Personalized methotrexate treatment in immune-mediated inflammatory diseases: a narrative review

Introduction

Background

Since its first use as a chemotherapeutic in the late 1950s, methotrexate (MTX) is now an indispensable drug in chemotherapy (high doses) and in low weekly doses, for the treatment of immune-mediated inflammatory diseases (IMIDs) (1). Currently, MTX is the first-line chemical/synthetic disease-modifying anti-rheumatic drug (csDMARD) of treatment in rheumatoid arthritis (RA), juvenile idiopathic arthritis (JIA), severe psoriasis, psoriatic arthritis (PsA), and next-in-line for mild Crohn’s disease (CD), pediatric inflammatory bowel disease (IBD), sarcoidosis, and myasthenia gravis (2-4).

MTX is an antifolate, prescribed weekly either orally or as subcutaneous injections. In low-dose MTX treatment, plasma concentration peaks after 1 hour (half-life of about 6 hours) and is undetectable by 24 hours (5-7). Oral MTX is absorbed by the proton-coupled folate transporter (PCFT) in the small intestines (5). Circulating MTX is taken up by the cells via the reduced folate carrier (RFC) and folate receptors depending on the cell type (1,8). Upon cellular entry, the enzyme folyl-polyglutamate synthetase (FPGS) converts MTX to its active metabolites MTX polyglutamate (MTX-PG_1–6) forms by adding up to five additional glutamate moieties (1,9,10). MTX-PG_1–6 inhibit dihydrofolate reductase (DHFR) and thymidylate synthase (TS), thereby affecting the folate pool and pyrimidine synthesis, respectively (9). Additionally, MTX-PG_1–6 inhibits 5-aminoimidazole-4-carboxamide ribonucleotide formyltransferase (ATIC), ultimately increasing extracellular adenosine concentrations, which, via interaction with adenosine receptor 2a/b confers potent anti-inflammatory effects (Figure 1) (1,9,11). MTX is eliminated via urine (~80%) or bile (~10%) (5). About 3% of MTX undergoes first-pass metabolism and is converted to 7-hydroxy MTX (7-OH-MTX) in the liver and is eliminated via urine (5). 7-OH-MTX can be polyglutamylated and is a very weak inhibitor of DHFR (12).

Figure 1 Mechanism of action of MTX. MTX, upon cellular entry, is polyglutamylated (MTX-PG_1–6) by FPGS while the reverse reaction is catalyzed by GGH. MTX-PG_1–3 can be effluxed by ABC transporters—ABCC1–5 and ABCG2. MTX-PG_1–6 are potent inhibitors of DHFR, TS, ATIC. Inhibition of TS affects pyrimidine synthesis. Inhibition of ATIC increases ATP which is effluxed from the cell. ATP is converted to ADP and AMP by CD39. CD73 further converts AMP to adenosine which enter the cell via AR2a/2b. Adenosine then acts as anti-inflammatory agent. ABC, ATP-binding cassette; ADP, adenosine diphosphate; AICAR, 5-aminoimidazole-4-carboxamide ribonucleotide; AMP, adenosine monophosphate; ATIC, 5-aminoimidazole-4-carboxamide ribonucleotide formyltransferase; ATP, adenosine triphosphate; DHF, dihydrofolate; DHFR, dihydrofolate reductase; dTMP, deoxythymidine monophosphate; dUMP, deoxyuridine monophosphate; FPGS, folyl-polyglutamate synthetase; FAICAR, 5-formamidoimidazole-4-carboxamide ribonucleotide; FR, folate receptor; GMP, guanosine monophosphate; IMP, inosine monophosphate; MTX, methotrexate; MTX-PG, methotrexate polyglutamate; PCFT, proton-coupled folate transporter; RFC, reduced folate carrier; THF, tetrahydrofolate; TS, thymidylate synthase.

Rationale and knowledge gap

Low drug costs with good safety profile, solidified the position of MTX as a key immunomodulator in the treatment of IMIDs. A recent large clinical study (NORDSTAR) compared the efficacy of MTX (with prednisolone or hydroxychloroquine ± sulfasalazine) against biologics (certolizumab-pegol, tocilizumab, and abatacept) in treatment naïve early RA patients and revealed that after 24 weeks of treatment, MTX was non-inferior to certolizumab-pegol and tocilizumab. Notably, radiographic progression was similar between all treatment arms (13). Nevertheless, 30–40% patients fail to respond to MTX with failure attributed to either inadequate response, or due to intolerance (14,15). A recent study suggested that ~65% of RA patients are underdosed with MTX (16). Multiple biomarkers and methods investigating prediction of MTX response exists. Combining different biomarkers can aid in personalizing MTX treatment, effectively treating diseases early with timely step-ups.

Objective

This narrative review explores recent developments in personalized medicine of MTX, firstly focusing on the clinical efficacy of MTX in IMIDs with strategies to improve it, and prediction of response to MTX, and secondly on therapeutic drug monitoring (TDM) of low-dose MTX, highlighting some of the tools currently available for this purpose. We present this article in accordance with the Narrative Review reporting checklist (available at https://jlpm.amegroups.com/article/view/10.21037/jlpm-25-7/rc).

Methods

Literature search was carried out in PubMed. Keywords used in the literature search included “methotrexate” AND “methotrexate-polyglutamates” AND disease entities, focusing on RA, JIA, CD/IBD, and sarcoidosis (Table 1). Papers published between 1980 and 2025 were included, including original research, meta-analysis, systematic reviews, clinical and observational studies, and clinical trials and references from selected articles. Case studies were excluded.

Clinical efficacy of MTX

Treatment guidelines

Efficacy

In RA, treatment is initiated with MTX, concomitantly with glucocorticoids and is escalated if there is insufficient response at 6 months (17,18). For polyarthritis and mono/oligoarthritis subtypes of PsA, a rapid start with csDMARD (MTX, leflunomide, or sulfasalazine). For psoriasis patients, MTX is started at 15 mg/week and escalated to 20–25 mg/week (19,20). For mild to moderate CD, the guidelines suggest use of subcutaneous/intramuscular MTX dosed up to 25 mg/weekly for induction and maintenance of response (21,22).

Toxicity

MTX is considered a safe drug with a toxicity profile including gastro-intestinal (GI) intolerances, elevated liver enzymes, and occurrences of cytopenia. Patients initially starting MTX are monitored every month for the first 3 months, followed by once every three to 4 months (17,20,23). Toxicities are more common in patients with liver or renal impairment (24,25). Folic acid supplementation reduces the toxicity of MTX with 1–2 mg supplemented daily starting 1 day after MTX intake (26). In the Netherlands, folic acid is prescribed at dosage of 5 mg/week (27). Japan College of Rheumatology recommends folic acid supplementation once a week dosed at 5 mg/week (18). There is no consensus between the folic acid dosing schemes and evidence for the most optimal scheme is lacking. Dosing folic acid at either 5 or 0.8 mg/week has similar effects on adverse events (AEs) in RA, while increasing to 30 mg/week did not offer additional benefits against toxicity (28,29). However, elevated unmetabolized folic acid in plasma is associated with long-term health risks (30).

Strategies to increase MTX efficacy

Route of administration

Route of administration impacts the bioavailability of a drug. The bioavailability of intravenous administration is 100% because of direct drug-delivery to circulation. Hoekstra et al. reported the bioavailability of a median oral MTX dose of 30 mg/week (range, 25–40 mg/week) was 64% (range, 21–96%) of subcutaneous route (7). Subcutaneous and intramuscular injections result in similar bioavailability, with both higher than oral (31). A recent study in RA patients reported after subcutaneous administration of MTX, accumulation of MTX-PG_1–5 in erythrocytes in the first 2 months of treatment was significantly higher than after oral dosing (32). A review from our group on oral and subcutaneous administration suggested that despite some initial advantages with subcutaneous administration, after prolonged (>3 months) treatment, there was no evidence to support its superiority over oral route (27). The change in the route of administration can be a shared-decision with the patients before changing treatment plans. Beyond route of administration, alternative dosing strategies can be considered.

Split-doses

The literature on the effects of splitting MTX dose is contradictory. One study showed increased bioavailability of oral MTX when the dose was split to two doses, 8 hours apart against once weekly (33). Another study did not show a clinical advantage of split MTX dose, but no increase in AEs due to the split dose was reported (34). Recently, a split MTX dose study (5 days/week or twice-on-the-same-day/week) demonstrated similar efficacies and AEs. Interestingly, this study observed reduced accumulation of MTX-PG₃ in erythrocytes in the daily-MTX group compared to the once-a-week-dose group of RA patients (35). In contrast, a study in 135 RA patients demonstrated that split oral dose (twice a week) had similar efficacy as single intramuscular dose, showed improved efficacy against single oral dose and came without any significant AEs (36). In a recently presented abstract at American College of Rheumatology (ACR), RA patients also showed better response in split dose (morning and evening) against single weekly dose, however increased transaminases emerged (at 16 weeks and persistent) along with low white blood cell count at 24 weeks (37). Splitting the dose of MTX can be an alternative before changing the route of administration, though patients need to be monitored for toxicities.

