Dr Peter Fitzgerald, Department of Psychiatry, Cork University Hospital, Wilton, Cork, Ireland. Tel: 353 21 4922593; fax: 353 21 4922584; e-mail: firstname.lastname@example.org
Abstract Irritable bowel syndrome (IBS) has been linked with abnormal serotonin functioning and immune activation. Tryptophan forms the substrate for serotonin biosynthesis, but it can alternatively be catabolized to kynurenine (Kyn) by the enzyme indoleamine 2,3-dioxygenase (IDO), the main inducer of which is interferon-gamma. The primary aim of this study was to test the hypothesis that IBS is associated with increased tryptophan (Trp) catabolism along the Kyn pathway due to increased IFN-γ levels. Plasma Kyn, Trp and IFN-γ levels were measured in 41 female IBS subjects and 33 controls. Indoleamine 2,3-dioxygenase activity was assessed using the Kyn to Trp ratio. Psychiatric co-morbidity was assessed using the Patient Health Questionnaire, and severity of IBS assessed using self-report ordinal scales. Irritable bowel syndrome subjects had increased Kyn concentrations compared with controls (P = 0.039) and there was a trend for Kyn:Trp to be increased in the IBS group (P = 0.09). There was a positive correlation between IBS severity and Kyn:Trp (r = 0.57, P < 0.001). Those with severe IBS symptoms had increased Kyn:Trp (P < 0.005) compared to those with less severe symptoms and controls, and were over twice as likely to have depression or anxiety compared to those with less severe IBS (RR = 2.2, 95% CI 1.2–3.9). No difference in IFN-γ levels was observed between groups; however, IFN-γ was positively correlated with Kyn:Trp in IBS (r = 0.58, P = 0.005) but not controls (r = 0.12, P = 0.5). Females with IBS have abnormal Trp catabolism. The Kyn:Trp is related to symptom severity, and those with severe IBS symptoms have increased shunting of Trp along the Kyn pathway which contributes to the abnormal serotonergic functioning in this syndrome.
Irritable bowel syndrome (IBS) is a common sensory and motility functional disorder of the gastrointestinal tract which affects approximately 15% of the Western population.1–3 The precise pathophysiology of IBS remains unclear, although it is generally regarded as a disorder of the brain-gut axis with associated disturbances in gut motor and sensory function. The concept of the brain-gut axis is a theoretical model describing bidirectional neural pathways which link cognitive and emotional centres in the brain to neuroendocrine centres, the enteric nervous system and the immune system.4
Serotonin (5-hydroxytrptamine (5-HT)) is a biogenic amine that functions as a neurotransmitter and is located predominantly in the gastrointestinal tract (80–90% of body stores). Peripherally, it is involved in the regulation of gastrointestinal secretion, motility, and sensation,5 while centrally, it plays a key role in the regulation of mood and cognition.6 Serotonin is, therefore, an important modulator of the brain-gut axis, and a dysfunctional serotonergic system may provide a plausible link between IBS and its high co-morbidity with psychiatric disorders such as major depression and anxiety in which a central serotonin deficiency plays a causative role.7–11
Serotonin is derived from the essential amino acid tryptophan (Trp). In addition to being the substrate for serotonin synthesis, Trp can alternatively be catabolized to kynurenine (Kyn).12 The majority of Trp is catabolized along the Kyn pathway, which is principally regulated by the enzyme tryptophan 2,3,-dioxygenase (TDO) and indoleamine 2,3-dioxygenase (IDO). While TDO is localized to the liver and is up-regulated by corticosteroids, IDO is expressed by a variety of cells and is inducible preferentially by the pro-inflammatory cytokine interferon-γ (IFN-γ).13
As the brain’s storage of Trp is limited, decreased Trp availability reduces the biosynthesis of central serotonin which can thereby increase susceptibility for psychiatric disorders such as depression.14 Increased Trp degradation by IDO may thus induce neuropsychiatric symptoms when the availability of Trp is insufficient for normal serotonin biosynthesis.15 Peripherally, abnormalities in serotonergic functioning within the gut have been described in those with IBS.16,17 In addition, a recent study has demonstrated abnormal concentrations of Kyn metabolites in a group of 37 IBS subjects, suggesting altered Trp degradation in this syndrome.18 As several studies have reported immune activation in IBS characterized by increased pro-inflammatory cytokine production,19–21 this provides a plausible biological mechanism for altered Trp degradation.
