Our goal was to investigate how biomarkers, including KYN metabolites, eCBs, inflammatory markers, and cortisol, relate to one another and potentially differ across groups with or without histories of CM or SUD. Beyond baseline levels of these molecules, we also included variables representative of stress-induced changes. These targets were selected based on existing literature suggesting crosstalk between these systems, particularly the KYN pathway and eCB system (Comai et al. 2024; Wences Chirino et al. 2023), as important contributors to the development of stress-related pathologies, including SUD (Morales-Puerto et al. 2021). The primary focus was a factor analysis, which reduced data from all 34 biomarkers into the 9 factors represented in Table 2. We subsequently explored the relationship between these factors and CM and SUD status.

Overall, we found that the majority of variance was explained by KYN metabolites at baseline and in response to stress exposure. Factor 1, a KYN metabolite factor, was comprised of baseline and stress-induced changes in KYN, QUIN, 3-HK, KYNA, as well as baseline CRP levels. Notably, the KYN metabolites that were most highly associated with Factor 1 are of the “inflammatory” or “neurotoxic” branch of the KYN pathway. For instance, 3-HK and QUIN have been shown to exhibit neurotoxic properties, including excitotoxicity (Guidetti and Schwarcz 1999; Lugo-Huitrón et al. 2013; Schwarcz et al. 1983; Schwarcz and Köhler 1983; Stone and Perkins 1981), and elevations in these metabolites have been observed in various psychiatric and degenerative disorders (Leclercq et al. 2021; Schwarcz et al. 2012). Similarly, recent research in human populations has implicated enhanced levels of KYN in mood disorders, such as depression (Comai et al. 2024) and SUD pathology, with the majority of the SUD studies conducted in alcohol use disorder populations (Jang et al. 2022; Morales-Puerto et al. 2021; Wang et al. 2023). The strong association of these neurotoxic metabolites with Factor 1 may additionally explain the loading of baseline CRP levels on this factor, as shifts in Trp metabolism along the KYN pathway are promoted by inflammation (O’Connor et al. 2009; Yoshida et al. 1979; Yoshida and Hayaishi 1978). Interestingly, the neuroprotective metabolite KYNA also loaded onto Factor 1 to a smaller but significant extent, somewhat complicating our interpretation of this factor. However, previous studies have reported elevations in KYNA in response to a stressor (Chiappelli et al. 2018; Herhaus et al. 2021), potentially related to heightened KYN production and metabolism during stress. Upregulated KYN production and metabolism contributes to increased activity along both the neuroprotective and neurotoxic branches of the KYN pathway and may, in part, explain the grouping of KYNA with neurotoxic KYN metabolites within our factor analysis.

A second KYN metabolite factor, Factor 3, also explained a significant portion of total variance. Factor 3 was comprised of baseline and stress-induced changes in Trp, KYNA, PIC, 3-HK. Distinct from Factor 1, the neuroprotective KYN metabolites were the most strongly associated with this factor. These included KYNA and PIC, which previous studies have demonstrated to have neuroprotective properties, potentially counteracting the toxic effects of QUIN (Beninger et al. 1994; Foster et al. 1984). Baseline and stress-induced changes in CRP levels additionally loaded onto this factor and displayed a negative relationship with these neuroprotective KYN metabolites. This grouping may reflect the relationship between neuroprotective KYN factors and an “anti-inflammatory” state. Interestingly, 3-HK at baseline and throughout stress exposure also loaded onto Factor 3. The balance between neurotoxic and neuroprotective KYN metabolites, as well as their relationship to inflammatory signaling, has been proposed as a potential mechanism underlying susceptibility to various psychiatric disorders and suicidality (Brundin et al. 2016; Marx et al. 2021). The present analysis found neurotoxic (i.e., QUIN, 3-HK) and neuroprotective (i.e., KYNA, PIC) KYN metabolites to partially group together. Several studies have reported KYN pathway dysfunction in AUD and SUD. However, our results did not find that stress-induced KYN pathway changes were moderated by SUD or CM status, suggesting that the KYN pathway may be involved in other, non-stress-related aspects of these disorders. Future studies are needed to more clearly elucidate the role of the KYN pathway in CM and SUD.