MTX with efflux transporter inhibitors

Increasing the cellular availability of MTX can theoretically space the time between consecutive doses or lower doses without impacting therapy efficacy. Conceivably, cellular retention of MTX can be increased by blocking cellular efflux transporters of MTX, i.e., ABCC1–5 and ABCG2 (Figure 1). Notably, increased ABCG2 expression on RA synovial macrophages correlated with reduced MTX response (38). Investigating the expression of six genes, four in the folate pathway—DHFR, FPGS, GGH, RFC and two efflux transporters—ABCC1 and ABCG2 in blood cells of RA patients at baseline and 12 weeks after MTX use, revealed that good responders had the highest decrease in the expression of ABCG2 between baseline and 12 weeks (39). Measuring levels of ATP-binding cassette (ABC) drug efflux transporters may help design treatment strategies using transporter inhibitors, though these inhibitors must still be tested for safety and effectiveness (39,40).

Aside from improving efficacy, predicting potential non-responders before treatment can be cost-effective and prevent worsening of disease.

Prediction of MTX response

Prediction modeling in IMIDs can either predict response or AEs to MTX. Lifestyle [smoking/alcohol use, body mass index (BMI)], clinical/genetic [alternative splicing (AS), presence of single-nucleotide polymorphisms (SNPs), kidney function], or other factors (age, sex, disease duration) play a role in prediction.

Biomarkers of response prediction

AS

During the processing of a pre-messenger RNA (pre-mRNA), the intronic and exonic regions are spliced differently, enabling the mRNA to transcribe protein variants. The first study describing the role of AS in FPGS described a loss-of-function variant of FPGS, leading to MTX resistance in leukemia. This novel transcript of FPGS was missing the portion coded by exon 12 (41). A following comprehensive PCR study on FPGS identified additional AS transcripts of FPGS, one of which was FPGS-8PR, a transcript with partial retention of the 8^th intronic region (42). A study in RA showed that the ratio of FPGS-8PR/FPGS-WT was associated with higher baseline disease activity score evaluated in 44 joints (DAS-44) at 13 and 26 weeks of MTX treatment. Predicting 13-week non-response with FPGS-8PR/FPGS-WT ratio had an area under the curve (AUC) of 0.89 and low-disease activity (LDA) was 0.74 (43). Further, ex-vivo monocytes from RA patients were found to have low FPGS activity due to increased FPGS-8PR. Monocyte-derived macrophages, however, had lower FPGS-8PR resulting in higher FPGS activity (44).

SNPs

Effects of SNPs on MTX (non)response and intolerance is an active area of research. SNPs in genes of different enzymes associated with MTX/folate influx/efflux (SLC19A1/RFC, ABCB1, ABCC2), enzymes (de)activating MTX (FPGS or GGH), and enzymes in the folate pathway (MTHFR, MTRR, ATIC) have been investigated. The results of a large-scale genome-wide association study with 1,424 early RA patients on MTX monotherapy did not identify any SNPs to be significantly associated with response. The study identified a SNP of NRG3 was associated with disease activity score evaluated in 28 joints (DAS-28) (45). A machine learning model combining 160 SNPs with other factors like age, sex, smoking, rheumatoid factor (RF), DAS-28, and EULAR response at 3 months had an AUC of 0.84. Without the SNPs, the AUC of model was 0.54 (46). Some SNPs that affect adult response do not necessarily have the same effects on response in children or in other IMIDs; for example, polymorphisms in FPGS were found to affect MTX response or toxicity in adults with RA (47-49). However, the SNPs in FPGS were found not to be associated with response or toxicity in a population of psoriatic patients or in JIA patients (50,51). Though studies associate SNPs and response, a validated panel of SNPs, rather than one maybe needed to predict response.

Plasma/serum biomarkers

In 2012, a multi-biomarker disease activity (MBDA) score was developed and validated in RA patients treated with DMARDs, including MTX. This score, aimed to monitor disease activity, comprised of 12 serum protein concentrations (52). The proteins evaluated include epidermal growth factor (EGF), vascular endothelial growth factor-A (VEGF-A), leptin, interleukin-6, (IL-6), serum amyloid A (SAA), C-reactive protein (CRP), vascular cell adhesion molecule 1 (VCAM-1), matrix metalloproteinase-1 (MMP-1), matrix metalloproteinase-3 (MMP-3), tumor necrosis factor receptor-1 (TNFR-I), human cartilage glycoprotein 39 (YKL-40), and resistin (52). Using the MBDA score on patients from the SWEFOT trial, Hambardzumyan et al. associated the individual components of the MBDA score at baseline to MTX response in DMARD-naïve RA patients. Patients reaching LDA at 3 months had significantly lower levels of CRP and IL-6, while TNFR-I and VCAM-1 were significantly higher. Logistic regression model using the protein biomarkers identified a combination of 4 protein biomarkers—VCAM-1, TNFR-I, CRP, and leptin, was associated with 3-month EULAR response (53).

A UK study with 100 MTX-naïve RA patients used baseline serum lipidomics to predict MTX response. Six-month MTX response was predicted using machine learning models with serum lipid profiles at pre-treatment and after 4 weeks of MTX treatment. The model with lipid profiles, however, only showed limited predictive ability and was not better than the clinical model (54). In psoriasis, hierarchical clustering of baseline serum metabolomics achieved separation between good and poor responders. After 16 weeks of MTX treatment, they identified that metabolites related to fatty acid, purine, androgen, and estrogen metabolisms were down-regulated in good responders while metabolites related to sugars and amino acids were up-regulated in poor responders (55). Semi-targeted metabolomics on plasma from 30 JIA patients collected at baseline and 3 months after MTX treatment identified alterations in 50 metabolites of which 12 were associated with disease activity. Reduction in three metabolites (dehydrocholic acid, biotin, and 4-picoline derived from microbiota) was associated with ACR response criteria (56). An untargeted metabolomics study using baseline serum from 82 early RA treatment-naïve patients, identified eight possible biomarkers that differed significantly between sufficient and insufficient responders after 3 months of MTX treatment (57). Homocystine, taurine, ATP, GDP, and uric acid that was significantly lower in insufficient responders. Sufficient responders had significantly higher, 1,3- or 2,3-bisphosphoglyceric acid (DPG), glycerol-3-phosphate, and phosphoenolpyruvate. The AUC of a model constructed using homocystine, 1,3- or 2,3-DPG, and glycerol-3-phosphate was 0.81 [95% confidence interval (CI): 0.72–0.91]. Overrepresentation analysis indicated that metabolites involved in glycolysis and Warburg effect at baseline were different between the two MTX response groups and could be related to activation of immune system (57). Metabolomic or lipidomic studies in other IMIDs are lacking compared to RA. Validating the findings from both these studies in external cohorts could help identify clinically useable metabolic biomarkers for MTX response in RA and other IMIDs.

Gut microbiome

Gut microbiome plays a key role in absorption of MTX and conversion to non-active metabolites, thus affecting bioavailability. The connection between gut microbiome and MTX response was first demonstrated in mice in 1972 (58). However, the field of gut microbiome and associations with MTX response is a relatively young field. A study in RA patients used baseline gut microbiome data to predict efficacy of oral MTX treatment, by constructing a random forest model. The model included microbiota Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways and clinical predictors with an AUC of 0.84. Notably, non-responders had higher bacterial biodiversity with upregulation in microbial pathway involving MAPK signaling pathway, DNA replication, fatty acid degradation, and ABC transporters (59).

Similarly in psoriasis patients, poor responders were found to have a higher bacterial diversity. After MTX treatment, although the diversity of bacterial community did not differ significantly, good responders showed an increase in relative abundance of Bacteroides vulgatus and family Veillonellaceae and a decrease in the relative abundance of Bacteroides caccae and family Erysipelotrichaceae. These findings need to be validated in larger patient cohorts (55). In a recent pre-print, a similar study was carried out by including treatment-naïve RA, PsA patients, and a control group which had IMIDs symptoms with a diagnosis unrelated to rheumatic diseases (60). Similar to the previous two studies, responders and non-responders had a significant difference in the biodiversity in PsA patients. Contrastingly, this study did not find a significant difference in species richness between responders and non-responders in both RA and PsA. A random forest-based prediction model distinguishing MTX responders and non-responders in PsA had a poor predictive power (AUC of 0.56), and a moderate AUC (0.74) for RA patients. Contrasting the observations in psoriasis, metabolic pathways related to amino acid and carbohydrate metabolism were upregulated in both groups. Specifically, in PsA MTX-responders, pathways in short-chain fatty acid metabolism was enriched. In RA MTX-non-responders, pathways in fatty acid biosynthesis and amino acid metabolism were enriched (60). These study results must be interpreted with caution as it awaits peer-review. In a cohort of JIA patients (78 Italian, 21 Dutch), a recent study found Italian patients had lower microbial abundance while Dutch patients had a weak increased abundance compared to healthy controls. This study noted that the disease activity, inflammation status were not drivers for dysbiosis while age, geographic origin were related (61). Microbiota are influenced by diet, geography, use of antibiotics to name a few. If the microbiota identified are validated in larger cohorts, gut microbiota can be investigated as a possible factor to explain the inter-individual differences in MTX-PG_1–6 accumulation in erythrocytes, seen after MTX treatment.

