Immune dysfunction in patients with functional gastrointestinal disorders


Jan Tack MD, PhD, Center for Gastroenterological Research, Catholic University Leuven, B-3000 Leuven, Belgium.
Tel: +32 16 34 57 61; fax: +32 16 34 59 39; e-mail:


Abstract  There is increasing evidence for involvement of the immune system in functional gastrointestinal disorder (FGID), including onset after acute gastrointestinal infections, genotypes resulting in altered cytokine expression and abnormal presence of immune cells. Our aim was to assess cellular and humoral immune responses in (i) FGIDs, compared to healthy subjects and (ii) acute vs unspecified onset FGIDs. Lymphocytic [interleukin (IL)-5, IL-10, IL-13 and interferon γ (IFN-γ)] and monocytic [IL-10, IL-12, tumour necrosis factor (TNF)-α] cytokine production was characterized at baseline and after stimulation with phytohemagglutinine and anti-CD28 or lipopolysaccharide (LPS) in controls (n = 32), irritable bowel syndrome (IBS) (n = 30), functional dyspepsia (FD) (n = 23) and non-cardiac chest pain (NCCP) (n = 15). Serum IL-6 and IL-10 concentrations were compared, and the immunophenotype was assessed using fluorescent-activated cell sorter. Findings were compared for acute vs unspecified onset FGID. Compared to controls, stimulated lymphocyte expression of IL-5 and IL-13 was enhanced in IBS, FD and NCCP (all P < 0.05). Conversely, the stimulated monocytic IL-12 and lymphocytic IL-10 expression were reduced in IBS and FD, while IFN-γ expression was also reduced in FD patients. Except for an increase in the numbers of CD3+CD45RA+CD45RO+ cells, no distinct cellular profile was detected. Patients with a presumed acute onset of their symptoms had higher serum IL-10 levels and more CD3+CD45RA+CD45RO+ cells, while TNF-α levels following stimulation with LPS were higher in FD patients reporting an acute onset. A shift towards a Th2 cytokine profile is present in FGID, while the cellular immunophenotype remains largely unchanged. Further research is indicated and could provide new therapeutic strategies for these disorders.


diarrhoea-predominant irritable bowel syndrome


enzyme-linked immunosorbent assay


fluorescent-activated cell sorter


functional dyspepsia


functional gastrointestinal disorder


irritable bowel syndrome




non-cardiac chest pain


natural killer


peripheral blood mononuclear cells




postinfectious irritable bowel syndrome


Functional gastrointestinal disorders (FGID) have a high prevalence in Western countries.1,2 Despite intensive research, the underlying pathophysiology remains unclear. It has long been recognized that an acute gastrointestinal infection may be the initiating factor in at least a subset of irritable bowel syndrome (IBS) patients.3–9 More recently, a similar aetiology has been proposed for functional dyspepsia (FD) based on the results from a population study following an outbreak of Salmonella gastro-enteritis and on distinct pathophysiological abnormalities observed in FD patients reporting an acute onset of their symptoms.8,10

Pathological studies showing the signs of persisting low grade inflammation in FGID patients also give credit to this theory. Gwee et al.11 observed an increased number of inflammatory cells on rectal biopsies of patients with postinfectious IBS (PI-IBS), compared to subjects who did not develop IBS. Other studies reported increased numbers of mast cells in IBS, and a significant correlation of the close relationship of these cells to nerve endings and IBS symptom severity.12 Our group also reported on the presence of lymphoid infiltrates in duodenal biopsies of acute onset FD patients, and a recent population-based case–control study found increased numbers of eosinophils in duodenal biopsies in FD.13,14

As inflammation seems to be involved, cytokines are also thought to play some role in the development of symptoms. Gwee et al.15 found higher interleukin (IL)-1β mRNA expression in the rectal mucosa of PI-IBS patients. Based on the measurement of unstimulated monocytic IL-10 and IL-12 production, a pro-inflammatory Th1-cytokine profile was observed in IBS.16 More recently, higher unstimulated tumour necrosis factor (TNF)-α, IL-1β, IL-6 and lipopolysaccharide (LPS)-induced IL-6 levels were shown in IBS, mostly in diarrhoea-predominant IBS (D-IBS) patients.17 Although concomitant cytokine levels were lacking, and although discrepancies between studies exit, genetic polymorphisms such as a high producer genotype of TNF-α and a low-producer genotype for IL-10 have also been associated with a higher prevalence of IBS.18,19

Therefore, the aims of this study were as follows: firstly, we investigated whether a different cellular and/or humoral immune response is present in FGIDs other than IBS, compared to healthy subjects. For this purpose, we studied a cohort of consecutive patients including IBS, FD and non-cardiac chest pain (NCCP). Secondly, we hypothesized that cellular and/or humoral immune responses would differ in patients with acute vs unspecified onset of FGIDs.

