Heterocyclic Amines Reduce Insulin-Induced AKT Phosphorylation and Induce Gluconeogenic Gene Expression in Human Hepatocytes

Abstract

Heterocyclic amines (HCAs) are well-known for their mutagenic properties. One of the major routes of human exposure is through consumption of cooked meat, as certain cooking methods favor formation of HCAs. Recent epidemiological studies reported significant associations between dietary HCA exposure and insulin resistance and type II diabetes. However, no previous studies have examined if HCAs, independent of meat consumption, contributes to pathogenesis of insulin resistance or metabolic disease. In the present study, we have assessed the effect of three HCAs commonly found in cooked meat (2-amino-3,4-dimethylimidazo[4,5-f]quinoline [MeIQ], 2-amino-3,8-dimethylimidazo[4,5-f]quinoxaline [MeIQx], and 2-amino-1-methyl-6-phenylimidazo[4,5-b]pyridine [PhIP]) on insulin signaling and glucose production. HepG2 or cryopreserved human hepatocytes were treated with 0–50 μM of MeIQ, MeIQx, or PhIP for 3 days. Treatment of HepG2 cells and hepatocytes with MeIQ and MeIQx resulted in a significant reduction in insulin-induced AKT phosphorylation, suggesting that HCA exposure decreases hepatic insulin signaling. HCA treatment also led to significant increases in expression of gluconeogenic genes, G6PC and PCK1, in both HepG2 and cryopreserved human hepatocytes. Additionally, the level of phosphorylated FOXO1, a transcriptional regulator of gluconeogenesis, was significantly reduced by HCA treatment in hepatocytes. Importantly, HCA treatment of human hepatocytes led to increases in extracellular glucose level in the presence of gluconeogenic substrates, suggesting that HCAs induce hepatic glucose production. The current findings suggest that HCAs induce insulin resistance and promote hepatic glucose production in human hepatocytes. This implicates that exposure to HCAs may lead to the development of type II diabetes or metabolic syndrome.

Introduction

Heterocyclic amines (HCAs) are known mutagens and carcinogens (Felton et al. 1999; Cheng et al. 2006). The main source of human exposure to HCAs is through cooked meat, although they are also found in other products including cigarette smoke and coffee (Zhang et al. 2011, 2020; Xian et al. 2019). They are primarily formed in muscle foods including meats and fish, which provide precursors, such as creatinine, amino acids, and aldehydes, required for the formation reaction (Eisenbrand and Tang 1993; Lan et al. 2004; Zamora and Hidalgo 2020). Cooking meat at high temperatures or for long periods of time favors HCA formation, and certain cooking methods, such as frying, grilling, and roasting, substantially enhance their formation (Lan et al. 2004). Thus far, more than 25 different HCAs have been isolated from food samples (Toribio et al. 2002). The most abundant HCAs found in cooked meat include PhIP (2-amino-1-methyl-6-phenylimidazo[4,5-b]pyridine), MeIQx (2-amino-3,8-dimethylimidazo[4,5-f]quinoxaline) and MeIQ (2-amino-3,4-dimethylimidazo[4,5-f]quinoline) (Fig. 1A). These HCAs have been detected in different types of cooked meat at levels ranging up to 150 ng/g (Knize et al. 1995). Many HCAs require hepatic bioactivation to yield mutagenic and carcinogenic effects (Guengerich 1992; Chou et al. 1995). The initial step of HCA metabolism is the generation of N-hydroxy-HCA derivatives, which is catalyzed by cytochrome P450, mainly isoenzyme CYP1A2 (Chou et al. 1995). Further metabolism and activation are carried out by arylamine N-acetyltransferase 2 (NAT2) and sulfotransferases, which result in esterification and formation of nitrenium ions. These products are highly genotoxic and capable of direct interaction with DNA by adduct formation. The resulting HCA-DNA adducts can cause errors in DNA replication and result in mutations, which contribute to carcinogenesis (Orzechowski et al. 1994; Chou et al. 1995).

Figure 1. Chemical structures of HCAs and relative cell viability of HepG2 and cryopreserved human hepatocytes following HCA treatment.

A, Chemical structures of the HCAs (MeIQ, MeIQx, and PhIP) used in the present study. B and C, Relative cell viability after 3-day treatment with HCAs. HepG2 cells (panel B) or cryopreserved human hepatocytes (panel C) were treated with varying concentrations of the indicated HCA for 3 days. Cells were fixed and stained with DAPI solution, and the total cell count was obtained using fluorescence imaging. Each cell count was presented as a percentage of the value in the vehicle-control group. Data points represent mean ± SEM.

Existing studies of HCAs have examined primarily their mutagenicity and carcinogenicity. However, a recent epidemiological study has linked HCA exposure via cooked meat to development of insulin resistance (Zelber-Sagi et al. 2018). This was a cross-sectional study evaluating non-alcoholic fatty liver disease and insulin resistance. Dietary HCA intake was estimated based on by questionnaires on meat type and cooking methods. The authors observed a significant association between HCA consumption and insulin resistance (odds ratio [OR] = 1.92; 95% confidence interval [CI] = 1.12–3.30) even after multivariate analyses adjusted for 1) age, gender, energy intake per day and BMI and 2) weekly hours of physical activity, smoking status, weekly alcohol portions, saturated fat and cholesterol intake. This study suggests that exposure to HCAs via consumption of cooked meat may contribute to development of insulin resistance.