Prediction models

RA

AE prediction models: few studies constructed models predicting AEs. In one such study, a model to predict AEs within 2 years of MTX use was constructed. This model included clinical [RF/anti-citrullinated protein antibody (ACPA) positive, MTX dose, and on MTX monotherapy] and genetic factors (SNPs of SLC19A1, ABCG2, ADORA3, and TYMS) had a predictive AUC of 0.78. However, the model lost its predictive power upon cross-validation (AUC of 0.57) (62). Another study in a population of 110 patients, using binomial logistic regression identified low BMI and FPGS SNP (rs10106) were associated with increased AEs. This model had a modest sensitivity of 44.4% and a specificity of 79% (63). A model constructed specifically for the prediction of hepatotoxicity, which included 12 clinical variables and seven SNPs, had an AUC of 0.902 with sensitivity of 85.1% and specificity of 87.8% (64).

Response prediction models for MTX: a least absolute shrinkage and selection operator (LASSO) regression model, genes predict DAS-28 after 6 months of MTX monotherapy, included baseline DAS-28, presence of erosion, MTX dose, and SNPs in the SLC19A1, SLCO1B1, TYMS, and AMPD1. The predicted 6-month DAS-28 values significantly correlated with the observed DAS-28 values (65). Predicting MTX efficacy in Japanese patients, a model incorporated clinical (DAS-28-CRP and folic acid use) and genetic factors (SNPs of SLCO3A1, CYP7A1, CHST2, GGH, SLC22A1, EPHX1, and ATP7B) and had a AUC of 0.841 (64). A similar study conducted in the Netherlands included different clinical (sex, baseline DAS-28, smoking status, RF positivity) and genetic (SNPs of AMPD1 34CC, ATIC 347G, ITPA 94A, and MTHFD1 1958AA genes) factors. The study had moderate AUC (71.5%) for the models predicting response after 3–4 months of MTX use and 74.6% for the model with the genetic parameters only (66). The RAMS prediction model was the first to include a psychological factor in the prediction of non-response (EULAR criteria) to MTX at 6 months. The predictors included RF negativity, Health Assessment Questionnaire (HAQ) score, 28-joint tender joint count (TJC28), Hospital Anxiety and Depression Scale (anxiety score) and disease activity. The AUC of the model was 0.77 (95% CI: 0.73–0.80). Excluding patients at remission or had LDA at baseline, model AUC was 0.72 (95% CI: 0.68–0.76) and 0.71 (95% CI: 0.66–0.75), respectively (67). A logistic regression prediction model built on the MTX arm of U-Act-Early cohort of RA patients identified baseline DAS-28, current smoking, and alcohol consumption with an AUC of 0.75 (95% CI: 0.66–0.84; sensitivity: 0.89; specificity: 0.52). Interestingly, this study suggested that consumption of 1–2 glasses of alcohol reduced risk of insufficient MTX response (68).

Using data from the tREACH cohort from the Netherlands, a prognostic prediction model for insufficient response was constructed (69). Final model included clinical factors like DAS-28 >5.1, HAQ >0.6, current smoking, BMI >25 kg/m², erythrocyte folate <750 nmol/L, and genetic factors (SNPs of ABCB1, ABCC3). The AUC of the test and validation cohorts was 0.80. External validation using a different cohort of RA patients resulted in an AUC of 0.75 (95% CI: 0.64–0.85) (69). This model was optimized and a new final model was proposed containing clinical factors only. The optimized model included baseline DAS-28, HAQ, BMI, erythrocyte folate and smoking for prediction and has an AUC of 0.75 (95% CI: 0.69–0.81) (available on https://www.evidencio.com/models/show/2191?v=1.31) (70). Different machine learning algorithms were tested and the logistic regression model was found to have the best AUC of 0.77 (95% CI: 0.68–0.86). The final logistic regression model contains BMI, HAQ, erythrocyte sedimentation rate (ESR), TJC28, DMARD or corticosteroid use, and smoking as the final predictors and is also available on Evidencio (https://www.evidencio.com/models/show/2415) (71).

Most of the prediction models published so far have not been validated or implemented in a clinical setting. Extended research in the form of observational and implementational studies are needed to evaluate the readiness for predictive application. Using prediction models to predict MTX (non)response over time can be cost-effective while saving patients time by starting them on predicted efficacious treatments. In patients predicted to respond to MTX, TDM can ensure they get timely corrections to their treatment course based on their present difficulties such as intolerances or AEs.

JIA

In JIA, the first MTX non-response prediction model was developed in 2012 by Bulatovic et al. Using 1-year MTX response data from 183 JIA patients recruited from two observational cohorts in the Netherlands, the prediction model was validated in a cohort of 104 patients. The final model consisted of ESR, SNPs for MTRR (rs1801394), ABCB1 (rs1045642), ABCC1 (rs35592), and PCFT (rs2239907), and had an AUC of 0.65 (95% CI: 0.54–0.77) (72). A follow-up MTX intolerance prediction model included JIA category, disease activity score, parent/patient assessment of pain, antinuclear antibody (ANA), alanine transaminase (ALT), thrombocytes, creatinine, and an interaction term between creatinine and JIA category and classified 77% of the patients correctly (73).

A recent study used a machine learning approach to predict MTX response in JIA using data from electronic medical records from 362 JIA patients (74). Due to the retrospective nature of the study, DAS-44-ESR at 3 months after MTX start was used to define response and non-response. The extreme gradient boosting (XGBoost) model performed the best with of 0.97 and included CRP, the absolute value of CD3⁺ T-cell, RF-immunoglobulin G (IgG), TJC, direct and indirect bilirubin, active partial thrombin time, prothrombin and thrombin times, and fibrinogen variables. The study also constructed a mixed variables model by including features collected 3 months after MTX start which unsurprisingly performed better at predicting 3-month response (AUC =0.99) (74).

TDM of MTX

To use MTX-PG_1–5 concentrations for TDM, they must be (I) reliably measurable from an accessible matrix, (II) have interindividual variations, and (III) there must be established relationships between drug concentration and clinical efficacy.

Monitoring of MTX and MTX-PG_1–5

Plasma/serum

Plasma-based monitoring of MTX in high-dose MTX treatment (as a chemotherapeutic) is common due to the elevated risk of side effects (75). Monitoring of low-dose MTX in plasma has been challenging due to rapid plasma MTX clearance by 24 hours. Recent studies in RA, JIA, and psoriasis patients suggest the possibility of measuring MTX in plasma on the ultra-performance liquid chromatography-tandem mass spectrometry (UPLC-MS/MS) to measure MTX drug adherence (6,76,77). These studies established adherence cut-off for MTX and 7-OH-MTX, based on the dose of MTX used (6,77). If validated, plasma-MTX can be used to identify potentially non-adherent patients. Quantifying MTX in plasma has a few caveats; firstly, factors like MTX dose, folate supplementation, malabsorption and fast metabolism, clearance, and bioavailability can influence the duration of MTX in plasma. Importantly, due to the low plasma quantities after 24 hours (~0.1 nmol/L), a sensitive device is mandatory. Plasma MTX is transient and dependent on the MTX administration, making it difficult to use for prediction of response or for TDM.

Erythrocytes and whole blood

To monitor efficacy of MTX, a suitable matrix was required. As MTX-PG_1–5 are the intracellular active metabolites, studies evaluated MTX-PG_1–5 concentrations in erythrocytes, which were chosen as a surrogate matrix due to the ease of collection. Multiple studies validated protocols to quantify MTX-PG_1–5 in erythrocytes. These employ different techniques like fluorescent polarization, high-pressure liquid chromatography (HPLC) post-column photo-oxidation, and UPLC-MS/MS (78-82). The UPLC-MS/MS methods using stable isotope labeled internal standards are preferred for their better sensitivity, specificity and accurate quantification of the MTX-PG_1–5 (71). This method can also quantify MTX-PGs in whole blood (correcting for hematocrit). Quantification in erythrocytes led to multiple studies evaluating possible associations between individual erythrocyte MTX-PG_1–5 concentration and therapy efficacy (83-86). After starting MTX, accumulation of MTX-PG_1–5 in erythrocytes was observed which stabilized by 3 to 6 months of MTX treatment, with MTX-PG₃ having the highest concentration after 3 months (Figure 2). A plausible explanation for this is due to the long lifespan of erythrocytes of about 120 days. By 3 to 6 months, all erythrocytes are exposed to MTX therefore no further accumulation is noted. The same pattern of erythrocyte MTX-PG_1–5 accumulation is identified in RA, JIA, CD, and psoriasis patients (83-86).

Figure 2 Accumulation pattern of MTX-PG_1–6 in different matrices. A schematic of the accumulation pattern of MTX-PG_1–6 in different matrices. The MTX-PG_1–6 were measured using UPLC-MS/MS. Distinct accumulation patterns can be observed between the representative cell type (erythrocytes), immune effector cells types (PBMCs) and target cells types (mucosa, sperm). MTX-PG, methotrexate polyglutamate; PBMCs, peripheral blood mononuclear cells; UPLC-MS/MS, ultra-performance liquid chromatography-tandem mass spectrometry.

