Targeting Tryptophan Metabolism for Tuberculosis Biomarkers and Host Directed Therapy

Abstract

Greater understanding of the role of tryptophan metabolism in the immune response to tuberculosis (TB) has provided promising avenues to explore new diagnostic and therapeutic modalities. Animal and human studies have demonstrated that host indoleamine 2,3-dioxygenase-1 (IDO1) is upregulated in response to infection with Mycobacterium tuberculosis resulting in increased tryptophan metabolism to kynurenine. In TB disease, this is evidenced by elevation of the plasma kynurenine to tryptophan ratio, which is reversed with effective TB treatment thus showing utility as a potential diagnostic and therapeutic biomarker. Kynurenine and downstream metabolites promote an immunosuppressive microenvironment in TB granulomas, which may facilitate immune evasion. IDO inhibition in non-human primates has highlighted its potential role as host-directed therapy by demonstrating increased T cell trafficking to the granuloma core, reduced bacterial burden, and decreased immunopathology. To realize the potential of exploiting the tryptophan to kynurenine metabolic pathway, innovative biomarker and host-directed therapy trials are needed.

INTRODUCTION

Tuberculosis (TB) continues to be a major public health challenge despite the introduction of rapid, accurate diagnostics and the implementation of new anti-TB drugs and regimens.[1] Regaining the status as the leading global cause of infectious disease death, the End TB Strategy goals for 2030 are far off track, and recent funding disruptions for TB care have the potential to reverse years of hard-won progress.[2, 3] As outlined by the World Health Organizations on World TB Day 2025, there is a need for giant and innovative steps forward to quell this devastating disease. One area of promise is leveraging our understanding of the host immune response against Mycobacterium tuberculosis (Mtb) to develop novel biomarkers and host-directed therapy (HDT) for TB disease.[4]

All commonly used diagnostics to confirm active TB disease were optimized for sputum samples, rely on direct detection of Mtb, and require a laboratory for implementation. Such requirements can be formidable barriers to TB diagnosis in primary care, resource-limited areas where persons with TB disease often first present for care. Thus, developing a lab-free, point-of-care (POC) test that uses blood or urine samples to diagnose or triage persons with presumed TB disease is a major priority in TB diagnostics.[5] Alterations in the host immune response offer an attractive non-bacteria-based option to identify active TB disease and potentially monitor response to therapy.

Regarding TB treatment, drug resistance continues to be a major threat limiting the effectiveness of antibiotics active against Mtb.[6] Even those who complete effective therapy can be left with life-long pulmonary impairment from lung cavitation and fibrosis.[7] By augmenting beneficial host immune responses and reducing pathological inflammation, HDT has the potential to be a valuable adjunct to antibiotics. Possible beneficial effects include enhancing immune-mediated Mtb clearance and mitigating damaging host responses that are associated with post TB lung disease (PTLD).[8, 9]

Over the last decade, advanced methods in metabolomics and transcriptional profiling have helped illuminate the role of tryptophan metabolism in TB disease. Through this increased understanding, ideas have emerged on how to utilize this knowledge to combat TB.[10] In this review, we discuss the association between human tryptophan metabolism and infection and disease caused by Mtb. Our goal is to highlight how this host response pathway can be targeted to improve TB diagnostics and develop novel HDTs that improve TB outcomes.

Tryptophan metabolism and the human immune system

Tryptophan is an essential amino acid, and its metabolism has profound implications for the human immune response. Tryptophan can be metabolized through multiple different pathways, but metabolism to kynurenine appears to have the greatest immunologic consequences.[11] This occurs via one of three enzymes: indoleamine -2,3-dioxygenase-1 (IDO1), indoleamine -2,3-dioxygenase-2 (IDO2), and tryptophan-2,3-dioxygenase (TDO). Kynurenine is subsequently metabolized to anthranilic acid, 3-hydroxyanthranilic acid, quinolinic acid, and picolinic acid, which are collectively referred to as kynurenines. Through agonism of the aryl hydrocarbon receptor,[12] these molecules promote differentiation of regulatory T cells at the expense of more inflammatory TH17 cells and lead to reduced secretion of chemokines crucial for T cell recruitment to the site of disease,[13, 14] thereby creating immunosuppressive microenvironments that impair T cell responses. The potential consequences of this immunologic shift were first observed in pregnancy, where tryptophan metabolism to kynurenines in the placenta is required to prevent a fatal maternal T cell response against the fetus.[15] However, these immunomodulatory effects are increasingly recognized to have consequences in a variety of pathologic states including cancer biology and infectious diseases such as TB. Impaired T-cell responses can promote immune evasion, and a variety of IDO inhibitors and kynureninases remain in clinical trials to determine their ability to promote efficacious immune responses against cancer cells.[16]

