HepG2 cells do not express xanthine oxidoreductase (XOR): Implications for XOR and uric acid-related research

Abstract

Although the relative extent of xanthine oxidoreductase (XOR) varies considerably in human tissues, the greatest specific activity is reported in the liver and intestines. Unlike murine models, where primary hepatocytes are readily available, human counterparts are not. As such, investigators often utilize the human carcinoma cell line HepG2 for in vitro experimentation as these cells proliferate well in culture medium. Some of the studies using HepG2 cells for proof-of-principal experimentation focus on uric acid (UA) and/or XOR activity. However, it has been reported that hepatocellular carcinoma diminishes XOR expression to nearly unmeasurable levels when compared to normal cell counterparts which posits the question of validity in the context of using HepG2 cells for XOR/UA assessments. As such, we closely examined XOR expression, protein abundance, and enzymatic activity in HepG2 cells and compared these results to an immortalized murine hepatocyte line (AML12) as well as murine liver homogenate. We report the absence of detectable XOR message, protein, and enzymatic activity in HepG2 cells. Since cellular XOR expression has been reported to be stimulated by hypoxia and serum starvation, we then exposed both AML12 and HepG2 cells to hypoxia (1% O₂) for 24 h or serum starvation which impacted XOR activity in AML12 cells, but not in HepG2 cells. Knowing that several studies report XOR activity in HepG2 cells, we compared a commonly used non-selective XOR assay kit to our sensitive and selective HPLC/electrochemical assay. The kit assay indicated XOR activity in HepG2 cells; however, XOR inhibition did not diminish this activity. To account for alternative enzymatic oxidization of hypo/xanthine to UA, we investigated HepG2 cells for aldehyde oxidase (AO) abundance and found immunodetectable AO protein as well as AO-inhibitable conversion of hypoxanthine to UA; albeit, at a much slower rate when compared to XOR. In toto, these data confirm that HepG2 cells do not demonstrate XOR activity and thus commonly used kit assays for XOR, in the absence of XOR inhibitor controls, may indicate a positive response induced by an alternative source of peroxide and UA such as AO.

1. Introduction

Uric acid (UA) has long been viewed as a terminal catabolite in the upper primates/humans that, under normal concentrations, may serve as an antioxidant [1–3]. Classically, elevated circulating levels of UA (hyperuricemia) are the primary risk indicator for gout where UA levels in the joints reach saturation, form monosodium urate crystals, induce severe inflammation and joint immobility. However, recent studies have shown that elevation in UA levels is associated with several metabolic and vascular disorders, including but not limited to obesity, diabetes, and stroke [4–8]. In some cases, the association between hyperuricemia and disease has generated considerable speculation that UA may be causative of these disease processes. As such, there has been considerable interest in investigating contributory roles for UA in several disease models. As a result of this renewed interest, there is an increase in the employment of cellular models, especially primary contributors to circulating UA concentration (e.g. hepatocytes).

Circulating UA levels are regulated by modulation of production and/or excretion via the kidneys and gut [6,9,10]. The currently recognized, sole enzymatic source of UA is the purine catabolizing enzyme, xanthine oxidoreductase (XOR). XOR catalyzes the terminal two steps of purine catabolism by serially oxidizing hypoxanthine to xanthine and then xanthine to UA [11,12]. XOR refers to two different forms of the same enzyme, xanthine dehydrogenase (XDH) and xanthine oxidase (XO). XOR is transcribed and translated from a single gene (Xdh) as XDH. Conversion to XO is accomplished via post-translational modification, which often occurs as a result of inflammation [13,14]. In the XDH form, the enzyme catalyzes the oxidation of purines to UA at the molybdenum cofactor (Mo-co) and reduces NAD⁺ to NADH at the FAD. In the XO form, the purine reaction is identical to XDH, yet, the terminal electron acceptor is molecular oxygen (O₂), which is reduced to oxidants; specifically, hydrogen peroxide (H₂O₂) and superoxide $\left({\text{O}}_2^{•-}\right)$ [12], (see scheme in Fig. 1A). However, it is critical to reinforce the fact that, regarding UA production, both forms of the enzyme perform this function.

