Interferon-Driven Tryptophan Metabolism Links Inflammation and Mental Health in Juvenile Dermatomyositis

Abstract

Background

Children with juvenile dermatomyositis (JDM) experience significant mental health burdens, including anxiety and depression. Inflammatory activation can alter tryptophan metabolism, particularly through IDO-driven kynurenine pathway induction, which has been implicated in diseases such as mood disorders and SLE. We therefore investigated skewed tryptophan metabolism and its association with disease activity and mental health symptoms in JDM.

Methods

Serum samples from JDM, juvenile idiopathic arthritis (JIA), and healthy controls (HC) were analysed for tryptophan metabolites by ELISA. Interferon-regulated genes were measured using qPCR. Ability of serum to induce indoleamine 2,3-dioxygenase (IDO) was assessed by flow cytometry in presence or absence of interferon antagonists. Cluster analysis was used to identify subgroups.

Results

In JDM patients, serum tryptophan and serotonin levels were lower, while kynurenine/tryptophan ratios, kynurenic acid, and quinolinic acid levels were higher compared to healthy controls. Metabolites from the kynurenine pathway were correlated with muscle inflammation (CRP, ESR, aldolase and CK), mental health outcomes (PSC-17 and PHQ-9 scores) and disease progression (PGA scores). Further, IDO1 mRNA levels correlated inversely with serotonin levels and positively with type I interferon signature marker MX2 in PBMCs. Serum from JDM patients induced IDO protein expression in monocytes through an IFN-dependent mechanism, which was significantly inhibited by both baricitinib and anifrolumab. Four clinical subgroups were identified.

Conclusions

Our study reveals a novel role of IFN in JDM pathogenesis, specifically in the upregulation of IDO and subsequent skewing of tryptophan metabolism towards kynurenine pathway, which may correspond to poor mental well-being and disease progression. Targeting this pathway may offer therapeutic potential for both disease activity and psychological outcomes for JDM children.

Supplementary Information

The online version contains supplementary material available at 10.1007/s10753-026-02469-8.

Background

Juvenile Dermatomyositis (JDM) is a rare heterogeneous idiopathic inflammatory vasculopathy in children, with an annual incidence of 2 to 4 per one million children, with a girl-to-boy ratio of around 2:1 [1, 2]. The pathogenesis of JDM remains unclear but is thought to be partially driven by innate immunity, with high expression of interferon (IFN) or IFN-regulated genes in peripheral blood, muscle, and skin [3, 4]. Further, children with JDM, as well as children with other inflammatory rheumatic diseases such as pediatric lupus and juvenile idiopathic arthritis (JIA) exhibit elevated levels of inflammatory type I and II IFNs, which correlate with disease severity [3–7]. Treatment for JDM patients is limited, primarily relying on high-dose steroids in combination with other immunosuppressive agents, often necessitating prolonged therapy associated with significant side effects [8, 9].

Children with JDM experience a persistent mental health burden and impaired quality of life, even during disease remission, with anxiety and depression rates ranging from 15 to 65% and suicidal ideation rates from 14 to 34% [10–12]. The underlying cause of mental illness is not known but thought to be linked to inflammation [13, 14]. We have previously demonstrated in SLE that type I IFNs can upregulate indoleamine 2,3-dioxygenase (IDO) [15], the rate-limiting enzyme in the conversion of tryptophan into kynurenine. Tryptophan is an essential amino acid, and dysregulation of its metabolic pathway has been well established in anxiety and depression [16, 17]. Beyond its role in protein synthesis, peripheral tryptophan (less than 5%) undergoes catabolism into serotonin, a key regulator of sleep and mood [18], whereas the majority (more than 95%) is metabolized into kynurenine via the enzyme IDO (Supplementary Figure S1) [19].

Kynurenine is a necessary building block for generating both neurotoxic (quinolinic acid) and neuroprotective (kynurenic acid) components that can interact with N-methyl-D-aspartate receptors (NMDARs), known to contribute to the pathophysiology of many neurological disorders [20, 21]. Altered kynurenine pathway has been linked to cognitive impairment, mental health disorders, such as schizophrenia, major depressive disorder, and bipolar disorder [22, 23], as well as other diseases with underlying inflammatory mechanisms, e.g. diabetes and cancer [24–26]. In addition, we and others have reported an altered kynurenine/tryptophan ratio in patients with adult SLE and JIA, while IDO activity in children with JIA correlates to disease activity [15, 27]. However, the potential role of tryptophan metabolism in the neuropsychiatric manifestations of JDM has not been assessed previously.

In brief, our novel study provides an exploration of the characteristics of tryptophan metabolites in JDM and their associations with disease pathogenesis and co-morbidity, particularly mental health burden. Additionally, we investigate the role of two potential anti-IFN therapeutic agents, baricitinib and anifrolumab, in inhibiting IDO expression.

