Marked elevations in pro-inflammatory polyunsaturated fatty acid metabolites in females with irritable bowel syndrome.

Irritable bowel syndrome (IBS) is the most common functional gastrointestinal disorder referred to gastroenterologists. Although the pathophysiology remains unclear, accumulating evidence points to the presence of low-level immune activation both in the gut and systemically. Circulating polyunsaturated fatty acids (PUFA) have recently attracted attention as being altered in a variety of disease states. Arachidonic acid (AA), in particular, has been implicated in the development of a pro-inflammatory profile in a number of immune-related disorders. AA is the precursor of a number of important immunomodulatory eicosanoids, including prostaglandin E(2) (PGE(2)) and leukotriene B(4) (LTB(4)). We investigated the hypothesis that elevated plasma AA concentrations in plasma contribute to the proposed pro-inflammatory profile in IBS. Plasma AA and related PUFA were quantified by gas chromatography analysis in IBS patients and controls. Both PGE(2) and LTB(4) were measured in serum using commercially available ELISA assays. AA concentrations were elevated in our patient cohort compared with healthy controls. Moreover, we demonstrated that this disturbance in plasma AA concentrations leads to downstream elevations in eicosanoids. Together, our data identifies a novel proinflammatory mechanism in irritable bowel syndrome and also suggests that elevated arachidonic acid levels in plasma may serve as putative biological markers in this condition.

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 satisfi ed Rome II criteria for IBS and in whom organic gastrointestinal diseases and clinically signifi cant systemic diseases had been excluded were considered for inclusion in the study. Pregnant women, individuals with known lactose intolerance or immunodefi ciency, or individuals who had any recent transient illness (i.e., within 2 weeks of participation in the study), such as viral illnesses or chest infections, were excluded.

Trial protocol
A total of 67 subjects, 41 patients with IBS and 26 healthy, sexmatched controls of comparable age and BMI, gave fully informed consent to take part in this study, which had local 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 signifi cant abnormalities leading to exclusion. The age (mean ± SD) of the patients was 45 ± 11.74 years, and the age of the comparison group was 39.04 ± 12.78. All patients and healthy comparison subjects were drug free, including anti-infl ammatory medications. The study was powered to detect differences in fatty acid concentrations at the P < 0.05 level between controls and IBS patients but not within patient subgroups.

Assessments
On arrival at the clinical investigation laboratory at 08.30 h, each subject completed the self report patient health questionnaire (PHQ) to assess the presence of major depression. This is a reliable and valid instrument that was developed as a diagnostic tool to be used in primary care ( 28 ). It tests for the presence of major depression using diagnostic criteria from the Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition (DSM-IV). The responses on the depression subscale of the questionnaire can also be used as a dimensional tool to rate the severity of depression ( 29 ). In addition to the PHQ, clinical severity of IBS was evaluated using self report ordinal scales in accordance with a previously published method ( 30,31 ). This involved subjects rating the severity of their IBS symptoms on a four-point ordinal scale (0--3) with regard to each abdominal complaint, interference with daily activities, and avoidance behavior as a result of complaints. A summarizing severity score for each patient was determined by taking the sum of the individual scores.

Sample collection
Whole blood was collected at 09.00 h in tubes that contained ethylenediametetraacetic acid (EDTA). Samples were centrifuged immediately and the plasma frozen at Ϫ 80°C until ready for analysis. Serum, where applicable, was generated similarly except that the collection tubes did not contain any anticoagulant.

Reagents
HPLC-grade methanol and chloroform were obtained from Alkem/Reagecon (Cork, Ireland). All other reagents were obtained from Sigma (Germany) unless otherwise stated.