Fig. 1. Schematic illustration of XOR-catalyzed reactions and in-silico analysis of XDH gene expression in HepG2 cells.

A) Diagrammatic representation of one monomer of the heterodimer holoenzyme depicting xanthine dehydrogenase (XDH) and xanthine oxidase (XO) substrates and reactions. B) The Depmap database shows 0.0145 log2 (TPM+1) for the XDH gene expression in HepG2 cells. C) The XDH copy number for HepG2 is 1.292. Depmap outlines gene-level copy number data is log2 transformed with a pseudo-count of 1; log2 (CN ratio + 1). A) Created in BioRender. King, R. (2025) https://BioRender.com/a32y810.

While XOR is expressed ubiquitously in mammalian tissues, the greatest specific activity is noted to be associated with the splanchnic system (e.g. liver and intestines) [15,16]. As such, investigators commonly utilize liver tissue homogenates or hepatocellular cell lines for proof-of-principle experiments involving XOR and more importantly, UA. Commonly used cell lines for these experimental endeavors include the immortalized murine hepatocyte (AML12) and the human hepatocellular carcinoma (HepG2) [17–21]. While the AML12 line has been useful in UA-related investigation, the HepG2 line is often employed to give a translational angle from mouse to human; however, the validity of this approach is questionable since hepatocellular carcinoma lacks Xdh protein expression and enzymatic activity [22]. Using HepG2 as an experimental model for XOR and UA is further questioned by in-silico data for HepG2 cells (Fig. 1B&C). While numerous reports demonstrate XOR protein and/or XOR activity in HepG2 cells, we posit that these determinations may be affected by several factors including the lack of specificity of the XOR activity assays used and the similarity between XOR and another hepatocyte enzyme, aldehyde oxidase (AO) which demonstrates significant sequence homology to XOR and shares overlap in several substrates and inhibitors [23,24].

Herein, we performed a series of experiments in HepG2 cells, with AML12 cells serving as a control/reference, to assess XOR message, protein, and enzymatic activity. Using sensitive and specific methods, we report that HepG2 cells do not demonstrate XOR message, protein, or activity. In addition, we show that a commonly used commercially available XOR kit, as well as underrecognized enzymatic activity of AO, can lead to the incorrect interpretation that HepG2 cells do indeed contain XOR and thus produce UA.

2. Materials and methods

Reagents:

All reagents were procured through Sigma-Aldrich unless otherwise specified.

Cell culture:

The HepG2 cells and AML12 hepatocytes were obtained from the American Type Culture Collection (ATCC). They were cultured at 37°C, under a humidified atmosphere of 5% CO₂. All cells tested negative for mycoplasma. HepG2 cells were grown in DMEM media and supplemented with heat-inactivated 10% Fetal Bovine Serum (FBS). AML12 hepatocytes were grown in DMEM media supplemented with 10% FBS, 5 μg/mL transferrin, 5 ng/mL selenium, 10 nM insulin, and 10 nM dexamethasone. All cell lines were supplemented with 1% penicillin/streptomycin. All media and supplements were obtained from ThermoFisher Scientific, Waltham, MA. All cell lines were measured and quantified before passage 15.

Hypoxia:

HepG2- and AML12- cells were grown to ~75% confluence before being placed in the incubation chamber of a BioSpherix XVivo system, model X3 hypoxic glovebox, with an incubation configuration. Oxygen was set and validated at 1%, CO₂ was 5%, and the temperature was 37°C. The cells were maintained in hypoxic conditions for 24 h before being processed for protein and RNA collection, as described below.

