Investigating the interaction between aryl hydrocarbon receptor and methotrexate: implications for treatment resistance in rheumatoid arthritis patients

Abstract

Background

The aryl hydrocarbon receptor (AHR) plays a key role in immune regulation and drug metabolism, potentially influencing methotrexate (MTX) treatment outcomes in patients with rheumatoid arthritis (RA). This exploratory study investigated the relationship between AHR activity and MTX responsiveness, and examined whether combination therapy with tocilizumab (TCZ), an interleukin (IL)-6 receptor inhibitor, could influence MTX resistance and treatment response.

Methods

We employed in silico docking to assess MTX binding to the AHR Per-Arnt-Sim (PAS)-B domain. Ex vivo and in vitro models using peripheral blood mononuclear cells (PBMCs) from RA patients and healthy donors were also used. Flow cytometry was used to analyze AHR expression across immune cell subtypes. Additionally, HepG2 cells served as a pharmacological model to study the interaction of MTX and TCZ with AHR and the expression of drug transporter genes.

Results

AHR expression was significantly higher in monocytes from good responders to MTX than in those from poor responders and MTX-intolerant patients, suggesting that monocytes were the PBMC subset most strongly associated with AHR-related patterns of MTX response. In silico analysis supported the binding of MTX to the PAS-B domain of AHR. The in vitro model confirmed that monocytes were the most responsive subset in the context of AHR-related changes. Treatment with TCZ tended to reduce the proportion of AHR-positive monocytes, whereas co-treatment with MTX shifted AHR toward a pattern comparable to that in good responders or under control conditions.

Conclusions

Our findings underscore the complexity of MTX pharmacodynamics and highlight AHR as a potential biomarker for predicting treatment response in RA patients. The combination of MTX and TCZ modulated AHR activity and could inform personalized therapeutic strategies, especially in patients exhibiting MTX resistance or intolerance. While preliminary, this multi-layered investigation—combining patient samples, 3D cultures, and molecular docking—supports further research into AHR-modulating therapies in RA.

Clinical trial number

Not applicable.

Supplementary Information

The online version contains supplementary material available at 10.1007/s43440-026-00875-1.

Introduction

Rheumatoid arthritis (RA) is a common health problem and one of the leading causes of disability worldwide [1]. Early, correct diagnosis and treatment with disease-modifying antirheumatic drugs (DMARDs), with methotrexate (MTX) remaining the gold standard in the treatment of RA, are crucial for preventing joint destruction and functional disability [2]. Unfortunately, MTX efficacy is variable, with more than one-third of patients failing to respond to MTX. In addition, 10–30% of patients discontinue therapy owing to adverse effects [3]. Although MTX has been in use for approximately 50 years, the mechanism of its action remains unclear. In general, its anti-inflammatory effect is mediated by adenosine release, which suppresses T-cell activation and cytokine production [2].

MTX is transported into cells by a solute carrier family 19 member 1/reduced folate carrier (SLC19A1/RFC1), after which MTX is polyglutamated by folylpolyglutamate synthase. Peripheral blood mononuclear cells (PBMCs), regardless of the route of administration, accumulate MTX and its polyglutamated form (MTX-PG) [4]. MTX inhibits several enzymes involved in purine and pyrimidine synthesis, including dihydrofolate reductase (DHFR), thymidylate synthase (TYMS), and aminoimidazole-4-carboxamide ribonucleotide (AICAR) transformylase [2]. In clinical practice, patients with a poor response to MTX are often treated with a combination of MTX and an interleukin (IL)-6 receptor inhibitor, tocilizumab (TCZ, RoActemra). MTX is frequently more effective in combination with biologics, but at this point, many mechanisms appear to be unclear, and the results of clinical trials are still inconsistent [5–7]. One of the potential reasons for the lack of response to MTX therapy is the activity of efflux drug transporters (e.g., ATP-binding cassette subfamily B and C members: ABCB1, ABCC1, and ATP-binding cassette subfamily G member 2 (ABCG2)), whose expression is regulated by transcription factors, such as the aryl hydrocarbon receptor (AHR) [8, 9].

AHR is not only an important transcription factor in drug metabolism but also plays a significant role in immune responses [10]. AHR has been linked with the wingless-related integration site (WNT)/β-catenin signaling pathway [11]. This may be important in the context of RA, where dickkopf-1 (DKK1), Wnt family member 5 A (Wnt5a), and β-catenin have been linked to inflammation and disease activity [12]. Regarding MTX transport, as mentioned above, AHR can regulate both influx (e.g., SLC19A1) [13] and efflux transporters (e.g., ABCG2) [8, 14, 15], which are essential in the development of drug resistance. Furthermore, MTX is structurally similar to folic acid, which interacts with AHR [16].

Our previous study revealed that RA patients with a poor response to MTX treatment (resistant or intolerant) had upregulated AHR and SLC19A1 mRNA levels in whole blood. Compared with good responders, poor responders also presented significantly greater ABCG2 mRNA levels. However, most of the patients in this group were treated with TCZ [17]. Therefore, we assume that treatment with this biologic may help overcome low-dose MTX resistance, but more detailed research using stricter patient exclusion criteria and protein-level analyses is needed. TCZ is an innovative biotechnological drug, a humanized monoclonal antibody that inhibits the action of the membrane-bound and soluble IL-6 receptors and consequently reduces the chronic inflammatory process and the symptoms of joint and systemic RA, systemic juvenile idiopathic arthritis (sJIA), and polyarticular juvenile idiopathic arthritis (pJIA) [18]. Notably, TCZ significantly decreases IL-17 and IL-10 but not IL-6, whereas combining TCZ with MTX results in a cytokine profile similar to that of MTX alone [19].

Additional evidence on the effects of TCZ emerged during the COVID-19 pandemic, when it was commonly used to block cytokine storm [20, 21], though IL-6 regulation is distinctive across cell lines and depends on many factors [22]. Current research indicates that IL-6 and AHR can mutually regulate each other’s expression [23–25].

Understanding molecular mechanisms of RA treatment efficacy may support the development of predictive tools. We hypothesized that AHR is associated with MTX response and may influence the treatment outcomes. We also considered whether adding TCZ might modulate AHR in a way that contributes to overcoming MTX resistance—without necessarily implying a direct mechanistic link. This exploratory study assessed AHR levels in immune cells of RA patients treated with MTX, TCZ, or both, and examined AHR-regulated gene expression responses using a 3D in vitro spheroid culture model.

Materials and methods

Patients

PBMCs were collected from 37 patients with RA who were undergoing different treatment regimens, including MTX, TCZ, or a combination of both (TCZ + MTX). Patients were recruited to the exploratory study from the National Institute of Geriatrics, Rheumatology, and Rehabilitation in Warsaw, Poland. Patients treated with MTX were categorized based on their response to therapy into three groups: good responders, poor responders, and those with intolerance. Patients were eligible for inclusion if they were adults (≥ 18 years old) with a confirmed diagnosis of RA established by a rheumatologist according to the 2010 ACR/EULAR classification criteria [26]. Additional inclusion criteria included ongoing treatment with MTX, TCZ, or a combination of both drugs, as well as the availability of complete clinical data. Unresponsive RA patients were defined as those taking MTX doses of ≥ 15 mg/week for at least 3 months but still exhibiting disease activity with a disease activity score 28 (DAS28) > 3.2 or no change in DAS28 since the last visit. The good responders had been on MTX for at least 3 months and had low or moderate disease activity, with a DAS28 score < 3.0. Patients classified as having MTX intolerance experienced adverse effects, including gastrointestinal symptoms (nausea, vomiting, abdominal pain), leukopenia, and elevated liver enzymes. Patients with active infections, malignancy, or other autoimmune or rheumatological diseases, pregnancy, or treatment with additional immunosuppressive therapies not included in the study protocol were excluded. All participants provided written informed consent before enrollment.

For in vitro studies, PBMCs were also obtained from healthy volunteers (HCs) with no clinical or laboratory evidence of autoimmune or inflammatory diseases.