Methods

Study Population and Ethical Statement

Children with JDM (n = 38), or JIA (n = 12) and healthy children (n = 21) were recruited at Seattle Children’s Hospital. Inclusion criteria for JDM included patients who met the 2017 European League Against Rheumatism/American College of Rheumatology (EULAR/ACR) juvenile idiopathic inflammatory myopathy classification [28], and inclusion criteria for JIA included patients who met the 2001 International League of Associations for Rheumatism (ILAR) JIA classification [29]. Banked serum samples from healthy controls (HCs) (n = 21) were obtained from the Stanley Manne Children’s Research Institute/Ann & Robert H. Lurie Children’s Hospital of Chicago Cure Juvenile Myositis Biorepository/Registry. Among the JDM children, 13 had follow-up visits ranging from 1 to 3 times. We included data from the first two visits in the longitudinal analysis, with a median duration of 18 months (range: 3 to 43 months). Blood samples were collected at the same study visit during which disease activity measures were assessed, including cutaneous dermatomyositis disease area and severity index (CDASI), childhood myositis assessment scale (CMAS), physician global assessment (PGA, including PGA muscle and PGA skin), Manual Muscle Testing 8 (MMT8), creatine kinase (CK), erythrocyte sedimentation rate (ESR), C-reactive protein (CRP), aldolase, aspartate aminotransaminase (AST), alanine aminotransferase (ALT) and lactate dehydrogenase (LDH). Mental health outcomes were determined by questionnaires at the time of biomarker sample collection, including Pediatric Symptom Checklist-17 (PSC-17) for psychosocial impairments and Patient Health Questionnaire-9 (PHQ-9) for depression. Myositis-specific antibodies (MSA) were assessed at the Oklahoma Medical Research Foundation Clinical Immunology Laboratory (immunoprecipitation) as well as Department of Laboratory Medicine and Pathology at University of Washington (euroimmun line blot). Baseline clinical characteristics are presented in Table 1.

Table 1.

Clinical and demographic characteristics of the studied population

	JDM (n = 38)	JIA (n = 12)	HC (n = 21)
Demographic features
Age (years), median (range)	9 (3–19)	12 (7–19)	13 (4–24)
Gender (female), %	60.50	66.70	57.10
Ethnicity, %
White	58.80	77.78	47.60
Hispanic	35.30	11.11	38.10
Black	5.90	11.11	9.50
Asian	2.90	22.22	9.50
Disease characteristics, median (range)
Disease duration (years)	3.5 (0–16)	4 (3–14.5)	N/A
Second visit duration (months)	18 (3–43)	N/A	N/A
CDASI (0–100)	1 (0–9)	N/A	N/A
CMAS (0–52)	50 (41–52)	N/A	N/A
MMT8 (0–150)	145 (126–150)	N/A	N/A
PGA total score (0–10)	2 (0–9)	2 (0–8)	N/A
Aldolase (1–7U/L)	4 (0.90–13)	N/A	N/A
CK (20–215U/L)	98 (29–289)	N/A	N/A
CRP (< 0.8 mg/L)	0.8 (0.5–1.2)	0.7 (0.2–1.1)	N/A
ESR (< 20 mm/h)	9.50 (1–31)	6 (1–20)	N/A
AST (25– 50U/L)	35 (22–678)	33.50 (25–63)	N/A
ALT (5–31U/L)	20 (12–286)	21 (12–92)	N/A
LDH (125– 345U/L)	390 (139–1726)	199.50 (159–240)	N/A
Mental health assessments, median (range)
PHQ-9 (0–27)	4 (0–15)	3 (1–6)	N/A
PSC-17 total scorec(0–34)	7 (0–17)	N/A	N/A
Myositis antibodiesa, %
anti-NXP2	21.05	N/A	N/A
anti-Mi2	23.68	N/A	N/A
anti-TIF1-γ	34.21	N/A	N/A
MSA negative	18.42	N/A	N/A
Treatment, %
immunomodulator	81.58	N/A	N/A
glucocorticoid	57.90	N/A	N/A

^a No patients with test result (n = 20) had anti-MDA5 antibodies

JDM, Juvenile dermatomyositis; JIA, Juvenile idiopathic arthritis; HC, Healthy controls; CDASI, Cutaneous dermatomyositis disease area and severity index; CMAS, Childhood myositis assessment scale; MMT8, Muscle memory test 8; CK, Creatine kinase; CRP, C-reactive protein; ESR, Erythrocyte sedimentation rate; PGA, Physician global assessment; PHQ-9, Patient health questionnaire-9; PSC-17, Pediatric symptom checklist-17

Enzyme-Linked Immunosorbent Assays (ELISAs)

Serum was freshly separated after collection and stored at − 80 °C until use. Serum levels of tryptophan, kynurenine, serotonin, kynurenic acid and quinolinic acid were analyzed using commercial enzyme-linked immunosorbent assay (ELISA) kits according to the manufacturer’s instructions (ImmuSmol) at the Lood laboratory.