Erythrocyte-MTX-PG_1–5 can be quantified via commercial sources. These reports also include the concentration of MTX-PG_1–5 expected for therapeutic efficacy (87,88). The limitations of these commercial tests are that they are laboratory-developed test without TDM implementation studies and can only be used in a research setting.

Alternate blood sampling systems

The abovementioned analytical methods require a phlebotomist to perform a venipuncture which is invasive to a certain degree. In 2014, the first method to quantify MTX-PG_1–5 from dried blood spots (Guthrie cards) was developed and validated. Dried blood spots use less volume of blood (12 µL). This study showed applicability in samples collected from JIA or juvenile dermatomyositis patients (89). This method though did not use internal standards for quantification. This method can also be used to monitor treatment compliance. Another study quantified MTX-PG₃ from blood sampled in fingerstick [10 µL volumetric absorptive microsampler (VAMS)] from RA patients. MTX-PG₃ was quantified on an LC-MS/MS with internal standards. A 1.3-fold higher concentration was noted in the VAMS after correcting for hematocrit when compared to concentrations in erythrocytes. This was attributed to greater extraction efficiency in the VAMS samples in the presence of plasma (90). Both these methods have potential applications in home sampling and for pediatric patient samples. If implemented, these methods can simplify TDM applications and compliance monitoring in IMID patients. The MTX-PG_1–5 accumulation pattern in whole blood is similar to erythrocytes.

Although there are advantages for quantifying MTX-PG_1–5 in erythrocytes, these cells are not involved in the pathogenesis of IMIDs. Methods to quantify in other matrices related to disease pathologies are being investigated.

Peripheral blood mononuclear cells (PBMCs)

The first matrix of interest are PBMCs. These are a collection of immune cells in the blood which include T- and B-lymphocytes and monocytes. Due to their relatively low numbers in blood and higher matrix effects (possibly due to phospholipids) than the erythrocyte matrix, quantification has, hitherto, been difficult. The first abstract quantifying MTX-PGs in PBMCs from JIA patients was presented in 2011 (91). Following this, Daraghmeh et al. in 2022 quantified MTX-PGs in PBMCs using LC-MS/MS without validating the method (82). A recent study compared MTX-PG_1–3 levels in erythrocytes and PBMCs from 40 early RA patients (32). Recently, we developed and validated a UPLC-MS/MS method quantifying low concentrations of MTX-PG_1–6. MTX-PG₅ was close to our limits of quantification. MTX-PG₆ was not detected in RA patient PBMC samples (92). The MTX-PG_1–3 distribution profile in PBMCs revealed that MTX-PG₁ represented the largest fraction (58%), followed by MTX-PG₂ (27%) and MTX-PG₃ (15%) (Figure 2) (32,92). This was confirmed in PBMCs isolated from CD (n=6) patients on MTX (93). This profile may be accountable by the short lifespan of PBMCs (1–4 days, depending on the cell type) and the relatively low FPGS activity in PBMCs (94-96).

Mucosa

Biopsies (n=28) from CD patients treated with MTX were taken during a routine endoscopy from different locations in the small intestine, colon or the rectum with different inflammation status and MTX-PG_1–6 levels were quantified. Due to the low number of samples, association study was not performed on mucosal MTX-PG_1–6 concentrations (93). The pattern of accumulation revealed high MTX-PG₁ and considerable accumulation of MTX-PG_2–5 with detectable concentration of MTX-PG₆ (Figure 2). Remarkably, the levels of MTX-PG_2–5 in mucosa were retained 3 weeks after stopping MTX treatment (93). Due to the invasive nature of the sample collection, mucosa may not be the most preferred matrix for TDM.

Sperm

Commonly, male patients on MTX, should they wish to have children, are advised to stop MTX 6 months prior. This led to the first investigation to study if MTX-PG_1–5 were quantifiable in sperm cells. They noted low levels of MTX-PG₁ and negligible concentrations of MTX-PG_2–5 in sperm cells. After checking other parameters related to sperm health, they concluding that male patients wishing to be fathers could continue MTX treatment (Figure 2) (97). Although not a TDM matrix, sperm MTX-PG_1–5 concentrations could be checked with great accuracy using UPLC-MS/MS methodology in male patients with a desire to have children.

Urine

Due to the action of aldehyde oxidase (AOX) in the liver, the catabolic product of MTX, 7-OH-MTX, can be measured in urine (98). Increased conversion of MTX to 7-OH-MTX, would imply that less MTX-PG_1–5 are sequestered in the cells. 7-OH-MTX has limited ability to block DFHR and may serve as a negative marker for MTX efficacy and retention. Additionally, it can also be hypothesized that adding drugs that inhibit AOX can have a synergistic effect on MTX. A preliminary study was performed using folic acid which is usually prescribed to patients on MTX (11). Folic acid was shown to block AOX leading to decrease in 7-OH-MTX, a similar effect identified in a 2003 study using cyclosporin A prescribed with MTX (99). In 2019, Morgan and Baggott hypothesized adding raloxifene (used to prevent osteoporosis in post-menopausal women) to MTX could decrease 7-OH-MTX formation, since raloxifene has a potent inhibitory effects on AOX (100,101). The caveats to this hypothesis however are that it needs to be investigated experimentally and the potential increase in MTX-related toxicity should not outweigh clinical benefits.

RA synovium

A study from 1989 suggests that by 4–24 hours after oral or intravenous MTX dosing, synovial fluid of RA patients had similar concentrations of MTX to that in plasma (102). Similar to mucosa, due to the invasive nature of synovial biopsies, it is underexplored and is not an ideal matrix for TDM.

Interindividual variations of MTX-PG_1–5 and response associations

Multiple prospective cohort studies have investigated associations between erythrocyte MTX-PG_1–5 concentrations and therapy efficacy (83,84,103-105). The large range of MTX-PG_1–5 concentrations indicated large inter-individual variations (106).

RA and JIA

Quantification of erythrocyte MTX-PG_1–5 concentrations identified inter-individual and longitudinal variations in the concentrations, enabling the use of erythrocyte MTX-PG_1–5 for TDM studies. Independent studies conducted in multiple countries associated erythrocyte MTX-PG_3&total accumulation with response after 3 to 6 months of treatment (83,84,107,108). A meta-analysis including 21 studies quantified erythrocyte MTX-PG_1–5 levels in RA, JIA, IBD (ulcerative colitis and CD), and atopic dermatitis (including psoriasis) (2). In most of these 21 studies, a higher accumulation of long-chain MTX-PG_3–5 correlated with lower disease activity. Conversely, a cross-sectional cohort of established RA patients on long-term MTX showed a higher erythrocyte MTX-PG₅ fraction in patients with high disease activity despite correcting for confounders such as age, kidney function [estimated glomerular filtration rate (eGFR)], and MTX dose differences (108). Similarly, a prospective study in 117 established RA patients, did not find an association with between erythrocyte MTX-PG_3&total concentrations and 6 months therapy efficacy (109). Interestingly, in a Japanese RA cohort (n=79), despite the maximum MTX dose of only 16 mg/week, concentrations of erythrocyte MTX-PG_total were similar to those in patients treated with a higher concentration of 25 mg/week, hinting at a possible influence of folate-pathway gene polymorphisms determining the accumulation of MTX-PG_1–5 (105). Despite inter-study variations, the meta-analysis study suggests using MTX-PG₃ or MTX-PG_total as TDM markers (2).

Recently, the “methotrexate monitoring” (MeMo) clinical study randomized 40 early RA patients into oral or subcutaneous treated with MTX (and prednisolone) (32). Though not powered for response correlations, higher concentrations of individual erythrocyte MTX-PG_3–5 concentrations were associated with lower disease activity. Significantly higher erythrocyte MTX-PG_3–5 concentrations at months 1 and 2 were found after subcutaneous MTX use, which equalized by month 3. This study suggested that the early higher intracellular MTX concentrations because of the higher bioavailability of subcutaneous MTX is short-lived. The MeMo study also measured PBMC-MTX-PGs in 40 RA patients. Notably, MTX-PG_total levels in PBMCs were 10–20-fold higher than in erythrocytes on a per cell basis. Interestingly, longitudinally (months 1, 2, 3, and 6), PBMC-MTX-PGs concentrations remained rather constant, which may be related with the relatively short lifespan of PBMCs and their repopulation in circulation after weekly MTX dosing. The MeMo study showed no associations of PBMC-MTX-PG levels with AEs (32).

CD

The results of a clinical trial similar to MeMo in patients with CD, treated with subcutaneous MTX were reported (85). Erythrocyte MTX-PG₃ after 3 months treatment was significantly associated with MTX drug survival. Forty-two percent of the CD patients experienced GI intolerances during MTX treatment. Notably, younger age, female sex, and no alcohol consumption were predictors of GI intolerance. High erythrocyte MTX-PG_3–5 levels were associated with decreased fecal calprotectin concentrations at 12, 25, and 51 weeks, while higher MTXPG_1&3 were associated with a biochemical response at week 12 (85). Furthermore, PBMC-MTX-PGs were measured in six CD patients and confirmed the PBMC-MTX-PG accumulation pattern observed in the MeMo clinical study for RA patients (93).