Insulin is an endocrine peptide hormone that elicits an anabolic response to nutrient availability. It instructs its target organs to uptake glucose and store excess nutrients in the form of glycogen and lipids (Petersen and Shulman 2018). The liver plays a crucial role in systemic regulation of glucose and lipid metabolism, and aberrant hepatic insulin action is considered a potential primary factor in insulin resistance (Santoleri and Titchenell 2019). Under a normal physiologic fasting state, high glucagon-to-insulin ratio decreases glucose consumption and shifts the liver into glucose production by consuming stored glycogen (i.e., glycogenolysis) and from glucogenic precursors (i.e., gluconeogenesis) (Lin and Accili 2011). In pathological insulin resistance, insulin fails to regulate hepatic metabolism, leading to altered glucose metabolism and excess glucose production, while increased lipid synthesis continues. This condition is known as selective hepatic insulin resistance (Brown and Goldstein 2008) and leads to hyperglycemia and fatty liver. Importantly, risk factors represented by metabolic syndrome include insulin resistance and hyperglycemia, in addition to hypertension, central obesity, and abnormal cholesterol or triglyceride levels (Saklayen 2018). An individual with metabolic syndrome is at an increased risk of serious health complications, including atherosclerosis, type II diabetes mellitus, heart attack, kidney disease, fatty liver, vascular disease, and stroke (Saklayen 2018). Some have argued that insulin resistance may be the underlying etiology of metabolic syndrome (Roberts et al. 2013). Nearly one-third of adults in the United States have metabolic syndrome, and 30.2 million adults (12.2% of US adults) have type II diabetes, and these numbers are growing (Saklayen 2018). Thus, understanding environmental factors that contribute to pathogenesis of insulin resistance and metabolic syndrome is necessary to combat this problem.

The aforementioned epidemiological study (Zelber-Sagi et al. 2018) suggested a causal link between HCA consumption and insulin resistance. However, no studies have examined the effect of HCA exposure on insulin sensitivity or glucose homeostasis. In the present study, we investigated the effects of MeIQ, MeIQx, and PhIP, three common HCAs generated when cooking meat, on insulin signaling and glucose production in a hepatocellular carcinoma cell line, HepG2, and cryopreserved human hepatocytes.

Materials and Methods

Heterocyclic amines (HCAs)

PhIP, MeIQx, and MeIQ were purchased from Toronto Research Chemicals. They were prepared into a solution with dimethyl sulfoxide (DMSO) at a stock concentration of 10 mM. The media containing the indicated concentrations of HCAs used for experiments was prepared freshly using the stock solution at the time of experiment.

Cell culture

Cryopreserved human hepatocytes were purchased from BioIVT and stored in liquid nitrogen until use. Hepatocyte samples were collected from consenting donors under IRB approved protocols at the FDA licensed donor center at BioIVT (http://www.bioivt.com/). Hepatocytes were prepared from fresh human tissue and were isolated and frozen within 24 hours of organ removal by BioIVT. The hepatocytes are from human transplant rejected livers and tested negative for hepatitis B and C and HIV1 and 2. Hepatocytes were thawed according to the manufacturer’s instructions by warming a vial of the hepatocytes at 37 °C for 90 seconds and suspending them in InVitroGRO HT medium (BioIVT) containing 1 mL TORPEDO^™ Antibiotic Mix (BioIVT) per 45 mL media. Hepatocytes were then plated on Biocoat^® collagen-coated plates (Corning) and remained in an incubator with a humidified air (95%) and carbon dioxide (CO₂, 5%) condition at 37 °C. HepG2 hepatocellular carcinoma cells were purchased from American Type Culture Collection (ATCC, HB-8065^™). The HepG2 cells were cultured in Dulbecco’s modified Eagle’s medium (DMEM, Gibco) containing 10% fetal bovine serum (FBS, Gibco BRL), 5.5 mM D-glucose, 1 mM pyruvate, 4 mM L-glutamine, and penicillin-streptomycin were kept in an incubator with a humidified air (95%) and carbon dioxide (CO₂, 5%) condition at 37 °C.