Quantitative Real-Time Polymerase Chain Reaction (qRT-PCR)

Total mRNAs from peripheral blood mononuclear cells (PBMCs) were extracted using Quick RNA miniprep Kit (ZymoResearch) following the manufacturer’s protocol. RNA concentrations and purities were measured on a NanoDrop spectrophotometer (Thermo Fisher Scientific). Subsequently, the isolated RNA was reverse-transcribed into cDNA using N High-Capacity cDNA Prep Kit (Thermo Fisher Scientific), and the cDNA library synthesis program was run automatically on a Thermo cycler (25 °C for 10 min, 37 °C for 2 h, 85 °C for 5 min and hold on 4 °C). The relative expression levels of various target genes were quantified by q-PCR employing the Applied Biosystems Fast SYBR green Master mix (Thermo Fisher Scientific). Each well was placed in duplicates, and a three-step Thermal cycling was performed (10 min at 95 °C, 40 cycles at 95 °C for 15 s, and annealing at 59 °C for 1 min). The relative abundance of mRNA was standardized with GAPDH as the internal control. According to the cycle threshold (Ct) value of the samples, the relative quantitative method was used to analyze the results of RT-PCR, and 2-ΔCt was calculated.

The primer pairs (5′−3′) used in this study were: GAPDH (F: GTGAAGGTCGGAGTCAAG; R: TGAGGTCAATGAAGGGGTC); IDO1 (F: GGCTTTGCTCTGCCAAATCC; R: TTCTCAACTCTTTCTCGAAGCTG).

Flow Cytometry

PBMCs from healthy donors were pre-incubated with or without 250 nM baricitinib or anifrolumab (MedChemExpress) for 30 min prior to the addition of 20μL serum from JDM or HC, or 250-unit (U) IFN-α (EMD Millipore) as a positive control. In some experiments, different concentrations of baricitinib or anifrolumab (50 nM and 250 nM) and IFN-α (250U, 1000U, and 2000U) were tested. The cells (1 × 10⁶/mL) were incubated overnight at 37 °C, followed by subsequent fixation, permeabilization (BD Biosciences) and staining of IDO (Invitrogen). Intracellular IDO levels were measured by flow cytometry (Beckman Coulter). FlowJo 10.4 (Tree Star Inc) was used for data analyses. Results are presented as the median fluorescent intensity (MFI) ratio.

Cluster Analysis Methodology

Cluster analysis was conducted using the Partitioning Around Medoids (PAM) algorithm with Gower distance, which accommodates the integration of both categorical and continuous variables. Fifteen variables included in the analysis were selected from all available variables based on their maximal discriminatory potential between clusters, while also considering their clinical relevance to JDM disease activity and severity (Supplementary Table S1). The optimal number of clusters was determined using the silhouette method based on Gower distance (Supplementary Figure S2).

Statistical Analysis

Data analyses were performed using SPSS 27.0 (IBM), GraphPad Prism 10.2.0 (for Mac) and R (version 4.4.0). Categorical variables were analyzed using Fisher’s exact test. Continuous variables were compared using the Mann–Whitney U test between two-groups and the Kruskal–Wallis test were used between more than two groups. Correlations between variables were determined by Spearman’s correlation coefficient. A statistically significant difference was considered when p < 0.05, and asterisks in all figures were used to indicate p values as follows: *p < 0.05, **p < 0.01, *** p < 0.001, and **** p < 0.0001.

Results

Altered Serum Tryptophan Metabolites in JDM

Serum levels of tryptophan and its metabolites in JDM and JIA patients, compared to healthy controls (HC), are presented in Fig. 1. Tryptophan levels were decreased in JDM (p < 0.05, Fig. 1A), kynurenine levels were increased in JIA (p < 0.05, Fig. 1B), with serotonin levels being reduced in both JIA (p < 0.05) and JDM (p < 0.05, Fig. 1C) as compared to HC. The kynurenine/tryptophan ratios were elevated in JDM patients as compared to HC (p < 0.05, Fig. 1D), indicating an enhanced IDO enzyme activity in JDM. Despite no difference in overall kynurenine levels, kynurenine metabolites, e.g. kynurenic acid levels were elevated in JDM patients compared to HC (p < 0.05, Fig. 1E), while quinolinic acid levels showed a marked increase in JDM patients relative to both HC and JIA (p < 0.01 and p < 0.05, respectively, Fig. 1F). Thus, these findings highlight that JDM patients have an altered tryptophan metabolism favoring the kynurenine pathway.

Fig. 1.

Serum levels of Tryptophan metabolites in JDM, JIA patients and healthy controls. Serum levels of (A) Tryptophan; (B) Kynurenine; (C) Serotonin; (D) the Kynurenine/Tryptophan ratio; (E) Kynurenic acid; and (F) Quinolinic acid, were analysed in healthy controls (HC, n = 21), JDM (n = 38), and JIA (n = 12). For statistical analyses, mean ± Standard Deviation (SD) were performed, *p < 0.05, **p < 0.01

Serum Levels of Tryptophan Metabolites are Associated With Inflammation, Mental Health Outcomes and Disease Progression

Serum tryptophan metabolites correlated with clinical markers of inflammation and muscle damage in JDM patients. As shown in Fig. 2A, kynurenine (KYN) demonstrated a modest positive correlation with CRP (r = 0.41, p < 0.05) and ESR (r = 0.50, p < 0.05), suggesting a potential association between elevated kynurenine levels and systemic inflammation. The kynurenine/tryptophan (KYN/TRP) ratio also showed moderate positive correlations with aldolase (r = 0.42, p < 0.05) and CK (r = 0.46, p < 0.05), which may suggest a potential link between increased IDO activity and muscle inflammation. Similarly, quinolinic acid (QUIN) and the quinolinic acid/kynurenic acid (QUIN/KYNA) ratio correlated with CRP (r = 0.45, p < 0.05 and r = 0.42, p < 0.05, respectively) and aldolase (r = 0.41, p < 0.05 and r = 0.59, p < 0.01, respectively), further linking neurotoxic tryptophan metabolites to disease activity. Due to the serum biomarker levels of some patients were within the normal range, we conducted a separate analysis to determine whether the serum concentrations of inflammatory markers and enzymes above the normal values were associated with tryptophan metabolites. The results were consistent with those mentioned above (supplementary file 2).

