The kynurenine pathway in relation to depression and anhedonia symptom severity among women with HIV

Abstract

Introduction

The kynurenine pathway (KP) has been implicated in cytokine-induced depression and anhedonia, a core depression symptom. We examined KP metabolites in relation to depression and anhedonia symptom severity in women with and without HIV (WWH and WWoH).

Methods

Participants included 239 women (55% WWH, mean age 52 years). Depressive symptoms were measured using the Center for Epidemiologic Studies Depression (CES-D) scale and high depressive symptoms were defined as a score ≥16; anhedonia severity was quantified on two CES-D items. KP metabolites, including quinolinic acid (QA, NMDA receptor agonist) and kynurenic acid (KA, NMDA receptor antagonist), were measured using liquid chromatography-tandem mass spectrometry. Adjusted multivariable linear regression was used to compare metabolites among groups and to assess relationships of KP metabolites with anhedonia and depressive symptoms.

Results

A greater proportion of WWH had high depression severity (CES-D ≥16) than WWoH (23% vs. 12%; p = .03). WWH with CES-D≥16 had greater QA compared to WWH with CES-D<16 (diff = 0.52 SD; p = .02). Using a dimensional quantitative approach, QA levels were positively associated with depressive ($\beta$ = 0.25; p = .01) and anhedonia symptom severity ($\beta$ = 0.24; p = .01) among WWH, but not in WWoH. Conversely, KA was inversely associated with anhedonia symptom severity in WWH ($\beta$ = −0.20; p =.03) and WWoH ($\beta$ = −0.23; p =.02).

Conclusion

This study extends understanding of the role of the KP in depression and anhedonia to WWH and suggests neuroactive metabolites as a possible underlying mechanism. The KP may affect reward neuro-circuitry through increased QA neurotoxicity and decreased KA neuroprotection.

1. Introduction

HIV infection remains a major public health issue with over 40 million people globally living with HIV as of 2024 (UNAIDS, 2024). Although antiretroviral therapy (ART) has led to near normal lifespans for many people living with HIV (Lewden et al., 2012; May et al., 2014; Wandeler et al., 2016), by some estimates, HIV will be the major driver of disability adjusted life years worldwide by 2030 with depression as the second leading cause (Mathers and Loncar, 2006). Depression is the most common neuropsychiatric complication among people with HIV (PWH), with a prevalence of 30-50% and up to 78% in some cohorts (Bengtson et al., 2016; Cook et al., 2018; Daubert et al., 2022; Mimiaga et al., 2013; Mudra Rakshasa-Loots et al., 2022; Tran et al., 2019; Uthman et al., 2014), compared to 10%-20% in the general U.S. population (Hasin et al., 2018). Among PWH, depression has been associated with worse overall HIV outcomes despite adjusting for ART adherence (Ammassari et al., 2004; Anastos et al., 2005; Babowitch et al., 2018; Berger-Greenstein et al., 2007; Chesney, 2003; Cook et al., 2004; Crockett et al., 2020; French et al., 2009; Horberg et al., 2008; Ickovics et al., 2001; Ironson et al., 2005; Kacanek et al., 2010; Nolan et al., 2017; Ogburn et al., 2019; Sinha et al., 2017; Springer et al., 2009; Tegger et al., 2008; Tucker et al., 2003).

There is evidence that suggests kynurenine pathway (KP) dysregulation may contribute to depression pathogenesis in PWH (Drivsholm et al., 2021; Martinez et al., 2014). The KP is principally induced by the enzyme indoleamine 2,3-dioxygenase (IDO), which is activated by immune mediators including inflammatory cytokines and HIV proteins such as Tat (Samikkannu et al., 2009; Wirleitner et al., 2003). The KP metabolizes tryptophan (TRP) into kynurenine (KYN), which is then compartmentally metabolized into neurotoxins 3-hydroxykynurenine (3-HK) and 3-hydroxyanthranilic acid (3-HAA), and then to quinolinic acid (QA), a glutamate receptor agonist (Fukui et al., 1991a; Goldstein et al., 2000). Conversely, KYN can also be metabolized into kynurenic acid (KA), which is a glutamate receptor antagonist with neuroprotective properties (Sapko et al., 2006). Evidence from Fukui et al. suggests that KP activity in the blood parallels its activity in the brain (Fukui et al., 1991a). Although the Fukui study was conducted in rodents and blood-brain barrier (BBB) permeability to peripheral KP metabolites may differ across species, peripheral KP metabolite levels in humans have been associated with clinical symptoms, treatment response, and neuroanatomical and glial changes, supporting their relevance as a proxy for central KP activity (Skorobogatov et al., 2021). Because HIV infection is characterized by chronic inflammation, even in the setting of suppressive ART, HIV-associated inflammation may induce the KP and shift KP metabolism toward increased production of neurotoxic metabolites and reduced production of neuroprotective metabolites, which may contribute to greater depressive symptoms.

Women with HIV (WWH) may represent a particularly vulnerable group with respect to depressive symptomatology. WWH have a higher prevalence of depression compared to men with HIV (MWH) (Ickovics et al., 2001). Depressive symptoms are associated with significantly higher risk of all-cause mortality in WWH (Ickovics et al., 2001; Todd et al., 2017). Research from the Women's Interagency HIV Study (WIHS) found that chronic depressive symptoms were associated with over three times greater mortality among WWH on ART, and over seven times greater mortality among WWH not on ART, compared with WWH on ART without depression (Todd et al., 2017). Among WWH in the WIHS, cumulative depressive symptom burden was also linked to missed HIV care visits, reduced ART adherence and virologic failure. (Mills et al., 2018, 2019). The neurobiological etiology for the high prevalence of depression among PWH, and particularly WWH, has not been completely characterized (Gonzalez et al., 2011; Zuniga et al., 2016.