More recent data indicate that the immunologic importance of tryptophan catabolism is likely even more far reaching than previously understood. IDO-mediated breakdown of tryptophan in dendritic cells is essential for the survival of long-lived plasma cells, which are responsible for durable antibody responses.[17] Additionally, the intracellular depletion of tryptophan that results from catabolism to kynurenine may have its own immunologic impacts. A lack of intracellular tryptophan leads to tryptophan to phenylalanine codon reassignments, which can both impair the activity of translated proteins and expand the diversity of antigens presented to immune cells.[18] The full implications of these more recently described immunologic effects have yet to be fully explored in the context of TB pathophysiology.

An overview of the association between tuberculosis and tryptophan

The ability of Mtb to induce tryptophan catabolism in a burden-dependent manner has been demonstrated across human cohorts (Figure 1 provides an overview of the impact of M. tuberculosis infection on tryptophan metabolism and related immune effects).[19–21] These observations mirror those in non-human primate [22, 23] and mouse models of TB disease.[24] Transcriptomics studies from humans and non-human primates suggest that dramatic upregulation of IDO1 in the lungs is the primary driver of this response.[19, 25, 26] So why might tryptophan catabolism have particular relevance in TB disease? For one, IDO1 is strongly induced by IFN-y secretion.[27] Patients with TB disease generally secrete high levels of IFN-y [28] and exposure to Mtb is defined by INF-y T-cell responses to Mtb antigens.[29] At the same time, Mtb is able to synthesize its own tryptophan, thereby avoiding any untoward consequences of IFN-y-induced tryptophan starvation.[30] The result is high IDO1 expression at the periphery of TB granulomas and accumulation of kynurenines in the lymphocytic cuff. [22] This strongly suggests that IDO1-mediated tryptophan catabolism is a ubiquitous host response to Mtb with a spatially defined role within the TB granuloma. It also highlights a key regulatory and counter-regulatory balance in TB pathogenesis. While INF-y deficiency can be a cause of TB disease[31], immunocompetent individuals with TB disease express high levels IFN-y yet fail to adequately control bacterial replication. One explanation for this apparent contradiction is that while some INF-y is necessary to control Mtb replication, an overabundance results in overexpression of IDO1 and impaired T-cell responses to Mtb. This could explain why PD-1 blockade, which results in increased IFN-y secretion increases the risk for TB disease.[32, 33]

Figure 1.

Overview of increased host tryptophan to kynurenine metabolism in Mycobacterium tuberculosis infection and related immune effects. IDO1, indoleamine 2,3-dioxygenase-1.

Whether accumulation of kynurenines and their associated effects on T-cells has a net benefit for host or pathogen likely depends on the site and stage of Mtb replication. In TB meningitis, enhanced tryptophan breakdown in the CSF has been reproducibly associated with lower mortality.[34, 35] Here the immunosuppressive effects of tryptophan breakdown may be necessary to mitigate intracranial inflammation and pressure. Yet in pulmonary TB disease, there is accumulating evidence that tryptophan breakdown serves as a mechanism for Mtb immune evasion. Mtb strains unable to synthesize tryptophan fail to establish infection, suggesting this represents an essential mechanism for TB pathogenesis.[36] Spatial studies of IDO1 expression in TB granulomas show it co-localizes with increased abundance of regulatory T-cells[25] and an immunosuppressive microenvironment.[37, 38] Inhibition of IDO in non-human primate models of TB disease results in enhanced tracking of T-cells to the granuloma core, where the vast majority of Mtb resides.[22] Further, IDO inhibition in non-human primates has not shown any evidence of pathologic lung inflammation and has been associated with reduced bacterial burden.[39, 40]. Together, these data suggest IDO-mediated tryptophan catabolism helps Mtb evade host immunity by limiting efficacious T-cell responses. Thus, this pathway represents a potential target for HDTs to facilitate immunological clearance in pulmonary TB.