The Kyn to Trp ratio (Kyn:Trp) i.e. the ratio of the concentration of the first product of IDO vs the concentration of its substrate, is an appropriate indicator of the degree of Trp degradation along this pathway, i.e. IDO activity. The Kyn:Trp provides a better and more normalized measurement than absolute Trp or Kyn concentration alone.22
The relationship between Trp degradation, IBS and psychiatric disorders has not yet been studied. Shunting of Trp metabolism down the Kyn pathway at the expense of serotonin biosynthesis may help to explain the abnormal serotonin signalling found in the syndrome, while also provide a plausible framework for understanding the strong association between the functional gastrointestinal disorder and its high co-morbidity with depression and anxiety.
In this study, we investigated plasma concentrations of Trp and Kyn and calculated the Kyn:Trp in females with IBS, while also assessing concentrations of IFN-γ and the presence of depression and anxiety disorders. The study is based on the hypothesis that there is increased Trp degradation along the Kyn pathway in IBS, which occurs due to increased immune activation (i.e. increased IFN-γ).
Materials and Methods
Female patients were recruited from a university database of IBS patients. The database comprised of people who had either attended gastroenterology clinics at Cork University Hospital or had responded to direct advertisement on the university campus or local newspaper regarding participation in IBS research. Individuals aged between 18 and 65 years who satisfied Rome II criteria for IBS2 and in whom organic gastrointestinal diseases and clinically significant systemic diseases had been excluded, were considered for inclusion in the study. Pregnant women, individuals with known lactose intolerance or immunodeficiency, or who had any recent transient illnesses (i.e. within 2 weeks of participation in the study), such as viral illnesses or chest infections etc. were excluded also.
A total of 74 subjects, 41 patients with IBS and 33 healthy, sex-matched controls of comparable age and BMI, gave fully informed consent to take part in this study, which had ethics committee approval.
Each potentially eligible patient was evaluated by a review of clinical history, performance of a physical examination, and measurement of full blood count and serum biochemistry with any clinically significant abnormalities leading to exclusion.
The age (mean ± SD) of the patients was 44.1 ± 11.3 years and of the comparison group was 41.3 ± 12.8 years. All patients and healthy comparison subjects were drug free.
Each subject, upon arrival at the clinical investigation laboratory at 8.30 am, completed the self-report Patient Health Questionnaire (PHQ) to assess for the presence of major depression and anxiety disorders. This is a reliable and valid instrument which was developed as a diagnostic tool to be used in primary care.23 It tests for the presence of major depression and anxiety disorders using diagnostic criteria from the Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition (DSM-IV). The responses on the depression sub-scale of the questionnaire can also be used as a dimensional tool to rate the severity of depression, while the anxiety sub-scale can only be used categorically to detect the presence or absence of an anxiety disorder.24
In addition to the PHQ, clinical severity of IBS was evaluated using self-report ordinal scales in accordance with a previously published method.25–27 This involved subjects rating the severity of their IBS symptoms on a four point ordinal scale (0–3) with regards to each of the following: reported frequency of abdominal complaints, interference with daily activities, and avoidance behaviour as a result of the complaints. A summarizing severity score for each patient was determined by taking the sum of the individual scores.
Six millilitres of whole blood was collected at 9 am in ethylenediaminetetraacetic acid (EDTA) tubes. Samples were centrifuged immediately and plasma frozen at −36 °C until ready for analysis.
Tryptophan and Kyn were determined by high performance liquid chromatography (HPLC), the system for which consisted of a Waters 510 pump (Waters Ireland, Dublin, Ireland), 717plus cooled autosampler, a 996 PDA detector, a Hewlett Packard 1046A Fluorescent Detector (Waters Ireland, Dublin, Ireland), a waters bus SAT/IN module and a croco-cil column oven. System components were used in conjunction with Waters Empower software (Waters Ireland, Dublin, Ireland). All samples were injected onto a reversed phase Luna 3u C18(2) 150 × 2 mm column (Phenomenex, Macclesfield, UK), which was protected by Krudkatcher disposable precolumn filters and security guard cartridges (Phenomenex). HPLC grade Acetonitrile, acetic acid and perchloric acid were obtained from Alkem/Reagecon (Cork, Ireland).