To date, there has been minimal KP research in the context of co-morbid HIV and depression. Our group showed that HIV is associated with KP activation, including a greater plasma KYN/TRP ratio, a surrogate for IDO activity, in WWH compared to women without HIV (WWoH) (Qi et al., 2018; Rogando et al., 2022). A recent study by Mudra Rakshasa-Loots and colleagues (2023) also confirmed increased KP activity in PWH but failed to find a relationship between KP metabolites and depressive symptoms (Mudra Rakshasa-Loots et al., 2023). However, a study by Drivsholm et al. documented a relationship of KP pathway activation with depression among PWH (Drivsholm et al., 2021). Furthermore, alterations in the KP may differ between men and women, with women showing greater decreases in neuroprotective KP metabolites and higher risk for depression (Nikkheslat et al., 2025). These conflicting findings of the KP in depression may be due to sex-specific differences in KP metabolites, differences in sample size, to the heterogenous nature of depression, as well as how depression is defined across studies.

The diagnosis of depression is based on a cluster of symptoms, and each symptom may involve a distinct biology. Addressing this challenge, the National Institute of Mental Health (NIMH) developed the Research Domain Criteria (RDoC) to emphasize the importance of studying behavioral constructs, i.e., quantitative dimensional analysis, rather than the categorical diagnosis of depression (Ely et al., 2021; Gabbay et al., 2012b). Our group has adapted such dimensional investigation documenting increased blood KP activity in relation to anhedonia — the decreased capacity to experience pleasure and a core symptom of depression presentation (DeWitt et al., 2017; Gabbay et al., 2010, 2012b) — as well as with suicidal thoughts and/or attempts (Bradley et al., 2015). In neuroimaging studies of adult populations with depression, relationships between blood KP metabolite levels and reward brain regions have also been reported (Meier et al., 2016; Savitz et al., 2015a, 2015b, 2015c). These findings suggest that the role of the KP in depression may be mediated through alterations in the reward neurocircuitry, which in turn lead to anhedonic behavior and depression.

Building upon the above observations, we leveraged data collected as part of the WIHS and the nested IDOze Sleep R01 to examine the association between KP metabolites and depressive symptoms, including anhedonia, in WWH and WWoH (Rogando et al., 2022). We hypothesized that KP metabolites would be associated with anhedonia severity among WWH regardless of any association with depressive symptoms.

2. Methods

2.1. Study population

The Women's Interagency HIV Study (WIHS, now part of the Multicenter AIDS Cohort Study/WIHS Combined Cohort Study) was a multicenter longitudinal cohort study of WWH and a comparison group of demographically similar WWoH enrolled across 10 U.S. study sites to investigate the treated course of HIV and associated conditions among women. IDOze is a study of the tryptophan/kynurenine pathways, sleep/circadian dysfunction, and cardiometabolic outcomes that enrolled women from the Chicago IL, and New York City (Brooklyn and Bronx) WIHS sites between October 2018 and January 2020. Both the WIHS and IDOze study procedures, methods, and participant characteristics have been previously described (Adimora et al., 2018; Rogando et al., 2022).

Inclusion criteria: English-speaking women aged 35-70 years. WWH had to be on stable antiretroviral therapy (ART), excluding efavirenz, with an HIV RNA level <200 copies/ml and CD4⁺ T Lymphocyte count ≥200 cells/μL at the visit 6 months before IDOze enrollment.

Exclusion criteria: severe chronic or acute medical or psychiatric illness, narcolepsy, illicit drug use (>1 day/week of self-reported use), use of psychotropic medication and/or sleep aids (prescription hypnotics, over the counter sleeping aids >2 nights per week, melatonin supplementation), night-shift work, pregnancy or lactation within past three months, and use of estrogen-containing contraceptives or hormone replacement therapy.

Written informed consent was obtained from all participants in accordance with Department of Health and Human Services guidelines and the institutional review board from each research study site.

2.2. Study procedures

The IDOze baseline study visit was coordinated with semiannual WIHS visits and involved the acquisition of plasma samples for KP metabolite analysis and completion of questionnaires utilizing WIHS survey instruments (Adimora et al., 2018; Bacon et al., 2005) including the Center for Epidemiologic Studies Depression Scale (CES-D). Women were asked to fast for the morning blood draw; 89% of the women provided fasted blood samples.

Blood Sampling for KP metabolite analysis: All blood was processed and frozen within 24 h of collection; plasma aliquots were stored at −70 °C until batch testing. Total (bound and unbound) tryptophan/kynurenine pathway metabolites of interest from plasma were measured using liquid chromatography-tandem mass spectrometry. Area under the curve was reported for each metabolite (Broad Institute Metabolomics Platform of MIT/Harvard). Specifically, a 6495 triple quadrupole mass spectrometer coupled to a 1290 Infinity II U-HPLC system (Agilent, Santa Clara, California) was used to quantify the relative abundance of plasma metabolites including tryptophan (TRP), kynurenine (KYN), kynurenic acid (KA), and quinolinic acid (QA). Raw data were analyzed using MassHunter software (Agilent) for automated peak integration, and the quality of integration was manually reviewed and compared against reference standards to confirm the identities of the metabolite. The metabolomics lab was blinded to HIV and depressive symptom status.