Fatty acid analysis
Lipids from 1 ml of blood plasma were extracted with 25 ml of chloroform:methanol 2:1 (v/v) containing 5 ppm butylated hydroxytoluene as an antioxidant ( 32 ), and the solvent was removed via gentle evaporation at 45°C under nitrogen gas. Phospholipids were then separated by solid phase extraction using 500 mg NH2 phase columns (Phenomenex, UK) as described However, it has not been extensively studied in this regard in the clinical setting even though it may provide some new insights into disease pathophysiology.
The fatty acid composition of the body is largely determined by dietary intake ( 16 ). Western society has a high ratio of -6 PUFA compared with -3 PUFA largely due to a high consumption of -6-rich vegetable oils in comparison to a low consumption of -3-rich foods such as oily fi sh ( 17 ). The dietary dominance of the -6 fatty acids favors the elaboration of pro-infl ammatory mediators produced along -6 metabolic pathways over less infl ammatory -3 pathway metabolites. This imbalance, with a shift in a pro-infl ammatory direction, has been implicated in a number of diseases, including cardiovascular disease ( 18 ), depression ( 19,20 ) and a variety of infl ammatory states ( 21 ). Moreover, the biological signifi cance of this ratio in these disease states is confi rmed by studies showing that dietary manipulations aimed at reducing -6 dominance can result in favorable disease outcomes with, for example, a reduction of the -6: -3 ratio to less than 4:1 being credited with reduced mortality in cardiovascular disease ( 17 ). This focus on the pro-infl ammatory potential of the -6 fatty acids has, in particular, been directed toward the role of arachidonic acid (AA) and its metabolites as infl ammatory mediators ( 22 ).
AA is a 20-carbon PUFA that is derived from -6 fatty acids and subsequently enters a biochemical cascade to give rise to potent immunomodulatory eicosanoids including prostaglandin E 2 (PGE 2 ) and leukotriene B 4 (LTB 4 ) ( 23 ). Given the dietary factors outlined above, AA is the major substrate for synthesis of these eicosanoids, produced by the action of cylcooxygenase and lipoxygenase enzymes, respectively, in Western societies ( 22 ). These prostaglandins and leukotrienes can be biologically active even at very low concentrations, and thus, minor alterations in AA status can induce profound downstream consequences ( 23 ). Interestingly, PGE 2 has been shown to induce IL-6 synthesis in macrophages ( 24 ). Elevated IL-6 levels have, to date, represented one of the most robust fi ndings supportive of immune activation in IBS ( 11,25,26 ). Despite the importance of the AA cascade, the limited data available on PUFA in IBS have not focused on either the parent molecule or the immunomodulatory metabolites produced along the pathway ( 27 ).
In this study we hypothesized that the infl ammatory signature in IBS is derived from an increased dominance of -6 PUFA over their -3 counterparts, leading to an increased input to the AA cascade. To test this theory, we measured the plasma PUFA profi le in IBS patients and healthy controls and also examined possible downstream alterations in eicosanoid production. The analysis endpoints for this study were total plasma -6 content, total plasma -3 content, the -6: -3 ratio, plasma AA concentrations, serum PGE 2 concentrations, and serum LTB 4 concentrations.

Subjects
Female patients were recruited from a university database of IBS patients. The database consisted of people who had either -3 levels. Plasma total -3 content was signifi cantly elevated in the IBS cohort compared with controls (4.94 ± 0.33 versus 3.88 ± 0.30 g/100 g FAME; t = 2.21, df = 65, P = 0.03). An ANOVA analysis did not reveal any differences between disease subtypes or status ( Table 1 ).

Correlation analysis
There was no correlation between IBS symptom severity and plasma AA levels (Pearson product moment correlation, r = 0.16, df = 39, P = 0.33). Neither was there a correlation between depression scores and plasma AA levels (Pearson product moment correlation, r = 0.17, df = 39, P = 0.30). Plasma AA levels were nonsignifi cantly elevated regardless of whether they are classifi ed according to IBS severity (8.49 ± 0.61 g/100 g FAME in the mild group, previously ( 33 ). Phospholipids were transesterifi ed as previously described ( 34 ), extracted with 4 mls of hexane and an aliquot taken for gas chromatography (GC) analysis.
Fatty acids were quantifi ed as fatty acid methyl esters (FAME) by GC analysis using a Varian 3400 gas liquid chromatograph (Varian 3400 capillary GC, Varian, Walnut Creek, CA) fi tted with a fl ame ionization detector. The results were expressed as a percentage of FAME (%, g/100 g FAME). Separation of the FAME was performed on a Chrompack CP Sil 88 column (Chrompack, Middlelburg, The Netherlands) 100 m × 0.25mm ID × 20 m fi lm thickness). Helium was used as a carrier gas at a pressure of 33.7 psi. The injector temperature was 225°C isothermal with a hold time of 5 min and the detector temperature was 250°C. The column temperature was programmed from an initial temperature of 80°C to a fi nal temperature of 200°C, with an initial delay of 8 min (hold time), at a rate of 8.5°C/min during each analysis. The column was held at the fi nal temperature of 200°C for 7 min (final hold time). Collected data were recorded and analyzed on a Minichrom PC system (VG Data Systems, Manchester, UK). Fatty acids were identifi ed based on the retention time of reference standards (Sigma).