Fig. 2.

Correlations between serum tryptophan metabolites, clinical biomarkers and mental health scores in JDM patients. (A) Correlation matrix between biomarkers relating to inflammation and muscle disease including C-reactive protein (CRP), erythrocyte sedimentation rate (ESR), creatine kinase (CK), and aldolase, with serum levels of kynurenine, the kynurenine/tryptophan ratio, quinolinic acid and the quinolinic acid/kynurenic acid ratio. Correlations between (B) serotonin with pediatric symptom checklist-17 total score (PSC-17); (C) kynurenine with PSC-17 total score; and (D) kynurenic acid with patient health questionnaire-9 (PHQ-9) depression score; Correlations between (E) ∆QUIN with ∆PGA muscle score; (F) ∆KYNA with ∆PGA skin score; and (G) ∆QUIN/KYNA ratio with ∆PGA skin score. For statistical analyses, *p < 0.05, **p < 0.01

Additionally, alterations in tryptophan metabolism were associated with mental health outcomes in JDM patients. As shown in Fig. 2B, serotonin levels were inversely correlated with Pediatric Symptom Checklist-17 (PSC-17) total scores (r = −0.74, p = 0.02), indicating that reduced serotonin availability may correspond to heightened anxiety symptoms. Unexpectedly, kynurenine levels also negatively correlated with the PSC-17 total scores (r = −0.61, p = 0.04, Fig. 2C). Upon further analysis of PSC-17 subscoring, serotonin levels were negatively correlated with PSC17 internalizing scores (r = −0.63, p = 0.04), while KYN levels were inversely correlated with PSC17 attention scores (r = −0.65, p = 0.03), as shown in Supplementary Figure S3A and B. Furthermore, higher kynurenic acid levels were associated with lower Patient Health Questionnaire-9 (PHQ-9) depression scores (r = −0.53, p = 0.03, Fig. 2D), suggesting a possible neuroprotective role of kynurenic acid in depressive symptoms. When comparing KYNA levels across different PHQ-9 score groups, the ‘no or minimal’ group exhibited the highest KYNA levels (Supplementary Figure S3C). Correlation analyses between other pairs of data were performed but did not reveal significant associations.

To evaluate the relationship between longitudinal changes in disease activity and tryptophan metabolite levels, paired samples from JDM patients were assessed for correlations between changes (Δ) in Physician Global Assessment (PGA) scores and corresponding shifts in key kynurenine pathway metabolites. These assessments were made across different time points within the same longitudinal cohort, with each data point representing the change observed during a single visit for each patient. As shown in Fig. 2E, improved PGA muscle scores over time were significantly correlated with reductions in QUIN concentrations (ΔQUIN, r = 0.64, p = 0.02), indicating that worsening muscle involvement is accompanied by increased levels of this neurotoxic metabolite. In contrast, changes in skin disease activity were inversely correlated with KYNA levels (ΔKYNA, r = −0.59, p = 0.03, Fig. 2F), while changes in QUIN/KYNA ratio-a surrogate marker of neurotoxic imbalance-demonstrated a significant positive correlation with changes in PGA skin (r = 0.66, p = 0.01, Fig. 2G). These longitudinal within-patient findings reveal a pattern of metabolic dysregulation skewed toward neurotoxic pathways during periods of heightened disease activity.

Taken together, these findings suggest a potential role of tryptophan metabolism in JDM pathogenesis and disease progression, including mental health outcomes.

IFN-mediated Upregulation of IDO1 Skews Tryptophan Metabolism in JDM

Our prior work [15] suggested that tryptophan metabolism is regulated by IFN-mediated upregulation of IDO1 in SLE. To determine whether a similar mechanism may occur in JDM, we first evaluated mRNA levels of IDO1, as well as the IFN-regulated gene MX2, in PBMCs of JDM patients. Consistent with the hypothesis of IDO1 limiting availability of tryptophan for serotonin generation, we observed an inverse correlation between IDO1 expression and serum levels of serotonin (r = −0.52, p = 0.04, Fig. 3A). Further, IDO1 mRNA levels correlated positively with MX2 expression (r = 0.62, p = 0.01, Fig. 3B), suggesting IFN-driven induction of IDO1 also in JDM. Finally, MX2 mRNA levels showed a strong positive correlation with physician global assessment (PGA) total score (r = 0.89, p < 0.01, Fig. 3C).

Fig. 3.