An overview of the study design and experimental workflow is presented in Fig. 1.

Fig. 1.

Study design and experimental workflow. The exploratory study combined patient-derived PBMC samples from RA patients, in vitro spheroid models of PBMCs from healthy donors, and in silico molecular docking to evaluate the interaction between MTX and the Per-Arnt-Sim B (PAS-B) domain of AHR. AHR protein expression in PBMCs from RA patients was analyzed by flow cytometry. PBMC spheroids derived from healthy donors and HepG2 spheroids were treated with MTX, TCZ, or TCZ + MTX. Gene expression was analyzed by ddPCR, and ATPase activity was measured as a marker of transporter activity. AHR protein expression was additionally assessed in PBMC spheroids by flow cytometry. Abbreviations: AHR, aryl hydrocarbon receptor; MTX, methotrexate; PBMCs, peripheral blood mononuclear cells; RA, rheumatoid arthritis; TCZ, tocilizumab; ddPCR, digital droplet polymerase chain reaction

AHR and MTX docking

The cryo-EM structure of the human indirubin-bound AHR complex (PDB code: 7ZUB [27]) was used for docking, which was performed with Flare™ software (Cresset, UK). All water molecules and co-crystallized ligands were removed from the PDB structure. The protein was prepared by adding hydrogen atoms, optimizing hydrogen bonds, removing atomic clashes, and assigning optimal protonation states. The docking accuracy was verified by docking the co-crystallized ligand, indirubin, to the protein binding site. The structure of methotrexate was taken from the PubChem database and then docked via a rigid receptor-flexible ligand docking approach.

PBMC isolation

PBMCs were isolated using BD^® Vacutainer CPT™ Mononuclear Cell Preparation Tubes with sodium citrate (cat. no. 362760, Becton Dickinson) following the manufacturer’s instructions, including two washing steps with HBSS (cat. no. ECB4007L, Euroclone). The cell pellet was re-suspended in Cryo-SFM medium (cat. no. C-29912, PromoCell) and stored in liquid nitrogen for further processing.

AHR in PBMCs - flow cytometry

For cytometric analyses, we used PBMCs from RA patients treated with MTX, TCZ, or TCZ + MTX, as well as cells from healthy individuals cultured in vitro as spheroids and treated with MTX, TCZ, or TCZ + MTX for 24 and 48 h. The PBMCs were stained for surface markers in 50 µL of Cell Wash (cat. no. 349524, Becton Dickinson) with 0.2% heat-inactivated fetal bovine serum (FBS, cat. no. F4135, Sigma‒Aldrich) for 30 min on ice and protected from light, using murine anti-human monoclonal antibodies against the following surface markers: 1 µL CD19-BV650; 2 µL each of CD14-APC-Cy7, CD3-FITC, CD4-BV421, CD8-APC, and CD45-V500; and 10 µL Brilliant Stain Buffer. For intracellular staining, the cells were first stained with Fixable Viability Dye 700 (cat. no. 564997, Becton Dickinson). The cells were incubated with 1x TF Fix/Perm Buffer (cat. no. 51-9008101, Becton Dickinson) for 40 min on ice, protected from light. After incubation, the cells were washed with 1x TF Perm/Wash Buffer and then stained with PE-CF594-conjugated AHR (cat. no. 565790, Becton Dickinson) and the appropriate isotype control for 40 min on ice protected from light. After the washing step, the cells were acquired on a FACS Celesta cytometer (BD Biosciences) and analyzed with FACS Diva (BD Biosciences) software. Compensation was performed with BD CompBeads Anti-Mouse Ig, k (cat. no. 51-90-9001229, Becton Dickinson). A detailed list of the reagents and the gating strategy is shown in Table S1 and Figure S1 in the Supplementary Materials.

Cell culture – PBMC and HepG2 spheroids

PBMCs, isolated as described above using CPT tubes, were cultured in U-shaped 96-well plates (cat. no. 781900, Brand) to form spheroids in RPMI 1640 medium (Invitrogen) supplemented with 10% heat-inactivated FBS, penicillin, and streptomycin (Sigma‒Aldrich). A detailed list of reagents is shown in Table S1.

HepG2 cells (a hepatocellular carcinoma human cell line, cat. no. HB-8065, ATCC) were also cultured as spheroids in U-shaped 96-well plates with DMEM and 10% FBS.

After 24 h, the cells were treated with methotrexate (50 nM, cat. no. M8407, Merck Life Science), TCZ (an anti-IL-6 receptor, 40 ng/µL, Roactemra, Roche), and, in the case of gene expression analysis and ATPase activity, with the cytokine IL-6 (50 ng/mL, cat. no. C-61625, PromoCell). The cells were further cultured at 37 °C for 24 and/or 48 h.

Viability and cytotoxicity

For viability assessment, the cells were harvested and stained for the appropriate membrane antigens with anti-CD3-FITC (cat. no. 555332, Becton Dickinson), anti-CD8-FITC (cat. no. 555366, Becton Dickinson), anti-CD4-APC-Cy-7 (cat. no. 341115, Becton Dickinson), anti-CD19-PE (cat. no. 340–364, Becton Dickinson), and anti-CD14-BV421 (cat. no. 567939, Becton Dickinson) antibodies. After the washing step, the cells were acquired and analyzed using a FACSAria cell sorter/cytometer and Diva software. Dead cells were excluded from analysis by 7-AAD staining (cat. no. 51-68981E, Becton Dickinson).

Cytotoxicity was measured using a colorimetric WST-1 kit (cat. no. 11644807001, Roche). The assay was carried out as described in the manufacturer’s instructions. PBMC spheroids were cultured as described above. The absorbance was measured at 420 nm (620 nm was used as the reference wavelengthand was subtracted) after 3 h in a multiplate reader (Infinite F200, Tecan^®). The results are reported as relative WST-1 activity, where 1.0 corresponds to the absorbance measured in control cultures.

IL-6 ELISA

The basal secretion of IL-6 by PBMCs was measured using an ELISA kit (cat. no. ab178013; Abcam) according to the manufacturer’s instructions. The ELISA standard range was 7.8–500 pg/mL. The sensitivity of the test was 1.6 pg/mL. The absorbance was measured at 450 nm using a multiplate reader (Infinite F200, Tecan^®).

ATPase activity

ATPase activity was measured using the ATPase/GTPase Activity Assay Kit (MAK113; Sigma‒Aldrich; Merck KGaA). PBMCs and HepG2 spheroids were cultured as described in the Cell Culture section. After 48 h of incubation with drugs or IL-6, the cells were lysed via a freeze‒thaw method involving freezing at -80 °C and subsequent thawing at room temperature. Three cycles of freezing and thawing were conducted, with each step lasting 20 min. Subsequently, the cell lysate was centrifuged at 10,000 × g for 10 min. The resulting supernatant was decanted, and the cell pellet was resuspended in 3.2 M ammonium sulfate (cat. no. 45216, Thermo Fisher Scientific). The cells were incubated for 20 min on ice and then centrifuged at 10,000 × g for 10 min at 4 °C. The supernatant was decanted, and the cells were resuspended in assay buffer from the ATPase/GTPase Activity Assay Kit. Phosphate standards and sample mixtures were prepared according to the kit guidelines. The reaction mixtures were incubated for 30 min at room temperature. A reagent (200 µl) was subsequently added to each well, and the mixture was incubated for 30 min at room temperature. The absorbance was read at 620 nm using a multiplate reader (Infinite F200, Tecan^®). The concentration [µM] of free phosphate (Pi) in the samples was determined from a standard curve. The formula used for computing enzyme activity [U/L] was as follows:

One unit denotes the quantity of enzyme catalyzing the production of 1 µmol of free phosphate per minute under the assay conditions.

