (v4.1.0; R Core Team, 2021). The level of statistical significance was set at 0.05.

CCL2 and FABP were measured in serum from 228 of 230 patients from the source cohort [5] with LSM measurement meeting quality requirements. Demographic and clinical characteristics of patients with clinically significant fibrosis, defined as LSM ≥8.2 [33], compared to those with no significant fibrosis are shown in Table 1. Serum CCL2 was weakly, positively correlated with LSM (Spearman r 0.17, p = 0.012), with median values of 172.0 (IQR 119.6–230.6) and 208.5 (IQR 152.4–270.4) in patients with LSM <8.2 and ≥8.2, respectively (Table 1; Fig. 1). Serum FABP2 was not associated with LSM. Logistic regression modelling demonstrated a modest but significant association between CCL2 and clinically significant fibrosis (OR 3.47, 95% CI: 1.0–12.0, p = 0.049); however, after adjustment for age, gender, BMI, and the presence of diabetes, no significant association was observed (Table 2). Adjusting the regression model for BMI alone reduced the OR for CCL2 by >20%, whereas replacing BMI with girth in the multivariable model further reduced the OR for CCL2 to 1.50 (95% CI: 0.35–6.53; p = 0.59). These data suggested the observed association between CCL2 and LSM may be related to an association between CCL2 and obesity, particularly visceral adiposity. Consistent with this, CCL2 was positively correlated with BMI and girth (see below).

Since CCL2 and FABP2 were not associated with liver fibrosis, we investigated associations between these serum analytes and other clinical parameters. This analysis included 242 of 250 patients in the source cohort (i.e., including an additional 14 patients omitted from the fibrosis analysis due to low quality LSM data). CCL2 was modestly, positively correlated with BMI, girth, and weight (Fig. 2). The relationship between CCL2 and girth appeared linear while there appeared to be a non-linear relationship between CCL2 and BMI, in particular for BMI >40, suggesting a potential relationship between CCL2 and visceral adiposity (Fig. 3). Consistent with this, logistic regression analysis showed that CCL2 was significantly associated with the presence of abdominal obesity, after adjusting for age and gender (adj. OR 34.26, 95% CI 4.17–281.67, p = 0.001). CCL2 was also significantly associated with the presence of metabolic syndrome (adj. OR 6.32, 95% CI 1.48–26.89, p = 0.013), but not hypertension, diabetes, or dyslipidemia (p = 0.089, 0.139, 0.270; respectively). FABP2 was modestly positively correlated with CCL2 and MMP7 and weakly negatively correlated with estimated glomerular filtration rate (eGFR), as previously reported [34‒36], but was not associated with any other clinical parameters (Fig. 2).

Given the global prevalence of NAFLD, there is a clear need for non-invasive biomarkers to identify patients at a risk of progressive fibrosis and liver-related complications. Patients with NAFLD with no or moderate fibrosis (histological stages F0-2) are mainly at a risk of cardiovascular or diabetes-related complications [37], whereas liver-related morbidity and mortality dramatically increase with advanced fibrosis and cirrhosis (F3-4) [38]. Biomarkers of fibrosis and its dynamics are highly desirable not only to enable early lifestyle intervention to prevent progression but to identify candidates for pharmacotherapy, and to monitor therapeutic efficacy. Algorithms utilizing indirect (liver function, inflammation, metabolic) or direct (fibrogenesis) biomarkers and combinations of these [33] have good diagnostic performance but are still sub-optimal for use as a screening test to diagnose or rule out advanced fibrosis. In this study, we report that the monocyte chemoattractant CCL2 that is implicated in NAFLD pathogenesis and progression in preclinical models is not independently associated with fibrosis in NAFLD patients, and is rather associated with visceral adiposity. Our results are broadly consistent with other studies that have found associations between circulating CCL2 and the presence of NAFLD and/or NASH [21‒24, 26] but not fibrosis [25]. Puengel et al. [18] recently reported an association between CCL2 and fibrosis stage in a cohort of biopsy-proven NAFLD patients and a significant correlation with FIB4 score. We similarly detected a positive correlation between CCL2 and fibrosis assessed by LSM, and a weak positive correlation with NAFLD fibrosis score; however, our data suggest a stronger association with adiposity than with fibrosis.