Associations between IDO1 and MX2 mRNA with serotonin levels and PGA in JDM patients. Correlation analyses between (A) IDO1 mRNA and serum serotonin levels; (B) MX2 mRNA and IDO1 mRNA; and (C) MX2 mRNA and physician global assessment total score (PGA)

Given the presence of IFN-mediated IDO1 expression in PBMCs and its association with reduced peripheral serotonin levels, we next sought to determine whether these findings could be replicated ex vivo. Incubation of PBMCs with recombinant IFN-alpha caused marked increase of IDO in monocytes, which was abrogated by inhibitors of IFNAR (anifrolumab) and JAK/STAT (baricitinib) (Fig. 4A). In addition, baricitinib and anifrolumab exhibited dose-dependent inhibitory effects on graded IFN-α–induced IDO expression in monocytes (Supplementary Figure S4).

Fig. 4.

IFN-alpha and serum-induced IDO expression in monocytes. (A) Inhibition of IFN-α-induced IDO expression in monocytes by baricitinib and anifrolumab (250 nM). (B) Induction of IDO expression in human monocytes stimulated by serum from JDM and HC; mean ± standard deviation (SD) were performed. (C) Inhibition of baricitinib and anifrolumab (250 nM) on serum-stimulated IDO expression. MFI, median fluorescence intensity; JDM, juvenile dermatomyositis; HC, healthy control. For statistical analyses, *p < 0.05, **p < 0.01, ***p < 0.001

Furthermore, healthy monocytes from donors treated with JDM serum exhibited significantly elevated IDO expression compared to HC (p < 0.01, Fig. 4B), and this effect was mitigated almost completely when the cells were pre-treated with either baricitinib or anifrolumab (p < 0.0001 and p < 0.001, respectively, Fig. 4C). These findings highlight the role of type I IFN-driven IDO induction in JDM pathogenesis and suggest that JAK inhibition or type I IFN blockade may be potential therapeutic strategies to modulate tryptophan metabolism in JDM.

Clinical Subgroups by Cluster Analysis

To better define the heterogeneity of tryptophan metabolites within the JDM cohort, we performed cluster analysis incorporating tryptophan metabolites as well as clinical components into the model, resulting in four distinct clusters (Fig. 5).

Fig. 5.

Different phenotypes and clinical features of JDM patients in four clusters. Four clusters of patients (clusters 1, 2, 3 and 4) and four clusters of variables (clusters A, B, C and D) were identified in JDM patients (n = 37) based on the partitioning around medoids algorithm employing the Gower distance. JDM, juvenile dermatomyositis; Anti-TIF1γ antibodies; CDASI, cutaneous dermatomyositis disease area and severity index; PGA, physician global assessment; QA/KA ratio, Quinolinic acid/Kynurenic acid ratio; CK, creatine kinase; KYN/TRP ratio, Kynurenine/Tryptophan ratio; ESR, erythrocyte sedimentation rate; CRP, C-reactive protein

Cluster 1 (n = 3) was defined as inflammation but less skin lesion/activity group, with patients (all girls) having relatively older age of 16.67 years (14.73–16.8) as well as the longest disease duration of 12.5 years (11.71–13.46). Patients in this cluster exhibited the highest levels of QUIN/KYNA ratio, KYN/TRP ratio, kynurenine, aldolase, ESR, CRP, QUIN and CK (all p < 0.05).

Cluster 2 (n = 7) was defined as active skin lesion group with higher PGA score (2 (0.5–3)) and CDASI score (5 (3.5–6.5)). These patients (all girls) had younger age of 7.53 years (5.89–10.61) and shorter disease duration of 2.08 years (1.38–4.29) with high prevalence of anti-TIF1γ (83.33%) and relatively higher QUIN and CK.

Cluster 3 (n = 12) had the highest male proportion (81.82%) and higher TRP and KYN levels, while with the shortest disease duration years (1.83 (0.21–5)). According to this finding of sex related cluster, after comparing the male and female data within each metabolite group, significant higher QUIN/KYNA ratio was found in female JDM compared to male JDM (Supplementary Figure S5).

Finally, Cluster 4 (n = 15), the largest cluster, was characterized by patients with quiescent disease, and no apparent abnormalities within the tryptophan metabolites.

Discussion

Children with juvenile dermatomyositis (JDM) have a high prevalence of mental health impairment, and the tryptophan metabolic pathway is well-established as being linked to mental illness. Our study is the first to identify a link between IFN-regulated alterations in tryptophan metabolism and mental health in JDM. Further, our data suggest that blocking type I IFNs may have potential clinical effects for JDM patients.

Though tryptophan metabolism has been implicated in several other inflammatory diseases, including rheumatic diseases such as SLE and JIA [27, 30, 31], the role of tryptophan metabolism in JDM has not been previously explored. In the context of mental health, dysregulation of the tryptophan–kynurenine pathway has been implicated in mood disorders, including depression and bipolar disorder, where a shift toward neurotoxic metabolites (e.g., QUIN) and away from neuroprotective metabolites (e.g., KYNA) is associated with symptom severity [26]. This imbalance may be linked to neuroinflammation, hypothalamic‑pituitary‑adrenal (HPA) axis activation, and cross‑talk with the gut–brain axis, all of which influence central neurotransmission and stress responses [32]. The interplay between autoimmunity and mental health may therefore be partly mediated by these metabolic shifts. Chronic immune activation in autoimmune diseases can increase IDO activity and kynurenine pathway flux, potentially reducing serotonin synthesis and increasing production of metabolites with neuroactive effects [33]. These biochemical changes could contribute to neuropsychiatric symptoms seen in systemic autoimmune diseases, and peripheral metabolite levels have been explored as correlates of clinical symptomatology in affective disorders [34].