Gene expression

RNA from PBMCs of healthy subjects and HepG2 spheroid cultures was isolated using the RNeasy Micro RNA kit (cat. no. 74004, Qiagen). The quantity and quality of isolated RNA were evaluated on a Denovix spectrophotometer. The reverse transcription reaction was performed using the High-Capacity cDNA Reverse Transcription Kit (cat. no. 4374967, Applied Biosystems) according to the manufacturer’s instructions. ddPCR Supermix for Probes no dUTP (cat. no. 186–3024, Bio-Rad) and TaqMan probes (Thermo Fisher) were used for gene expression analysis (ID assays are listed in Table S1): AHR, ARNT, AHRR, SLC19A1, ABCG2, MTRR, DHFR, TYMS, ABCC1, ABCC2, ABCC3, ABCC4, and ABCC5. Each reaction mixture was then loaded into the sample well of an eight-well disposable cartridge (cat. no. 1864007, Bio-Rad) along with 70 µL of droplet generation oil (cat. no. 1864006, Bio-Rad), and droplets were formed in 40 µL with the Droplet Generator (Bio-Rad). The droplets were then transferred to a 96-well PCR plate (cat. no. 12001925, Bio-Rad), heat-sealed with foil, and amplified in a C1000 Touch Thermal Cycler (Bio-Rad). The ddPCR results expressed as copy/µL were transformed into copy/ng.

Statistical analysis

The normality of data distribution was assessed using the Shapiro–Wilk test.

Demographic and clinical characteristics of patients assigned to the respective study arms were compared using the Kruskal-Wallis test or a one-way ANOVA, followed by the Dunn’s test or the Games-Howell test, respectively. Post hoc p-values were adjusted using the Holm-Bonferroni method. Continuous variables are presented as means and standard deviations or medians and IQRs, depending on the distribution assessed by the Shapiro–Wilk test. Categorical variables are presented as numbers (percentages) and were compared using the chi-square (χ²) test, with Monte Carlo simulation (2000 replicates) when appropriate. Correlations between continuous variables were evaluated using Pearson’s or Spearman’s correlation coefficients, depending on data distribution. All tests were two-tailed. In the case of missing data, analyses were performed on available data (complete-case analysis); no imputations were performed. Flow cytometry data from RA patients were analyzed using a one-way ANOVA with Tukey’s multiple comparison test, except for CD8⁺ T cells, for which the Kruskal–Wallis test was applied.

Flow cytometry data from PBMC spheroids are expressed as median fluorescence intensity (MFI) and the percentage of AHR-positive cells. MFI values were analyzed using a mixed-effects model followed by Tukey’s post hoc test for monocytes and B cells, while the Kruskal–Wallis test was applied to CD4 + and CD8 + T cells. The percentage of AHR-positive cells was analyzed using a mixed-effects model followed by Tukey’s post hoc test for B cells, and the Kruskal-Wallis test for monocytes, CD4+, and CD8 + T cells.

ATPase activity in HepG2 cells was analyzed using one-way ANOVA with Tukey’s post hoc test, whereas PBMC samples were analyzed using the Kruskal–Wallis test.

Gene expression data were analyzed using the Kruskal–Wallis test. Differences in gene expression between PBMC and HepG2 spheroids were evaluated using the Mann–Whitney U test. Correlations between gene expression levels were assessed using Spearman’s rank correlation.

Differences were considered statistically significant at p < 0.05.Analyses were conducted using R software (version 4.4.1; R Foundation for Statistical Computing, Vienna, Austria). Data processing and visualization were conducted using tidyverse (version 2.0.0), ggplot2 (version 4.0.2), ggpubr (version 0.6.2), rstatix (version 0.7.3), and related packages. Figure preparation and statistical analyses of flow cytometry results and ATPase activity experiments were performed using GraphPad Prism software (version 10.4.2).

Results

MTX binds to AHR in the PAS-B domain

To explore the potential interaction between MTX and AHR suggested by our previous findings, we performed molecular docking simulations using an in silico model. AHR comprises three principal domains: a highly conserved N-terminal bHLH domain; a pair of degenerate Per-Arnt-Sim (PAS) repeats (designated PAS A and PAS B); and a poorly conserved C-terminal transactivation domain. PAS A is essential for heterodimerization with the aryl hydrocarbon nuclear translocator (ARNT), while the PAS B domain is responsible for ligand binding. Docking analysis showed that MTX binds specifically within the PAS-B domain of AHR (Fig. 2). MTX interacts with amino acids in the binding site, such as Gln383, Leu353, His291, Phe295, Ile325, Ser346, Phe324, Pro297, and Leu308.

Fig. 2.

The complex formed by MTX and AHR. A) MTX in the ligand binding site and the interacting amino acid residues are shown. B) Two-dimensional binding model of MTX and the protein complex, showing important interactions such as hydrogen bonds between MTX and AHR residues, as well as hydrophobic interactions. Abbreviations: AHR, aryl hydrocarbon receptor; MTX, methotrexate

Hydrogen bonds were detected between methotrexate and Ser346, as well as between MTX and Gln383. Hydrophobic interactions were observed between MTX and AHR residues such as Ile325, Phe324, Leu353, Leu380, Pro297, and Phe295. Additionally, aromatic interactions were observed between the aromatic moiety of methotrexate and the aromatic rings of Phe351 and Phe295.

Higher expression of AHR in RA monocytes from good responders to MTX

Considering the significant role of AHR in drug transport and metabolism, we investigated its associations with immune cell subtypes in RA patients undergoing different treatments and showing varying responses to MTX. We assessed AHR expression in PBMCs using flow cytometry. Patients received MTX (median dose 25 mg/week), TCZ (8 mg/kg body weight every 4 weeks), or combined TCZ + MTX therapy (median MTX dose 15 mg/week), all supplemented with folic acid. Additionally, those on MTX were categorized as good responders, poor responders, or intolerant. The clinical characteristics of all patients are presented in Table 1and Table S2. Statistical analysis showed no significant differences in age, sex, or disease duration among groups. The duration of therapy in the studied patients spanned many years, with a median disease duration of 10 years.

Table 1.

Demographic and clinical characteristics of the cohort of patients^# with rheumatoid arthritis (RA), stratified by treatment regimen and response to MTX

	All patients (N = 37)	MTX-good response (N = 8)	MTX-poor response (N = 8)	MTX-intolerance (N = 8)	TCZ (N = 8)	TCZ + MTX (N = 5)	Test statistics	p-value
Age,mean (SE)	59.86 (1.82)	66.62 (2.51)	58.88 (5.64)	58.50 (4.77)	59.25(2.10)	53.80 (2.99)	F4, 32 = 1.2	p = 0.331
Sex
Female, n (%)	33 (89.19%)	6 (75.00%)	7 (87.50%)	7 (87.50%)	8 (100.00%)	5 (100.00%)	χ² = 3.3	0.7882
Male, n (%)	4 (10.81%)	2 (25.00%)	1 (12.50%)	1 (12.50%)	0 (0.00%)	0 (0.00%)
Disease duration, median (IQR)	36; 10.00 (12.00)	4.00 (4.00)	4.00 (4.00)	7; 14.00 (9.50)	13.00 (4.50)	4.00 (7.00)	H4 = 6.5, n = 36	0.1643
Larsen scale, n/N* (%)
0-I	7/29 (24.14%)	1/2 (50.00%)	2/7 (28.57%)	1/7 (14.29%)	1 (12.50%)	2 (40.00%)	χ² = 8.2	0.4722
II-III	12/29 (41.38%)	1/2 (50.00%)	2/7 (28.57%)	2/7 (28.57%)	6 (75.00%)	1 (20.00%)
IV-V	10/29 (34.48%)	0/2 (0.00%)	3/7 (42.86%)	4/7 (57.14%)	1 (12.50%)	2 (40.00%)
RF, n (%)	23 (62.16%)	3 (37.50%)	5 (62.50%)	6 (75.00%)	5 (62.50%)	4 (80.00%)	χ² = 3.3	0.5152
Anti-CCP	25 (67.57%)	4 (50.00%)	6 (75.00%)	5 (62.50%)	6 (75.00%)	4 (80.00%)	χ² = 1.98	0.8052
DAS28, median (IQR)	3.03 (3.46)	2.30 (0.88)	5.43 (1.25)	5.85 (1.25)	2.04 (0.64)	1.58 (0.11)	H4 = 26.9, n = 37	< 0.001 3
CRP (mg/L), N*; median (IQR)	36; 2.00 (8.50)	7; 3.00 (8.00)	3.00 (13.25)	11.00 (16.00)	1.00 (0.06)	0.91 (0.50)	H4 = 22.1, n = 36	< 0.001 3
ESR (mm/h), median (IQR)	12.00 (19.00)	8.50 (4.50)	14.00 (6.75)	28.50 (22.00)	4.50 (3.50)	5.00 (4.00)	H4 = 19.8, n = 37	< 0.001 3
VAS (mm), median (IQR)	30.00 (51.00)	25.00 (17.50)	82.00 (23.25)	75.50 (19.00)	19.00 (14.25)	24.00 (1.00)	H4 = 22.5, n = 37	< 0.001 3
WBC (103/µL), mean (SD)	6.51 (0.46)	6.94 (0.80)	8.70 (0.81)	7.88 (1.05)	3.71 (0.37)	4.61 (0.29)	F4, 32 = 7.6	< 0.001 1
Neutrophiles (103/µL), N*; mean (SD)	36; 3.98 (0.39)	4.41 (0.70)	6.12 (0.67)	4.68 (0.78)	1.69 (0.17)	4; 2.06 (0.22)	F4, 31 = 8.5	< 0.001 1
Lymphocytes (103/µL), N*; median (IQR)	36; 1.69 (0.84)	1.61 (1.03)	1.66 (0.30)	2.24 (0.81)	1.68 (0.21)	4; 1.66 (1.16)	H4 = 3.2, n = 36	0.5293