Other IMIDs

The evidence for erythrocyte MTX-PGs and response associations in other IMIDs is meager compared to RA. A retrospective study in patients with sarcoidosis noted a significant negative association between MTX-PG₃, MTX-PG₄ and changes in maximum standardized uptake value (SUV_max) on the fluorodeoxyglucose (FDG)-positron emission tomography (PET)/computed tomography (CT). Patients with ≥30% decrease in SUV_max are responders (103). In myasthenia gravis patients treated for 12 months with MTX, Pasnoor et al. showed erythrocyte MTX-PG_4&5 concentrations correlated with the different outcome measures (110). In psoriasis or PsA patients, no association between erythrocyte MTX-PG₃ and response was noted (86,111). Interestingly, in a retrospective cohort of children with atopic dermatitis (n=30), erythrocyte MTX-PG₃ level (at 12 weeks) was higher in responders, correlating with response (111).

Therapeutic window for MTX-PG_1–5

To derive useful clinical input from the erythrocyte MTX-PG_1–5 concentrations, a reliable TDM cut-off and which individual MTX-PGs species should be measured must be chosen. MTX-PG_1&5 are subject to either variations due to plasma contaminations or very low concentrations, respectively. Among MTX-PG_2–4, MTX-PG_3&total are good options for clinical implementation (2). The clinical studies building our meta-analysis had a wide the cut-off for erythrocyte MTX-PG_total ranging from 27.3 to 83.3 nmol/L, due to number of responders, MTX-PGs detection methods, and time of MTX-PGs measurement among others (2). In the commercial center at Exagen, the therapeutic range of MTX-PG_total was defined between 20 to ≥60 nmol/L packed erythrocytes (87). At Labcorp, the other commercial center for quantifying MTX-PGs, the 3-month cut-off for therapeutic effect was set higher at ≥74 nmol/L packed erythrocytes of MTX-PG_total (88).

A prospective study in MTX-naïve Japanese RA patients (n=79) indicated the cut-off for MTX treatment efficacy (DAS-28 improvement ≥1.2) between 80 and 100 nmol/L packed erythrocytes of MTX-PG_total (105). This study identified MTX-PG_total >105 nmol/L packed erythrocytes as a cut-off for risk of liver damage (transaminases of ≥50 IU/L) (105). NONMEM modeling of MTX-PG_1–5 using accumulation using data from 395 RA patients (including 3,401 MTX-PG_1–5 concentrations and 1,377 DAS-28 measurements), identified an optimal therapeutic response window for MTX-PG_3–5-sum at between 47 and 78 nmol/L packed erythrocytes after 3 months of MTX treatment (112). In case of insufficient response despite an erythrocyte-MTX-PG_total concentration >78 nmol/L packed erythrocytes, other treatment options can be considered. If concentrations of MTX-PG_3–5-sum are below 9.15 nmol/L packed erythrocytes after 1 month of MTX treatment, a dose increase should first be considered (112). This model-informed precision dosing should first be clinically validated, but serves a promising strategy to guide personalized MTX dosing schemes. Although the therapeutic window was identified in RA, it is also promising to evaluate the potential use of this therapeutic window in other IMIDs (Figure 3).

Figure 3 Response prediction and TDM of MTX-PGs. Prediction models can predict response at baseline, helping to personalize the most optimal treatment plan while TDM offers real-time guide of treatment. Using model-informed precision dosing, treatment plans can be tailored to each patient to obtain the most optimal efficacy while balancing toxicity. For example, if the erythrocyte MTX-PG_3–5-sum is above or below the therapeutic window, the dose can be adjusted to reach the target window, while also monitoring for hepatotoxicity/sub-optimal erythrocyte MTX-PG_3–5-sum concentrations. BMI, body mass index; DMARD, disease-modifying anti-rheumatic drug; ESR, erythrocyte sedimentation rate; HAQ, Health Assessment Questionnaire; MTX-PG, methotrexate polyglutamate; MTX-PG_3–5-sum, sum of concentrations of MTX-PG₃ to MTX-PG₅; TDM, therapeutic drug monitoring; TJC28, 28-joint tender joint count.

TDM implementation

Implementing proactive-TDM can maximize benefits for the patients. Prospective randomized clinical trials are needed to evaluate the clinical efficacy of model-informed precision dosing strategies against standard dosing. Integration of efficacy and toxicity monitoring with artificial intelligence (AI) algorithms can increase the role of patients and make them central to their treatment. With machine learning approaches, patients can get personalized advice and possible diagnostic tests to alter dose, switch medication or monitor intolerances (Figure 3).

Conclusions

Extensive research has been performed and is currently ongoing to improve MTX treatment in IMIDs. Being both cost-effective and clinically effective, MTX is a primary choice in many IMIDs. Many studies have investigated TDM of MTX and have also tried to determine what factors can influence cellular MTX-PG accumulation and how this impacts response. Importantly, state-of-the-art analytical tools as LC-MS/MS allow accurate measurements of very low levels of MTX-PGs not only in abundant erythrocytes but also in low numbers (a few million) of immune cells (PBMCs/tissues) involved in IMID disease pathology. These analytical advances can be positioned to guide future clinically-directed laboratory studies aiming at personalized treatments for IMIDs. Additionally, these analytical tools can assist in experimental studies aiming at increasing MTX-PG cellular retention, e.g., by therapeutic interventions with drugs preventing deactivation of MTX or blocking MTX efflux. Further studies are needed to demonstrate the feasibility of utilizing PBMC-MTX-PG_1–5 concentrations for TDM approaches. Being the relevant target cells in IMIDs, investigating effects of MTX and MTX-PG_1–5 on PBMCs can reveal novel insights in response determinants. Ideally, the implementation of these TDM strategies of predicting response as well as timely course of drug dose adaptations and/or drug switch may assist in reducing disease progression for individual patients. To this end, clinical implementation of MTX (non)response prediction models and TDM cut-offs are warranted, initially in RA and at a later stage for other IMIDs treated with MTX.

Acknowledgments

None.

Footnote

Reporting Checklist: The authors have completed the Narrative Review reporting checklist. Available at https://jlpm.amegroups.com/article/view/10.21037/jlpm-25-7/rc

Funding: None.

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://jlpm.amegroups.com/article/view/10.21037/jlpm-25-7/coif). R.d.J. has a non-financial advisory relationship with the start-up Amplio Pharma. The other authors have no conflicts of interest to declare.

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.

Open Access Statement: This is an Open Access article distributed in accordance with the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License (CC BY-NC-ND 4.0), which permits the non-commercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the formal publication through the relevant DOI and the license). See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