The analysis method was based on that by Herve et al.28 The mobile phase consisted of 50 mmol L−1 acetic acid, 100 mmol L−1 Zinc Acetate with 3% (v/v) acetonitrile and was filtered through a 0.45 μm Millipore filter (AGB, Dublin, Ireland) and vacuum degassed prior to use. Separations were achieved by isocratic elution at 0.3 mL min−1. The fluorescent detector was set to an excitation wavelength of 254 nm and an emission wavelength of 404 nm. The PDA detector start wavelength was 210 nm and the end wavelength was 400 nm with chromatogram extraction at 330 nm. Working standard dilutions were prepared from millimolar stock solutions of each standard and stored at −80 °C until required for analysis.
Plasma samples were deproteinized by the addition of 20 μL of 4 mol L−1 perchloric acid to 200 μL of plasma spiked with 3-nitro-l-Tyrosine as internal standard. Twenty microlitres of either sample or standard was injected onto the HPLC system and chromatograms generated were processed using Waters Empower software.
Analytes were identified based on their characteristic retention time and the concentrations were determined using Analyte:Internal standard peak height ratios that were measured and compared with standard injections. Results were expressed at ng analyte per mL of plasma.
Measurement of plasma IFN-γ was performed using an electro-chemiluminescence multiplex system Sector 2400 imager from Meso Scale Discovery (Gaithersburg, MD, USA) where antibodies labelled with Sulfo-tag reagents emitted light upon electrochemical stimulation. This is an ultra-sensitive method which has a detection limit for IFN-γ of 0.8 pg mL−1.
Group mean differences of Trp and Kyn concentrations and of the Kyn:Trp were analyzed by one-way analysis of variance with appropriate comparisons. Due to the normal distribution of our data, correlation analyses were performed using a Pearson product-moment correlation co-efficient, and analysis of an association between IBS severity and psychiatric disorder was performed using the Fisher’s Exact Test. All statistical calculations were performed using Graph Pad Prism version 4.0 for windows (Graph Pad Software, San Diego, CA, USA).
At least one psychiatric disorder was found to be present in approximately half of the IBS subjects (21/41 cases). Major depression was the most common disorder, with 41% (17/41) of patients meeting criteria for this disorder. Anxiety disorders were present in 26% (11/41) of patients, and ∼17% (7/41) met criteria for both depression and anxiety.
None of the control group (n = 33) met criteria for current depression or an anxiety disorder.
Twenty-two per cent (9/41) of patients rated their IBS symptoms as mild (a sum score of 3 or less on the severity scale); ∼41% (17/41) reported symptoms of a moderate severity; and 37% (15/41) reported symptoms which were severe in nature (i.e. a score of 6 or greater on the severity scale).
There was a significant association between the presence of a psychiatric disorder (i.e. those who met criteria for depression and/or anxiety disorders) and the experience of severe IBS symptoms (i.e. those with an IBS severity score of at least 6; Fisher’s Exact Test P = 0.02). Eleven of the 15 patients with severe IBS had a psychiatric illness: only one had an anxiety disorder alone, while the remaining 10 had major depression (with four of those having co-morbid anxiety). The relative risk of psychiatric disorder in those with severe IBS symptoms was greater than twice that of those with less severe gastrointestinal symptoms (RR = 2.2, 95% CI 1.2–3.9).
Although there was no difference in plasma Trp levels between IBS subjects and controls (P = 0.32), Kyn concentrations were significantly increased in the IBS group (600 ± 20.3 ng mL−1vs 534 ± 21.7 ng mL−1, respectively; P = 0.039) and there was a trend for the Kyn:Trp ratio to be increased (0.060 vs 0.053; P = 0.09).
There was a significant correlation observed between the Kyn:Trp of the IBS patients and the IBS severity score (Pearson r =0.57, P <0.001; Fig. 1). There was no correlation found between Kyn:Trp and severity of depressive symptoms (r =0.24, P =0.12), and while those with a co-morbid anxiety disorder had an elevated mean Kyn:Trp ratio compared to those IBS subjects without anxiety (0.060 vs 0.053 respectively) this did not reach significance (P = 0.13).