Assessments of Depression and Anhedonia: Depressive symptomatology was assessed using the 20-item CES-D Scale (Radloff, 1977), with participants reporting the frequency of each symptom experienced in the past week (“rarely/none,” “some/a little,” “occasionally/moderate,” “most or all of the time,” scored 0-3 respectively). Scores are summed across all items, and total scores range from 0 to 60. Higher scores indicate greater depression severity with a score of ≥16 indicative of clinically significant depressive symptoms. The CES-D cutoff of ≥16 has been widely used and validated as an indicator of elevated depressive symptoms in people with HIV and other chronic illnesses, with good sensitivity, specificity, and internal consistency, including in large cohorts involving thousands of participants and in other WIHS analyses (Adams et al., 2018; Lewinsohn et al., 1997; Perez et al., 2025; Radloff, 1977; Siddaway et al., 2017).

Anhedonia was quantified by two specific items from the CES-D that demonstrate decreased motivation and capacity to experience pleasure. Anhedonia scores were assessed by summing the scores from CES-D question 7 (“I felt that everything I did was an effort” demonstrating loss of interest in activities) with the reversed score of question 16 (“I enjoyed life” scored 3(rarely) – 0(most of the time) indicating hedonic capacity (Carleton et al., 2013). Anhedonia scores range from 0 to 6, with higher scores reflecting greater anhedonia severity. We have used a similar approach with different depression scales in other studies of the KP and depression/anhedonia (Gabbay et al., 2012c, 2015).

2.3. Statistical methods

We performed a cross-sectional secondary analysis using data from the IDOze study to evaluate associations between KP metabolites, depressive symptoms, and covariates of interest (enumerated below), measured at similar time points. All associations between KP metabolites and HIV status, depression status, and depressive symptom severity were examined using multivariable linear regression models, with a consistent set of covariates applied across analyses.

Descriptive Statistics. Demographic and clinical characteristics were summarized by HIV status. Continuous variables are presented as means with standard deviations (SD), and categorical variables as frequencies and percentages. Comparisons between WWH and WWoH were performed using Wilcoxon rank sum test for continuous variables and Chi-square or Fisher's exact tests for categorical variables, as appropriate.

KP Metabolite Analyses. Primary outcomes of interest were plasma concentrations of TRP, KYN, KA, and QA, as well as the kynurenine/tryptophan (KYN/TRP) ratio. Metabolite concentrations were log-transformed prior to analysis to approximate normality.

To compare KP metabolite levels by HIV status, each metabolite was modeled as the dependent variable with HIV status as the primary independent variable. Differences in metabolite levels between WWH and WWoH are reported as adjusted mean differences.

To compare KP metabolite levels by depression status, analyses were stratified by HIV status. High depressive symptom groups were defined using a CES-D score ≥16, and separate models were fit for WWH and WWoH. Estimated marginal means (EMMs) were calculated for high depressive symptom (CES-D score ≥16) and low/no depressive symptom (CES-D < 16) groups, and group differences are presented as contrasts of the EMMs (high vs. low/no depressive symptoms). To facilitate comparison of effect sizes across metabolites, metabolite values were standardized (Z-scored) within each HIV status group (mean = 0, SD = 1) for both analyses.

Continuous symptom associations. Associations between depressive symptom severity and KP metabolite levels were examined using continuous measures. CES-D total score, anhedonia score, and metabolite values were standardized (Z-scored) within each HIV status group. Multivariable linear regression models were used to estimate standardized beta coefficients, representing the change in metabolite levels (in SD units) associated with a one-standard-deviation increase in symptom severity.

Covariates. All models were adjusted for age, body mass index (BMI), hypertension, diabetes mellitus, alcohol use, and smoking status. Models restricted to WWH were additionally adjusted for nadir CD4⁺ T-cell count.

Software. All statistical analyses were performed using R Statistical Software (v4.5.0). The gtsummary, emmeans, and tidyverse packages were utilized for data summary, estimation of marginal means, and data manipulation, respectively. Statistical significance was defined as a two-sided p-value <0.05.

3. Results

3.1. Participant demographic and clinical characteristics

All patient demographic and clinical characteristics are in Table 1. The study sample included 239 women, of whom 131 (55%) were women with HIV (WWH) and 108 (45%) were women without HIV (WWoH). The mean age of the cohort was 52 years (SD = 7.8). The majority of participants identified as Non-Hispanic Black (71%), followed by Hispanic (25%).

Table 1.

Participant demographics and clinical characteristics.