Sample collection:

Liver samples were harvested from 10-week-old C57Bl/6J mice after whole-body effusion with 1x PBS. Fifty (50) mg pieces of liver were homogenized in 300 μL RIPA (ThermoFisher, #J63306.AP) with a protease inhibitor cocktail, HALT (ThermoFisher, #1862209). The samples were centrifuged for 10 min at 13,200×g at 4°C. The resulting supernatant was used in all assays. Cells were washed in sterile PBS to collect protein and gently scraped in 5 mL ice-cold PBS into a 50 mL falcon tube, followed by a 5-min spin at 1800 rpm. The supernatant was aspirated, and the pellet was resuspended in 600 μL RIPA (ThermoFisher, #J63306.AP) with a protease inhibitor cocktail, HALT (ThermoFisher, #1862209). The pellet was stored at −20°C until analyzed. Total RNA was isolated according to TRIzol manufacturer’s instructions. In brief, the working area was sprayed with RNaseZap (ThermoFisher, #AM9780), cell media was aspirated, followed by a sterile 1xPBS washing step. Then 3 mL TRIzol reagent (ThermoFisher, #15596018) was added to the cell culture plate. The lysate was pipetted into sterile snap-close tubes containing 600 μL of chloroform and incubated for 3 min. The tubes were centrifuged for 15 min at 12,000×g at 4°C. The aqueous phase was transferred to a new tube, and 1.5 mL of isopropanol was added and left to incubate for 10 min at 4°C. The solution was centrifuged at 12,000×g for 10 min at 4°C, where the result was a small white pellet. The supernatant was discarded, and 3 mL of 75 % ethanol was added before vortexing, followed by centrifugation at 7,500×g for 5 min at 4°C. The supernatant was discarded, and the pellet was allowed to dry in the hood for 10 min. The dried pellet was resuspended in nuclease-free water. RNA purity was assessed using a Nano-Drop spectrophotometer, with a 260/280 reading of at least 2.0.

cDNA synthesis and Quantitative PCR:

Total RNA was converted to cDNA according to High-Capacity cDNA Transcription Kit (ThermoFisher, #4368814) manufacturer’s instructions. In brief, a 2X RT master mix was made so that each well would have the following components: 2.0 μL of 10X RT Buffer, 0.8 μL of 25X dNTP Mix (100 mM), 2.0 μL of 10X RT Random Primers, 1.0 μL of Multiscribe Reverse Transcriptase, and 4.2 μL of Nuclease-Free water. For a 20 μL reaction, the 2X RT master mix comprises 10 μL and the remaining 10 μL is 500 ng of total RNA in nuclease-free water. The tubes are loaded into a thermal cycler with the following settings: step one 25°C for 10 min, step two 37°C for 120 min, step three 85°C for 5 min, and step four 4°C hold until collection. Relative gene expression was analyzed using quantitative PCR (qPCR) with an SYBR Green PCR Master Mix (ThermoFisher) with the Applied Biosystem QuantStudio 6 Flex (ThermoFisher), using 50 ng of cDNA per reaction. The gene expression was normalized to the housekeeping gene β-actin. The relative change in mRNA expression was calculated using the ▴▴Ct method [25]. The qPCR was performed with technical and biological triplicates. Integrated DNA Technologies (IDT) PrimerQuest Tool was used to design and assess primer efficiencies. For a list of primers used, see (Suppl. Table 1).