Tryptophan Pathway as TB Biomarker

Increased tryptophan breakdown to kynurenine in TB disease is reflected in the plasma concentrations of both molecules and summarized by the kynurenine/tryptophan (Kyn/Trp) ratio. Studies from multiple human cohorts suggest the plasma Kyn/Trp ratio reflects disease activity in human TB, increasing gradually prior to TB disease diagnosis[20, 21] and declining with appropriate TB treatment.[19, 21] This has raised the possibility it could be used as a blood-based biomarker to diagnose pulmonary TB. Indeed, many studies have shown the plasma Kyn/Trp ratio to be an accurate biomarker of TB disease, with area under the receiver operator characteristic curve (AUC) estimates ranging from 0.67 to 0.99.[19, 21, 41–44] Accuracy appears to be similar in important sub-populations including people living with HIV (PLHIV)[21, 43] and pregnant women,[45] though the threshold for TB diagnosis may need to be adjusted in these groups given the potential for both states to increase basal tryptophan breakdown.[15, 46, 47] In PLHIV, the Kyn/Trp ratio is correlated with the size of the viral reservoir [46] and may therefore have lower accuracy in PLHIV not virally suppressed or require adjustment for the degree of HIV viremia.[48] Limited data exists in children, with one study finding the plasma Kyn/Trp had a lower AUC (0.67) for the diagnosis of pulmonary TB in persons <15 years old.[49] While many adult studies indicate the plasma Kyn/Trp ratio meets the World Health Organization (WHO) Target Product Profile (TPP) for a TB triage test (>90% sensitivity at >70% specificity),[50] it falls short of the TPP for a new non sputum POC TB diagnostic.[51] Thus, combination with other biomarkers would be needed to meet these accuracy standards.

Whether the Kyn/Trp ratio can have utility as a TB triage test will likely depend on whether the detection method used can fill a practical niche for TB control programs. It is unlikely to replace current microbiologic testing in hospitals and referral centers but could potentially have a role in community screening programs and rural health centers not equipped with a laboratory.[52] Yet to fill this role, detection would need be translated to a POC platform similar to what has been done with CRP.[53] If POC, lab-free measurement of the plasma Kyn/Trp ratio were possible, this could hold practical advantages over other triage test such as Mtb PCR detection using tongue swabs [54] or computer-aided diagnosis of chest x-rays,[55] which generally have greater equipment and reagent needs. Importantly, ELISA kits have shown similar accuracy to mass spectrometry for measuring the Kyn/Trp ratio[43], suggesting translation to immune-based POC platforms is possible.

While most biomarker studies of the Kyn/Trp ratio come from analysis of plasma, it has also shown utility as a biomarker of extrapulmonary TB when measured in other biofluids. The Kyn/Trp ratio is elevated in pleural fluid in patients with pleural TB[56] and in the CSF of patients with TB meningitis.[34, 35, 57] However, additional study is needed to determine its clinical utility in these contexts. In the case of TB meningitis, other causes of lymphocytic meningitis show similar CSF biochemical changes,[58] potentially limiting its clinical value.

Other potential clinical biomarker applications for the plasma Kyn/Trp ratio include monitoring response to TB treatment and predicting TB disease progression. In pulmonary TB, the plasma Kyn/Trp ratio is markedly elevated at diagnosis and declines in a stepwise fashion with appropriate treatment.[19, 21, 59] Higher plasma Kyn/Trp ratio values at pulmonary TB diagnosis have been associated with longer time to culture clearance[19] and higher mortality.[44] Yet no study to date has definitively assessed whether it can predict treatment failure or relapse, which would have the greatest clinical utility. Conversely, higher tryptophan is predictive of increased mortality in TB meningitis. [34, 35] Thus, a lower CSF Kyn/Trp ratio may be a marker of favorable treatment response in TB meningitis. The plasma Kyn/Trp ratio is also elevated in latent TB infection,[19] decreasing with latent TB treatment, and gradually increasing in those who progress to TB disease.[20, 21] However, while these differences are evident at the population level, they are small in magnitude, and it is unclear whether they could have clinical significance such as predicting TB disease progression.