In addition to its crucial impact on the nervous system and psychological health, the tryptophan metabolic pathway also plays a significant role in immune regulation. Kynurenine has been shown to enhance monocyte chemotaxis and exert proinflammatory effects on co-cultured astrocytes in vitro. It also exacerbates systemic inflammation–induced monocyte trafficking, neuroimmune disturbances, and depressive-like behavior in mice [35]. IDO is an important component of the immune system, serving as a natural defense against various pathogens. It is produced by cells in response to inflammation and exerts an immunosuppressive function by limiting T-cell activity and promoting immune tolerance mechanisms [36]. In this study, we found that compared to HC serum, JDM serum had higher KYN/TRP ratios, reflecting increased IDO activity. Additionally, increased kynurenine metabolites KYNA and QUIN, and decreased serotonin levels in JDM as compared to HC, all illustrate that tryptophan metabolism is skewed towards the kynurenine pathway in JDM. Moreover, JDM is characterized by a strong interferon gene signature, which promotes increased IDO activity and shifts tryptophan metabolism away from serotonin synthesis toward the kynurenine pathway and its downstream metabolites.

It is important to note that these measurements were done in peripheral blood, and not in cerebrospinal fluid (CSF) which is a limitation of the current study. However, this emphasizes the clinical utility of serum as a future biomarker given the ease and lower invasiveness of collection over CSF. According to existing evidence, TRP, KYN, and QUIN can cross the blood–brain barrier (BBB). Although free tryptophan transferred from blood to the brain serves as a precursor for neurotransmitter synthesis (e.g., serotonin), levels in the brain are noticeably lower than in the blood [37–39]. Sixty percent of kynurenine in the brain is directly from peripheral blood and can be taken up by astrocytes and microglia, and then respectively converted into KYNA (a neuroprotective NMDA receptor antagonist) and QUIN (a neurotoxic NMDA receptor agonist) [40, 41]. Previous studies have shown that a disrupted blood–brain barrier (BBB) in patients with autoimmune diseases allows increased infiltration of inflammatory and neurotoxic factors into the brain [42]. However, there are scant evidence demonstrating BBB disruption in JDM. Moreover, Raison and colleagues found that peripheral administration of IFN-alpha activated IDO alongside central cytokine responses, leading to increased brain levels of KYN, QUIN, and KYNA, and contributing to depressive symptoms [43]. Thus, future studies should measure tryptophan metabolites in CSF as well as in serum to evaluate which would be the most appropriate and of highest clinical significance in evaluating mental health, as well as investigate the function and penetration of BBB in JDM children, thereby outlining the whole picture of tryptophan metabolism in JDM.

Next, we investigated the cause of the increased IDO activity in JDM patients. IDO is inducible preferentially by type I and type II IFNs [44, 45], which are increased in JDM patients. Consistently, we found IDO1 mRNA levels in PBMCs moderately correlating with MX2, an IFN-regulated gene, suggesting a possible IFN-driven induction of IDO in JDM. To study causality, we developed an in vitro system to interrogate the ability and mechanism of JDM serum to support de novo induction of IDO1. Similar to the metabolic profiles demonstrating enhanced IDO1 activity in JDM and to what we had observed previously in SLE, serum from JDM patients more prominently induced IDO1 expression in monocytes as compared to serum from healthy individuals (Fig. 4B) [15]. Though primarily induced by type I and type II IFNs, IDO1 can also be upregulated by other inflammatory cytokines. To determine which signaling pathways were involved in upregulating IDO1 in JDM, we evaluated two key pathways, e.g. the commonly used signaling pathway, Janus kinase/signal transducer and activator of transcription (JAK/STAT) pathway, signaling downstream of type I and type II IFNs, as well as the type I IFN specific receptor, interferon-α/β receptor (IFNAR) [46]. These pathways are currently targeted in several rheumatic diseases, with anifrolumab, a monoclonal antibody medication which targets IFNAR, being approved by the FDA for the treatment of moderate to severe lupus. Baricitinib, a JAK1/JAK2 inhibitor, is used to treat autoimmune diseases such as rheumatoid arthritis and has shown promise in dermatomyositis, SLE and myasthenia gravis [47–49]. In addition, several studies showed that patients with JDM-associated calcinosis responded to Baricitinib [50, 51]. Of note, these medications are not currently approved for clinical use in JDM, although they are under investigation. We confirmed the ability of both drugs to dose-dependently block IFN-mediated upregulation of IDO1 (Supplemental Figure S4), as well as JDM serum-dependent upregulation of IDO1. Taken together, these findings suggest that the primary initiators of IDO1 in JDM, at least in the circulation, are likely type I interferons, given the near-complete inhibition observed with the selective IFNAR inhibitor, which specifically targets type I IFN signaling.