Materials and methods


Thirty-two healthy volunteers and 68 consecutive patients with FGID were recruited for the study. None of the healthy subjects had symptoms or a history of gastrointestinal disease, nor were they taking any medication. Based on the presenting symptoms and negative results from standard investigations, including upper and lower endoscopy, abdominal ultrasound, oesophageal manometry and 24-h pH-metry as indicated, patients were classified according to the Rome II criteria for FGID (IBS, FD or NCCP). The protocol was approved by the ethical committee of the Leuven University.

Healthy volunteers and patients were excluded from the study when a major psychiatric-disorder-influencing symptom perception was diagnosed (e.g. major depression, anorexia nervosa, psychosis), when an inflammatory disorder indicated by routine biochemical analysis was present or when an underlying organic disease contributed to the complaints.

Additionally, patients reporting an acute onset together with one or more symptoms suggestive for an acute gastrointestinal infection (fever, headache, muscle ache, vomiting and diarrhoea) were noted to have acute onset FGID. The remaining patients were classified as unspecified onset FGID. The Hospital Anxiety and Depression Scale (HADS) was used to quantify anxiety and depression levels.20

Allergy screening

Finally, both patients and healthy volunteers were screened for atopy based on a validated questionnaire.21 Only those items which a senior allergologist judged compatible with an atopical condition, and not merely mimicking it, were considered for a positive allergic status.

In addition to the questionnaire and to the percentage of eosinophils, specific immunoglobulin E antibodies for grass mix (gx3), birch tree (t3), cat dander (e1) and dust mite(d1) were assessed in peripheral blood using a commercially available radioallergosorbent test (RAST). The RAST was considered positive when the concentration of at least one of the tests exceeded 1 U mL−1.

Measurement of cytokine production

For cytokine production by lymphocytes, peripheral blood mononuclear cells (PBMC) were isolated from heparinized blood on Ficoll–Hypaque gradients, and cultured in RPMI 1640 with 10% bovine calf serum, and phytohemagglutinine (PHA) (0.5 μg mL−1) with or without anti-CD28 monoclonal antibody (Clone 9.3) for 72 h. Supernatants were collected, frozen, and cytokines (IL-5, IL-10, IL-13 and INF-γ) were determined at the end of the study by enzyme-linked immunosorbent assay (ELISA).

For cytokine production by monocytes, whole blood was diluted in RPMI 1640 to obtain 50 000 monocytes in 2 mL, followed by stimulation with LPS (100 ng mL−1) for 72 h. TNF-α, IL-10 and IL-12 were measured in the supernatants using ELISA.

For ELISA, a sandwich assay was used with monoclonal antibodies for coating, and biotin-marked monoclonal antibodies for detecting, followed by streptavadin-peroxidase and TMB as substrate.

Circulating serum cytokines IL-6 and IL-10 were determined by ELISA using commercial EASIA and FLEXIA kit respectively. LPS from Escherichia coli 026:B6 was purchased from Sigma (St Louis, MO, USA). TNF antibodies were purchased from Biosource; IL-5, IL-13, IL-10 and interferon γ (IFN-γ) antibodies from BD Pharmingen (San Diego, CA, USA) and IL-12-antibodies from R & D Systems Europe (Abingdon, UK). EASIA and FLEXIA kits were purchased from Biosource (Nivelles, Belgium).


Immunophenotyping was performed by flow cytometry on whole blood samples drawn on EDTA. One-hundred microlitres of whole blood was incubated during 10 min with labelled monoclonal antibodies followed by red blood cells lysis with FACSlyse solution, according to the manufacturer’s procedure. After washing, analyses were performed on a FACScan flow-cytometer (Becton Dickinson, Erembodegem, Belgium). The ‘back-gating’ strategy based on CD14/CD45-labelling was used to confirm the gating of the lymphocytes on the basis of the scatter characteristics. Cell Quest Software was applied for the analysis of the list mode data. Results were expressed as a percentage of the total population of lymphocytes or CD3(+) cells and as an absolute value.