Eicosanoid analysis
PGE 2 and LTB 4 were measured in serum from a reduced subject group of the trial subjects outlined above. From 25 of the patient group (47 ± 10.63 years) and 19 of the control group (36.21 ± 11.87 years), serum samples in addition to the plasma samples were prepared. Separate Assay Designs EIA assay kits (Cambridge Biosciences, UK) were used to measure the analytes, and the assays were performed as per the manufacturer's instructions.

Data analysis
Data were expressed as mean values ± SEM. Data were analyzed by Student-test, one way ANOVA, ANCOVA, and by Dunnets multiple comparison posthoc tests as appropriate. Correlations were assessed according to the Pearson product moment correlation.

Subject characteristics
DSM-IV major depression was comorbid in 41% (17 out of 41) of IBS patients. None of the control group (n = 26) met criteria for current depression. Twenty-two per cent (9 out of 41) of patients rated their IBS symptoms as mild (a sum score of 3 or less on the severity scale); 41% (17 out of 41) reported symptoms of moderate severity; and 37% (15 out of 41) reported symptoms which were severe in nature (i.e., a score of 6 or greater on the severity scale). Twelve of the IBS cohort were classifi ed as having diarrheapredominant IBS (D-IBS), 9 as constipation-predominant (C-IBS), and 20 had an alternating bowel habit (A-IBS). In addition, 17 of the IBS group had currently active symptoms (CA), 17 were categorized as recently active (RA), and 7 were in a quiescent disease phase (Q).
The principal fi nding in this study is that plasma AA concentrations were signifi cantly elevated in the IBS cohort compared with controls. Although the study was not powered to detect IBS subgroup differences, an analysis of the data according to disease status revealed that the data was robust enough to indicate that it was the currently active IBS cohort that made the largest contribution to the elevated plasma AA concentrations. Although a further analysis according to disease subtype (C-IBS, A-IBS, or D-IBS) did not yield any statistically signifi cant results, there appeared to be a uniform increase across all disease subtypes. Interestingly, a previous study that examined plasma fatty acid profi les in a mixed gender IBS cohort did not report any alterations ( 27 ). However, although that study did report AA concentrations, it was not one of its statistical endpoint measures nor did it examine the infl ammatory mediators produced along the AA cascade. Moreover, it 8.49 ± 0.43 in the moderate group, and 8.57 ± 0.63 in the severe group; F = 1.60, df = [3,63], P = 0.20) ( Fig. 4A ) or ± depressive comorbidity (8.78 ± 0.42 in the depressed subgroup and 8.33 ± 0.34 in the nondepressed subgroup; F = 2.78, df = [2, 64], P = 0.07) ( Fig. 4B ). Additional statistical analysis revealed no correlation between depression scores and either serum PGE 2 or LTB 4 concentrations. Nor was there a correlation between IBS symptom severity and serum PGE 2 or LTB 4 concentrations (unpublished observations).