Western blotting:

Electrophoresis used 4–12% Bis-Tris gradient Criterion^™ XT Precast Gels (Bio Rad, #3450123). Samples were resolved by SDS-PAGE and transferred to PVDF. Membranes were blocked with 5% bovine serum albumin. The primary antibody xanthine oxidase A-3 mouse monoclonal antibody (Santa Cruz, #398548) was used at a 1:500 dilution. The Beta-actin primary antibody was a mouse monoclonal antibody (Santa Cruz, #47778), 1:1000 dilution. The secondary antibody was mouse IgG Fc BP-HRP (Santa Cruz, #525409), 1:3000 dilution. The aldehyde oxidase rabbit monoclonal antibody (ThermoFisher, #19495–1-AP) was added at a 1:500 dilution. Rabbit secondary antibody IgG Fc BP-HRP (Jackson Immuno, #711-035-152) was added at 1:5000 dilution, courtesy of Dr. Bradley Webb. Blots were imaged with Syngene G-Box, model Chemi XX6. All primary antibodies were incubated at 4°C overnight. All secondary antibodies were incubated at RT for 1 h.

XOR activity assay:

XOR activity assay is described in detail [26]. Cells were homogenized in RIPA buffer with a protease inhibitor cocktail and concentrated to a soft pellet via centrifugation. Cell lysates were processed for XOR activity as follows. Potential urate oxidase (uricase) activity was inhibited by the addition of oxonic acid (100 μM) to ensure UA levels were not altered and thus enzyme activity was not underestimated. Total UA production in 60 min at 37°C in the presence of xanthine (75 μM) served as the basis for the quantification of total xanthine oxidoreductase (XOR) activity. To ensure XOR dependence on urate formation as well as establish base-line UA levels, allopurinol (100 μM) was used in parallel samples to inhibit XOR. To determine XDH specific activity, a third tube was added with the addition of NAD+ and no allopurinol. Following incubation for 60 min at 37°C, protein was precipitated by addition of ice cold acetonitrile. The samples were then centrifuged for 12 min at 13,200×g, at 4°C. Following centrifugation, the supernatant was removed, placed in borosilicate glass tubes, dried (60 min), and resuspended in isocratic mobile phase (300 μL) and filtered through a 0.20 μm nylon membrane filter unit into 11 mm plastic snap-top auto-sampler vials. The UA content was measured by electrochemical detection (Vanquish UltiMate 3000 ECD-3000RS) coupled to reverse-phase HPLC using a C18 column (150 × 4.6 mm, 3 μm particle size, Luna Phenomenex) and isocratic mobile phase (50 mM sodium dihydrogen phosphate, 4 mM dodecyltrimethylammonium chloride, 2.5% methanol, pH 7.0 using a flow rate of 1 mL/min). XO and XDH activity are denoted as μUnits/mg of protein.

AO-dependent UA production capacity:

HepG2 lysate samples were assessed for capacity to oxidize hypoxanthine to UA by using a modified version of the XOR assay described above. For AO-dependent UA production, lysates were incubated with hypoxanthine (500 μM) for 1 h at 37°C with or without the AO inhibitor Raloxifene, at 50 nM. The presence of UA was assessed as above via reverse phase HPLC coupled to an electrochemical detector.

Colorimetric XO assay kit:

Invitrogen Amplex Red Xanthine/Xanthine Oxidase Assay Kit (ThermoFisher, #A22182) was used for XO assay analysis following the manufacturer’s protocol. In brief, the 10 U/mL xanthine oxidase stock solution was diluted to make a xanthine oxidase standard curve ranging from 0 to 10 mU/mL. A positive control was prepared by diluting 20 mM of hydrogen peroxide in deionized water. The working solution was concocted by combining a 100 μM Amplex Red reagent, 0.4 U/mL HRP, 200 μM xanthine, and 4.88 mL of 1X reaction buffer. In a flat bottom 96 well plate, 50 μL of the Amplex Red working solution was loaded into each well followed by 50 μL of either sample, positive hydrogen peroxide control, diluted liver homogenate, or xanthine curve. The plate was incubated at 37°C for 30 min. The absorbance was measured at 560 nm, on a BioTek Synergy H4 Plate Reader. When necessary, febuxostat was used at a concentration of 10 μM.