FABP2 has previously been suggested to be a marker of mucosal damage in the setting of CLD and diabetes [29‒31], but we did not find any association between FABP2 and a range of clinical parameters reflecting liver disease severity. FABP2 was positively correlated with CCL2, however, which may reflect inflammation, but we did not assess other inflammatory markers. Elevated circulating FABP2 has previously been reported as a marker of diabetic nephropathy and chronic renal failure and shown to negatively correlate with eGFR in patients with kidney disease and the general population [34‒36]. Although FABP2 is released from damaged enterocytes, it is rapidly cleared by the kidneys, so elevated circulating levels are thought to reflect renal insufficiency [36]. Elevated FABP2 was very weakly, but significantly, associated with poorer kidney function (low eGFR) in this cohort, although the association was not strong enough to suggest utility as a biomarker.

Monocytes and macrophages are key cellular mediators of hepatic inflammation and fibrogenesis in NAFLD, as well as extrahepatic pathologies such as metabolic syndrome, renal, and cardiovascular disease. The CCL2/CCR2 axis regulates monocyte recruitment systemically, so the lack of a specific association with liver fibrosis is perhaps not surprising. CCL2 signalling through CCR2 impacts multiple pathways in addition to those leading to cell migration and has been suggested to have broader roles in the regulation of inflammation and metabolism [39]. Increased CCL2 mRNA expression and infiltrating CCR2+ macrophages have been observed in both the liver and visceral adipose tissue of patients with NAFLD [11, 19, 22, 26]. Serum CCL2 has previously been reported to be elevated in obese patients compared to lean controls [40, 41] and reduced following bariatric surgery [42]. CCL2 was produced by isolated adipocytes from morbidly obese individuals and was reduced after weight loss [43]. In mice, overexpression of CCL2 was sufficient to promote insulin resistance, macrophage infiltration into adipose tissue and hepatic steatosis [44]. Serum and urinary CCL2 are also associated with renal disease in both obesity and diabetic nephropathy [45‒47]. Interestingly, Kawano et al. [48] found that the first organ expressing high levels of CCL2 in response to a high-fat diet in mice was the colon, leading to macrophage recruitment and increased gut permeability, followed by adipose inflammation and insulin resistance. Our findings suggest that adipose tissue is a key source of CCL2 in obese patients with NAFLD, which may be expected, given the relative size of the organ. CCL2 expression in subcutaneous adipose tissue has recently been linked to senescence in these depots and increased visceral adiposity, in obese individuals, which may contribute to inflammation and limit healthy adipose expansion [49].

Strengths of this study are that it includes a well-characterised cohort of patients and a broad range of clinical features and parameters related to NAFLD and comorbidities. Key limitations of this study include a lack of histological assessment of fibrosis and the relatively small size of the cohort, although this is the largest analysis of serum CCL2 in a NAFLD cohort reported to date. Non-invasive assessment of fibrosis by elastography is well established in hepatology centres, and its implementation as a screening tool in primary care settings is an active area of research. Our threshold for clinically significant fibrosis is a conservative estimate, at the upper end of values reported to positively identify fibrosis [50, 51]. In conclusion, we have demonstrated that circulating CCL2 and FABP2 are associated with NAFLD comorbidities but not liver disease severity or fibrosis in patients with NAFLD.

The Enhanced Liver Fibrosis (ELF) test was provided by Siemens Healthineers (Erlangen, Germany).

This research was conducted ethically in accordance with the World Medical Association Declaration of Helsinki. Informed written consent was obtained from each participant, and the protocol was approved by the Metro South Health and The University of Queensland Human Research Ethics Committee (HREC/15/QPAH/301; UQ2015001047).

The authors have no conflicts of interest to declare.

No specific funding was obtained for this study. We are grateful for core funding support from the Mater Foundation, Brisbane, Australia. The Mater Foundation did not contribute to the preparation of the data or the manuscript.

Michelle Ferrari-Cestari: data collection, analysis, and manuscript preparation. Satomi Okano: statistical analysis and table and figure preparation. Preya J. Patel, Suzanne Williams, Anthony Russell, and Leigh U. Horsfall: patient recruitment and clinical data collection. Sahar Keshvari: data collection. David A. Hume: funding support. Elizabeth E. Powell: study conception, patient recruitment, clinical data collection, and manuscript preparation. Katharine M. Irvine: study conception, data analysis, supervision, and manuscript preparation.

Data relating to individual participants in this study are not publicly available as this information may compromise their privacy. De-identified primary data will be made available upon request to the corresponding author (K.M.I), subject to HREC approval. Further enquiries can be directed to the corresponding author.