However, it should be noted that the two inhibitors, anifrolumab and baricitinib, also had a minor, but statistically significant, effect on reducing IDO1 induction by serum from HC. The reason for this is not fully clear, but we hypothesize that serum may contain either low levels of inflammatory cytokines sufficient to induce baseline IDO1 expression [52], and/or that serum contain low levels of danger-associated molecular patterns released upon clotting sufficient to induce TLR activation and subsequent IDO1 expression [53]. It is also plausible that there is a low level of constitutive signaling through either IFNAR or JAK/STAT resulting in baseline levels of IDO1 in isolated monocytes. Granted that IDO, through regulating tryptophan metabolism, also governs physiological processes necessary for maintaining our body's proteins, muscles, enzymes, and neurotransmitters, future studies are warranted to evaluate the role of drugs targeting these pathways, e.g. IDO1 activity, in affecting normal growth and metabolism.

Though we have assumed above that IDO1 is the main enzyme regulating tryptophan metabolism in JDM, it needs to be noted that there are three enzymes that can catalyze the first rate-limiting step in the catabolism of tryptophan to kynurenine: tryptophan 2,3-dioxygenase (TDO), IDO1, and IDO2. Expression of TDO in humans is normally restricted to the liver. As for IDO1 and IDO2, they are two closely related enzymes originating by gene duplication, but IDO2 has less tryptophan catabolizing activity and is more narrowly expressed than IDO1 [54]. Further, they seem to play opposite roles in immune responses. IDO1 is primarily known for suppressing T cell responses and promoting immune tolerance, which can contribute to tumor immune evasion, while IDO2 may have a pro-inflammatory role, particularly in B cell-mediated autoimmunity [55]. Over the past decade, an increasing number of studies have focused on IDO2. Future research may benefit from investigating how altered tryptophan metabolism within specific IDO2-expressing cells, such as antigen presenting cells (APC), influence their functional properties.

Cluster analysis combining tryptophan metabolites and clinical features identified four distinct subgroups in JDM, highlighting disease heterogeneity and altered immune metabolism. Cluster 1, defined by long disease duration and systemic inflammation despite limited skin activity, showed markedly elevated KYN/TRP and QUIN/KYNA ratios, consistent with sustained IDO activation and a shift toward proinflammatory and neurotoxic kynurenine metabolites, potentially reflecting chronic interferon-driven immune activation. Although Cluster 1 patients exhibited low contemporaneous PGA scores and minimal skin involvement, their marked elevation of kynurenine pathway metabolites and muscle inflammation markers may reflect persistent or subclinical immune activation rather than only momentary clinical activity. Cluster 2 comprised younger patients with active cutaneous disease and high anti-TIF1γ prevalence, in whom elevated QUIN suggests a link between kynurenine pathway dysregulation and skin-predominant inflammation in early disease. Future studies incorporating skin and muscle tissue samples are needed to further explore the relationship between anti-TIF1γ, IDO, and tryptophan metabolism in local tissues. Cluster 3, enriched for male patients with short disease duration and higher TRP and KYN levels, may represent a distinct early or sex-influenced immunometabolic phenotype. In contrast, Cluster 4 showed quiescent disease with preserved tryptophan homeostasis. Together, these findings support a role for dysregulated tryptophan metabolism in JDM pathogenesis, inflammatory persistence, and neuroimmune crosstalk, with potential implications for disease stratification and targeted intervention.

Interestingly, we observed a sex‑specific difference in the QUIN/KYNA ratio, with female JDM patients showing a significantly higher ratio than male JDM patients, whereas no such difference was seen in the healthy control group. The QUIN/KYNA ratio is often interpreted as a proxy for the balance between neurotoxic and neuroprotective branches of the kynurenine pathway, with QUIN associated with excitotoxic and pro‑inflammatory effects and KYNA more linked to neuroprotective and anti‑inflammatory roles [41]. Alterations in this balance have been implicated in immune and neuropsychiatric conditions and may reflect underlying sex differences in immune regulation or tryptophan metabolism. For example, sex‑specific kynurenine pathway alterations have been reported in mood disorders, suggesting differential modulation of downstream metabolites by biological sex [56].

Additional limitations include the small sample size of the JDM cohort, with many patients exhibiting inactive or low disease activity, the reliance on self- or caregiver-reported mental health questionnaires, and the presence of missing clinical data. For the low disease activity of patients, a separate correlation analysis was conducted on those tryptophan metabolites with serum concentrations of inflammatory markers and enzymes above the normal values. The results were consistent with data shown in Fig. 2A (Supplementary File 2). Moreover, most children had low PSC-17 scores. Considering of this, further analysis of the PSC-17 subscoring revealed that serotonin levels were negatively correlated with PSC-17 internalizing scores (e.g., feelings of sadness, worry, and self-doubt), as shown in Supplementary Figure S3. Taken together, relatively low disease activity and mild mental health issues may limit the generalizability of the study. Therefore, future research should include patients with varying levels of disease activity. The lack of distinction between parent-reported and self-reported mental health assessments may introduce bias, as parents typically score assessments more negatively than their children, which could affect the interpretation of the data. Meanwhile, ethnic diversity was limited, and certain minority groups were underrepresented due to regional factors. These may limit the ability of correlation analyses to fully represent the overall condition, as most observed correlations between serum tryptophan metabolites, clinical biomarkers, and mental health scores were of moderate strength. Also, the mental health assessments in this study were cross-sectional and mainly conducted in JDM patients; follow-up visits and integrated assessments for control groups are needed to establish causality or stability over time. Larger longitudinal cohorts with a high level of ethnic diversity and integrated clinical data containing inflammatory cytokines, muscle enzymes, and neuropsychological tests are necessary for future studies to validate these findings.