References

1. Brown PM, Pratt AG, Isaacs JD. Mechanism of action of methotrexate in rheumatoid arthritis, and the search for biomarkers. Nat Rev Rheumatol 2016;12:731-42. [Crossref] [PubMed]
2. van de Meeberg MM, Hebing RCF, Nurmohamed MT, et al. A meta-analysis of methotrexate polyglutamates in relation to efficacy and toxicity of methotrexate in inflammatory arthritis, colitis and dermatitis. Br J Clin Pharmacol 2023;89:61-79. [Crossref] [PubMed]
3. Obi ON, Saketkoo LA, Russell AM, et al. Sarcoidosis: Updates on therapeutic drug trials and novel treatment approaches. Front Med (Lausanne) 2022;9:991783. [Crossref] [PubMed]
4. Prado MB Jr, Adiao KJB. Methotrexate in generalized myasthenia gravis: a systematic review. Acta Neurol Belg 2023;123:1679-91. [Crossref] [PubMed]
5. Inoue K, Yuasa H. Molecular basis for pharmacokinetics and pharmacodynamics of methotrexate in rheumatoid arthritis therapy. Drug Metab Pharmacokinet 2014;29:12-9. [Crossref] [PubMed]
6. Bluett J, Riba-Garcia I, Verstappen SMM, et al. Development and validation of a methotrexate adherence assay. Ann Rheum Dis 2019;78:1192-7. [Crossref] [PubMed]
7. Hoekstra M, Haagsma C, Neef C, et al. Bioavailability of higher dose methotrexate comparing oral and subcutaneous administration in patients with rheumatoid arthritis. J Rheumatol 2004;31:645-8.
8. Frigerio B, Bizzoni C, Jansen G, et al. Folate receptors and transporters: biological role and diagnostic/therapeutic targets in cancer and other diseases. J Exp Clin Cancer Res 2019;38:125. [Crossref] [PubMed]
9. Cronstein BN, Aune TM. Methotrexate and its mechanisms of action in inflammatory arthritis. Nat Rev Rheumatol 2020;16:145-54. [Crossref] [PubMed]
10. Raz S, Stark M, Assaraf YG. Folylpoly-γ-glutamate synthetase: A key determinant of folate homeostasis and antifolate resistance in cancer. Drug Resist Updat 2016;28:43-64. [Crossref] [PubMed]
11. Baggott JE, Morgan SL. Methotrexate catabolism to 7-hydroxymethotrexate in rheumatoid arthritis alters drug efficacy and retention and is reduced by folic acid supplementation. Arthritis Rheum 2009;60:2257-61. [Crossref] [PubMed]
12. Seither RL, Rape TJ, Goldman ID. Further studies on the pharmacologic effects of the 7-hydroxy catabolite of methotrexate in the L1210 murine leukemia cell. Biochem Pharmacol 1989;38:815-22. [Crossref] [PubMed]
13. Hetland ML, Haavardsholm EA, Rudin A, et al. Active conventional treatment and three different biological treatments in early rheumatoid arthritis: phase IV investigator initiated, randomised, observer blinded clinical trial. BMJ 2020;371:m4328. [Crossref] [PubMed]
14. Hazlewood GS, Barnabe C, Tomlinson G, et al. Methotrexate monotherapy and methotrexate combination therapy with traditional and biologic disease modifying antirheumatic drugs for rheumatoid arthritis: abridged Cochrane systematic review and network meta-analysis. BMJ 2016;353:i1777. [Crossref] [PubMed]
15. van de Meeberg MM, Seinen ML, Fidder HH, et al. Subcutaneous administration, higher age and lower renal function are associated with erythrocyte methotrexate accumulation in Crohn's disease: a cross-sectional study. BMC Gastroenterol 2022;22:365. [Crossref] [PubMed]
16. Muller IB, Hebing RF, Jansen G, et al. Personalized medicine in rheumatoid arthritis: methotrexate polyglutamylation revisited. Expert Rev Precis Med Drug Dev 2018;3:331-4.
17. Smolen JS, Landewé RBM, Bergstra SA, et al. EULAR recommendations for the management of rheumatoid arthritis with synthetic and biological disease-modifying antirheumatic drugs: 2022 update. Ann Rheum Dis 2023;82:3-18. [Crossref] [PubMed]
18. Kameda H, Yamaoka K, Yamanishi Y, et al. Japan College of Rheumatology guidance for the use of methotrexate in patients with rheumatoid arthritis: Secondary publication. Mod Rheumatol 2023;34:1-10. [Crossref] [PubMed]
19. Gossec L, Kerschbaumer A, Ferreira RJO, et al. EULAR recommendations for the management of psoriatic arthritis with pharmacological therapies: 2023 update. Ann Rheum Dis 2024;83:706-19. [Crossref] [PubMed]
20. Nast A, Smith C, Spuls PI, et al. EuroGuiDerm Guideline on the systemic treatment of Psoriasis vulgaris - Part 1: treatment and monitoring recommendations. J Eur Acad Dermatol Venereol 2020;34:2461-98. [Crossref] [PubMed]
21. Lichtenstein GR, Loftus EV, Isaacs KL, et al. ACG Clinical Guideline: Management of Crohn's Disease in Adults. Am J Gastroenterol 2018;113:481-517. Erratum in: Am J Gastroenterol 2018;113:1101. [Crossref] [PubMed]
22. Łodyga M, Eder P, Gawron-Kiszka M, et al. Guidelines for the management of patients with Crohn's disease. Recommendations of the Polish Society of Gastroenterology and the Polish National Consultant in Gastroenterology. Prz Gastroenterol 2021;16:257-96. [Crossref] [PubMed]
23. Onel KB, Horton DB, Lovell DJ, et al. 2021 American College of Rheumatology Guideline for the Treatment of Juvenile Idiopathic Arthritis: Recommendations for Nonpharmacologic Therapies, Medication Monitoring, Immunizations, and Imaging. Arthritis Rheumatol 2022;74:570-85. [Crossref] [PubMed]
24. Ezhilarasan D. Hepatotoxic potentials of methotrexate: Understanding the possible toxicological molecular mechanisms. Toxicology 2021;458:152840. [Crossref] [PubMed]
25. Hamed KM, Dighriri IM, Baomar AF, et al. Overview of Methotrexate Toxicity: A Comprehensive Literature Review. Cureus 2022;14:e29518. [Crossref] [PubMed]
26. Jansen G. Folate supplementations for methotrexate therapies in cancer and arthritis: rationales revisited. J Mol Clin Med 2022;5:1.
27. Vermeer E, Hebing RCF, van de Meeberg MM, et al. Oral Versus Subcutaneous Methotrexate in Immune-Mediated Inflammatory Disorders: an Update of the Current Literature. Curr Rheumatol Rep 2023;25:276-84. [Crossref] [PubMed]
28. Stamp LK. A Pilot Randomized Controlled Double-Blind Trial of High- Versus Low-Dose Weekly Folic Acid in People With Rheumatoid Arthritis Receiving Methotrexate. J Clin Rheumatol 2019;25:284-7. [Crossref] [PubMed]
29. Dhir V, Sandhu A, Kaur J, et al. Comparison of two different folic acid doses with methotrexate--a randomized controlled trial (FOLVARI Study). Arthritis Res Ther 2015;17:156. [Crossref] [PubMed]
30. Liu M, Zhang Z, Zhou C, et al. Relationship of several serum folate forms with the risk of mortality: A prospective cohort study. Clin Nutr 2021;40:4255-62. [Crossref] [PubMed]
31. Brooks PJ, Spruill WJ, Parish RC, et al. Pharmacokinetics of methotrexate administered by intramuscular and subcutaneous injections in patients with rheumatoid arthritis. Arthritis Rheum 1990;33:91-4. [Crossref] [PubMed]
32. Hebing RC, Lin M, Bulatovic Calasan M, et al. Pharmacokinetics of oral and subcutaneous methotrexate in red and white blood cells in patients with early rheumatoid arthritis: the methotrexate monitoring trial. Ann Rheum Dis 2023;82:460-7. [Crossref] [PubMed]
33. Hoekstra M, Haagsma C, Neef C, et al. Splitting high-dose oral methotrexate improves bioavailability: a pharmacokinetic study in patients with rheumatoid arthritis. J Rheumatol 2006;33:481-5.
34. Pandya S, Aggarwal A, Misra R. Methotrexate twice weekly vs once weekly in rheumatoid arthritis: a pilot double-blind, controlled study. Rheumatol Int 2002;22:1-4. [Crossref] [PubMed]
35. Rodríguez-Báez AS, Huerta-García AP, Medellín-Garibay SE, et al. Disease activity and therapeutic drug monitoring of polyglutamates of methotrexate after daily or weekly administration of low-dose methotrexate in patients recently diagnosed with rheumatoid arthritis. Basic Clin Pharmacol Toxicol 2022;130:644-54. [Crossref] [PubMed]
36. Dhaon P, Das SK, Srivastava R, et al. Oral Methotrexate in split dose weekly versus oral or parenteral Methotrexate once weekly in Rheumatoid Arthritis: a short-term study. Int J Rheum Dis 2018;21:1010-7. [Crossref] [PubMed]
37. Prasad CB, Dhir V, Gupta R, et al. Comparison of two dosing schedules for oral methotrexate (split-dose versus single-dose) once weekly in patients with active rheumatoid arthritis: a multicenter, open label, parallel group, randomized controlled trial (SMART study). Arthritis Rheumatol 2023;75:3123-5.
38. van der Heijden JW, Oerlemans R, Tak PP, et al. Involvement of breast cancer resistance protein expression on rheumatoid arthritis synovial tissue macrophages in resistance to methotrexate and leflunomide. Arthritis Rheum 2009;60:669-77. [Crossref] [PubMed]
39. Muto S, Minamitani N, Ogura T, et al. Good response to methotrexate is associated with a decrease in the gene expression of ABCG2, a drug transporter, in patients with rheumatoid arthritis. Mod Rheumatol 2021;31:1079-86. [Crossref] [PubMed]