In addition to KYN metabolites, eCBs also accounted for a significant portion of the total variance. Specifically, Factor 2 was comprised of AEA, OEA, and PEA at baseline and across stress exposure. Baseline and stress-induced changes in 2-AG separated from other eCBs and loaded onto Factor 9. Furthermore, inflammatory markers separated onto three factors. Factor 5, a pro-inflammatory factor, was comprised of baseline and stress-induced changes in CRP and IL-6. Interestingly, levels of TNF at baseline and following stress exposure loaded onto Factor t 6 with various KYN metabolites. Baseline and stress-induced changes in levels of IL-10 loaded onto Factor 8 and were inversely related to baseline 2-AG levels. Stress-induced changes and baseline levels of 2-AG loaded onto Factor 9. Cortisol levels at baseline and across stress exposure loaded onto Factor 4, separating from other biomarkers.

We found a main effect of CM exposure on Factor 5, a pro-inflammatory factor comprised of baseline and stress-induced changes in CRP and IL-6. Specifically, individuals with a history of CM exhibited higher levels of Factor 5 than their non-exposed counterparts. Our findings align with the robustly established relationship between exposure to chronic stressors, such as CM, and enhanced pro-inflammatory signaling. Meta-analyses assessing immune signaling in CM populations have found that CM exposure significantly increases baseline peripheral levels of CRP, IL-6, and TNF (Baumeister et al. 2016; Brown et al. 2021). Likewise, longitudinal studies prospectively assessing CM status have found adults with a history of maltreatment to display elevated baseline levels of CRP independent of co-occurring early life risks (including low birth weight, socioeconomic disadvantage, and low intelligence quotient), adulthood stress, and adult health and health behaviors (Danese et al. 2007, 2008). Previous work has additionally demonstrated enhanced pro-inflammatory signaling in CM populations in response to an acute laboratory stressor. Individuals with a history of CM exhibit enhanced reactivity of IL-6 during psychosocial stress tests compared to healthy controls (Carpenter et al. 2010; Schreier et al. 2020). Our findings support the positive relationship between CM exposure and baseline and stress-induced pro-inflammatory signaling, particularly of CRP and IL-6.

We also found a difference relative to CM and SUD status on Factor 9, an eCB factor. Baseline and stress-induced changes in levels of the eCB ligand 2-AG were negatively associated with Factor 9. In the absence of CM (CM-), levels of Factor 9 were significantly increased if SUD pathology was present (SUD+), reflecting reductions in baseline and stress-induced changes in 2-AG levels. Conversely, when both CM exposure (CM-) and SUD pathology were absent (SUD-), decreases in Factor 9 and thereby enhancements of 2-AG were observed. As a phasic regulator of stress, 2-AG increases in response to acute stressors, adaptively blunting the magnitude and duration of the stress response (Hill et al. 2010, 2011; Wang et al. 2012). Exposure to chronic stressors such as CM, however, are associated with dysregulated eCB signaling and modulation of the different arms of stress response, including the HPA axis (Hill et al. 2009; Patel et al. 2009; Rossi et al. 2008; Wamsteeker et al. 2010). Consequently, CM-induced alterations of eCB signaling may contribute to the development of stress-related disorders such as SUD (Fuentes et al. 2024).