3. Results

The Cancer Dependency Map (DepMap) is a public database that determines which genomic modifications are integral for cancer cell growth [27]. Querying the database with a HepG2 input revealed extremely low to no expression of XDH, Fig. 1B. In addition, the XDH gene copy number varies between individual cells, resulting in highly variable XDH expression levels despite already low transcription (as indicated in Fig. 1C). The AML12 line is non-cancerous and thus it is not included in this database.

To validate the in-silico data in Fig. 1, Xdh mRNA was measured in both AML12- and HepG2- cells. Cells were grown in standard conditions to establish a baseline of Xdh mRNA expression and protein content. To induce amplification of XOR levels, cells were cultured in a hypoxic environment (1% O₂) for 24 h, as we and others have shown [28,29], hypoxia increases XOR activity in cultured cells. Likewise, serum starvation was also employed to potentially modulate XOR levels. HepG2 cells demonstrated little to no detectable XDH mRNA expression regardless of the stimuli, Fig. 2A. While AML12 cells exhibited significantly increased Xdh mRNA expression in response to serum starvation, hypoxic conditions had no discernible effect (Fig. 2C). This difference in the serum-starved group could speak to the differential translation of mRNA. The increase in Xdh mRNA could be due to a survival pathway being activated, as there has been documentation of Xdh translation altering in response to growth conditions [29]. Assessment of XOR protein expression via Western blot revealed no detectable bands in HepG2 cells regardless of the experimental conditions (Fig. 2B). Conversely, AML12 cells consistently expressed immuno-detectable XOR protein across all experimental conditions, with no statistical differences observed (Fig. 2D). An alternative blot for Fig. 2D, run under identical conditions is provided in Suppl. Fig. 1.

Fig. 2. XOR expression remains undetectable in HepG2 cells despite hypoxic and serum-starved conditions.

A) XDH mRNA expression (qPCR) was not altered by serum starvation or hypoxia (1% O2 for 24 h). B) XOR protein expression was not detectable in HepG2 cells regardless of culture conditions. RAW cells were utilized as a positive control. C) Xdh mRNA expression (qPCR) detection was minimal under normal growth conditions and elevated by serum starvation in AML12 cells. D) XOR protein expression in AML12 cells. Data are the mean ± STD, and these studies were performed independently a minimum of three times, n = 3. Data were analyzed by one-way analysis of variance (ANOVA) with Tukey’s multiple comparison test. A multiple comparison test was performed on all groups and significance is indicated by ***, p < 0.005. (R = RAW cells and M = molecular weight standard)

To translate the observed findings of XOR expression and protein levels in HepG2- and AML12- cells into functional relevance, we assessed XOR enzymatic activity using two complementary methods: a sensitive and selective HPLC-electrochemical approach, and a commercially available XOR assay kit. Typical colorimetric XOR activity kits and their detectable products are represented in Table 1. While XOR can produce many products, most colorimetric XOR assay kits detect the presence of H₂O₂. As such, this method is not optimal for assessing XOR activity, as several enzymatic sources of H₂O₂ exist as alternatives to XOR, including aldehyde oxidase (AO), superoxide dismutase (SOD), and glucose oxidase (GO). Additionally, there may be differences in baseline H₂O₂ concentrations. This concept, along with an overview of both the kit assay and the HPLC method, is represented in Fig. 3A. To facilitate comparisons, XOR enzymatic activity was concurrently measured using both the HPLC and a commercially available XO activity assay kit. HepG2- and AML12- cells grown in either hypoxic or normoxic conditions demonstrated XO activity when analyzed by the Invitrogen Amplex Red Xanthine Oxidase Assay Kit. To assess the specificity of the kit for XO-derived H₂O₂, samples were treated with febuxostat, a selective XO inhibitor [30]. In Fig. 3B, treatment with febuxostat resulted in decreased enzymatic activity in both the control mouse liver homogenate and AML12 samples. However, in contrast, HepG2 samples exhibited no response in enzymatic activity in the presence of febuxostat using the kit-based method, suggesting that XO was not the source of the detected signal. This absence of XO activity in HepG2 cells was validated by the HPLC-EC detection method (Fig. 3C). It is also important to note that the kit assay amplifies the XO activity (mU/mg protein) by 1000-fold when compared to the sensitive and selective HPLC assay (μU/mg protein) indicating a gross overestimate of enzymatic activity. Furthermore, the extent of difference in amplitude of the signal between the liver and the cells in the kit versus HPLC-EC method is striking and suggests that the kit does not accurately report liver homogenate XO activity. Additionally, assessment of UA levels in cells and cell culture medium (Fig. 3D and E, respectively) revealed robust UA levels both intracellularly and in the medium of AML12 cells. In contrast, HepG2 cells exhibited low UA levels in the medium and undetectable levels intracellularly. This raises the question: might there be an unknown source of UA in HepG2 cells?