IDO inhibitors for HDT in pulmonary TB

The potential benefit of modulating host tryptophan metabolism to reprogram granuloma microenvironments and enhance control of Mtb has been elegantly demonstrated in a series of non-human primate studies.[22, 39, 40] The same group of investigators administered the IDO inhibitor D-1 methyl tryptophan (D1MT) in an high Mtb dose acute infection model, a low dose infection model as an adjunct to anti-TB treatment, and a Mtb/SIV co-infected model administered with antiretroviral therapy (ART). When given alone, D1MT was found to reduce bacterial burden and lung damage and lead to increased survival.[22] Importantly, D1MT was also associated with increased trafficking of CD4+ T cells to the core of granulomas where there exists a higher burden of bacteria. The additional studies found that when given along with TB treatment or ART, D1MT was associated enhanced recruitment of effector T cells to the granuloma core, improved Mtb specific CD4+ and CD8+ T cell responses, and reduction of granuloma necrosis (Figure 2).[39, 40] None of the studies found an increase in Mtb burden to organs outside the lung indicating IDO suppression did not lead to an increase in disease dissemination or increased immunopathology.

Figure 2.

Schematic representation of IDO1 inhibition and impact on M. tuberculosis granulomas. Based on studies conducted in rhesus macaques. IDO1, indoleamine 2,3-dioxygenase-1.

Our accumulating knowledge of the effect of Mtb hijacking of host tryptophan metabolism to subvert the host immune response and promising data on IDO inhibition in NHP TB models posit that disrupting the metabolism of tryptophan to kynurenine may be a promising HDT strategy.

Regarding implementing a HDT clinical trial targeting tryptophan metabolism among persons with pulmonary TB, the field of cancer has laid the groundwork in developing numerous IDO inhibitors that could be considered for use. IDO1 upregulation in cancer is thought to foster an environment of immune escape and subverting this process is meant to enhance immune control of malignant cells.[60] Numerous ongoing clinical trials have been or are being conducted for various types of cancer utilizing IDO inhibitors that preferentially target IDO1.[16] The best studied oral IDO inhibitors in cancer to date include Indoximod (competitive IDO1 inhibitor), Epacadostat (Trp competitive inhibitor with selectively for IDO1), and BMS-986205 (specific for ID01). Thus far, the anti-tumor efficacy of IDO inhibitors has been limited when used as single agents, slightly enhanced when used adjunctively with chemotherapy and/or immune checkpoint blockade but overall underwhelming. However, in terms of safety, and important for moving forward in the field of TB, IDO inhibitors have been shown to be well tolerated and safe in adults and children in the various cancer trials.[61–63] Additionally, the various studies provide a critical foundation for pharmacokinetic profiles and dose optimization of IDO inhibitors. Some agents, such as PF-0684003, a selective IDO1 inhibitor, would be less suitable for TB disease given its ability to cross the blood-brain barrier and potential to worsen underlying subclinical TB disease in the central nervous system.[16]

Future clinical research priorities

Biomarker studies

Despite promising initial results, nearly all studies examining the plasma Kyn/Trp ratio as a TB biomarker have been case control studies, with follow up measures performed only in those successfully completing TB treatment. Prospective studies are warranted to further assess the utility of the plasma Kyn/Trp ratio as a biomarker of TB disease and treatment response either alone or in combination with other analytes. Data published to date indicate it is unlikely to meet the TPP for a confirmatory TB diagnostic (≥65% sensitivity and ≥98% specificity for a POC, lab-free test).[51] Thus, study designs should focus on its potential for implementation as a TB triage test. This could include evaluation in community screening settings in participants with and without TB symptoms (Figure 3A) or in outpatient settings in high-TB prevalence areas where patients might present for TB evaluation (Figure 3B). Treatment response studies should include prospectively enrolled participants with and without drug resistance with plasma collected frequently (i.e. biweekly or monthly) at early clinical time points when changing treatment in non-responders could improve clinical outcomes (Figure 3C).