Western blot

HepG2 cells or cryopreserved human hepatocytes were plated the day before the treatment on 24-well plates. Cells were incubated for 2 days with complete DMEM containing indicated HCAs (PhIP, MeIQ, or MeIQx) at varying concentrations (0, 10, 25, and 50 μM), then incubated overnight with serum-free DMEM containing indicated HCA treatment conditions. Prior to lysis, cells were treated with insulin (Sigma-Aldrich) at 100 nM (freshly prepared in PBS) for 10 minutes. The cells were lysed immediately in Laemmli buffer (50 mM Tris-Cl [pH 6.8], 2% sodium dodecyl sulfate [SDS; w/v], 0.1% bromophenol blue, 10% [v/v] glycerol) and boiled for 10 minutes. Protein concentrations were determined using Pierce BCA Assay kit (Thermo Scientific) per manufacturer’s instructions. One to two μL of 2-mercaptoethanol was added to each sample and boiled again for 5 minutes. Twenty-five to fifty μg of protein per sample was loaded and separated on 4–12% gradient Bis-Tris Plus polyacrylamide gels (Invitrogen). The gel was transferred to a PVDF membrane and blocked in 5% (w/v) skim milk in tris-buffered saline containing 0.1% Tween 20 (TBST) for 30 minutes. Membranes were incubated with primary antibodies (1:3,000) overnight at 4°C. Membranes were washed 3 times for 5 minutes each with TBST. Membranes were incubated for 1 hour at room temperature with an HRP-conjugated secondary antibody (1:5,000). Membranes were washed with TBST 3 times and then protein-antibody complex was detected with chemiluminescent substrate (Thermo Scientific). Densitometry was performed using Image J analysis software (NIH). Insulin-stimulated phospho-AKT (p-AKT) (Ser473) protein band was quantified and normalized to total AKT. Values presented represent the p-AKT/total-AKT ratio in response to insulin following treatment with indicated HCAs, relative to the vehicle-treated control group. Phospho-FOXO1/FOXO3a/FOXO4 protein band was quantified relative to total FOXO1 protein and normalized to GAPDH. The following antibodies were purchased from Cell Signaling Technology: phosphorylated protein kinase B (p-AKT Ser473; Cat. No. 4060); and protein kinase B (AKT) (Cat. No. 4091); and anti-rabbit IgG, HRP-linked (Cat. No. 7074); and phosphorylated FOXO1 (T24)/FOXO3a (T32)/FOXO4 (T28) (Cat. No. 42022); and FOXO1 (Cat. No. 2880); and GAPDH (Cat. No. 2118). Each experiment was conducted three times with three biological replicates per experiment. For experiments using cryopreserved human hepatocytes, three different donor batches were used with three biological replicates from each batch.

Gene expression analysis by RT-qPCR

HepG2 cells or cryopreserved human hepatocytes were plated the day before the treatment on 12-well plates. Cells were incubated for 3 days with complete media containing indicated HCAs (PhIP, MeIQ, or MeIQx) at varying concentrations (0, 10, 25, and 50 μM). RNA was isolated from the treated cells using E.Z.N.A. Total RNA Kit 1 (Omegabiotek) per manufacturer’s protocol. cDNA was synthesized using High Capacity cDNA Reverse Transcriptase PCR (Thermo Scientific) per manufacturer’s instructions. Gene-specific cDNA was amplified and detected using iTaq Universal SYBR Green Supermix (Bio-Rad) and gene-specific primers and StepOne real-time PCR system (Applied Biosystems). The following primers were used: G6PC, forward 5’- ACGAATCTACCTTGCTGCTCA-3’; reverse 5’-AAAATCCGATGGCGAAGCTG3’. PCK1 or PEPCK, forward 5’-TGATGAGCCGCTAGCTTCAG-3’; reverse 5’-GCCTTTATGTTCTGCAGCCG-3’. FOXO1, forward 5’-AGTGGATGGTCAAGAGCGTG-3’, reverse 5’-TTTGAGCTAGTTCGAGGGCG-3’. PPARα, forward 5’-GATTTCGCAATCCATCGGCG-3’, reverse 5’-AAACGAATCGCGTTGTGTGAC-3’. PGC1α, forward 5’-CACGGACAGAACTGAGGGAC-3’, reverse 5’-TTCGTTTGACCTGCGCAAAG-3’. 18S rRNA, forward 5’-GGAAGGGCACCACCAGGAGT-3’; reverse 5’-TGCAGCCCCGGACATCTAAG-3’. GAPDH, forward 5’-GGTGAAGCAGGCGTCGGAGG-3’; reverse 5’-GAGGGCAATGCCAGCCCCAG-3’. Results in HepG2 cells were normalized to 18S rRNA, and results in cryopreserved human hepatocytes were normalized to GAPDH. The relative fold change was calculated using the delta-delta Ct (2^–ΔΔCt) method with StepOne software (Applied Biosystems). Each experiment was conducted three times with three biological replicates per experiment. For experiments using cryopreserved human hepatocytes, three different donor batches were used with three biological replicates from each batch.

Extracellular glucose measurements

Twenty-four hours after plating, hepatocytes were incubated for 2 days with media containing the indicated HCA at varying concentrations. Then, hepatocytes were incubated for 6 hours with serum-free, glucose-free DMEM containing varying concentrations of indicated HCAs. Hepatocytes were then incubated for 24 hours with serum-free, glucose-free DMEM containing gluconeogenic substrates (2 mM pyruvate, 20 mM lactate, 0.5 mM L-lysine, and 5 mM glycerol) in addition to the indicated HCAs. Media was collected from individual wells following HCA treatment for extracellular measurement of glucose. Glucose levels from the collected media were measured using a colorimetric glucose oxidase assay kit (Sigma-Aldrich, GAGO-20). Three different donor batches of hepatocytes were used with three biological replicates from each batch.

At the end of treatment, cells were washed twice with PBS, fixed with 3.7% formaldehyde (in PBS) for 10 minutes at room temperature, permeabilized with 0.25% Triton X-100 in PBS for 10 minutes at room temperature, and stained with 1 μg/mL DAPI (4′,6-diamidino-2-phenylindole) solution for 5 minutes at room temperature. Cells were observed under a fluorescence microscope (BioTek Cytation 5) using the DAPI channel to measure the total number of cells per well. Glucose measurements were normalized to the total number of cells.