^#Patients were recruited for this exploratory study from the National Institute of Geriatrics, Rheumatology, and Rehabilitation in Warsaw, Poland

The patients were assigned to the following study arms: MTX-good response, MTX-poor response, MTX-intolerance, tocilizumab (TCZ), and TCZ + MTX

Continuous variables are presented as mean (SD) or median [IQR], depending on data distribution. Categorical variables are presented as a number (percentage)

Statistical analysis was performed using one-way ANOVA¹, chi-square (χ²) test², or Kruskal–Wallis test³, as appropriate. All tests were two-tailed. Analyses were performed on available cases, missing values were not imputed

N* – number of patients with available clinical data; n – number of patients with a given characteristic

Abbreviations: Anti-CCP, anti-cyclic citrullinated peptide; CRP, C-reactive protein; DAS28, Disease Activity Score 28; ESR, erythrocyte sedimentation rate; RF, rheumatoid factor; VAS, visual analogue scale; WBC, white blood cells

Patients with poor response and intolerance to MTX had significantly higher disease activity than the other groups (Table 1, Figure S2A).

Post-hoc analysis using Dunn’s test with Holm correction revealed the strongest DAS28 score differences between the MTX-intolerance and TCZ + MTX groups, as well as between the MTX-intolerance and TCZ groups.

Similarly, analysis of visual analogue scale (VAS) pain scores demonstrated significant differences between MTX poor-response and MTX good-response patients (Figure S2F), as well as between MTX intolerance and MTX good-response patients. Significant differences were also observed between MTX poor-response and TCZ + MTX groups (Figure S2F), and between MTX intolerance and TCZ + MTX groups. The strongest differences were found between MTX poor-response and TCZ groups and between MTX intolerance and TCZ groups.

Moreover, patients with intolerance also had significantly increased mean values of C-reactive protein (CRP) and erythrocyte sedimentation rate (ESR) (Figure S2B, Figure S2C). For CRP levels, post-hoc analysis revealed significant differences between the MTX intolerance group and both the TCZ + MTX groups and TCZ groups (Figure S2B). Similarly, for ESR, significant differences were observed only between the MTX intolerance group and both the TCZ + MTX and TCZ groups (Figure S2C). No significant differences were found among the treatment-effective groups (MTX-good response, TCZ + MTX, and TCZ).

MTX-intolerant and resistant patients were mostly treated with glucocorticoids (GCS) (Table 2). Half of the resistant RA patients also received sulfasalazine. One MTX-resistant patient was treated with rituximab, a monoclonal antibody targeting CD20 on B lymphocytes; however, including this patient did not affect the overall statistical inference. In the whole cohort, 18 of 37 patients (49%) received GCS, 17 (46%) did not, and for 2 patients (5%) this information was unavailable.

Table 2.

Concomitant treatment characteristics of the cohort of patients^# with rheumatoid arthritis (RA), stratified by treatment regimen and response to methotrexate (MTX)

Treatment	MTX-good response (N = 8)	MTX-poor response (N = 8)	MTX-intolerance (N = 8)	TCZ (N = 8)	TCZ + MTX (N = 5)
GCS, n/N* (%)	3/8 (37.50%)	6/8 (75%)	6/8 (75%)	2/6 (33.33%)	1/5 (20%)
hydroxychloroquine (plaquenil), n/N* (%)	0/8 (0.00%)	2/8 (25%)	0/8 (0%)	2/4 (50%)	0/5 (0%)
sulfasalazine, n/N* (%)	1/8 (12.50%)	4/8 (50%)	2/8 (25%)	0/6 (0%)	0 (0%)
biologics n/N* (%)	0/8 (0.00%)	1/8 (12.50%)	0/8 (0%)	6/6 (100%)	5 (100%)

^#Patients were recruited for this exploratory study from the National Institute of Geriatrics, Rheumatology, and Rehabilitation in Warsaw, Poland

Data are presented as number (percentage)

Analyses were performed on available cases, missing values were not imputed

N* – number of patients with available clinical data; n – number of patients with a given characteristic

Abbreviations: GCS, glucocorticoids; MTX, methotrexate; TCZ, tocilizumab

Analysis of AHR expression in PBMCs revealed a high percentage of AHR-positive monocytes and B lymphocytes. The highest median AHR fluorescence was observed in monocytes.

A one-way ANOVA revealed a significant treatment effect on AHR expression in monocytes (F_{4, 32} = 7.38, p = 0.0002) (Fig. 3A). Post hoc comparisons showed lower AHR expression in the MTX-poor response, MTX-intolerance, and TCZ groups than in MTX good responders.

Fig. 3.

AHR protein expression in peripheral blood mononuclear cell subsets from patients^# with rheumatoid arthritis under different treatment and response conditions. The patients were assigned to the following study arms: MTX good response, n = 8, MTX poor response n = 8; MTX intolerance, n = 8; TCZ (8 mg/kg body weight every 4 weeks) n = 8; and combined TCZ + MTX therapy (TCZ 8 mg/kg body weight every 4 weeks plus MTX at a median dose of 15 mg/week) n = 5). For MTX-treated patients, the median MTX dose was 25 mg/week. AHR protein expression was assessed by flow cytometry and presented as median fluorescence intensity (MFI) in monocytes (A), B cells (B), CD8 + T cells (C), and CD4 + T cells (D) in peripheral blood mononuclear cells isolated from patients with rheumatoid arthritis. The results are presented as mean ± SEM (panels A, B, and D) or medians with IQR and analyzed using a one-way ANOVA followed by Tukey’s multiple comparison test or the Kruskal–Wallis test, respectively. * p < 0.05; ** p < 0.01; *** p < 0.001. #Patients were recruited for this exploratory study from the National Institute of Geriatrics, Rheumatology, and Rehabilitation in Warsaw, Poland. Abbreviations: AHR, aryl hydrocarbon receptor; IQR, interquartile range; MFI, median fluorescence intensity; MTX, methotrexate; SEM, standard error of the mean; TCZ, tocilizumab

No significant treatment effect was observed for AHR expression in B lymphocytes (one-way ANOVA, F_{4, 30} = 1.9, p = 0.14) (Fig. 3B).