		HIV Status
	Combined samplea (n = 239)	Women without HIVa (n = 108)	Women with HIVa (n = 131)	p-valueb
Age in years, mean (SD)	52.1 (7.8)	51.1 (8.1)	52.9 (7.5)	0.06
Race/ethnicity, n (%)				0.01
Hispanic	60 (25%)	29 (27%)	31 (24%)
Non-Hispanic Black	170 (71%)	79 (73%)	91 (69%)
Non-Hispanic Other	9 (3.8%)	0 (0%)	9 (6.9%)
Menopausal, n (%)				0.04
No	85 (36%)	46 (43%)	39 (30%)
Yes	154 (64%)	62 (57%)	92 (70%)
BMI, mean (SD)	31.9 (7.5)	31.8 (7.2)	31.9 (7.8)	0.88
Hypertension, n (%)	139 (58%)	70 (65%)	69 (53%)	0.06
Diabetes, n (%)	79 (33%)	33 (31%)	46 (35%)	0.46
Smoking status, n (%)				0.01
Non-smoker	153 (64%)	59 (55%)	94 (72%)
Smoker	86 (36%)	49 (45%)	37 (28%)
Any alcohol usec, n (%)	112 (47%)	59 (55%)	53 (40%)	0.03
Marijuana/hash usec, n (%)	60 (25%)	33 (31%)	27 (21%)	0.08
Crack, cocaine or heroin usec, n (%)	10 (4.2%)	6 (5.6%)	4 (3.1%)	0.35
CES-D ≥ 16, n (%)	43 (18%)	13 (12%)	30 (23%)	0.03
CES-D total score, mean (SD)	9.0 (9.3)	7.4 (7.7)	10.3 (10.4)	0.06
Anhedonia score, mean (SD)	1.7 (1.6)	1.6 (1.6)	1.7 (1.6)	0.67
Current ART usec, n (%)	127 (53%)	NA	127 (97%)
HIV RNA, % undetectable4, n (%)	105 (80%)	NA	105 (80%)
CD4 T lymphocyte count, mean (SD)	758.3 (347.7)	NA	758.3 (347.7)
Nadir CD4 T lymphocyte count, mean (SD)	226.6 (161.7)	NA	226.6 (161.7)

² HIV RNA viral load <20 copies/ml.

^aMean (SD); n (%).

^bWilcoxon rank sum test; Fisher's exact test; Pearson's Chi-squared test.

^cUse within the last 6 months.

There were significant demographic and behavioral differences between the groups. Racial/ethnic distribution varied significantly (p =.01), with a higher proportion of participants identifying as “Other” race in the WWH group. Substance use patterns differed too; WWoH had significantly higher rates of current smoking (45% vs. 28%; p = .01) and any alcohol use (55% vs. 40%; p =.03). No significant differences were observed in BMI (p = .88) or the prevalence of diabetes (p =.46).

Regarding psychiatric characteristics, WWH had a significantly higher prevalence of clinically significant depressive symptoms (CES-D ≥ 16) compared to WWoH (23% vs. 12%; p = .03). Anhedonia scores were comparable between the two groups (mean 1.7 WWH vs. 1.6 WWoH; p = .67). As shown in Fig. 1, anhedonia severity scores were highly variable in both WWH and WWoH.

Fig. 1.

Distribution of anhedonia severity scores. WWH and WWoH with high depressive symptoms and low/no depressive symptoms showed a wide range of anhedonia severity scores. In both WWH and WWoH, anhedonia severity scores overlapped among women with and without depression. Anhedonia severity scores were highly variable regardless of HIV or depressive symptom status.

Among WWH, 97% were receiving antiretroviral therapy (ART), and 80% had an undetectable HIV viral load. The mean current CD4⁺ T-cell count was 758.3 cells/μL (SD = 347.7), and the mean nadir CD4⁺ count was 226.6 cells/μL (SD = 161.7).

3.2. Comparisons of KP metabolites by HIV status

We used multivariable linear regression to examine whether metabolite levels differed between WWH and WWoH. In line with our previous work (Rogando et al., 2022), WWH showed higher levels of several KP metabolites compared with WWoH. WWH had higher concentrations of KYN (diff = 0.45 SD; p = .001) and QA (diff = 0.51 SD; p = .001), as well as an elevated KYN/TRP ratio (diff = 0.48 SD; p = .001). KA and TRP showed nonsignificant differences (Table 2).

Table 2.

Adjusted multivariable linear regression models comparing KP metabolites by HIV status.

Metabolite	Difference (SD) (95% CI)	P-value
Kynurenic acid	0.23 (−0.04, 0.50)	0.09
Kynurenine	0.45 (0.19, 0.70)	0.001
Quinolinic acid	0.51 (0.25, 0.77)	0.001
Tryptophan	−0.02 (−0.29, 0.25)	0.88
KYN/TRP	0.48 (0.23, 0.72)	0.001

Differences in metabolite levels are reported as adjusted mean differences of Z-scored metabolites (WWH vs. WWoH). Positive values indicate higher levels in WWH. Adjusted models include age, BMI, hypertension, diabetes, alcohol use, and smoking status.

3.3. Group comparisons of KP metabolites by HIV and depression status

WWH and WWoH were stratified into those with high depressive symptoms (CES-D score ≥16) and those with low/no depressive symptoms (CES-D < 16). We used EMMs and standardized contrasts to assess adjusted differences in KP metabolite levels by depression status, stratified by HIV status (Table 3 and Fig. 2). Covariate adjustments were applied as specified in the Methods.

Table 3.

Adjusted multivariable linear regression models comparing KP metabolites by HIV and depressive symptom status.