Statistical analysis

Differences in relative gene expression, glucose production, and p-AKT/total AKT ratio between the treatment groups vs. control groups were tested for significance by one-way ANOVA followed by Dunnett’s Comparison Test. Dose-response linear trend was tested for significance by the post test for linear trend using a linear regression model. All statistical analyses were performed using GraphPad Prism v8.2.1 (GraphPad Software). The results are expressed as the mean ± the standard error of the mean (SEM). Statistical significance was determined per the following p-values: * indicates p < 0.05, ** indicates p < 0.01, *** indicates p < 0.001, **** indicates p < 0.0001.

Results

Measuring cytotoxicity and determining treatment concentrations of HCAs

We first determined optimal concentrations for HCA treatment by treating HepG2 cells or cryopreserved human hepatocytes with varying concentrations (0 – 50 μM) of MeIQ, MeIQx, or PhIP (Fig. 1A). We selected these concentrations initially, for we and others have previously used similar concentrations for in vitro experiments with HCAs (Knasmüller et al. 1999). Following 3 days of treatment, there was no significant cell death at any of the concentrations tested in both HepG2 and cryopreserved human hepatocytes (Fig. 1B and C). Thus, we proceeded with a dosing regimen of 0 – 50 μM of HCAs for following experiments.

HCAs decrease insulin signaling in HepG2 cells

Decreased insulin signaling is one of the key features of insulin resistance and type II diabetes (Petersen and Shulman 2018). Insulin-insulin receptor interaction activates an intrinsic tyrosine protein kinase, which auto-phosphorylates the receptor and downstream substrates (Petersen and Shulman 2018). AKT (protein kinase B) is a major downstream effector of insulin signaling and regulates a variety of insulin-mediated responses downstream (e.g., suppression of gluconeogenesis and promotion of lipid synthesis) (Petersen and Shulman 2018; Zhang et al. 2019). For this reason, the level of insulin-induced AKT phosphorylation is often used as an index of insulin sensitivity (Zhang et al. 2019). Insulin signaling leads to AKT phosphorylation at Thr308 and Ser473. The Thr308 site is phosphorylated first by upstream kinases, and then the Ser473 site is phosphorylated to fully activate the AKT protein (Gonzalez and McGraw 2009; Mackenzie and Elliott 2014).

To evaluate the effects of HCAs on insulin sensitivity, we assessed insulin-induced phosphorylation of AKT at Ser473 following HCA treatment in HepG2 cells. HepG2 cells have been used to study hepatic insulin signaling (Yang et al. 2016; Sefried et al. 2018), and so we conducted initial experiments using this model and later confirmed these findings in cryopreserved human hepatocytes. HepG2 cells were treated with varying concentrations of MeIQ, MeIQx, or PhIP for 3 days prior to the insulin treatment. Phosphorylated AKT (at Ser473) and total AKT levels were measured using Western blot. HepG2 cells treated with MeIQ and MeIQx showed a concentration-dependent decrease in phosphorylated AKT to total AKT ratio (p-AKT/AKT) in response to insulin, compared to vehicle-control group (Fig. 2). There was a 25.0% (p < 0.01), 32.0% (p < 0.01), and 56.5% (p < 0.001) significant reduction in insulin-stimulated p-AKT/AKT following treatment with MeIQ at 10 μM, 25 μM and 50 μM, respectively (Fig. 2A). MeIQx treatment caused a dose-dependent reduction (linear trend, p = 0.0002) in insulin-stimulated p-AKT/AKT, with a 27.7% (p < 0.05), 45% (p < 0.01), and 51.3% (p < 0.001) significant reduction at 10 μM, 25 μM, and 50 μM, respectively (Fig. 2B). PhIP treatment at 10 μM caused a 19% (p < 0.05) significant reduction in insulin-stimulated AKT activation, but this reduction was not observed at higher concentrations of PhIP (Fig. 2C).

Figure 2. Insulin-induced AKT phosphorylation following HCA treatment in HepG2 cells.

HepG2 cells were cultured with the indicated concentration of MeIQ (panel A), MeIQx (panel B), and PhIP (panel C) for 2 days, followed by a serum-starvation overnight. Prior to harvest, cells were treated with 100 nM insulin for 10 minutes. The relative levels of phospho-AKT (p-AKT) (Ser473) and total AKT (AKT) were measured using Western blot. The insulin-stimulated p-AKT/AKT ratio was expressed as a percentage of the vehicle-treated control cells. Data points represent mean ± SEM. Significance of linear trend was determined using the post test for linear trend (linear regression model). *, p < 0.05; **, p < 0.01; ***, p < 0.001.

HCAs decrease insulin signaling in cryopreserved human hepatocytes

Next, we sought to confirm the findings in HepG2 cells using cryopreserved human hepatocytes. We repeated the study using pooled cryopreserved human hepatocytes which represented a mixture of hepatocytes from 10 or 20 donors. Cryopreserved human hepatocytes were treated with 25 μM MeIQ, MeIQx, or PhIP for 3 days prior to the insulin treatment.