Immunophenotyping was used for the following markers: CD3, CD5, CD4, CD8, CD4, CD4-to-CD8-ratio, CD3+CD25+, CD3+HLADR+, CD3+CD11b+, CD3+CD69+, CD3+CD28+, CD3+CD45RA+, CD3+CD45RO+, CD4+CD45RO+, CD4+CD45RA+, CD3+CD45RA+CD45R0+, CD3+CD16+CD56+, CD19, CD19+CD5+, CD3CD16+CD56+.

Antibodies, FACSlyse solution, FACScan and Cell Quest Software were purchased from Becton Dickinson (Erembodegem, Belgium).


Despite logarithmic transformation, no normal distribution could be obtained for most of the immunological parameters; therefore, non-parametric tests were applied (Wilcoxon test). Additionally, to test the possible confounding effect of age on the results from univariate analysis, we applied logistic regression with backward elimination including age and the cytokines revealed during univariate analysis as independent variables, and the subject group (control vs patients) as the dependent variable. Prevalence was tested with Chi-square. Data are expressed as median [first–third quartile]. A P-value of <0.05 was considered significant. Categorical data were analysed by Fischer’s exact test. All statistical analyses were performed in sas 9.1 (SAS Institute, Tervuren, Belgium).


Study subjects

In the patient group, 30 patients had IBS, 23 had FD and 15 suffered from NCCP with 80%, 83% and 73% women respectively (ns). In the control group (32 subjects), 77% were women (not different from patients). Patients were significantly older than control subjects [40.5 (30.5–50.5) vs 30.5 (24.5–44.5) years; P = 0.01]. In the patient group, the NCCP group was significantly older [51.0 (43.0–56.0) vs 39.0 (30.0–45.0) years for IBS and 37.0 (27.0–50.0) years for FD, anovaP = 0.02].

Detailed data on the onset of symptoms and on anxiety and depression were available for a subset of the patients (n = 40; 20 IBS, 13 FD and seven NCCP), of which 9, 6 and 2 patients reported an acute onset in respectively IBS, FD and NCCP. No difference was noted in gender distribution (14/17 vs 20/23 for respectively acute and unspecified onset, P = ns) or age (43.2 ± 3.2 vs 40.6 ± 3.2 for respectively acute and unspecified onset, P = ns).

Healthy subjects vs patients

Allergy screening  The prevalence of allergy was comparable between healthy volunteers and patients (25/32, 23/30, 14/23 and 7/15 for healthy volunteers, IBS, FD and NCCP respectively, P = ns). The percentage of eosinophils in peripheral blood was comparable between groups [3.96 (2.40–4.80), 2.71 (1.75–3.05), 2.95 (1.50–3.40) and 3.1 (1.90–3.40) for controls, IBS, FD and NCCP respectively, P = ns]. Additionally, a RAST was performed in about half of the subjects. We did not notice any difference in RAST positivity between groups (6/17, 3/12, 6/12, 1/6, P = ns).

Anxiety and depression  Anxiety and depression scores were significantly higher in the patient group compared to healthy volunteers (anxiety: 4.1 ± 1.1 vs 10.7 ± 1.2, P = 0.002 and depression: 1.6 ± 0.8 vs 8.4 ± 1.4, P = 0.003). Amongst the different patient groups, comparable scores were obtained for anxiety (10.8 ± 1.4 in IBS, 10.8 ± 2.9 in FD and 10.3 ± 5.0 in NCCP) and for depression (7.4 ± 1.6 for IBS, 9.8 ± 3.6 for FD and 9.0 ± 4.5 for NCCP).

Cytokine production  No difference in basal plasma IL-6 [0.0 (0.0–0.0) vs 0.0 (0.0–0.0) pg mL−1, P = 0.33] and IL-10 [0.0 (0.0–91.0) vs 32.0 (0.0–55.0) pg mL−1, P = 0.79] was noted between patients (n = 45) and controls (n = 25). However, following stimulation with PHA and anti-CD28 the lymphocytic production of IL-5 and IL-13 significantly increased in patients (n = 44) when compared to controls (n = 25) [259 (136–462) vs 92 (49–151) pg mL−1, P < 0.01 for IL-5 and 2274 (1056–3624) vs 1072 (407–1703) pg mL−1, P < 0.01 for IL-13], while IFN-γ production was significantly decreased [1301 (1161–1875) vs 2507 (1425–3424) pg mL−1, P < 0.01]. A significant correlation was found between both anxiety and depression scores and stimulated IL5 (respectively r = 0.55, and r = 0.52, both P = 0.01), IL13 (respectively r = 0.61, P = 0.003 and r = 0.57, P = 0.006) and TNF levels (respectively r = −0.51, P = 0.02 and r = −0.42, P = 0.05) for the whole group, including controls. No significant correlations were detected in the separate groups. Surprisingly, a weak but significant correlation was found between basal plasma IL-10 levels and depression scores (r = 0.45, P = 0.03), while anxiety was not correlated.