Table 1.

Commonly used XO activity kits.

Kit Name	Catalog #	Detection Method
Amplex™ Red Xanthine/Xanthine Oxidase Assay Kit	A22182	Fluor/Abs for Resorufin
Amplex™ Red Hydrogen Peroxide/Peroxidase Assay Kit	A22188	Fluor/Abs for Resorufin
OxiRed™	Ab142045	Abs for H2O2/OxiRed

Fig. 3. Verification of XO enzymatic activity and UA levels by two independent assay methods.

A) Scheme demonstrating workflow and detection strategies for peroxide-based XO activity assay kits versus HPLC-associated, UA-dependent assessment of enzymatic activity. B) XO activity measured by the peroxide-based kit. C) XOR activity measured by HPLC-EC. D) UA levels in cell lysates via HPLC-EC. E) UA levels in cell culture medium. Data are the mean ± STD, and these studies were performed independently a minimum of three times, n = 3. Data analyzed by student’s t-test due to comparison in each group with or without Febux (B–C) or analyzed by one-way analysis of variance (ANOVA) with Tukey’s multiple comparison test (D–E), *p < 0.05, **p < 0.005. (FBS = Fetal Bovine Serum, Febux = Febuxostat, BLD = below limit of detection) Created in BioRender. King, R. (2025) https://BioRender.com/m92u864.

We next assessed HepG2 cells for aldehyde oxidase (AO) protein expression and capacity for AO to convert hypoxanthine to UA, Fig. 4. It is importantly to note that unlike XO, AO does not assume a dehydrogenase form and thus divalently shuttles all substrate-derived electrons to H₂O₂. Previous observations indicate that AO can oxidize both hypoxanthine and xanthine, albeit at a rate ~1000-fold slower than XOR [31]. Notably, AO protein expression was detected in HepG2 cells, migrating as a ~150 kD monomer, Fig. 4A. As is common with anti-AO antibodies, this antibody also exhibited cross-reactivity with XOR isolated from fresh cow’s milk. Importantly, incubation of the HepG2 lysate for 1 h with 500 μM hypoxanthine generates UA, which is abrogated by Raloxifene, a commonly used inhibitor of AO, indicating the presence of AO activity as well as capacity to oxidize hypoxanthine to UA.

Fig. 4. Aldehyde oxidase (AO) is expressed and enzymatically active in HepG2 cells, catalyzing the oxidation of hypoxanthine to UA.

A) On the left, 30 μg of HepG2 cell lysate protein was electrophoresed and transferred to a PDVF membrane then stained for total protein (Ponceau) and the (right) probed for AO using a polyclonal anti-AO antibody: note AO bands were detected at the 150 kD range where a full monomer (~150 kD) of the homodimeric (~300 kD) AO would be expected to migrate. Control lane of purified XOR protein was added to indicating nonspecific binding for AO antibody. This antibody exhibited cross-reactivity with XOR isolated from fresh cow’s milk, which is likely due to the considerable homology between XOR and AO. B) HepG2 lysates were exposed to hypoxanthine for 1 h at 37°C with or without the AO inhibitor, Raloxifene, and assessed for UA production. These studies were performed a minimum of three independent times. (M = molecular weight standard, BLD = below limit of detection)