40. Oerlemans R, Blits M, Dijkmans BAC, et al. Expression profiling of ABC transporters in peripheral blood lymphocytes and monocyte-derived macrophages of rheumatoid arthritis patients. J Mol Clin Med 2020;3:47-60.
41. Stark M, Wichman C, Avivi I, et al. Aberrant splicing of folylpolyglutamate synthetase as a novel mechanism of antifolate resistance in leukemia. Blood 2009;113:4362-9. [Crossref] [PubMed]
42. Wojtuszkiewicz A, Raz S, Stark M, et al. Folylpolyglutamate synthetase splicing alterations in acute lymphoblastic leukemia are provoked by methotrexate and other chemotherapeutics and mediate chemoresistance. Int J Cancer 2016;138:1645-56. [Crossref] [PubMed]
43. Muller IB, Lin M, Lems WF, et al. Association of altered folylpolyglutamate synthetase pre-mRNA splicing with methotrexate unresponsiveness in early rheumatoid arthritis. Rheumatology (Oxford) 2021;60:1273-81. [Crossref] [PubMed]
44. Muller IB, Lin M, de Jonge R, et al. Methotrexate Provokes Disparate Folate Metabolism Gene Expression and Alternative Splicing in Ex Vivo Monocytes and GM-CSF- and M-CSF-Polarized Macrophages. Int J Mol Sci 2023;24:9641. [Crossref] [PubMed]
45. Taylor JC, Bongartz T, Massey J, et al. Genome-wide association study of response to methotrexate in early rheumatoid arthritis patients. Pharmacogenomics J 2018;18:528-38. [Crossref] [PubMed]
46. Myasoedova E, Athreya AP, Crowson CS, et al. Toward Individualized Prediction of Response to Methotrexate in Early Rheumatoid Arthritis: A Pharmacogenomics-Driven Machine Learning Approach. Arthritis Care Res (Hoboken) 2022;74:879-88. [Crossref] [PubMed]
47. Owen SA, Hider SL, Martin P, et al. Genetic polymorphisms in key methotrexate pathway genes are associated with response to treatment in rheumatoid arthritis patients. Pharmacogenomics J 2013;13:227-34. [Crossref] [PubMed]
48. Moya P, Salazar J, Arranz MJ, et al. Methotrexate pharmacokinetic genetic variants are associated with outcome in rheumatoid arthritis patients. Pharmacogenomics 2016;17:25-9. [Crossref] [PubMed]
49. Muralidharan N, Sundaram R, Kodidela S, et al. Folyl polyglutamate synthethase (FPGS) gene polymorphisms may influence methotrexate adverse events in South Indian Tamil Rheumatoid Arthritis patients. Pharmacogenomics J 2020;20:342-9. [Crossref] [PubMed]
50. Warren RB, Smith RL, Campalani E, et al. Outcomes of methotrexate therapy for psoriasis and relationship to genetic polymorphisms. Br J Dermatol 2009;160:438-41. [Crossref] [PubMed]
51. Hinks A, Moncrieffe H, Martin P, et al. Association of the 5-aminoimidazole-4-carboxamide ribonucleotide transformylase gene with response to methotrexate in juvenile idiopathic arthritis. Ann Rheum Dis 2011;70:1395-400.
52. Curtis JR, van der Helm-van Mil AH, Knevel R, et al. Validation of a novel multibiomarker test to assess rheumatoid arthritis disease activity. Arthritis Care Res (Hoboken) 2012;64:1794-803. [Crossref] [PubMed]
53. Hambardzumyan K, Bolce RJ, Wallman JK, et al. Serum Biomarkers for Prediction of Response to Methotrexate Monotherapy in Early Rheumatoid Arthritis: Results from the SWEFOT Trial. J Rheumatol 2019;46:555-63. [Crossref] [PubMed]
54. Maciejewski M, Sands C, Nair N, et al. Prediction of response of methotrexate in patients with rheumatoid arthritis using serum lipidomics. Sci Rep 2021;11:7266. [Crossref] [PubMed]
55. Qiu Q, Deng J, Deng H, et al. Association of the characteristics of the blood metabolome and gut microbiome with the outcome of methotrexate therapy in psoriasis. Front Immunol 2022;13:937539. [Crossref] [PubMed]
56. Funk RS, Becker ML. Metabolomic Profiling Identifies Exogenous and Microbiota-Derived Metabolites as Markers of Methotrexate Efficacy in Juvenile Idiopathic Arthritis. Front Pharmacol 2021;12:768599. [Crossref] [PubMed]
57. Gosselt HR, Muller IB, Jansen G, et al. Identification of Metabolic Biomarkers in Relation to Methotrexate Response in Early Rheumatoid Arthritis. J Pers Med 2020;10:271. [Crossref] [PubMed]
58. Valerino DM, Johns DG, Zaharko DS, et al. Studies of the metabolism of methotrexate by intestinal flora. I. Identification and study of biological properties of the metabolite 4-amino-4-deoxy-N 10 -methylpteroic acid. Biochem Pharmacol 1972;21:821-31. [Crossref] [PubMed]
59. Artacho A, Isaac S, Nayak R, et al. The Pretreatment Gut Microbiome Is Associated With Lack of Response to Methotrexate in New-Onset Rheumatoid Arthritis. Arthritis Rheumatol 2021;73:931-42. [Crossref] [PubMed]
60. AmendLHuangKDSivapornnukulPComparative analysis of the treatment-naive microbiome across rheumatic diseases to predict MTX treatment response.bioRxiv:2024.07.31.605913 [Preprint]. 2024. Available online: https://www.biorxiv.org/content/10.1101/2024.07.31.605913.abstract
61. van Dijkhuizen EHP, Del Chierico F, Malattia C, et al. Microbiome Analytics of the Gut Microbiota in Patients With Juvenile Idiopathic Arthritis: A Longitudinal Observational Cohort Study. Arthritis Rheumatol 2019;71:1000-10. [Crossref] [PubMed]
62. Jenko B, Lusa L, Tomsic M, et al. Clinical-pharmacogenetic predictive models for MTX discontinuation due to adverse events in rheumatoid arthritis. Pharmacogenomics J 2017;17:412-8. [Crossref] [PubMed]
63. Sandhu A, Dhir V, Ahmad S, et al. Clinico-genetic model to predict methotrexate intolerance in rheumatoid arthritis. Clin Rheumatol 2020;39:201-6. [Crossref] [PubMed]
64. Onishi A, Kamitsuji S, Nishida M, et al. Genetic and clinical prediction models for the efficacy and hepatotoxicity of methotrexate in patients with rheumatoid arthritis: a multicenter cohort study. Pharmacogenomics J 2020;20:433-42. [Crossref] [PubMed]
65. Jenko B, Tomšič M, Jekić B, et al. Clinical Pharmacogenetic Models of Treatment Response to Methotrexate Monotherapy in Slovenian and Serbian Rheumatoid Arthritis Patients: Differences in Patient's Management May Preclude Generalization of the Models. Front Pharmacol 2018;9:20. [Crossref] [PubMed]
66. Eektimmerman F, Allaart CF, Hazes JM, et al. Validation of a clinical pharmacogenetic model to predict methotrexate nonresponse in rheumatoid arthritis patients. Pharmacogenomics 2019;20:85-93. [Crossref] [PubMed]
67. Sergeant JC, Hyrich KL, Anderson J, et al. Prediction of primary non-response to methotrexate therapy using demographic, clinical and psychosocial variables: results from the UK Rheumatoid Arthritis Medication Study (RAMS). Arthritis Res Ther 2018;20:147. [Crossref] [PubMed]
68. Teitsma XM, Jacobs JWG, Welsing PMJ, et al. Inadequate response to treat-to-target methotrexate therapy in patients with new-onset rheumatoid arthritis: development and validation of clinical predictors. Ann Rheum Dis 2018;77:1261-7. [Crossref] [PubMed]
69. de Rotte MCFJ, Pluijm SMF, de Jong PHP, et al. Development and validation of a prognostic multivariable model to predict insufficient clinical response to methotrexate in rheumatoid arthritis. PLoS One 2018;13:e0208534. [Crossref] [PubMed]
70. Gosselt HR, Verhoeven MMA, de Rotte MCFJ, et al. Validation of a Prognostic Multivariable Prediction Model for Insufficient Clinical Response to Methotrexate in Early Rheumatoid Arthritis and Its Clinical Application in Evidencio. Rheumatol Ther 2020;7:837-50. [Crossref] [PubMed]
71. Gosselt HR, Verhoeven MMA, Bulatović-Ćalasan M, et al. Complex Machine-Learning Algorithms and Multivariable Logistic Regression on Par in the Prediction of Insufficient Clinical Response to Methotrexate in Rheumatoid Arthritis. J Pers Med 2021;11:44. [Crossref] [PubMed]
72. Bulatovic M, Heijstek MW, Van Dijkhuizen EH, et al. Prediction of clinical non-response to methotrexate treatment in juvenile idiopathic arthritis. Ann Rheum Dis 2012;71:1484-9. [Crossref] [PubMed]
73. van Dijkhuizen EH, Bulatović Ćalasan M, Pluijm SM, et al. Prediction of methotrexate intolerance in juvenile idiopathic arthritis: a prospective, observational cohort study. Pediatr Rheumatol Online J 2015;13:5. [Crossref] [PubMed]
74. Mo X, Chen X, Li H, et al. Early and Accurate Prediction of Clinical Response to Methotrexate Treatment in Juvenile Idiopathic Arthritis Using Machine Learning. Front Pharmacol 2019;10:1155. [Crossref] [PubMed]
75. Mulder MB, Huisman R, Engels FK, et al. Therapeutic Drug Monitoring of Methotrexate in Plasma Using Ultra High-Performance Liquid Chromatography-Electrospray Ionization-Tandem Mass Spectrometry: Necessary After Administration of Glucarpidase in Methotrexate Intoxications. Ther Drug Monit 2018;40:383-5. [Crossref] [PubMed]