Further exploratory analysis was performed in IBS subjects according to classification of IBS severity (i.e. mild, moderate and severe groups). The Kyn:Trp (mean ± SEM) in the severe IBS group was 0.068 ± 0.004 (Fig. 2) which was significantly greater than those with moderate (0.054 ± 0.002) or mild IBS (0.051 ± 0.002) or healthy controls (0.053 ± 0.002; F = 5.02, P = 0.0036). To compensate for multiple comparisons, a Tukeys multiple comparison test was performed which revealed the severe IBS group to have a significantly higher Kyn:Trp compared with all other groups (P < 0.05).
While there was no difference observed in mean Trp concentrations between the four groups (10 131 ± 519 ng mL−1 in mild IBS group, 10 544 ± 423 ng mL−1 in moderate IBS patients, 10 174 ± 546 ng mL−1 in severe IBS patients and 9930 ± 312 ng mL−1 in controls; F =0.17, not significant), there was a significant difference found in the mean Kyn concentrations: 514 ± 34 ng mL−1 in the mild IBS group, 564 ± 26 ng mL−1 in moderate IBS group, 669 ± 35 ng mL−1 in severe IBS group and 534 ± 21.7 ng mL−1 in controls [F =4.7, P =0.0047; tukey’s test demonstrating that the severe IBS group had significantly greater (P < 0.01) mean Kyn concentration than the controls]. Plasma kyn concentrations were positively correlated with IBS severity score (r = 0.38, P = 0.013), while there was no significant relationship between Trp concentration and IBS symptom severity (P = 0.8).
There was no difference found between the mean IFN-γ concentration in the IBS group vs that of the controls (1.07 ± 0.09 pg mL−1vs 1.05 ± 0.18 pg mL−1, respectively; P = 0.9). Irritable bowel syndrome patients which had a co-morbid affective disorder (n = 21) had elevated IFN-γ levels compared to those without such co-morbidity (n = 20; 1.25 ± 0.15 pg mL−1vs 0.9 ± 0.09 pg mL−1; P = 0.05), but there was no significant correlation between IFN-γ concentration and depressive symptom severity in either the total IBS group (P > 0.8) or within the depressed subgroup (P = 0.4).
Analysis according to IBS severity did not reveal any differences in IFN-γ concentrations between those with mild, moderate or severe symptoms (F = 0.13, P = 0.9).
However, a correlation analysis of IFN-γ concentration and Kyn:Trp in IBS subjects revealed a significantly positive association (r = 0.34, P = 0.03) which was strengthened when those with psychiatric illness were excluded (r = 0.58, P = 0.005; Fig. 3). No such correlation was found in the control subjects (r = 0.12, P = 0.5).
The main finding of our study was of increased Kyn levels in a group of female IBS patients and a positive correlation between Kyn:Trp ratio and IBS symptom severity.
Those which were classified as having severe IBS had a significantly increased Kyn:Trp, indicative of increased IDO activity, the consequences of which are an increased rate of Kyn production and consequently reduced 5-HT biosynthesis. Furthermore, though IFN-γ levels were found to be similar in IBS subjects compared to healthy controls, we observed a significant association between IFN-γ and Kyn:Trp in the IBS group which was not present in controls.
The implication of an increased Kyn to Trp ratio is that Trp is being preferentially catabolized to Kyn due to increased activity of IDO. Increased Kyn production occurs at the expense of serotonin biosynthesis, thus increased IDO activity may contribute to the serotonergic dysfunction in IBS while also providing a biologically plausible explanation for the increased incidence of major depression and anxiety disorders in this syndrome (via predisposing to a central 5-HT deficit). In agreement with this, it is of note that we observed a 50% incidence of depression or anxiety disorder in the IBS sample, with the majority of cases occurring in those with severe IBS symptoms.
Tryptophan depletion paradigms have been used in previous studies to investigate the effect of altered Trp availability on depression and anxiety disorders29–32 in addition to IBS.33,34 Rapid Trp depletion is accomplished through the ingestion of an amino acid beverage void of Trp after adherence to a low Trp diet during the previous 24 h.30 In depression, Trp depletion leads to a deterioration in mood,29 while in patients with an anxiety disorder, it leads to worsening of anxiety symptoms.31 With regards to IBS, the two small Trp depletion studies which have been carried out to date demonstrated worsening of gastrointestinal symptoms due to acute alteration in Trp availability.33,34 While such studies support the hypothesis that abnormal serotonin modulation may be involved in IBS, it should be emphasized that such Trp depletion experiments are brief (a matter of hours only) and therefore do not address the possible consequences of a more prolonged yet less severe reduction of Trp availability as may be seen in those with ongoing increased activation of the Kyn pathway.