Figure 3.

Study designs for evaluating tryptophan biosignatures in tuberculosis. CAD, computer-assisted detection; CXR, chest x-ray; DS, drug-susceptible; Kyn, kynurenine; MDR, multidrug-resistant; TB, tuberculosis; Trp, tryptophan.

HDT Clinical Trials

The concept of HDTs, designed to improve host response against infection, are older than antibiotics, with Koch attempting to use tuberculin injections as therapy in 1890.[64] Clinical trials are critical to demonstrate whether our increased understanding of the host immune response to TB can translate into pathways that can be targeted for effective HDTs. As one example highlighting this point, based on pre-clinical data there was much promise that Vitamin D supplementation could improve Mtb killing in humans;[65] however, clinical trials failed to show an impact on treatment outcomes.[66, 67]

While enthusiasm in the scientific field for HDT is high, the roll out for clinicals trials has been slow. There are only ten HDT trials for TB disease currently listed in clinicaltrials.gov with just three actively recruiting participants for evaluation of either adjunctive doxycycline, n-acetylcysteine, acetylsalicylic acid, and/or ibuprofen. In regard to recently completed studies, a phase 2 trial evaluating 4 adjunctive HDTs among drug-susceptible pulmonary TB participants provided not only encouraging findings on the utility of HDT to preserve lung function but a framework for future such studies.[8] While primary study endpoints were safety and tolerability, secondary efficacy outcomes included sputum culture conversion, spirometry, and inflammatory markers including PET CT and serum CRP. No differences in bacterial outcomes were found but those in the CC-11050 (type 4 phosphodiesterase inhibitor) and everolimus (serine/threonine protein kinase mTOR inhibitor) groups had an ~6% increased recovery of FEV₁ with everolimus also being associated with max SUV reduction on PET CT. [68] This study highlights the importance of including lung health and inflammatory marker outcomes in HDT clinical trials.

In terms of optimal clinical trial designs, further guidance can be gleaned from two large, ongoing TB HDT clinical trials including incorporation of both lung function and bacterial clearance as outcomes [69]; however, many questions and uncertainties remain. Utilizing various international trial sites through existing trial networks would help ensure a geographically and demographically diverse study population and enhance generalizability. An open question is whether early HDT trials should target pulmonary TB patients with moderate to severe disease [8] given a higher likelihood to see an improvement or be more inclusive of all persons with disease. Beyond traditional measures of safety and microbiological outcomes; serial measurement of lung function should be a key primary outcome of HDT aimed at limiting immunopathology. For bacterial clearance measures, more nuanced measures including time to culture positivity, time to detection, or time to extinction would allow for a great ability to detect difference in microbiological effect.[70] Additionally, the inclusion of outcomes such as quality of life, radiological improvement, inflammatory biomarkers, metabolic responses, and measure for pharmacodynamics will provide key metrics to provide a more comprehensive evaluation of HDTs (see Supplemental Figure 1). Particularly for IDO inhibitors, it would be prudent to choose an agent that has limited ability to cross the blood brain barrier and/or rule out subclinical CNS disease among study participants. If early phase 2 adaptive RCTs show promise incorporating plans for a seamless transition to phase 3 RCTs would be critical. [71, 72]

CONCLUSION

Recent insights gained through the advent of new scientific approaches including sensitive – omics and spatial methods have helped illuminate the role of IDO1 upregulation in TB disease. Harnessing and exploiting this information, investigators have found the Kyn/Trp ratio to be an attractive biomarker for disease detection and response to treatment and that IDO inhibition can enhance the effectiveness of the host immune response to Mtb in granulomas. This preliminary data provides a strong impetus for the research community to confirm and validate the findings through innovative clinical studies and trials. Only then will we determine the role and potential impact of measuring and disrupting tryptophan metabolism to enhance the detection and management of TB.

Table 1.