Phosphorylated AKT (at Ser473) and total AKT levels were measured using Western blot. Cryopreserved human hepatocytes treated with MeIQ and MeIQx showed a significant decrease in phosphorylated AKT to total AKT ratio (p-AKT/AKT) in response to insulin, compared to vehicle-control group (Fig. 3). There was a 53.5% (p < 0.05), 58.5% (p < 0.01), and 13.5% reduction in insulin-stimulated p-AKT/AKT following treatment with MeIQ, MeIQx, or PhIP respectively (Fig. 3), although the reduction in PhIP-treated cells was not statistically significant, likely because PhIP treatment alone induced p-AKT even in the absence of insulin. These findings are consistent with observations made in HCA-treated HepG2 cells.

Figure 3. Insulin-induced AKT phosphorylation following HCA treatment in cryopreserved human hepatocytes.

Cryopreserved human hepatocytes were cultured with 25 μM of MeIQ, MeIQx, or PhIP for 2 days, followed by a serum-starvation overnight. Prior to harvest, cells were treated with 100 nM insulin for 10 minutes. The relative levels of phospho-AKT (p-AKT) at Ser473 and total AKT (AKT) were measured using Western blot. The insulin-stimulated p-AKT/AKT ratio was expressed as a percentage of the vehicle-treated control cells. Data points represent mean ± SEM. *, p < 0.05; **, p < 0.01.

HCAs induce expression of gluconeogenic genes in HepG2 cells

The liver plays a key role in regulation of glucose homeostasis by controlling various pathways of glucose metabolism, including glycogenolysis, glycolysis and gluconeogenesis (Han et al. 2016). Regulation of enzymes involved in these pathways is required to maintain proper glucose homeostasis. The gluconeogenesis metabolic pathway generates glucose from non-carbohydrate carbon substrates. Key regulatory enzymes in this pathway are induced under fasting conditions (Han et al. 2016) and include glucose 6-phosphatase (G6PC) and phosphoenolpyruvate carboxykinase (PCK1; PEPCK). In insulin resistance, insulin fails to properly phosphorylate AKT, which can lead to dysregulation of gluconeogenic enzymes like G6PC and PCK1, ultimately causing sustained glucose production, despite insulin secretion (Santoleri and Titchenell 2019; Zhang et al. 2019).

To investigate the effects of HCAs on glucose homeostasis, we examined if HCA treatment altered expression of gluconeogenic, insulin receptor-target genes, G6PC and PCK1, in HepG2 cells. G6PC catalyzes the final step of gluconeogenesis, hydrolyzing glucose-6-phosphate to generate free glucose and inorganic phosphate (Puigserver et al. 2003). PCK1 is the first rate-limiting enzyme of gluconeogenesis that converts oxaloacetate and GTP into phosphoenolpyruvate (PEP) and CO₂ (Puigserver et al. 2003). Treatment of HepG2 cells with MeIQ, MeIQx, or PhIP resulted in an increase in G6PC gene transcripts (Fig. 4A–C). Each of the HCAs showed a statistically significant increase in G6PC gene transcript levels at the 50 μM concentration, compared to vehicle-control (MeIQ, p < 0.05; MeIQx, p < 0.0001; PhIP, p < 0.0001). In particular, MeIQx treatment at 50 μM resulted in a nearly 10-fold increase in G6PC mRNA compared to the vehicle-control (Fig. 4B) and exhibited a dose-response (linear trend, p = 0.0003). Similarly, PhIP exhibited a dose-dependent (linear trend, p < 0.0001) increase in G6PC mRNA (Fig. 4C). Additionally, treatment with MeIQ, MeIQx, or PhIP resulted in a significant increase (p < 0.05) in PCK1 gene transcripts in HepG2 cells at the 50 μM concentration, and MeIQx and PhIP both exhibited significant concentration-dependent increases in PCK1 mRNA (MeIQx, p = 0.0022; PhIP, p = 0.0152) (Fig. 4A–C).

Figure 4. Increases in insulin receptor-target genes involved in gluconeogenesis following HCA treatment.

HepG2 cells (panels A-C) and cryopreserved human hepatocytes (panels D-F) were cultured with varying concentrations of the indicated HCA (MeIQ, MeIQx, or PhIP) for 3 days. The relative mRNA level of G6PC or PCK1 was measured by RT-qPCR using 18S ribosomal RNA as an internal control for HepG2 cells and using GAPDH as an internal control for hepatocytes. The mRNA levels were expressed as changes relative to that in the vehicle-control. Data points represent mean ± SEM. Significance of linear trend was determined using the post test for linear trend (linear regression model). *, p < 0.05; **, p < 0.01; ***, p < 0.001; ****, p < 0.0001.

HCAs induce expression of gluconeogenic genes in cryopreserved human hepatocytes

Next, we determined if HCAs altered expression of gluconeogenic, insulin receptor-target genes, G6PC and PCK1, in cryopreserved human hepatocytes as well. Similar to our findings in HepG2 cells, we observed significant increases in G6PC and PCK1 mRNA levels following MeIQ, MeIQx or PhIP treatment in cryopreserved human hepatocytes (Fig. 4D–F). MeIQ, MeIQx, or PhIP treatment each exhibited statistically significant linear dose-response trend in G6PC gene transcript levels (MeIQ, p = 0.0002; MeIQx, p < 0.0001; PhIP, p = 0.0066), and MeIQ or MeIQx treatment each exhibited a statistically significant linear dose-response trend in PCK1 gene transcript levels (MeIQ, p = 0.0005; MeIQx, p = 0.0002). These results agreed with the observations made in HepG2 cells.