DISCUSSION
Here we show, what is to our knowledge, the fi rst demonstration of altered circulating PUFA metabolites in IBS.   contribution to the altered AA concentrations. There was no difference between these patients and the cohort that did not meet the criteria for depression.
We have also demonstrated that the consequences of the elevated AA levels include signifi cant elevations in both PGE 2 and LTB 4 in our IBS cohort. Of further interest is that the elevation in serum PGE 2 concentrations is largely due to the D-IBS subtype, whereas the trend toward elevated LTB 4 levels is equally apparent across all disease subtypes and status. Our laboratory and other research groups have recently reported elevated IL-6 levels in IBS ( 11,26 ). The source of such alterations has heretofore not been identifi ed, but it is possible that the fi ndings reported here might shed some light on the matter. It has previously been reported that PGE 2 can elevate IL-6 levels ( 24 ). Furthermore, it has been shown that PGE 2 itself can activate COX-2, the inducible enzyme responsible for the formation of its immediate, unstable precursor prostaglandin H 2 (PGH 2 ) ( 37 ). It is certainly plausible that the altered biological cascade described here could foster the disturbed basal IL-6 profi le reported in IBS due to a self-sustaining, but low level, chronic increase in PGE 2 production. Interestingly, although PGE 2 does possess anti-infl ammatory properties, they do not appear to extend to IL-6 ( 38 ).
Of course, these alterations in PUFA metabolites may have additional functional consequences. AA itself has a role in intestinal barrier function ( 39 ), and the increases described here are likely to have an impact in that regard. A role for elevated PGE 2 levels in altered GIT muscle activity is also possible ( 23 ). Of further note is that circulating PGE 2 can readily cross the blood brain barrier and thus impact the CNS component of the brain-gut axis, as well as the ENS ( 40 ). An increased number of GIT immune cells has been proposed as an indicator of immune activation in IBS ( 41,42 ). Interestingly, LTB 4 is regarded as a neutrophil chemoattractant and a promoter of both leukocyte adhesion and infi ltration ( 22,43 ). Moreover, the elevated LTB 4 we report here may have relevance to the morphological and functional changes that have been detected in mast cells in IBS ( 44 ), as this leukotriene is both an activation product of mast cells and a chemoattractant for their progenitors ( 45 ). It is also known that alterations in basal prostaglandin levels can impact on the brain-gut axis through the effects they exert on HPA-axis secretory activity ( 40 ). In short, these eicosanoids and their PUFA precursor are critical agents in the normal functioning of the GIT, and the perturbations in the system we have described here could directly or indirectly impact motility, secretion, nociception, cytoprotection, and the immunological milieu of the gut ( 46,47 ). Moreover, we cannot exclude that there are other AA metabolites produced along this complex metabolic cascade that contribute to the physiological and behavioral symptoms associated with IBS.
Contrary to our expectations, the increases in AA levels occurred despite an apparent shift toward the -3 PUFA arm of the metabolic cascade. We found that total -3 levels were signifi cantly elevated in our IBS cohort compared with controls without any alteration in total -6 levels. The factors behind this shift toward the -3 fatty acids are diffi cult to ex-did not take account of possible gender differences in PUFA profi les that have been previously reported ( 35 ).
Correlation analyses revealed that there was no relationship between plasma AA concentrations and IBS symptom severity, suggesting that elevated AA is not a direct cause of IBS. Additionally, if the data is grouped according to those with mild, moderate, or severe IBS, the increase is evident in all subgroups, albeit at a nonsignifi cant level. Given the complexity of AA metabolism ( 23 ), it is not surprising that our correlation analysis did not yield a simple linear relationship between levels of this fatty acid and symptom severity in our IBS patients. It has previously been shown that altered fatty acid profi les can have multiple downstream effects, confi rming the challenges that confront attempts at correlating elevated AA levels directly with IBS symptom scores ( 36 ). Because of the recent association between AA and depression ( 19 ), we also examined whether those patients with a depressive comorbidity made a greater  plain from the current dataset and are not in line with reports from other disorders that have been associated with elevated immune parameters such as depression ( 19 ) and Crohn's disease ( 48 ). Of note is that in the latter study, the elevated -6: -3 ratio was related to specifi c cytokine genotypes. We did not record subject dietary habits in this study and consequently cannot rule out an altered dietary consumption of fatty acids in the IBS cohort. Despite this limitation, data from previous studies suggests that such an alteration would have an anti-infl ammatory outcome ( 49 ). It is noteworthy that an apparent trend toward a reduction in the -6: -3 ratio did not reach statistical signifi cance in this study and that the magnitude of the -3 shift described may not be large enough to be of immunological importance. It is also interesting that the IBS is not thought to be more common in countries with traditionally high fi sh consumption rates, like Japan, than in Western societies ( 5 ). However more detailed epidemiological studies are required to probe the potential associations that have been demonstrated for other disorders with an infl ammatory component such as depression and cardiovascular diseases ( 18,19 ). The results presented here suggest that even if dietary factors come into play, they are insuffi cient to counteract the sequalae of events leading to the increased input to the AA cascade. While the precise mechanism behind these alterations is unclear at the moment, the activity of phospholipase A 2 (PLA 2 ), the enzyme responsible for the release of arachidonic acid from the cell membranes, may be worth investigating in future studies ( 50 ).

CONCLUSION
We have comprehensively characterized the PUFA profi le in a female IBS population. Our results indicate that AA levels are increased in the clinical setting. We have demonstrated a clear and robust increase in pro-infl ammatory markers downstream of AA in our IBS cohort that may have relevance for previously described alterations in pro-infl ammatory markers such as IL-6. Although further studies are required to elucidate the mechanism behind this phenomenon, the alterations described here may represent a novel biomarker candidate panel in IBS that is especially needed given the current reliance on a symptom-based diagnostic scheme.