A one-way ANOVA revealed a significant treatment effect on AHR expression in CD4 + T cells (F_{4, 31} = 4.2, p = 0.0075) (Fig. 3D). Post hoc comparisons showed higher AHR expression in the MTX-intolerance group than in MTX good responders.

According to the Kruskal–Wallis test, no significant treatment effect on AHR expression in CD8 + T cells was observed (H = 8.2, N1 = 8, N2 = 8, N3 = 8, N4 = 8, and N5 = 5, p = 0.08) (Fig. 3C).

MTX combined with tocilizumab has a distinct effect on AHR expression in both B cells and monocytes – in vitro model

To complement our patient-derived data and model AHR regulation under more controlled conditions, we next cultured PBMCs from healthy donors as 3D spheroids, which better reflect physiological environments.

AHR protein expression in monocytes, B cells, CD4 + T cells, and CD8 + T cells was assessed via flow cytometry, analyzing both median fluorescence intensity (MFI) (Fig. 4A-D) and the percentage of AHR-positive cells (Fig. 4E-H). Cytotoxicity assays showed reduced PBMC viability, particularly in monocytes treated with MTX and TCZ + MTX (Figure S3).

Fig. 4.

AHR protein expression in PBMC spheroids from healthy donors under control and drug treatment conditions. PBMC spheroids derived from healthy donors were cultured for 48 h in an exploratory study and assigned to the following study arms: CTRL, n = 6; MTX (50 nM), n = 6; TCZ (40 ng/µL), n = 6; and combined TCZ + MTX treatment (TCZ 40 ng/µL plus MTX 50 nM), n = 3. AHR protein expression was assessed by flow cytometry as median fluorescence intensity (MFI) in monocytes (A), B cells (B), CD4 + T cells (C), and CD8 + T cells (D), and as the percentage of AHR-positive cells in monocytes (E), B cells (F), CD4 + T cells (G), and CD8 + T cells (H). The results are presented as mean ± SEM (panels A, B, and F) or medians with IQR (panels C, D, E, G, and H) and analyzed using mixed-effects analysis followed by Tukey’s post hoc test or the Kruskal–Wallis test, respectively. No significant treatment effects were observed. Abbreviations: AHR, aryl hydrocarbon receptor; CTRL, control; IQR, interquartile range; MFI, median fluorescence intensity; MTX, methotrexate; PBMC, peripheral blood mononuclear cell; SEM, standard error of the mean; TCZ, tocilizumab

At baseline, AHR protein expression assessed as MFI was highest in monocytes (range: 569–1484), compared with B cells (76–100), CD8 + T cells (114–197), and CD4 + T cells (154–292).

Although AHR MFI in monocytes tended to be lower under TCZ treatment, according to the mixed-effects analysis, no significant treatment effect on AHR MFI in these cells was observed (F_{2.1, 8.4} = 3.7, p = 0.07) (Fig. 4A). According to the mixed-effects analysis, no significant differences in the treatment effect on AHR median fluorescence intensity (MFI) in B cells were observed either (F_{2.8, 11.1} = 1.7, p = 0.23) (Fig. 4B). Likewise, no significant treatment effect on AHR MFI was observed in CD4 + T cells (Kruskal–Wallis test, H = 2.2, total n = 20, p = 0.53) (Fig. 4C) and CD8 + T cells (Kruskal–Wallis test, H = 0.6, total n = 20, p = 0.90) (Fig. 4D).

No significant treatment effect was observed in AHR-positive monocytes (Kruskal–Wallis test, H = 2.1, total n = 21, p = 0.55) (Fig. 4E), B cells (mixed-effects analysis, F_{2.1, 8.2} = 4.3, p = 0.05) (Fig. 4F), CD4 + T cells (Kruskal–Wallis test, H = 2.4, total n = 27, p = 0.50) (Fig. 4G), or CD8 + T cells (Kruskal–Wallis test, H = 4.5, total n = 27, p = 0.21) (Fig. 4H). However, the percentage of AHR-positive B cells tended to decrease following MTX treatment, and a similar pattern was observed in B cells (Fig. 4F) and monocytes (Fig. 4E) under TCZ treatment.

ATPase activity different response to MTX, increased activity after MTX combined with TCZ

AHR regulates drug transporters. ABC transporters use energy derived from ATP hydrolysis, which is proportional to transporter activity. To assess this, we measured basal ATPase activity in PBMCs and, due to their heterogeneity, also in HepG2 cells, a widely used model for drug metabolism studies. In this experimental setup, PBMCs and HepG2 cells were treated for 48 h with MTX, TCZ, or both, and exogenous IL-6 was added to simulate inflammatory conditions.

As expected, ATPase activity was lower in PBMCs (range: 2.08–9.23 U/I) compared to HepG2 cells (range: 7.18–23.86 U/I), likely reflecting baseline differences in transporter expression (Fig. 5). A one way ANOVA revealed a significantly reduced ATPase activity in HepG2 spheroids (F_{4, 19} = 6.6, p = 0.002) (Fig. 5A). Post hoc comparisons showed significantly lower ATPase activity in the MTX group than in the control and IL-6-stimulated groups. Moreover, co-treatment with TCZ + MTX significantly increased activity relative to MTX alone.

Fig. 5.

ATPase activity in HepG2 and PBMC spheroids under control, inflammatory, and drug treatment conditions. ATPase activity was measured in HepG2 spheroids (A) and PBMC spheroids derived from healthy donors (B), cultured for 48 h in an exploratory study and assigned to the following study arms: CTRL, n = 3 (A) and n = 15; IL-6 (50 ng/mL), n = 3 (A) and n = 9 (B); MTX (50 nM), n = 6 (A) and n = 15 (B); TCZ (40 ng/µL), n = 6 (A) and n = 15 (B); and combined TCZ + MTX treatment (TCZ 40 ng/µL plus MTX 50 nM), n = 6 (A) and n = 15 (B). The results are presented as mean ± SEM (panel A) or medians with IQR (panel B) and analyzed using one-way ANOVA followed by Tukey’s multiple comparison test or the Kruskal–Wallis test followed by Dunn’s post hoc test, respectively. * p < 0.05; ** p < 0.01; *** p < 0.001. Abbreviations: CTRL, control; IL-6, interleukin-6; IQR, interquartile range; MTX, methotrexate; PBMC, peripheral blood mononuclear cell; SEM, standard error of the mean; TCZ, tocilizumab

A similar upward trend was observed in PBMC spheroids (Fig. 5B). In PBMC spheroids, a Kruskal–Wallis test also revealed a significant overall treatment effect (H = 10.6, N1 = 15, N2 = 6, N3 = 15, N4 = 15, N5 = 15, p = 0.03) (Fig. 5B), although Dunn’s post-hoc test did not show significant differences between individual groups.

Gene expression changes in PBMC spheroids under inflammatory and drug treatment conditions

In a preliminary experiment, we checked basal secretion of IL-6 by PBMC spheroids. The median basal IL-6 concentration in PBMC supernatants was 10.23 pg/mL at 24 h and 14.52 pg/mL at 48 h Moreover, in PBMC spheroids treated with MTX and 40 ng/µl TCZ, IL-6 levels were below the detection limit (Figure S4). According to the Mann-Whitney U test, no significant differences in basal IL-6 concentration in PBMC supernatants between 24 h and 48 h of incubation were observed (U = 17, N1 = 7 and N2 = 7, p = 0.37). The median IL-6 concentration was 10.23 pg/mL at 24 h and 14.52 pg/mL at 48 h.