HCAs alter expression of genes that regulate gluconeogenesis in cryopreserved human hepatocytes

Since HCA treatment induced expression of G6PC and PCK1, in cryopreserved human hepatocytes, we investigated if HCAs altered the expression of additional genes that regulate gluconeogenesis, including FOXO1, PPARα, and PGC1α. Forkhead box O1 (FOXO1) gene encodes a transcription factor protein that regulates metabolic homeostasis and insulin signaling (Xing et al. 2018). FOXO1 regulates expression of gluconeogenic genes, G6PC and PCK1, and increased expression of FOXO1 correlates with increased expression of G6PC and PCK1 (Gross et al. 2008). Our study also showed an increased FOXO1 expression following treatment with HCAs in hepatocytes (Fig. 5 A–C). Cryopreserved human hepatocytes treated with MeIQx showed a statistically significant increase (p < 0.05) in FOXO1 mRNA levels, and hepatocytes treated with MeIQ showed a slight increase in FOXO1 mRNA, though not statistically significant, while PhIP-treated hepatocytes did not show a change in FOXO1 levels.

Figure 5. Changes in expression of genes and protein that regulate gluconeogenesis following HCA treatment in cryopreserved human hepatocytes.

Cryopreserved human hepatocytes were cultured with 25 μM of the indicated HCA (MeIQ, MeIQx, or PhIP) for 3 days. The relative mRNA level of the indicated gene was measured by RT-qPCR using GAPDH as an internal control (panels A-C). The mRNA levels were expressed as changes relative to that in the vehicle-control. Relative levels of phospho-FOXO1 (p-FOXO1) at Thr24, FOXO1, and GAPDH were measured using Western blot (panel D). The p-FOXO1 protein level was measured relative to total FOXO1 protein level and normalized to GAPDH protein level, then presented as relative change compared to the vehicle-control (panel E). Bars represent mean ± SEM. *, p < 0.05; **, p < 0.01; ***, p < 0.001; ****, p < 0.0001.

Furthermore, peroxisome proliferator activated receptor alpha (PPARα) regulates expression of genes involved in glucose and lipid metabolism and transport (Kersten and Stienstra 2017). Although hepatocytes treated with MeIQ showed only a marginal increase in PPARα mRNA levels that was not statistically significant, hepatocytes treated with MeIQx and PhIP showed a significant increase (p < 0.01, p < 0.0001, respectively) in PPARα mRNA levels (Fig. 5 A–C). These results indicate that HCA treatment of hepatocytes lead to changes in gene expression of some of the key regulators of gluconeogenesis.

PPARG Coactivator 1 Alpha (PGC1α) is a transcriptional co-activator that regulates expression of genes involved in glucose and fatty acid metabolism and is a known regulator of hepatic gluconeogenesis (Yoon et al. 2001; Goffart and Wiesner 2003). Cryopreserved human hepatocytes treated with MeIQ, MeIQx, or PhIP expressed a significant decrease (p < 0.001) in PGC1α mRNA levels (Fig. 5 A–C). These results contradict what was expected since a reduction in gene expression of this co-activator of gluconeogenic genes does not correlate with increased expression of G6PC and PKC1 genes observed after exposure to HCAs. However, PGC1α is not the sole regulator of these genes. Additionally, type 2 diabetic patient livers have been shown to have decreased PGC1α levels (Westerbacka et al. 2007; Ahrens et al. 2013; Koliaki et al. 2015), and mice with low hepatic PGC1α accumulate excess lipids in their livers (Leone et al. 2005; Estall et al. 2009; Morris et al. 2012), which is a hallmark of insulin resistance. Thus, decreased PGC1α expression may be associated with the metabolic changes observed after HCA treatment, and this association and its mechanism need to be explored further.

To further investigate the mechanism by which HCAs lead to increased gluconeogenic gene expression in hepatocytes, we investigated changes in FOXO1, a transcription factor for gluconeogenic genes, at the protein level. FOXO1 is transcriptionally active only in its non-phosphorylated form (Puigserver et al. 2003). FOXO1 protein is phosphorylated by AKT, which alters DNA binding activity and thus halts transcription of gluconeogenic genes (Puigserver et al. 2003). Failure to phosphorylate FOXO1 renders the protein active and thus induces gluconeogenesis (Zhang et al. 2006). Cryopreserved human hepatocytes treated with MeIQ, MeIQx, or PhIP showed a statistically significant decrease in phospho-FOXO1 (Thr24) relative to total FOXO1 protein when compared to the vehicle-treated control (MeIQ, p < 0.001; MeIQx, p < 0.01; PhIP, p < 0.001) (Fig. 5 D–E). This indicates that increased FOXO1 activity may be partially responsible for increased gluconeogenic gene expression in HCA-treated hepatocytes.