Metabolite	Low/no depressive symptoms (CES-D < 16) (EMM, 95% CI)	High depressive symptoms (CES-D score ≥16) (EMM, 95% CI)	Difference (SD)*	P-value
Women With HIV
Kynurenic acid	−0.09 (−0.32, 0.14)	0.05 (−0.34, 0.44)	0.14	0.53
Kynurenine (KYN)	−0.09 (−0.31, 0.13)	−0.01 (−0.38, 0.36)	0.08	0.69
KYN/TRP	−0.11 (−0.32, 0.10)	0.09 (−0.28, 0.45)	0.20	0.33
Quinolinic acid	−0.11 (-0.33, 0.11)	0.41 (0.02, 0.79)	0.52	0.02
Tryptophan (TRP)	0.02 (−0.21, 0.25)	−0.17 (−0.56, 0.22)	−0.19	0.39
Women Without HIV
Kynurenic acid	−0.10 (−0.34, 0.14)	−0.17 (−0.76, 0.42)	−0.07	0.82
Kynurenine	0.08 (−0.14, 0.31)	−0.29 (−0.85, 0.26)	−0.37	0.20
KYN/TRP ratio	0.11 (-0.11, 0.32)	−0.45 (-0.97, 0.08)	−0.55	0.04
Quinolinic acid	−0.01 (−0.24, 0.22)	−0.13 (−0.69, 0.43)	−0.12	0.69
Tryptophan	−0.04 (−0.27, 0.19)	0.28 (−0.29, 0.85)	0.32	0.28

Values are estimated marginal means (EMM) of Z-scored metabolites. All models are adjusted for BMI, age, hypertension, diabetes, alcohol use, and smoking status; models for women with HIV are additionally adjusted for nadir CD4⁺ cell count. *Difference represents the contrast of estimated marginal means (High depressive symptoms vs. Low/no depressive symptoms) in standard deviation units.

Fig. 2.

Adjusted standardized differences in KP metabolite levels between participants with high depressive symptoms and low/no depressive symptoms, stratified by HIV status. Women with HIV (WWH) and women without HIV (WWoH) were stratified by depressive symptom severity (CES-D ≥16 vs. <16). Adjusted differences in KP metabolites by depression status (High depressive symptoms – Low/no depressive symptoms), stratified by HIV status, were assessed using multivariable linear regression with estimated marginal means and standardized contrasts. All models were adjusted for age, BMI, hypertension, diabetes, alcohol use, and smoking status, with additional adjustment for CD4⁺ nadir in WWH. Among WWH, higher depressive symptoms (CES-D ≥ 16) were associated with significantly higher quinolinic acid (QA) levels compared with lower depressive symptoms CES-D < 16), with no significant differences observed for kynurenic acid (KA), kynurenine (KYN), tryptophan (TRP), or the KYN/TRP ratio. Among WWoH, higher depressive symptoms were associated with a significantly lower KYN/TRP ratio, with no significant differences in individual KP metabolite levels.

Women with HIV: In WWH, the high depressive symptom group had elevated levels of QA, with QA levels that were 0.52 SD higher than WWH with low/no depressive symptoms (p = .02). No statistically significant differences were observed for KA (diff = 0.14 SD, p = .53), KYN (diff = 0.08 SD, p = .69), TRP (diff = −0.19 SD, p = .39), or the KYN/TRP ratio (diff = 0.20 SD, p = .33).

Women without HIV: In WWoH, the high depressive symptom group had a significantly lower KYN/TRP ratio compared to the low/no depressive symptom group (diff = −0.55 SD, p = .04). There were no significant differences in individual metabolite levels between WWoH with high depressive symptoms and WWoH with low/no depressive symptoms.

3.4. Association of metabolites with CES-D depression and anhedonia scores

Separate linear regression models were created to examine the relationship between each metabolite and total CES-D score, as well as between each metabolite and anhedonia score (Table 4 and Fig. 3). These analyses were conducted for both WWH and WWoH. All models were adjusted for age, BMI, hypertension, diabetes, alcohol use, and smoking status, and additionally for CD4⁺ nadir in models restricted to WWH.

Table 4.

Adjusted multivariable linear regression models correlating KP metabolites with continuous CES-D and anhedonia scores in WWH and WWoH.

Metabolite (Z-score)	Std. Beta coefficient (95% CI)	P-value
Women With HIV - Anhedonia Score
Kynurenic acid	−0.20 (-0.38, -0.02)	0.03
Kynurenine (KYN)	−0.06 (−0.24, 0.11)	0.47
Quinolinic acid	0.24 (0.06, 0.41)	0.01
Tryptophan (TRP)	−0.05 (−0.23, 0.13)	0.60
KYN/TRP	−0.04 (−0.21, 0.13)	0.62
Women With HIV - CES-D Total Score
Kynurenic acid	−0.02 (−0.21, 0.16)	0.79
Kynurenine	−0.02 (−0.19, 0.15)	0.82
Quinolinic acid	0.25 (0.07, 0.43)	0.01
Tryptophan	−0.12 (−0.30, 0.06)	0.21
KYN/TRP	0.04 (−0.13, 0.21)	0.62
Women Without HIV - Anhedonia Score
Kynurenic acid	−0.23 (-0.42, -0.03)	0.02
Kynurenine	−0.12 (−0.31, 0.07)	0.20
Quinolinic acid	0.01 (−0.18, 0.20)	0.93
Tryptophan	0.01 (−0.19, 0.21)	0.92
KYN/TRP	−0.13 (−0.31, 0.05)	0.15
Women Without HIV - CES-D Total Score
Kynurenic acid	−0.05 (−0.25, 0.15)	0.60
Kynurenine	−0.03 (−0.22, 0.16)	0.77
Quinolinic acid	−0.03 (−0.22, 0.16)	0.76
Tryptophan	0.15 (−0.04, 0.35)	0.12
KYN/TRP	−0.11 (−0.29, 0.07)	0.23

All models are adjusted for BMI, age, hypertension, diabetes, alcohol use, and smoking status; models for women with HIV are additionally adjusted for nadir CD4⁺ cell count. Both outcome and predictor were standardized (z-scored), so coefficients represent SD change in metabolite per 1 SD change in score.