Collectively, these findings conclusively demonstrate that HepG2 cells lack Xdh mRNA and protein expression, and consequently, exhibit no XOR enzymatic activity. Moreover, our results suggest that any detectable H₂O₂ signal in the HepG2 cells using the commercially available assay kit could be attributed to alternative sources rather than XO. Furthermore, under standard growth conditions, HepG2 cells produce negligible amounts of UA.

4. Discussion

Recent reports have clearly shown that elevation in UA levels is associated with several metabolic and cardiovascular disease processes including hypertension, diabetes, and metabolic syndrome [1–8]. Emerging evidence suggest that UA may play a causal or contributory role in the pathogenesis of these diseases, underscoring the importance of understanding UA’s mechanistic involvement [4–8]. Given that hepatocytes are primary contributors to circulating UA concentration, investigating UA production mechanisms in hepatocytes is crucial for understanding disease pathology and identifying potential therapeutic targets. Of the hepatocyte cell lines currently available, AML12 (murine, immortalized) and HepG2 (human, carcinoma) have been used for this type of investigation extensively. However, previously reported observations in hepatocellular carcinoma [22] as well as in silico database information (Fig. 1) questions the validity of using HepG2 cells for UA-associated studies as they may have limited to no XOR expression. When combined with our unpublished observations over the past two decades, which indicate undetectable XOR activity in HepG2 cells, and considering that this cell line is widely used to conduct UA-related studies, we believed it was critically important to closely examine XOR expression and UA content. This study aims to inform the scientific community about the limitations of using HepG2 cells to investigate UA metabolism. Rather than referencing specific studies, we encourage readers to explore the existing literature, which includes numerous publications utilizing HepG2 cells to study UA metabolism. This approach allows us to address a critical issue while maintaining a neutral and constructive tone.

Herein, we show the absence of XOR expression, protein, and enzymatic activity of HepG2 cells when compared to the immortalized murine AML12 hepatocyte line, and murine liver homogenates, Figs. 2–3. While no epigenetic factors repressing XOR have been reported, certain factors, such as NC2 and AREB6, have been speculated to negatively regulate XDH expression; however, these mechanisms have not been rigorously validated [32]. As such, we exposed our cells to stimuli that typically induces XOR expression in epithelial and endothelial cells (hypoxia) and these exposures did not alter XOR expression, protein or enzymatic activity in HepG2 cells. Furthermore, serum starvation, a common stressor used to modulate protein expression, also failed to induce XOR expression in HepG2 cells [28,29,33,34]. To ascertain why published studies report XO enzymatic activity in HepG2 cells, we addressed the more commonly used assay employed by these studies. This commercial available assay is a peroxide-dependent colorimetric kit, commonly employed without validation controls using XO inhibitors. Using this commercially available kit assay, a signal suggestive of XO activity was detected in HepG2 cells. However, this signal’s source was shown not to be XO as the presence of the potent XO inhibitor, febuxostat did not alter the color change indication. Collectively, these findings demonstrate that HepG2 cells lack XOR. This conclusion is firmly supported by our data, which show the absence of detectable XOR mRNA, protein, and enzymatic activity in HepG2 cells. Additionally, the identification of a false-positive signal from a commonly used XOR assay kit further reinforces the validy of this conclusion. Of note, we detected UA in the HepG2 cell medium (Fig. 3E), that was absent in the cell lysate and not entirely attributable to the 10% FBS, which contains 0.2 μM UA (data not shown). This observation presents a paradox: despite the absence of detectable XOR, the enzyme currently recognized as the sole source of UA, we still observe UA production. This raises a fundamental question: what alternative source is responsible for UA production in HepG2 cells? An initial consideration might be the role of urate transporters that could potentially shuttle UA into or out of the cell. This may be a contributing factor as HepG2 cells express the ATP-binding cassette family G2 transporter protein (ABCG2, aka, Breast Cancer Resistance Protein, BCRP) which also is an effective UA efflux transporter [35]. Another possibility would be an alternative enzymatic process that could oxidize hypoxanthine, xanthine, or both to generate UA.