Stimulated IL-5 [225.5 (79–425), 311 (167–703) and 407 (242–667) pg mL−1 for respectively IBS, FD and NCCP, respectively P = 0.02, P < 0.01 and P = 0.02] and IL-13 production [1816 (994–3113), 3171 (1771–3679) and 3063 (1991–5964) pg mL−1, for respectively IBS, FD and NCCP, respectively P = 0.04, P < 0.01 and P = 0.02] were significantly elevated in all separate patient groups vs controls (Fig. 1).

Figure 1.

 (A) Stimulated lymphocytic (with PHA and anti-CD28) cytokine production in controls, irritable bowel (IBS), functional dyspepsia (FD) and non-cardiac chest pain (NCCP) expressed in pg mL−1. *P < 0.05. (B) Stimulated monocytic (with LPS) cytokine production in controls, IBS, FD and NCCP expressed in pg mL−1. *P < 0.05, P = 0.06.

In contrast, the production of IL-10 after incubation with PHA and anti-CD28 was significantly reduced in patients compared to controls [1084 (786–1869) vs 1439 (1186–2635), P = 0.03]. This difference was attributable to a significant decrease in FD [944 (713–2824), P = 0.02] and a borderline significant decrease in IBS [1072 (733–1704), P = 0.06], while no differences were detected within the patient groups (Fig. 2).

Figure 2.

 Lymphocytic IL-10 production after incubation with PHA and anti-CD28 expressed in pg mL−1. A significant decrease was detected in FD and a nearly significant decrease in IBS. *P = 0.02, P = 0.06.

Lipopolysaccharide-stimulated monocytic production of IL-12 was more frequently suppressed in patients than in controls (40/44 vs 10/25, P < 0.01). Twenty-two of 23 FD, 13 of 14 IBS and five of seven NCCP patients showed the suppressed IL-12 levels after LPS (respectively < 0.01, P = 0.03 and P = 0.58 vs controls).

The strongest correlation in cytokine productions after mitogen stimulation was found for IL-5 with IL-13 (P = 0.93, < 0.0001) (Table 2). Production of IL-12 was significantly correlated with TNF-α (Ρ = 0.60, P < 0.0001) and IFN-γ (Ρ = 0.40, < 0.001). Significant negative correlations were present between IFN-γ and the production of the Th2-type cytokines IL-13 (P = −0.56, P < 0.0001) and IL-5 (P = −0.52, P < 0.0001) as well as between IL-12 and these Th2-type cytokines (P = −0.35, P < 0.01 for IL-5; P = −0.36, P < 0.01 for IL-13). Therefore, and despite the decreased IL-10 production, we conclude to a shift from a Th1 to a Th2 type of cytokine production. When performing a logistic regression analysis including age, stimulated IL-5 was the main contributor to the model (P = 0.004, R2 = 0.1727).

Table 2.   Spearman Rho coefficients and corresponding P-values of the correlations between different cytokines
  1. IL-5, IL-13 and IFN-γ following stimulation with PHA and anti-CD28. TNF-α and IL-12 following stimulation with LPS.