Fig. 3.

Adjusted standardized associations between depressive symptom severity (anhedonia and CES-D total score) and KP metabolites, stratified by HIV status. Separate adjusted linear regression models were constructed to examine associations between each KP metabolite and total CES-D score, as well as between each KP metabolite and anhedonia score, in women with HIV (WWH) and women without HIV (WWoH). All models were adjusted for age, BMI, hypertension, diabetes, alcohol use, and smoking status, with additional adjustment for CD4⁺ nadir in WWH. Higher CES-D total score was associated with higher QA levels in WWH only. Greater anhedonia was associated with higher QA and lower KA levels in WWH, while in WWoH, greater anhedonia was associated with lower KA levels only. No other significant associations were observed.

CES-D total score. Higher CES-D total score was significantly associated with higher QA levels in WWH ($\mathrm{\beta }$ = 0.25, p = .01). No other significant associations were observed between the CES-D total score and metabolite levels in either WWH or WWoH (all p > .05), including the KYN/TRP ratio in WWoH ($\mathrm{\beta }$ = −0.11, p = .23).

Anhedonia Score. In WWH, greater anhedonia was significantly associated with higher levels of QA ($\mathrm{\beta }$ = 0.24, p = .01) and lower levels of KA ($\mathrm{\beta }$ = −0.20, p = .03). In WWoH, higher anhedonia was similarly associated with lower levels of KA ($\mathrm{\beta }$ = −0.23, p = .02), but no significant association was found with QA ($\mathrm{\beta }$ = 0.01, p = .93.

4. Discussion

This study examined the relationship between KP metabolites and depressive symptoms in women with and without HIV. Our findings support our hypothesis that KP activation is increased in WWH who report higher depressive symptoms. Specifically, we observed higher levels of QA, a neurotoxic KP metabolite, in WWH with high depressive symptoms compared to WWH with low/no depressive symptoms. In addition, we found that KP metabolites, namely QA and KA, were significantly associated with anhedonia and overall depressive symptom severity in WWH.

Group comparison findings: Our group comparison finding of increased QA in WWH with high depressive symptoms suggest activity of the KP and its neurotoxic branch in WWH. These findings are consistent with prior evidence demonstrating KP activation in PWH with greater depressive symptoms. Previous studies have reported higher plasma QA and KYN/TRP ratios associated with greater depressive symptom severity (Drivsholm et al., 2021; Martinez et al., 2014). Moreover, ART-mediated reductions in KP activity, as measured by decreases in the KYN/TRP ratio, have been found to partially mediate improvements in depressive symptoms, further supporting a mechanistic link between KP and depressive symptoms (Martinez et al., 2014). In this study, KP activation and its association with depressive symptoms persists in a cohort of women that are virally suppressed after covariate adjustment, further implicating the KP in depression in PWH.

In our group comparisons, we also found decreased KP activity, as measured by KYN/TRP, in WWoH with depressive symptoms compared to WWoH with low/no depressive symptoms, which contradicts the hypothesis that KP activity is higher in those with greater depressive symptoms. However, we had a low number of WWoH with depressive symptoms (n = 13), limiting our ability to draw reliable conclusions from this comparison.

In many studies in PWoH and PWH, group comparisons have been inconsistent. These discrepancies may reflect differences in sample size, variation in how depression is measured, or the heterogeneous nature of depression itself. Our study addresses several of these limitations by leveraging a large, well-characterized cohort of women with and without HIV. Strict inclusion criteria and adjustment for relevant covariates reduce confounding by social and behavioral factors and restricting the sample to women minimizes variability related to sex differences.

In addition, our dimensional approach allows us to examine how the KP relates to depression severity and to a specific behavioral construct, namely anhedonia. Because depression is defined by a cluster of symptoms, heterogeneity in symptom profiles can obscure biological relationships. The KP may be more directly involved in the pathogenesis of particular behaviors or symptom domains that are not fully captured by a categorical diagnosis of depression, which requires endorsement of five of nine possible symptoms and can vary substantially across individuals.

Another explanation for discrepant findings, particularly in PWoH, may be lower baseline levels of inflammation in cohorts without inflammatory conditions such as HIV. It is possible that the KP plays a larger role in inflammation-associated depression, of which HIV-associated depression may be one subtype.

Correlational analyses: Examining specific relationships between the KP and depression symptomatology, we documented that KP metabolites KA and QA were significantly correlated with anhedonia severity. Our findings are in agreement with other independent studies in youth and adults with depression (Anderson et al., 1990; Cowen et al., 1989; Gabbay et al., 2012a; Maes et al., 1996; Savitz et al., 2015d). This finding also fits our earlier study in adolescent depression where we identified increased KP activity (indexed by KYN/TRP ratio) in melancholic depression subtype, for which anhedonia is a core symptom, compared to non-melancholic and healthy controls (Gabbay et al., 2010). Since anhedonia reflects deficits in reward processes, this finding suggests that the KP may induce alterations within the neural reward circuitry.

Neuroimaging studies in adult participants with depression have documented relationships between blood KP metabolite levels and volumes of reward-related brain regions (Meier et al., 2016; Savitz et al., 2015a, 2015b, 2015c). Our group also reported relationships between KP metabolites and connectivity within salience/reward neurocircuitry in depressed youth (DeWitt et al., 2017), supporting the idea that the KP may interact with the reward neurocircuitry. Our findings with anhedonia underscore the importance of a dimensional investigative approach focusing on behavioral constructs rather than the categorical diagnostic scheme of psychiatric conditions.