HCAs increase glucose production in cryopreserved human hepatocytes

Development of insulin resistance leads to altered glucose homeostasis, and as a result, increased and sustained glucose production is a hallmark of insulin resistance and type II diabetes (Petersen and Shulman 2018). Since HCA treatment on HepG2 cells and hepatocytes lead to an increase in expression of gluconeogenic, insulin receptor-target genes, G6PC and PCK1, we tested if HCA treatment leads to increased glucose production in cryopreserved human hepatocytes. During the HCA treatment, hepatocytes were serum- and glucose-starved before being provided with gluconeogenic substrate media containing pyruvate, lactate, L-lysine, and glycerol. The level of extracellular glucose in the media was measured using a glucose oxidase-based assay and then normalized to total cell count. The HCA treatments resulted in an increase in extracellular glucose in cryopreserved human hepatocytes, suggestive of an increase in glucose production (Fig. 6). Although the differences in the extracellular glucose level between the control group and individual MeIQ treatment groups were not found statistically significant, there was a statistically significant linear trend (i.e., concentration-dependent increase, p = 0.0032) (Fig. 6A). There was a similar concentration-dependent effect (p = 0.0228) by MeIQx, and the highest concentration of MeIQx caused nearly a 2-fold increase in extracellular glucose, compared to the vehicle-control (Fig. 6B). PhIP treatment also caused an averaged 34% increase in extracellular glucose although there was no dose-response observed (Fig. 6C). These results suggest that HCAs can induce glucose production in human hepatocytes at variable levels.

Figure 6. Increases in glucose production following HCA treatment in cryopreserved human hepatocytes.

Cryopreserved human hepatocytes were cultured with varying concentrations of the indicated HCA MeIQ (panel A), MeIQx (panel B), or PhIP (panel C) for 2 days, and glucose- and serum-starved overnight. The cells were then incubated in a gluconeogenic media containing 2 mM pyruvate, 0.5 mM lysine, 20 mM lactate, and 5 mM glycerol for 24 hours. Media was collected, and extracellular glucose was measured using a glucose oxidase-based assay. Glucose concentration was normalized to total cell count and expressed as a percentage of the value found in the vehicle-control cells. Data points represent mean ± SEM. Significance of linear trend was determined using the post test for linear trend (linear regression model).

Discussion

The current study demonstrated, for the first time, that common HCAs found in cooked meat can cause insulin resistance and increases in glucose production and gluconeogenic gene expression in human hepatocytes. Notably, three different HCAs commonly found in cooked meats (i.e., MeIQ, MeIQx, and PhIP) and two different cell model systems (i.e., HepG2 and cryopreserved human hepatocytes) were used to establish these findings, indicating that they are not unique to a particular HCA or exclusive to one type of cell population.

The present findings imply that exposure to HCAs via consumption of cooked meat could potentially lead or contribute to development of insulin resistance and hyperglycemia, which are key hallmarks of type II diabetes and metabolic syndrome. This is supported by several epidemiological studies. As previously mentioned, Zelber-Sagi and colleagues reported that consumption of meat cooked via “unhealthy” methods (e.g., grilling or broiling to well-done level or frying) is related to insulin resistance and found that there is a significant association between estimated HCA consumption via cooked meat and insulin resistance, among both general population and patients with non-alcoholic fatty liver disease (odds ratio [OR] = 1.92; 95% confidence interval [CI] = 1.12–3.30) (Zelber-Sagi et al. 2018). Importantly, after multivariate adjustment, the association was independent of cholesterol and saturated fat intake. A higher frequency of unhealthy cooking methods, such as broiling, barbequing, and roasting, was each independently associated with a higher risk of type II diabetes. In contrast, the frequency of stewing or boiling red meats was not associated with type II diabetes risk (Liu et al. 2018). These studies suggest that unhealthy cooking methods, which favor production of HCAs, significantly increase the risk of developing insulin resistance or type II diabetes. Accordingly, the authors suggested exposure to polyaromatic hydrocarbons and HCAs commonly present in cooked meat as a potential underlying, biological mechanism for this phenomenon (Liu et al. 2017). In the follow-up study of three prospective cohorts, the authors reported that open-flame and/or high-temperature cooking >15 times/month, compared with <4 times/month, the hazard ratio and 95% CI of type II diabetes was 1.28 and 1.18–1.39, respectively (Liu et al. 2018). In addition, estimated intake of HCAs was also associated with an increased type II diabetes risk, and this was the first study linking higher estimated dietary HCA intake with an increased risk of type II diabetes (Liu et al. 2018). Taken together, these previous and our present findings support the hypothesis that exposure to HCAs produced during unhealthy cooking of meat contributes to pathogenesis of insulin resistance and type II diabetes.