Among the analyzed genes, AHR and ABCC1 showed the highest baseline expression in PBMC spheroids. Exogenous IL-6 was associated with the high AHR expression, whereas MTX showed the lowest values at both time points (24 h: IL-6 median 12.44 pg/mL, IQR 24.05 vs. MTX median 7.09 pg/mL, IQR 8.48; 48 h: IL-6 median 14.36 pg/mL, IQR 25.53 vs. MTX median 6.15 pg/mL, IQR 8.98); however, these differences did not reach statistical significance (Kruskal–Wallis test: 24 h, χ² = 4.2, df = 4, p = 0.377; 48 h, χ² = 4.4, df = 4, p = 0.353; Figure S5A–B). TCZ alone was associated with increased AHR expression, particularly at 48 h (median 23.95 copy/ng, IQR 22.29), whereas its combination with MTX corresponded to lower AHR levels (24 h: median 6.42 copy/ng, IQR 10.35; 48 h: median 7.51 copy/ng, IQR 7.12). This observation may reflect the influence of MTX on AHR transcriptional regulation, although further validation is needed; results did not reach statistical significance. As with MTX alone, the combination of TCZ with MTX also appeared to downregulate AHR regulatory genes (the translocator – ARNT and the repressor - AHRR) and selected efflux transporters, although none of the changes reached statistical significance either time point (ARNT: 24 h, χ² = 2.65, df = 4, p = 0.618; 48 h, χ² = 5.58, df = 4, p = 0.233; AHRR: 24 h, χ² = 2.03, df = 4, p = 0.731; 48 h, χ² = 4.43, df = 4, p = 0.351). TCZ alone induced a trend toward higher expression of the analyzed drug transporters, particularly after extended incubation (48 h, median 23.95 copy/ng, IQR 22.29) (Figure S6).

The proinflammatory environment (IL-6 treatment) tended to induce the expression of drug transporters, AHRR (24 h: median 2.30 copy/ng, IQR 1.39; 48 h: median 3.54 copy/ng, IQR 1.10) and ARNT (24 h: median 4.99 copy/ng, IQR 6.04; 48 h: median 5.64 copy/ng, IQR 2.98).

Genes encoding drug transporters responded differently to MTX; for SLC19A1 and ABCG2, we observed a non-significant trend toward downregulation (Figure S6).

TCZ + MTX treatment showed a trend toward downregulation of MTX metabolism-related genes (MTRR, DHFR, and TYMS), although differences were not statistically significant (Kruskal–Wallis test; all p > 0.05; Figure S5C-D).

Gene expression in HepG2 cells

Following the reasoning provided earlier, we also assessed the expression of selected genes in HepG2 spheroids. We focused on previously implicated genes (AHR, SLC19A1, ABCG2), and also examined STAT3A as a shared target for TCZ and MTX. HepG2 spheroids were treated for 48 h, but the differences in analyzed gene expression were not statistically significant (Kruskal–Wallis test; all p > 0.05; Figure S7). IL-6 induced the highest AHR expression, followed by MTX exposure. Both IL-6 and MTX showed a downregulated trend SLC19A1 and ABCG2 compared to control (Kruskal–Wallis test; all p > 0.05). MTX slightly upregulated AHR (median 5.29 copy/ng, IQR 6.91), whereas TCZ reduced it (median 1.76 copy/ng, IQR 1.45).

Comparative expression of AHR, SLC19A1, and ABCG2 in PBMCs and HepG2 cells

Under control conditions, AHR expression was high and did not significantly differ between PBMCs and HepG2 cells (Wilcoxon test, W = 87, p = 0.20, Fig. 6A). AHR expression was significantly upregulated in PBMCs compared with HepG2 cells (Wilcoxon test, W = 35, p = 0.026) in TCZ conditions. In HepG2 cells, the expression of SLC19A1 and ABCG2 under all treatment conditions was significantly higher than that in PBMCs (Wilcoxon test, all p < 0.05, Fig. 6B-C).

Fig. 6.

Expression of selected AHR-related and transporter genes in PBMC and HepG2 spheroids under control, inflammatory, and drug treatment conditions. The expression of AHR (A), SLC19A1 (B), and ABCG2 (C) was assessed in PBMC and HepG2 spheroids. PBMC spheroids derived from healthy donors and in HepG2 spheroids were cultured for 48 h of in an exploratory study and assigned to the following study arms: CTRL, n = 15 (HepG2) and n = 16 (PBMC, panels A and B) or n = 9 (PBMC, panel C); IL-6 (50 ng/mL), n = 15 (HepG2) and n = 14 (PBMC, panels A and B) or n = 9 (PBMC, panel C); MTX (50 nM), n = 15 (HepG2) and n = 12 (PBMC, panels A and B) or n = 9 (PBMC, panel C); TCZ (40 ng/µL), n = 15 (HepG2) and n = 10 (PBMC, panels A and B) or n = 9 (PBMC, panel C); and combined TCZ + MTX treatment (TCZ 40 ng/µL plus MTX 50 nM), n = 15 (HepG2) and n = 10 (PBMC, panels A and B) or n = 9 (PBMC, panel C). The results are presented as box-and-whisker plots, where the central line indicates the median, the box represents the interquartile range (IQR), whiskers indicate the range of non-outlier values, and dots represent outliers, and analyzed using the Mann–Whitney U test. * p < 0.05; ** p < 0.01; *** p < 0.001; **** p < 0.0001 Abbreviations: ABCG2, ATP-binding cassette subfamily G member 2; AHR, aryl hydrocarbon receptor; CTRL, control; IQR, interquartile range; IL-6, interleukin-6; MTX, methotrexate; PBMC, peripheral blood mononuclear cell; SLC19A1, solute carrier family 19 member 1; TCZ, tocilizumab

Differential gene regulation under combined MTX and TCZ treatment

We next investigated whether AHR expression correlated with drug transporters and how these relationships varied with cell type and treatment. AHR expression was correlated with ABCG2 (Spearman’s rank correlation test, CTRL: R = 0.491, p = 0.056; MTX: R = 0.846, p < 0.001; IL: R = 0.701, p = 0.007; TCZ: R = 0.875, p < 0.001; TCZ + MTX: R = 0.683, p = 0.05) and SLC19A1 (Spearman’s rank correlation test, CTRL: R = 0.774, p = 0.001; MTX: R = 0.853, p < 0.001; IL: R = 0.833, p < 0.001; TCZ: R = 0.770, p = 0.014; TCZ + MTX: R = 0.515, p = 0.133); however, these associations were dependent on the cell culture conditions and cell line. In PBMCs, the correlation between AHR and transporter genes was lost under TCZ + MTX treatment (Figure S8A, S9A). In HepG2, AHR and SLC19A1 correlations were observed only in the control samples (Spearman’s rank correlation test, CTRL: R = 0.667, p = 0.007; p > 0.05 under MTX, IL-6, TCZ and TCZ + MTX) (Figure S9B). In contrast, in HepG2, the correlation between AHR-ABCG2 was observed across all treatments, including MTX + TCZ (Spearman’s rank correlation test, CTRL: R = 0.614, p = 0.015; MTX: R = 0.686, p = 0.006; IL: R = 0.529, p = 0.045; TCZ: R = 0.918, p < 0.001; TCZ + MTX: R = 0.796, p < 0.001) (Figure S8B).

Moreover, in PBMCs, TCZ + MTX disrupted the correlation between TYMS-DHFR, TYMS-SLC19A1, and DHFR-SLC19A1 (Figure S10 A-C), which were otherwise present across treatments (Spearman’s rank correlation test; TYMS-DHFR: CTRL: R = 0.806, p = 0.008; MTX: R = 0.687, p = 0.031; IL-6: R = 0.786, p = 0.028; TCZ: R = 0.905, p = 0.005; TCZ + MTX: R = 0.607, p = 0.167; TYMS-SLC19A1: CTRL: R = 0.818, p = 0.007; MTX: R = 0.806, p = 0.006; IL-6: R = 0.810, p = 0.020; TCZ: R = 0.881, p = 0.007; TCZ + MTX: R = 0.571, p = 0.20; DHFR-SLC19A1: CTRL: R = 0.942, p = 0.004; MTX: R = 0.703, p = 0.028; IL-6: R = 0.833, p = 0.015; TCZ: R = 0.929, p = 0.002; TCZ + MTX: R = 0.464, p = 0.302).