Characteristics of Clinical Studies evaluating the use of the Kynurenine/Tryptophan ratio as a biomarker for Active Tuberculosis^*

Study Years Location	Study Focus	Total N	Study Group	Control Group	TB Disease Type	Findings#
1996–2023 Thailand[42]	1. Diagnostic biomarker	26	Adults with TB/HIV (n=13)	PLHIV without TB (n=13)	Clinical & confirmed pulmonary TB	Diagnostic: 77% sensitivity at 77% specificity
2008–2012, South Africa[21]	1. Diagnostic biomarker 2. Predictive biomarker 3. Treatment response	139^	Adults with TB/HIV (n=32)	PLHIV without TB (n=70) HIV-infected with pneumonia (n=37)	Clinical & confirmed pulmonary TB	Diagnostic: 97% sensitivity at 99% specificity Prediction: 61% sensitivity at 99% specificity Response: Decreased to levels similar to controls after treatment
2008–2012, South Africa[43]	1. Diagnostic biomarker	90*	Adults with TB/HIV (n=20) and TB disease alone (n=25)	PLHIV with latent TB (n=20) HIV-uninfected with latent TB (n=25)	Clinical & confirmed pulmonary TB	Diagnostic: 85% sensitivity at 92% specificity in HIV-negative and 90% sensitivity at 80% specificity in PLHIV
2009–2015 South Africa, Georgia[19]	1. Diagnostic biomarker 2. Treatment response	231	Adults with TB/HIV (n=85) and TB disease alone (n=89)	Asymptomatic adults with (n=20) and without (n=37) latent TB	Confirmed pulmonary TB	Diagnostic: 88% sensitivity at 70% specificity Response: Decreased to levels similar to controls after treatment. Baseline level correlated with time to culture conversion.
2010–2011 Japan[44]	1. Treatment response	259	Elderly adults (n=174)	Age and gender matched controls (n=85)	Confirmed pulmonary TB (80%) and extrapulmonary TB (20%)	Response: High levels at TB diagnosis associated with mortality, 87% sensitivity at 67% specificity for predicting mortality
2010–2011 Japan[56]	1. Diagnostic biomarker (pleuralfluid)	92	Elderly adults (n=34)	Carcinomatous pleurisy (n=36), para pneumonic effusion (n=15), other pleural effusion (n=7)	Pleural TB confirmed by biopsy or microbiology	Diagnostic: 88% sensitivity at 69% specificity
2011–2013 Ethiopia[48]	1. Diagnostic biomarker	249	Adults with TB/HIV (n=124)	PLHIV without TB (n=125)	Confirmed pulmonary TB	Diagnostic: 73% sensitivity at 61% specificity
2011–2014, South Africa[45]	1. Diagnostic biomarker	189*	Pregnant women with TB/HIV (n=72)	Pregnant women living with HIV (n=117)	Clinical & confirmed pulmonary TB (92%) and extrapulmonary TB (8%)	Diagnostic: 97% sensitivity at 77% specificity
2014–2019 India[49]	1. Diagnostic biomarker	51	Children with TB disease alone (n=19)	Children that were household contacts to persons with TB (n=32)	TB confirmed by Xpert MTB, Mtb culture, or biopsy	Diagnostic: 82% sensitivity at 56% specificity
2017–2018 China[73]	1. Diagnostic biomarker	22&	Adults (n=16)	Lung cancer (n=6)	Confirmed pulmonary MDR-TB	Diagnostic: 94% sensitivity at 83% specificity
2019–2022 Ethiopia[59]	1. Diagnostic biomarker 2. Treatment response	186	Adults (n=82)	Household contacts with a positive TB symptom screen (n=104)	Confirmed pulmonary drug susceptible TB	Diagnostic: 95% sensitivity at 70% specificity Response: Decreased to levels similar to controls after treatment

MDR, multidrug-resistance; PLHIV, persons living with human immunodeficiency virus; TB, tuberculosis

^*Studies included where those mentioned in the review, and which utilized the kyurenine/tryptophan ratio.

^{^}Repeat measures from the same controls at different time points were used to calculate test characteristics

^&Test characteristics were only calculated for participants with MDR-TB versus those with lung cancer

^#Up-to-date targeted product profiles (TPPs) for non-sputum-based, point of care (POC) tests envisage ≥65.0% sensitivity and >98.0% specificity for a diagnostic test.[74]