Specific metabolite contributions to depressive symptoms: In the current study, increased QA was associated with increased anhedonia and depressive symptom severity, while KA was associated with decreased anhedonia severity. This relationship may be mediated through the interaction of QA and KA with the NMDA receptor. QA is an NMDA receptor agonist, hypothesized to contribute to depression through excitotoxity and oxidative stress. KA is NMDA receptor antagonist with putative neuroprotective properties that may decrease depressive symptoms (illustrated in Fig. 4). Indeed, ketamine, which is also an NMDA receptor antagonist with rapid antidepressant properties, was shown to improve anhedonia (Nogo et al., 2022; Wilkowska et al., 2021). It is possible that the effects of KA in depression may also be mediated through other receptors including the α7-nicotinic receptor, G protein-coupled receptor GPR35, and the aryl hydrocarbon receptor (AhR) (Hilmas et al., 2001; Savitz, 2020; Wang et al., 2006). Both GPR35 and AhR may play a role in the modulation of inflammation, which contribute to depression pathogenesis (Grishanova and Perepechaeva, 2024; Wu et al., 2023.

Fig. 4.

Proposed relationship of HIV-associated factors and KP activation in anhedonia pathogenesis in PWH. HIV-associated inflammation and viral proteins increase IDO activity, converting TRP to KYN. In HIV, KYN may be primarily metabolized through the neurotoxic branch of the KP, leading to higher QA and lower KA. QA, an NMDA agonist, causes excitotoxicity and neuronal death by over-activating NMDA receptors, while reduced KA, an NMDA antagonist, enhances QA's effects. This neuronal damage, especially in areas related to reward circuitry, may contribute to anhedonia in PWH. TRP, tryptophan; IDO, Indoleamine 2,3-dioxygenase; KYN, kynurenine; BBB, blood-brain barrier; CNS, central nervous system; 3-HK, 3-Hydroxykynurenine; 3-HAA, 3-hydroxyanthranilic acid; QA, quinolinic acid; KA, kynurenic acid.

It is plausible that central levels of QA and KA in PWH can contribute to excitotoxic effects. KYN and 3-hydroxykynurenine (3-HK) can cross the blood-brain barrier (BBB) via active transport mechanisms and can be metabolized within the CNS into either QA or KA. A recent meta-analysis demonstrated consistent concordance between peripheral and central measures of KYN and 3-HK (Skorobogatov et al., 2021). Although QA and KA themselves poorly cross the BBB via passive diffusion, greater availability of KYN and 3-HK can increase production of QA and/or KA in the brain.

Additionally, BBB disruption - which can be induced by chronic inflammation, HIV proteins, and QA itself (Cheng et al., 2018; Kanmogne et al., 2005; Raymond et al., 2016; Reynolds and Morton, 1998, 1998t’Astný et al., 2000) and has been implicated in depression pathogenesis (Cheng et al., 2018; Dion-Albert et al., 2022; Medina-Rodriguez and Beurel, 2022; Menard et al., 2017) – may enable greater passage of peripheral KP metabolites, further augmenting brain QA and KA. In parallel, activated CNS cells such as brain macrophages, microglia, and astrocytes, which are known to be activated during HIV infection, can locally produce QA and KA, further contributing to their CNS concentrations.

In our study and others, PWH have been shown to have higher levels of both peripheral and central QA compared to PWoH (Drivsholm et al., 2021; Eggertsen et al., 2023; Heyes et al., 1991, 2001). In one study, elevated cerebrospinal fluid QA in PWH was associated with cerebral atrophy in regions vulnerable to excitotoxic injury and implicated in depression, including the striatum and limbic cortex. In contrast, such atrophy was not observed in regions relatively resistant to excitotoxicity, such as the non-limbic cortex, thalamus, and white matter (Heyes, 2001 Although we observed associations with absolute measures of QA and KA with depressive symptom severity, the ratio of QA to KA may be relevant to neuropsychiatric disease pathogenesis. The ratio of QA to KA peripherally and centrally likely varies with inflammatory state and disease process. Future studies are needed to determine whether the relative balance of KP metabolites is associated with depression pathogenesis, and how this may differ in other neuropsychiatric disorders.

Relevant to the pathogenesis of anhedonia, a large body of evidence has tied KP neurotoxicity to alterations within the reward neurocircuitry. Animal studies showed that the toxicity of 3-HK and 3-HAA depends on transport-mediated cellular uptake, which varies by brain region (Okuda et al., 1998). Striatal and cortical neurons are especially susceptible to these toxins (Okuda et al., 1998). QA directly affects the mesolimbic dopaminergic system, inducing dopaminergic and GABAergic neuronal death (Araujo et al., 2000; Beskid and Finkiewicz-Murawiejska, 1992; Kurachi et al., 2000; Sumiyoshi et al., 2004). Conversely, KA regulates the activity of midbrain dopaminergic neurons, and decreased endogenous KA levels result in reduced dopamine release (Erhardt et al., 2009; Wu et al., 1994.