The mechanism by which HCAs induce insulin resistance in hepatocytes is unknown. Pathological features that are commonly associated with cellular insulin resistance include inflammation (de Luca and Olefsky 2008), mitochondrial dysfunction (Sangwung et al. 2020), ER stress (Salvadó et al. 2015), and lipotoxicity (Alves et al. 2011). One of the common underlying attributes of these features is increased reactive oxygen species (ROS) (Petersen and Shulman 2018) and ROS have been shown to play a causal role in different forms of insulin resistance (Houstis et al. 2006). Multiple mechanisms by which elevated ROS may induce or contribute to insulin resistance have been proposed. For example, ROS are known to activate JNK and IKKβ (Storz and Toker 2003; Kamata et al. 2005) which are considered key contributors to the development of insulin resistance in obesity and diabetes (Gual et al. 2005). JNK and IKK can inhibit insulin receptor signaling by phosphorylating insulin receptor substrate 1 (IRS1) at serines. Serine-phosphorylation of IRS1 inhibits its association with the insulin receptors, promotes its degradation and suppresses binding of PI3K (Gual et al. 2005; Petersen and Shulman 2018), thus, blunting the effects of downstream insulin signaling. Likewise, a mechanism by which HCAs induces insulin resistance in hepatocytes may be contributed by increased ROS. In fact, consumption of HCA from cooked meat has been associated with increased systemic level of oxidative stress. A study of 561 adults in Brazil found that estimated intake of HCAs from meat is associated with the level of oxidative stress, estimated by malondialdehyde concentration in the plasma (OR 1.17; 95% CI 1.01–1.38) (Carvalho et al. 2015). In support, we have observed that HCAs, including PhIP and MeIQx, cause increases in DNA damage and ROS level in Chinese hamster ovary (CHO) cells that express human CYP1A2 and NAT2 (Metry et al. 2010, Salazar-González et al., unpublished data). Moreover, the level of DNA damage and ROS production was higher in cells expressing a rapid allele of NAT2 (NAT2*4), compared to a slow allele, NAT2*5B (Metry et al. 2010). This data indicates that HCAs induce cellular ROS levels, and this effect is dependent on their metabolism by CYP1A2 and NAT2. Based on this, we speculate that a potential mechanism for HCA-induced insulin resistance may be related to the formation of ROS during or following HCA metabolism. It is well-known that mutagenicity of HCAs requires metabolism. Metabolic activation of HCAs occurs primarily in the liver by CYP1A2-mediated N-oxidation of the exocyclic amine groups to form the N-hydroxy-HCA derivatives (Nebert 1991; Woziwodzka et al. 2010). The N-hydroxy-HCA metabolites can undergo phase II conjugation reactions to form highly reactive esters. These esters may undergo heterolytic cleavage to generate reactive nitrenium ions, which are the ultimate carcinogenic metabolite (Turesky 2007). NAT2 is one of the primary enzymes responsible for catalyzing this phase II metabolism. NAT2 expresses a well-defined genetic polymorphism in humans (Hein 2002). Depending on the combination of NAT2 alleles they carry, individuals can be categorized into rapid, intermediate, or slow acetylators who exhibit differential metabolism of aromatic amines, including HCAs (Hein 2002). However, it is currently unknown if metabolism of HCAs is also required for their effects on hepatic insulin resistance and glucose production. Additionally, our lab has reported that human NAT2 is transcriptionally regulated by glucose and insulin in liver cancer cell lines, including HepG2 cells (Hong et al. 2022). We speculate that NAT2-mediated metabolism of HCAs is an important determinant of their metabolic effects in hepatocytes. A further investigation is required to assess the contribution of HCA metabolism to their effects on insulin sensitivity and glucose production.

A notable observation from the current study is that MeIQ and MeIQx caused a dose-dependent reduction in insulin sensitivity and a dose-dependent increase in glucose production, while the effects of PhIP were either not dose-dependent or absent. Although PhIP did, however, lead to a significant decrease (p < 0.05) in insulin sensitivity at 10 μM in HepG2 cells, it failed to produce a similar response in cryopreserved hepatocytes and its ability to induce glucose production was marginal. The differences in structure of these HCAs (Fig. 1A) and their metabolism are potentially responsible for this observation. The structure of PhIP is notably different from other compounds in its classification. Most HCAs, including MeIQ and MeIQx, have fully planar aromatic structures with no bulky out-of-plane functionalities (Felton et al. 1999). PhIP, in contrast, possesses a phenyl moiety that is not necessarily co-planar with the main bicyclic imidazopyridine, and so the metabolism of PhIP is different from other HCAs. CYP1A2 primarily catalyzes the detoxification of MeIQx by oxidation of the 8-methyl group, whereas it catalyzes the bioactivation of PhIP by oxidation of the exocyclic amine group (Langouët et al. 2001; Gu et al. 2010), and so it is speculated that alternative enzymes may be involved in PhIP-induced toxicity. In the neonatal mouse model, higher incidences of lymphoma and hepatocellular adenoma occurred in female P4501A2-knockout mice than in wild-type mice exposed to high doses of PhIP (11 or 22 mg/kg), indicating that PhIP-induced carcinogenesis is independent of P450 1A2 expression (Kimura et al. 2003). It was also observed that PhIP-DNA adduct formation was independent of NAT2 acetylator activity in adult female congenic rapid and slow rats, while MeIQx-DNA adducts, particularly in the liver, were significantly lower in slow acetylators (Metry et al. 2009). The structural and metabolic differences responsible for discrepancies in HCA-induced genotoxicity and carcinogenicity could also potentially be responsible for the discrepancies reported here (i.e., PhIP not exhibiting a linear dose-response in the results reported).

In summary, HCAs commonly found in cooked meat caused insulin resistance and increased glucose production in HepG2 and human hepatocytes. We have also shown that insulin receptor-target genes involved in gluconeogenesis are upregulated in human hepatocytes following exposure to HCAs, indicating that upregulation in gluconeogenesis may, at least in part, contribute to HCA-induced increase in glucose production. Taken together, the current findings imply that, independent of meat consumption, dietary exposure to HCAs contribute to pathogenesis of insulin resistance and hyperglycemia. Additional studies are required to determine the precise mechanism by which HCAs elicit the responses reported here. Future studies should also explore the role of metabolism of HCAs as well as NAT2 genetic polymorphisms on metabolic effects of HCAs.