While our finding of increased Kyn production in patients with IBS may prompt one to speculate that the clinical implications of this derive from a deficit in serotonin biosynthesis, an alternative (or complimentary) possibility exists that the increased production of Kyn and its downstream derivatives (such as kynurenic acid or quinoloinic acid, N-methyl-D-aspartic acid (NMDA) antagonists or agonists, respectively) may also play an active role in gastrointestinal symptom generation. NMDA receptors have been detected in the myenteric plexus and are believed to play a role in gut motility and visceral nociception,35,36 with activation by glutamate leading to increased contractile activity.37 The balance between quinolinic and kynurenic acid may be involved in the control of enteric neuronal excitability and gut motility38 and it is of note that a recent study by Wollny et al. demonstrated decreased levels of kynurenic acid in addition to increased levels of the precursors to quinolinic acid in 37 patients with IBS.18
The question as to the cause of the increased rate of Kyn production from Trp in our IBS sample remains unanswered. In our original hypothesis, we speculated that IBS would be associated with increased IDO activity due to increased expression of the pro-inflammatory cytokine IFN-γ, which is the primary inducer of IDO,13 and thereby preferentially direct Trp catabolism along the Kyn pathway. However, we did not find any difference in mean IFN-γ levels between the IBS and control groups. We did, however, demonstrate a positive correlation between Kyn:Trp and IFN-γ in the IBS group which was not evident in the healthy controls, and which strengthens when subjects with depression are excluded (which is known to be associated with increased IFN-γ expression). This suggests an increased sensitivity in IBS subjects to the effect of IFN-γ on IDO induction. The cause of such increased sensitivity is a matter of speculation, with one possibility being a synergistic interaction between IFN-γ and other pro-inflammatory molecules on IDO induction and activity, such as TNF-alpha or prostaglandins.39–41
One should also consider the possibility of the increase in IDO activity being an epiphenomenon of major depression, which is known to be associated with an increased Kyn:Trp.42,43 However, a number of analyses on our data provide supportive evidence that the finding of increased IDO activity in IBS is related to IBS per se rather than depression: firstly, we did not observe an association between severity of depressive symptoms and Trp catabolism in the IBS group, which one would have expected if the increased IDO activity was due to depression rather than IBS, while we did find an association between IBS symptom severity score and increased Trp catabolism. Secondly, the positive correlation between Kyn:Trp and IFN-γ concentration in the IBS group was strengthened by excluding those with co-morbid depression, while no such correlation was found in the healthy controls.
There are limitations to our study which need to be acknowledged. A larger sample of patients needs to be studied to confirm our findings, and the heterogeneity of the IBS group recruited in terms of severity and setting should be reduced. Our sample came from a mixture of community, primary and secondary care sources and consequently severity of illness ranged from mild to severe. It is of note that while we found a significant difference in Kyn:Trp between those with severe IBS and controls, this reduced to trend level when those with mild to moderate IBS were included in the analysis, possibly as a result of the above heterogeneity. Ideally, replication of the study in a group of IBS patients limited to each of these severity classifications alone is warranted to address this issue.
The lack of a structured psychiatric interview to verify psychiatric diagnoses is a further limitation; however, it was decided to use a validated self-report instrument for pragmatic study design reasons. There is good agreement, however, between the PHQ diagnoses and those of independent mental health professionals (κ = 0.65; overall accuracy, 85%; sensitivity, 75%; specificity 90%).23
In summary, the data provides evidence for increased shunting of Trp along the Kyn catabolic pathway in females with IBS, and suggests this may be due to increased sensitivity of the enzyme IDO to IFN-γ. Such findings provide a biological basis for the high co-morbidity with depressive and anxiety disorders found in this condition. Further studies should seek to replicate this finding on a larger cohort of patients, and investigate potential causes of such increased sensitivity.
This work was supported in part by Science Foundation Ireland in the form of a centre grant (Alimentary Pharmabiotic Centre), the Health Research Board of Ireland, the Higher Education Authority of Ireland, and the Wellcome Trust.
The authors have no competing interests to declare.