In the present investigation, baseline and stress-induced changes in 2-AG were reduced in the presence of SUD (SUD+) in individuals without CM (CM-), indicated by greater levels of Factor 9. These findings may therefore suggest a mediating role of 2-AG signaling in SUD pathology. This theory is supported by recent preclinical work demonstrating pharmacological enhancement of 2-AG via inhibition of its primary degrading enzyme, monoacylglycerol lipase, attenuates the negative withdrawal effects of ethanol (Morgan et al. 2022). Interestingly, pharmacological inhibition of DAGL, the primary synthesizing enzyme of 2-AG, which results in reduced levels of 2-AG was found to reduce alcohol consumption in a variety of mouse models (Winters et al. 2021). Taken together, 2-AG signaling appears to be implicated in different aspects of SUD pathology, specifically in models of alcohol use disorder (for a comprehensive review see (Elliott et al. 2025). The present findings expand upon this work, demonstrating an interaction effect between CM and SUD status, not limited to AUD, on 2-AG signaling in humans. Interestingly, in the presence of CM (CM+), no effect of SUD status was observed on Factor 9. It is possible that CM may play an interactive role in the relationship between SUD and 2-AG signaling. However, due to the cross-sectional design of the present study, mechanistic conclusions cannot be drawn. Future research should further examine the nature of this relationship to understand the mechanistic role of 2-AG signaling in SUD pathology and CM populations.

We did not, however, observe a significant impact of CM or SUD status on the KYN metabolite factors, contrasting existing research demonstrating an association between KYN pathway metabolism, chronic stress exposure, and stress-related psychopathologies such as SUD (Comai et al. 2024; Contella et al. 2024; Leclercq et al. 2021; Wences Chirino et al. 2023) as discussed earlier. However, research translating the impact of chronic stress on KYN metabolic function in humans is generally lacking. Thus, our more complex population, including both current or lifetime SUD diagnosis and CM history, may contribute to the differences in results. Specifically, the current project included participants with both previous and ongoing SUD, which may have unique biomarker profiles and contribute to the discrepancy between our results and previous findings. The present study additionally assessed CM exposure prospectively, which may explain why our findings do not validate previous work in retrospectively assessed CM populations (Löfberg et al. 2023). Furthermore, while previous work in healthy humans has utilized the MAST to assess stress responsivity of KYN metabolites (La Torre et al. 2021), the temporal response of KYN metabolites to acute stressors may be altered in clinical populations. Thus, the current stress paradigm may not have been able to capture stress-induced changes in KYN metabolites and thereby the effect of CM and SUD status on KYN metabolite-driven factors.

Our study has several limitations. While our sample size was considerable overall (N = 101), the 2 × 2 factorial design resulted in small subgroups (n = ~ 25), limiting our statistical power. Furthermore, we were underpowered to assess the contributions of CM subtype. Additionally, due to the prospective assessment of CM status in the study population, the current findings are limited in their generalizability to retrospectively assessed CM populations (Baldwin et al. 2024). Furthermore, the clinical control “SUD-only” group, comprised of individuals with current or lifetime SUD diagnosis but no prospectively documented CM, retrospectively reported CM exposure on the CTQ at rates comparable to the CM groups. While the CTQ has been shown to be a limited measure of maltreatment in SUD populations (Löfberg et al. 2023), self-reported CM exposure in our “SUD-only” group nuances the interpretation of our results.

Potential confounding effects of psychiatric comorbidities and medications should also be considered in the interpretation of the current results. However, participants prescribed psychotropic medications were required to have stable use for at least three months to be included in the study, and diagnoses of current or lifetime schizophrenia, bipolar or psychotic disorder, or current suicidality were excluded. The present analyses did not investigate potential confounding effects of other comorbid psychiatric conditions, which should be considered for future investigations, as KYN metabolites, eCBs, stress, and inflammatory biomarkers are impacted in various psychiatric diagnoses.

Finally, the cross-sectional approach means we are unable to determine if changes in the biomarkers reported are a consequence of exposure or representative of inherent differences existing prior to exposure. Regardless, we provide novel insights into the relationship between a variety of baseline and stress-induced biomarkers and have pro-inflammatory biomarkers, specifically CRP and IL-6, and 2-AG signaling as potential avenues worth further exploring to better understand the mechanistic underpinnings of vulnerability to SUD development in CM populations.