Discussion

The present study characterized the role of AHR in MTX and TCZ treatment in relation to MTX response in RA patients. We utilized a multi-level approach across clinical and experimental models. Our findings demonstrate that: (1) in silico molecular docking indicates a direct binding interaction between MTX and the AHR PAS-B domain; (2) ex vivo analysis of RA patients revealed significantly lower AHR expression in monocytes and B cells of poor responders and MTX-intolerant patients, while CD4 + T cells showed a significant opposing trend of upregulated AHR levels; and (3) 3D in vitro models, despite lacking statistical significance, provided supportive evidence for a distinct effect of TCZ co-treatment on restoring AHR protein expression and AHR regulation. This study suggests that monocyte-specific AHR expression could be a potential predictive biomarker for MTX response in RA.

A better understanding of MTX response is particularly relevant in RA, where timely initiation of effective therapy is essential for preventing irreversible joint damage and long-term disability. In the present study, we assessed AHR protein levels in PBMCs of RA patients with good or poor response to MTX under different treatment regimens: MTX, TCZ, or both. Additionally, we analyzed samples from patients with MTX intolerance [28] separately, given the clinical importance of this multifactorial phenomenon, which remains underexplored.

The most important finding of the present study was that AHR expression in monocytes of RA patients with MTX intolerance and poor response was significantly lower than that observed in good responders. AHR expression in monocytes was also significantly lower in patients treated with TCZ. Interestingly, the combination of MTX and TCZ resulted in AHR expression in monocytes comparable to that in good responders. These findings point to monocytes as the PBMC subset most strongly associated with the AHR-related pattern of MTX response.

Patients with intolerance were characterized by the highest levels of inflammatory markers and disease activity. Patients who were resistant to MTX presented a similar clinical picture, although with normal ESR and CRP.

The monocyte-centered pattern observed in our study is biologically plausible, given the established role of monocytes in RA pathogenesis and treatment response. AHR plays a role in monocyte differentiation and function; however, its specific contribution to RA remains unclear. It has been shown that AHR shapes the monocyte response to cytokines [29], promotes their differentiation into dendritic cells (DCs) while impairing macrophage development [30], and that MTX induces AHR expression in monocytes before differentiation [31]. Monocytes are key players in the pathological development of autoimmune diseases, including RA. They produce inflammatory cytokines and contribute to joint destruction via macrophage- and fibroblast-mediated pathways [32]. As a bridge between innate and adaptive responses, DCs trigger the differentiation of proinflammatory Th1 and Th17 lymphocytes. Therefore, they play a key role in immune homeostasis and tolerance [33]. In treatment-naïve RA, elevated monocyte counts predict an ineffective or poor response to MTX therapy [34, 35], and monocyte phenotypes may serve as markers of therapeutic efficacy or disease prognosis [36–39].

Consistent with patient data, AHR was also predominantly expressed in monocytes in our PBMC spheroid model. In this 3D spheroid model using PBMCs from healthy donors, monocytes appeared to be the most responsive to MTX, not only in terms of AHR-related changes but also in showing the strongest reduction in viability. TCZ slightly reduced AHR protein levels in monocytes compared to control samples, in contrast to the mRNA data, whereas the combination of MTX and TCZ restored AHR protein expression to control levels. The discrepancy between the mRNA and protein levels may suggest differences in the regulation of transcriptional and translational processes or differences in mRNA or protein stability. Although these observations remain preliminary, the spheroid model supported the relevance of monocytes as the PBMC subset most sensitive to treatment-related AHR modulation. Interestingly, monocytes also exhibit limited but relevant drug-metabolizing capacity [40]. While less effective than hepatocytes, under specific conditions they can differentiate into “neohepatocytes” with metabolic properties resembling those of liver cells [41].

In addition to monocytes, patients with MTX intolerance also presented the lowest AHR expression in B cells. This may reflect impaired B-cell regulation under heightened inflammatory conditions. In particular, B regulatory cells (Bregs, a subtype of CD19 + B cells), responsible for anti-inflammatory IL-10 secretion, have been shown to depend on AHR signaling [42]. The depletion of AHR in these cells may contribute to the proinflammatory clinical profile observed in MTX-intolerant patients. Monocyte-derived cytokines affect B-cell activation and differentiation [43], whereas B cells produce autoantibodies such as rheumatoid factor (RF) and anti-citrullinated protein antibodies (ACPAs). The generation of such autoantigens is linked to post-translational modifications, including citrullination and carbamylation, which may occur during autophagy and contribute to the breakdown of immune tolerance in RA. In addition, extracellular vesicles may transport citrullinated and carbamylated proteins to neighboring cells, thereby amplifying immune activation and perpetuating synovial inflammation [44]. In our previous study, the presence of RF was a significant predictor of poor MTX response [17], and B cells in RA patients showed increased activation markers [45]. AHR may also influence antibody production [46, 47], as B cell-specific AHR knockout mice develop more severe arthritis [42].

In contrast to monocytes and B cells, MTX-intolerant patients in our study exhibited significantly elevated AHR expression in CD4 + T cells. AHR activation in CD4 + T cells may enhance proinflammatory IL-17 and IL-22 secretion [48], possibly contributing to the intolerance mechanism. MTX modulates the autoimmune response by affecting CD4 + T cell subsets such as Th1/Th2 and Th17/Treg, and also influences CD73-expressing Tregs associated with anti-inflammatory adenosine production [49]. Thus, the increased AHR expression observed in CD4 + T cells in MTX-intolerant patients may indicate that the relationship between AHR activity and treatment outcome is cell-specific rather than uniform across PBMC populations.

Beyond its immunological role, AHR may also participate directly in the pharmacological response to MTX. Our in silico data suggested binding of MTX within the PAS-B domain of AHR. This is notable because MTX shares structural similarity with folic acid, another known AHR ligand [16]. These findings support the possibility that MTX may influence AHR not only indirectly through inflammatory and metabolic pathways but also through direct molecular interaction.

In line with this concept, our transcriptional data suggest that AHR is linked to MTX handling through transporter-related pathways associated with drug resistance. It has been shown that downregulation of SLC19A1 mRNA depends on AHR activation by its ligand TCDD [13]. In our study, MTX exposure resulted in decreased mRNA levels of several transporters, including SLC19A1. Additionally, we confirmed a correlation between the activation of AHR and ABCG2 upregulation, consistent with previous studies linking AHR signaling with drug resistance mechanisms [8, 15]. Notably, baseline AHR expression was higher in PBMCs than in HepG2, in agreement with single-cell RNA databases [50], likely reflecting the immune-specific function of AHR. This difference may stem from the altered transcriptional landscape of HepG2, a hepatocarcinoma-derived cell line, as previously discussed [51].

We also observed a significant induction of AHR in PBMCs compared to HepG2 cells, particularly after TCZ treatment. This contrasted with our initial hypothesis, as TCZ—an IL-6 receptor inhibitor—was expected to reduce AHR expression by blocking IL-6 signaling. IL-6 stimulation upregulated AHR, consistent with earlier studies [24]. However, TCZ only reduced AHR mRNA in HepG2 cells, not in PBMCs. Both hepatocytes and many PBMC subtypes produce IL-6 [52], but their sensitivity to cytokine modulation likely differs. These findings suggest that the effects of TCZ on AHR regulation are cell-context-dependent and may differ between immune and hepatic compartments.

Importantly, the combination of MTX with TCZ also disrupted the correlation between mRNA levels of AHR and ABCG2 in both PBMCs and HepG2 cells. Because ABCG2 is linked to MTX efflux and drug resistance, this observation suggests that TCZ co-treatment may modify AHR-associated resistance pathways rather than simply suppress inflammation. Drug sensitivity is another important factor. Inhibition of DHFR expression, for example, has been associated with improved MTX efficacy in combination therapy with metformin [53]. In our study, DHFR transcription was downregulated in PBMCs treated with the combination of MTX and TCZ. At the same time, the combination of TCZ and MTX led to significantly higher ATPase activity in HepG2 cells compared with MTX alone, which may indicate enhanced drug efflux. In PBMCs, the increase was not significant. This aligns with the liver’s dominant role in drug metabolism. Moreover, prolonged TCZ + MTX exposure increased SLC19A1 expression, potentially enhancing MTX uptake. However, uptake and retention of MTX also depend on the complex coordination of transport, polyglutamylation, and intracellular trafficking [4, 54].