While we did not examine KP metabolites in the CSF, converging evidence suggests that KP activity in the blood parallels its activity in the brain. Findings include increased plasma IDO activity coupled with parallel increases of KYN, QA, and KA in the CSF in response to treatment with IFN-α (Raison et al., 2010). Preclinical data have specifically linked peripheral and central IDO activation to anhedonic behaviors (Henry et al., 2008; Moreau et al., 2008). Additionally, peripheral inhibition of IDO in mice blocked the central transcription of IDO in the brain and the development of depressive-like and anhedonic behaviors following immunological stimulation (O'Connor et al., 2009). Furthermore, Fukui et al. demonstrated that KYN and 3-HK cross the BBB and may each significantly contribute to their cerebral pools (Fukui et al., 1991b).

General inflammation in depression and HIV: This study contributes to the limited literature examining the relationship between the KP and depression in PWH (Drivsholm et al., 2021; Mudra Rakshasa-Loots et al., 2023). Prior work suggests that increased peripheral and neuroinflammation in HIV infection contributes to a higher incidence of depressive symptoms among PWH (Mudra Rakshasa-Loots et al., 2022). In this context, the chronic inflammation that persists in HIV may be associated with alterations in KP metabolism, including a relative shift toward greater production of neurotoxic and reduced production of neuroprotective KP metabolites, which may relate to depressive symptom severity. The IDOze study did measure inflammatory biomarkers; however, these analyses were outside the scope of the current report. These biomarkers are currently being examined in relation to CES-D scores as part of a separate analysis. Additional studies are needed to further delineate the interrelationships among inflammation, KP metabolism, and depression in the context of HIV.

Several limitations are important to note. First, the study did not use comprehensive or diagnostic depression or anhedonia assessments. The CES-D is an epidemiologic assessment of depressive symptomatology rather than a clinical tool used to diagnose depression, and anhedonia was assessed using a subset of two CES-D items, rather than a specific anhedonia survey. The duration of these symptoms was also not assessed. Second is the small sample size of depressed WWoH, which may have contributed to Type II error and the lack of significant associations between anhedonia and depression severity and KP metabolites in the WWoH group. Another limitation is not assessing KP metabolites in the CSF, however, as presented above, evidence suggests that KP activity in the blood correlates with KP activity in the brain.

In addition, it is important to note that these findings are based on secondary analyses of existing data from the IDOze Study. The primary goal of the IDOze study was to examine the bidirectional association between sleep and KP metabolites and their impact on cardiometabolic indices among WWH and a demographically and behaviorally similar comparison group of WWoH. To reduce potential confounding, IDOze excluded women who self-reported provider diagnosed significant psychiatric illness (both treated and untreated), had uncontrolled hypertension or diabetes, used hormonal contraception or hormonal replacement therapy, were unable or unwilling to abstain from using various substances during the 7-10 day sleep/actigraphy period, and WWH without virologic suppression. Consequently, our study sample is not representative of WWH at large but rather reflects a subset who met criteria for the sleep study, specifically selecting for women with limited medical and psychiatric comorbidities, no substance use, and well-controlled HIV. While this limits generalizability, the comparison of demographically and behaviorally similar WWH and WWoH may reduce confounding by social and behavioral factors, allowing the analyses to more directly interrogate biological mechanisms that may contribute to depression risk in WWH.

Taken together, this study provides additional evidence for the possible role of the KP in depressive symptoms in PWH. Specifically, our findings support previous work suggesting that QA may play a role in anhedonic and depressive symptoms, and KA may act as neuroprotector. Future studies in PWH should focus on behavioral constructs and utilize neuroimaging studies focusing on neurocircuits such as anhedonia and the reward neurocircuitry.

CRediT authorship contribution statement

Vilma Gabbay: Writing – review & editing, Writing – original draft, Supervision, Conceptualization. Caitlin Hills: Writing – review & editing, Visualization. Francisco Cardozo: Formal analysis. Aaron B. Chance: Writing – original draft. Anjali Sharma: Writing – review & editing, Supervision, Conceptualization. Wenzhu Mowrey: Writing – review & editing, Formal analysis. Zoe Baker: Writing – review & editing, Data curation. Qibin Qi: Writing – review & editing, Methodology. Helen J. Burgess: Writing – review & editing, Resources, Funding acquisition, Data curation. Joan W. Berman: Writing – review & editing, Methodology. Stephen J. Gange: Writing – review & editing. Deborah R. Gustafson: Writing – review & editing. Kathleen M. Weber: Writing – review & editing, Project administration, Methodology, Funding acquisition, Conceptualization. Audrey L. French: Writing – review & editing, Resources, Methodology, Funding acquisition.

Role of the funding source

This study was supported by the National Institutes of Health (NIH) under Awards R01HL142116 (A French and H Burgess); R01MH128878 (V Gabbay and A Sharma); R01DA054885 (V Gabbay, A Sharma, J Starrels); R01MH131207 (V Gabbay, J Berman, A Sharma, C Hills); R21MH126501, R01MH120601, R21MH121920, and R01MH126821 to V Gabbay; MSTP Training Grant T32GM149364 (C Hills); U01 HL146204 (A Sharma, K Anastos, D Hanna); U01 HL146202 (T Wilson, D Gustafson); U01HL142645 (A French, M Cohen, R Ross); U01-HL146193 (S Gange). All data utilized for these analyses were collected from NHLBI IDOze R01HL 142116 and the core Women's Interagency HIV Study (WIHS), now the MACS/WIHS Combined Cohort Study (MWCCS) at the Chicago, Brooklyn, and Bronx study sites. The content is solely the responsibility of the authors and does not necessarily represent the official views of the NIH.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Data availability

Data will be made available on request.