76. Ra AG, Littlewood Z, Barton A, et al. The use of serum methotrexate level as an assessment of biochemical drug adherence in psoriasis. Br J Dermatol 2024;190:927-9. [Crossref] [PubMed]
77. Möhlmann JE, de Roock S, Egas AC, et al. Adherence to low-dose methotrexate in children with juvenile idiopathic arthritis using a sensitive methotrexate assay. Pediatr Rheumatol Online J 2024;22:52. [Crossref] [PubMed]
78. Inoue S, Hashiguchi M, Kawai S, et al. Erythrocyte methotrexate-polyglutamate assay using fluorescence polarization immunoassay technique: application to the monitoring of patients with rheumatoid arthritis. Yakugaku Zasshi 2009;129:1001-5. [Crossref] [PubMed]
79. Dervieux T, Orentas Lein D, Marcelletti J, et al. HPLC determination of erythrocyte methotrexate polyglutamates after low-dose methotrexate therapy in patients with rheumatoid arthritis. Clin Chem 2003;49:1632-41. [Crossref] [PubMed]
80. van Haandel L, Becker ML, Williams TD, et al. Comprehensive quantitative measurement of folate polyglutamates in human erythrocytes by ion pairing ultra-performance liquid chromatography/tandem mass spectrometry. Rapid Commun Mass Spectrom 2012;26:1617-30. [Crossref] [PubMed]
81. den Boer E, Meesters RJ, van Zelst BD, et al. Measuring methotrexate polyglutamates in red blood cells: a new LC-MS/MS-based method. Anal Bioanal Chem 2013;405:1673-81. [Crossref] [PubMed]
82. Daraghmeh DN, Moghaddami M, Bobrovskaya L, et al. Quantitation of methotrexate polyglutamates in human whole blood, erythrocytes and leukocytes collected via venepuncture and volumetric absorptive micro-sampling: a green LC-MS/MS-based method. Anal Bioanal Chem 2022;414:6029-46. [Crossref] [PubMed]
83. Ćalasan MB, den Boer E, de Rotte MC, et al. Methotrexate polyglutamates in erythrocytes are associated with lower disease activity in juvenile idiopathic arthritis patients. Ann Rheum Dis 2015;74:402-7. [Crossref] [PubMed]
84. de Rotte MC, den Boer E, de Jong PH, et al. Methotrexate polyglutamates in erythrocytes are associated with lower disease activity in patients with rheumatoid arthritis. Ann Rheum Dis 2015;74:408-14. [Crossref] [PubMed]
85. van de Meeberg MM, Fidder HH, Oldenburg B, et al. Therapeutic drug monitoring of methotrexate in patients with Crohn's disease. Aliment Pharmacol Ther 2023;58:1151-62. [Crossref] [PubMed]
86. Woolf RT, West SL, Arenas-Hernandez M, et al. Methotrexate polyglutamates as a marker of patient compliance and clinical response in psoriasis: a single-centre prospective study. Br J Dermatol 2012;167:165-73. [Crossref] [PubMed]
87. Exagen. AVISE MTX The Methotrexate Polyglutamates Test. Accessed February 5, 2025. Available online: https://exagen.com/tests/mtx/
88. Labcorp. Methotrexate Polyglutamates. Accessed February 5, 2025. Available online: https://www.labcorp.com/tests/504104/methotrexate-polyglutamates
89. Hawwa AF, Albawab A, Rooney M, et al. A novel dried blood spot-LCMS method for the quantification of methotrexate polyglutamates as a potential marker for methotrexate use in children. PLoS One 2014;9:e89908. [Crossref] [PubMed]
90. Brady K, Qu Y, Stimson D, et al. Transition of Methotrexate Polyglutamate Drug Monitoring Assay from Venipuncture to Capillary Blood-Based Collection Method in Rheumatic Diseases. J Appl Lab Med 2019;4:40-9. [Crossref] [PubMed]
91. van Haandel L, Leeder JS, Becker ML. Measurement of methotrexate metabolites in peripheral blood mononuclear cells in juvenile idiopathic arthritis: a more relevant cellular biomarker for drug response? Arthritis Rheum 2011;63:S95. [Crossref] [PubMed]
92. Sundaresan J, Lin M, Jansen G, et al. Unique methotrexate polyglutamates distributions in peripheral blood mononuclear cells of rheumatoid arthritis patients: Development and validation of a UPLC-MS/MS method. J Pharm Biomed Anal 2025;262:116882. [Crossref] [PubMed]
93. van de Meeberg MM, Sundaresan J, Lin M, et al. Methotrexate accumulation in target intestinal mucosa and white blood cells differs from non-target red blood cells of patients with Crohn's disease. Basic Clin Pharmacol Toxicol 2024;135:308-20. [Crossref] [PubMed]
94. Tough DF, Sprent J. Lifespan of lymphocytes. Immunol Res 1995;14:1-12. [Crossref] [PubMed]
95. Patel AA, Zhang Y, Fullerton JN, et al. The fate and lifespan of human monocyte subsets in steady state and systemic inflammation. J Exp Med 2017;214:1913-23. [Crossref] [PubMed]
96. Muller IB, Lin M, Struys EA, et al. Development and Validation of a Sensitive UHPLC-MS/MS-Based Method for the Analysis of Folylpolyglutamate Synthetase Enzymatic Activity in Peripheral Blood Mononuclear Cells: Application in Rheumatoid Arthritis and Leukemia Patients. Ther Drug Monit 2019;41:598-606. [Crossref] [PubMed]
97. Perez-Garcia LF, Röder E, Krijthe BP, et al. Is methotrexate safe for men with an immune-mediated inflammatory disease and an active desire to become a father? Results of a prospective cohort study (iFAME-MTX). Ann Rheum Dis 2023;82:1068-75. [Crossref] [PubMed]
98. Klapkova E, Kukacka J, Kotaska K, et al. The influence of 7-OH methotrexate metabolite on clinical relevance of methotrexate determination. Clin Lab 2011;57:599-606.
99. Fox RI, Morgan SL, Smith HT, et al. Combined oral cyclosporin and methotrexate therapy in patients with rheumatoid arthritis elevates methotrexate levels and reduces 7-hydroxymethotrexate levels when compared with methotrexate alone. Rheumatology (Oxford) 2003;42:989-94. [Crossref] [PubMed]
100. Morgan SL, Baggott JE. The importance of inhibition of a catabolic pathway of methotrexate metabolism in its efficacy for rheumatoid arthritis. Med Hypotheses 2019;122:10-5. [Crossref] [PubMed]
101. Obach RS, Huynh P, Allen MC, et al. Human liver aldehyde oxidase: inhibition by 239 drugs. J Clin Pharmacol 2004;44:7-19. [Crossref] [PubMed]
102. Tishler M, Caspi D, Graff E, et al. Synovial and serum levels of methotrexate during methotrexate therapy of rheumatoid arthritis. Br J Rheumatol 1989;28:422-3. [Crossref] [PubMed]
103. Janssen Bonás M, Sundaresan J, Keijsers RGM, et al. Methotrexate Polyglutamate Concentrations as a Possible Predictive Marker for Effectiveness of Methotrexate Therapy in Patients with Sarcoidosis: A Pilot Study. Lung 2023;201:617-24. [Crossref] [PubMed]
104. Dervieux T, Zablocki R, Kremer J. Red blood cell methotrexate polyglutamates emerge as a function of dosage intensity and route of administration during pulse methotrexate therapy in rheumatoid arthritis. Rheumatology (Oxford) 2010;49:2337-45. [Crossref] [PubMed]
105. Takahashi C, Kaneko Y, Okano Y, et al. Association of erythrocyte methotrexate-polyglutamate levels with the efficacy and hepatotoxicity of methotrexate in patients with rheumatoid arthritis: a 76-week prospective study. RMD Open 2017;3:e000363. [Crossref] [PubMed]
106. Stamp LK, O'Donnell JL, Chapman PT, et al. Determinants of red blood cell methotrexate polyglutamate concentrations in rheumatoid arthritis patients receiving long-term methotrexate treatment. Arthritis Rheum 2009;60:2248-56. [Crossref] [PubMed]
107. Dervieux T, Greenstein N, Kremer J. Pharmacogenomic and metabolic biomarkers in the folate pathway and their association with methotrexate effects during dosage escalation in rheumatoid arthritis. Arthritis Rheum 2006;54:3095-103. [Crossref] [PubMed]
108. Stamp LK, O'Donnell JL, Chapman PT, et al. Methotrexate polyglutamate concentrations are not associated with disease control in rheumatoid arthritis patients receiving long-term methotrexate therapy. Arthritis Rheum 2010;62:359-68. [Crossref] [PubMed]
109. Sandhu A, Dhir V, Bhatnagar A, et al. High Methotrexate Triglutamate Level Is an Independent Predictor of Adverse Effects in Asian Indian Rheumatoid Arthritis Patients-A Preliminary Study. Ther Drug Monit 2017;39:157-63. [Crossref] [PubMed]
110. Pasnoor M, Heim AJ, Herbelin L, et al. Methotrexate Polyglutamation in a Myasthenia Gravis Clinical Trial. Kans J Med 2020;13:10-3.
111. Rahman SI, Siegfried E, Flanagan KH, et al. The methotrexate polyglutamate assay supports the efficacy of methotrexate for severe inflammatory skin disease in children. J Am Acad Dermatol 2014;70:252-6. [Crossref] [PubMed]
112. Hebing RCF, Bartelink IH, Gosselt HR, et al. Methotrexate Polyglutamates Exposure - Response Modeling in a Large Cohort of Rheumatoid Arthritis Patients Starting Methotrexate. Clin Pharmacol Ther 2023;114:893-903. [Crossref] [PubMed]