Taken together, our data suggest that transcriptional dysregulation of key efflux transporters such as ABCG2 may contribute to the modulation of MTX resistance under TCZ co-treatment. Furthermore, TCZ treatment may modulate AHR signaling and potentially contribute to overcoming MTX resistance, supporting the rationale for combined MTX-TCZ therapy in selected patients. The combination of MTX and TCZ may have synergistic effects on AHR normalization; however, further mechanistic studies are required to clarify this interaction. Considering AHR’s role in the modulation of immune responses [55, 56], its restoration may contribute to rebalancing immune responses and holds potential for therapeutic application. This proposed relationship is summarized in Fig. 7. Although preliminary, these findings suggest that AHR-related mechanisms may help in the development of personalized therapeutic strategies in RA, potentially improving clinical outcomes while optimizing treatment costs. The potential clinical significance of the present study lies in the possible use of AHR expression in monocytes as a predictive biomarker of MTX response in RA patients. This may be particularly relevant, as reliable biomarkers of MTX response are still lacking in routine clinical practice. Identification of patients unlikely to benefit from MTX could allow earlier optimization of therapy and reduce unnecessary exposure to potential adverse effects of the drug.

Fig. 7.

A proposed working model linking aryl hydrocarbon receptor (AHR), IL-6 receptor signaling, and the response to methotrexate (MTX). Poor response or intolerance to MTX was associated with reduced monocyte AHR activity and an IL-6-associated inflammatory environment, which may contribute to dysregulation of AHR-related pathways involved in MTX transport and resistance. Combination therapy with tocilizumab (TCZ) and MTX blocks IL-6R signaling, which may affect AHR reactivity, shift monocyte AHR toward a responder-like pattern, and promote a more favorable MTX transport-related balance, potentially contributing to improved MTX sensitivity. Abbreviations: ABCG2, ATP-binding cassette subfamily G member 2; AHR, aryl hydrocarbon receptor; IL-6, interleukin-6; IL-6R, interleukin-6 receptor; MTX, methotrexate; SLC19A1, solute carrier family 19 member 1; TCZ, tocilizumab

Study limitations

The present study has several limitations. In vitro experiments only partially reflect the in vivo RA environment, and the metabolic heterogeneity of immune cells hampers direct extrapolation across tissues. Gene expression is dynamic and time-dependent, and despite the inclusion of healthy controls, the lack of ex vivo conditions limits comparability with patient samples. Additionally, patient cohort size was limited, particularly in the TCZ + MTX group, reducing statistical power. Moreover, most of the patients were also treated with GCS and concomitant therapies, which may affect outcomes. However, in this study, patients were strictly selected, and concomitant therapies are inevitable due to the clinical course of the disease. Although we observed transcript-level correlations between AHR and drug transporters, functional validation—such as AHR inhibition or transporter assays—was not performed. ATPase activity provides only an indirect indication of ABC transporter function. The absence of direct functional assays limits conclusions about drug transport and resistance mechanisms.

It is important to emphasize that this study adopts an exploratory rather than confirmatory approach. The aim was to generate mechanistic hypotheses through the integration of patient data, in vitro models, and in silico analysis. Larger cohorts and mechanistic validation are needed to confirm AHR’s role in MTX response and to clarify the extent to which TCZ modulates AHR-dependent resistance-related pathways.

Conclusions

In conclusion, cell-specific MTX responses reflect complex AHR-related mechanisms. Further studies may clarify these pathways. MTX likely interacts with AHR via the PAS-B domain, affects AHR-related processes, and its metabolism may be AHR-dependent. Monocyte AHR expression may predict MTX efficacy, while TCZ appears to modulate AHR signaling. The role of AHR across PBMC subtypes warrants further investigation.

Supplementary Information

Below is the link to the electronic supplementary material.

Abbreviations

7-AAD

7-aminoactinomycin D

ABCB1

ATP-binding cassette subfamily B member 1

ABCC 1–5

ATP-binding cassette subfamily C member 1–5

ABCG2

ATP-binding cassette subfamily G member 2

ACPAs

Anti-citrullinated protein antibodies

ACR

American College of Rheumatology

AHR

Aryl hydrocarbon receptor

AHRR

Aryl hydrocarbon receptor repressor

AICAR

Aminoimidazole-4-carboxamide ribonucleotide

Anti-CCP

Anti-cyclic citrullinated peptide

ARNT

Aryl hydrocarbon receptor nuclear translocator

bHLH

Basic helix–loop–helix

Bregs

Regulatory B cells

CPT

Cell preparation tube

CRP

C-reactive protein

CTRL

Control group

DAS28

Disease Activity Score 28

DCs

Dendritic cells

ddPCR

Digital droplet polymerase chain reaction

DHFR

Dihydrofolate reductase

DKK1

Dickkopf-1

DMARDs

Disease-Modifying Antirheumatic Drugs

DMEM

Dulbecco’s Modified Eagle Medium

ELISA

Enzyme-linked immunosorbent assay

ESR

Erythrocyte sedimentation rate

EULAR

European Alliance of Associations for Rheumatology

EV

Extracellular vesicle

FBS

Fetal bovine serum

GCS

Glucocorticoids

HAQ

Health Assessment Questionnaire

HBSS

Hank’s Balanced Salt Solution

HC

Healthy control(s)

HepG2

Hepatocellular carcinoma human cell line

IL

Interleukin

MFI

Median fluorescence intensity

MTX

Methotrexate

MTRR

Methionine synthase reductase

PAS-B

Per-Arnt-Sim B domain of AHR

PBMCs

Peripheral blood mononuclear cells

PDB

Protein Data Bank

pJIA

Polyarticular juvenile idiopathic arthritis

PTM

Post-translational modification

RA

Rheumatoid arthritis

RF

Rheumatoid factor

sJIA

Systemic juvenile idiopathic arthritis

SLC19A1/ RFC1

Solute carrier family 19 member 1/ Reduced folate carrier 1

STAT3

Signal transducer and activator of transcription 3

TCZ

Tocilizumab

TNF-α

Tumor necrosis factor-alpha

TYMS

Thymidylate synthase

VAS

Visual analogue scale

WBC

White blood cells

WNT

Wingless-related integration site

WST-1

Water-soluble tetrazolium salt-1

XRE

Xenobiotic response element

Author contributions

Conceptualization – A.W., A.G-P.; methodology- A.W., E.K-W.; formal analysis – B.S., A.W., E.K-W.; investigation –G.F., Y.K., T.K., E.K-W., A.E-M., D.B., E.M.; docking analysis- M.J.; patient classification and data curation –A.M., K.B., B.J., M.S., A.F-G., M.O.; writing—original draft preparation – A.W.; writing—review and editing, A.G-P., B.S., A.M.

Funding

This work was supported by the Polish Ministry of Science and Higher Education through statutory funding for the National Institute of Geriatrics, Rheumatology and Rehabilitation, Warsaw, Poland, by grant S/1 “Prevalence of Genetic and Epigenetic Factors in Patients with Rheumatoid Arthritis and Their Impact on Treatment Efficacy and Safety”, as well as statutory funding for the National Medicines Institute, Warsaw, Poland.

Data availability

The datasets generated and analysed during the current study are not publicly available due to their limited size and experimental nature, which do not warrant deposition in a public repository. However, they are available from the corresponding author on reasonable request.

Declarations

Ethical approval

This study was performed in accordance with the Helsinki declaration and was approved by the Research Ethics Committee of the National Institute of Geriatrics, Rheumatology and Rehabilitation in Warsaw, Poland (approval number KBT-5/8/2022).

Consent to participate

All participants provided written informed consent to participate in the study.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.