Further research is needed to elucidate the relationship between tryptophan metabolites and JDM pathogenesis, particularly the molecular mechanisms underlying muscle damage and neuropsychiatric effect. Moreover, while baricitinib and anifrolumab have demonstrated efficacy in other autoimmune diseases, additional clinical and pharmacological studies are required to establish their effectiveness in JDM.

Conclusion

In conclusion, our study reveals a significant shift of tryptophan metabolism in JDM, characterized by decreased serum serotonin levels and increased kynurenine pathway metabolites. These metabolic alterations correlate with anxiety and depression scores, suggesting a potential link between tryptophan metabolism and mental health in JDM. Moreover, we have reported a novel role of type I IFN in driving IDO upregulation in JDM. Notably, baricitinib and anifrolumab effectively inhibited IFN-mediated IDO induction, highlighting a potential therapeutic strategy to mitigate both disease activity and neuropsychiatric symptoms in JDM patients.

Supplementary Information

Below is the link to the electronic supplementary material.

Abbreviations

JDM

Juvenile dermatomyositis

JIA

Juvenile idiopathic arthritis

HC

Healthy controls

IDO

Indoleamine 2,3-dioxygenase

IFN

Interferon

NMDARs

N-methyl-D-aspartate receptors

EULAR

European league against rheumatism

ACR

American college of rheumatology

ILAR

International league of associations for rheumatism

CDASI

Cutaneous dermatomyositis disease area and severity index

CMAS

Childhood myositis assessment scale

PGA

Physician global assessment

MMT8

Manual muscle testing 8

CK

Creatine kinase

ESR

Erythrocyte sedimentation rate

CRP

C-reactive protein

PSC-17

Pediatric symptom checklist-17

PHQ-9

Patient health questionnaire-9

ELISA

Enzyme-linked immunosorbent assay

PBMCs

Peripheral blood mononuclear cells

Ct

Cycle threshold

qRT-PCR

Quantitative real-time polymerase chain reaction

PAM

Partitioning around medoids

KYN

Kynurenine

KYN/TRP

Kynurenine/tryptophan

QUIN

Quinolinic acid

QUIN/KYNA

Guinolinic acid /kynurenic acid

CSF

Cerebrospinal fluid

BBB

Blood-brain barrier

IFNAR

Interferon-α/β receptor

TDO

Tryptophan 2,3-dioxygenase

APC

Antigen presenting cells

Author contributions

Conceptualization, ML, XZ, SS, CL; methodology, YW, SS, AL, PH, JAG-C, CL; software, YW, PH, JS, TW, AJ; validation, YW, CL; formal analysis, YW, PH, AJ; investigation, YW, PH, JAG-C, AJ, JS, CL; resources, SS, AL, PH, JAG-C, TW, CL; data curation, YW, JS, CL; writing—original draft preparation, YW, CL; writing—review and editing, all authors; visualization, YW; supervision, ML, XZ, SS, CL; project administration, CL; funding acquisition, XZ, CL. All authors read and approved the final manuscript.

Funding

Dr. Lood and the Seattle Children’s Hospital Juvenile Myositis Center of Excellence received funding from the CureJM Foundation. Dr. Lood received funding from The Myositis Association (TMA) and is supported as the Herndon and Esther Maury Endowed Professor of Rheumatoid Arthritis. This study was also supported by the Chinese National Key Technology R&D Program, Ministry of Science and Technology (2024YFC2510301), National High Level Hospital Clinical Research Funding (2025-PUMCH-C-010).

Data Availability

All data generated or analysed during this study are included in this published article and its supplementary information files.

Declarations

Ethics Approval and Consent to Participate

In accordance with the Declaration of Helsinki, this study is IRB-approved (PIROSTUDY14426) at Seattle Children’s Hospital and accessed through the Juvenile Myositis Seattle Children’s Hospital Biorepository, headed by Dr. Lood and Shenoi, under IRB approval STUDY00003100 (November 9, 2017) and STUDY00003804 (December 21, 2022). Informed consent or assent was obtained from all patients/caregivers as needed.

Conflict of Interest

Dr. Lood received research funding from Boehringer Ingelheim, Pfizer, Gilead Sciences, Eli Lilly, Redd Pharma, Archon, Neutrolis, and Amytryx. Dr. Lood is a scientific advisor for Citryll and Redd Pharma. Dr. Shenoi has served as a consultant for Cabaletta Bio. The other authors declare no conflict of interest.

Clinical Trial Number

Not applicable.