To address the issue of an alternative enzymatic source of UA, we explored the possibility that HepG2 cells express AO protein. The rationale for this choice is based on the fact that XOR and AO share abundant homology, characterized by both containing the same co-factors, considerable sequence similarity, similar size (~300 kD), and promiscuous active sites, facilitating crosstalk and overlap in substrate specificity. For example, AO has been reported to have the capacity to oxidize both hypoxanthine and xanthine at its Mo-co whereas XO oxidizes some aldehydes [31]. The key difference between AO and XOR regarding hypo/xanthine oxidation is the rate constants (10⁻³ M⁻¹s⁻¹, for AO vs. 10⁻⁶ M⁻¹s⁻¹ for XOR) clearly demonstrating AO’s oxidizing capacity for these purines to be much slower [31]. Regardless, of the speed of reaction, AO would be a prime candidate for the enzymatic production of UA in HepG2 cells. More importantly, AO turnover would also provide the H₂O₂ necessary to produce the positive results observed with the commercially available kit assay (Fig. 3) since AO, like XO, uses O₂ as a terminal electron acceptor.

Immunodetection of AO in HepG2 cell lysates revealed a positive band at approximately 150 kD, consistent with the expected molecular weight of AO (Fig. 4A). However, this experiment is limited as a positive control with purified AO was not possible, as this protein is not commercially available. Moreover, the polyclonal antibody used in these studies exhibits a considerable degree of nonspecific binding, as indicated by the multiple bands detected on the blot. In addition, the control lanes reveal that the anti-AO antibody cross-reacts with purified XOR, which highlights the similarity between the two enzymes, despite complicating the interpretation of the results. Regardless, assessment of the capacity of these cell lysates to oxidize hypoxanthine to UA was clearly inhibited by a commonly used AO inhibitor, Raloxifene, at concentrations that we have previously shown do not affect XOR activity [36]. When taken together, these data support the hypothesis that UA can be generated in HepG2 cells by AO. However, as we note, further study is warranted to firmly conclude that AO is indeed the principal contributor to UA production by HepG2 cells. For example, while HepG2 cells have been used as a 2-D model for hepatocellular carcinoma as well as many other studies, previous reports suggest that a 3-D model of growth with fibroblast scaffolds may better represent the molecular complexities of liver cancers as well as other endogenous milieus [37]. As such, future studies with HepG2 cells and XOR should recognize this culture scenario. However, we are unaware of reports whereby this 3-D growth format was used to investigate XOR or UA production using HepG2 cells.

In summary, our findings led to a crucial conclusion: HepG2 cells do not express XOR. Notably, we highlight a critical caveat for researchers using commercial colorimetric assay kits to detect XOR activity, as these kits rely on H₂O₂ as an intermediate, often leading to false positives if proper controls are not implemented. To ensure accuracy, rigor, and reproducibility, it is crucial to implement XOR inhibitors as appropriate controls to verify specificity. Furthermore, we hypothesize that UA may be generated in HepG2 cells via AO activity, which does not represent the canonical purine degradation pathway. In toto, we warn that HepG2 cells should not be used for in vitro work designed to test questions related to UA production.

Abbreviations

AO

aldehyde oxidase

XOR

xanthine oxidoreductase

XO

xanthine oxidase

XDH

xanthine dehydrogenase

UA

uric acid

FBS

fetal bovine serum

Appendix A. Supplementary data

Supplementary data to this article can be found online at https://doi.org/10.1016/j.freeradbiomed.2025.02.023.