IL-5ρ = 0.93
P < 0.0001
ρ = −0.52
P < 0.0001
ρ = −0.24
P = 0.04
ρ = −0.35
P < 0.01
IL-13 ρ = −0.56
P < 0.0001
ρ = −0.24
P = 0.05
ρ = −0.36
P < 0.01
IFN-γ  ρ = 0.27
P = 0.02
ρ = 0.40
P = < 0.001
TNF-α   ρ = 0.60
P < 0.0001

Immunophenotyping   Apart from an increased number of CD3+CD45RO+CD45RA+ cells in patients expressed as absolute number [388.0 (320.5–520.7) vs 245.5 (197.6–405.3), P < 0.01] or relative to CD3+ cells [20.4% (15.4–25.0) vs 14.5% (10.2–20.5), P < 0.01], no differences were found between controls and patients, especially when considering the number of CD3+ [1505.5 (1149.2–1800.1) vs 1517.3 (1161.1–1815.7), P = ns], CD3+CD45RO+ (memory T-cells) [832.0 (717.3–936.9) vs 758.8 (651.8–923.4), P = ns] and CD3+CD45RA+ cells (naïve T cells) [767.0 (474.4–995.4) vs 638.4 (527.2–926.6), P = ns] (Fig. 3). Comparing each subgroup of patients separately with controls, we identified similar findings (data not shown).

Figure 3.

 Results of immunophenotyping for CD3+, CD3+CD45RA+, CD3+CD45RO+ and CD3+CD45RA+CD45RO+ cells in controls and patients. *P < 0.01.

Between controls and patients, no differences were detected either in the numbers of lymphocytes, B cells (CD19+), NK (natural killer) cells (CD3+CD16+CD56+) nor in the numbers of CD4+ and CD8+ cells (Table 1).

Table 1.   Demographic and haematological data for controls, irritable bowel syndrome, functional dyspepsia and non-cardiac chest pain
  1. IBS, irritable bowel syndrome; FD, functional dyspepsia; NCCP, non-cardiac chest pain; na, not applicable. aP < 0.01 between IBS and NCCP. bP = 0.05 between FD and NCCP. cP = 0.01 between controls and patients.

No. patients32302315
Women (%)23 (77)24 (80)19 (83)11 (73)
Age30.5 (24.5–44.5)c39.0 (30.0–45.0)a37.0 (27.0–50.0)b51.0 (43.0–56.0)ab
Acute onsetna9/206/132/7
Percent lymphocytes35.5 (33.7–41.0)32.3 (30.6–43.7)37.6 (26.4–39.8)28.2 (27.1–30.6)
T cells1517 (1161–1816)1574 (1153–1770)1262 (1101–1649)1689 (1352–1966)
CD4+ cells925 (769–1260)992 (824–1134)966 (713–1336)1172 (943–1484)
CD8+ cells637 (508–787)585 (516–800)571 (528–637)570 (519–636)
NK cells187 (132–356)267 (180–342)275 (194–339)172 (107–241)
B cells236 (159–295)240 (199–283)236 (188–293)306 (199–484)

Acute vs non-acute onset of symptoms

Cytokine production  Independently of the patient group, an increased plasma IL-10 concentration was present in acute onset FGID [91.0 (32.0–106.0) vs 0.0 (0.0–0.0), P < 0.001]. No difference in IL-6 concentration was found (Fig. 4), neither in stimulated cytokine expression from PBMC or monocytes. In FD, the LPS-stimulated expression of TNF-α was significantly enhanced in the group reporting an acute onset as compared to unspecified onset [105.5 (80–118) vs 52 (42–64) pg mL−1, P = 0.02].

Figure 4.

 Serum concentration of IL-10 and IL-6 in acute and unspecified onset FGID. *P < 0.001.

The strongest correlation in cytokine productions after mitogen stimulation was found for IL-5 and IL-13 (P = 0.89, P < 0.0001). Production of IL-12 was significantly correlated with TNF-α (P = 0.63, P < 0.0001) and IFN-γ (P = 0.67, P < 0.001). Significant negative correlations were present between IFN-γ and the production of the Th2-type cytokines IL-13 (P = −0.538, P = 0.03) and IL-5 (P = −0.35, P = 0.04). However, correlation of IL-12 and these Th2-type cytokines were not significant anymore (P = −0.24, P = ns for IL-5; P = −0.28, P = ns for IL-13).

The strongest correlation in cytokine productions after mitogen stimulation was found for IL-5 with IL-13 (P = 0.93, P < 0.0001) (Table 2). Production of IL-12 was significantly correlated with TNF-α (P = 0.60, P < 0.0001) and IFN-γ (P = 0.40, P < 0.001). Significant negative correlations were present between IFN-γ and the production of the Th2-type cytokines IL-13 (P = −0.56, P < 0.0001) and IL-5 (P = −0.52, P < 0.0001) as well as between IL-12 and these Th2-type cytokines (P = −0.35, P < 0.01 for IL-5; P = −0.36, P < 0.01 for IL-13). Therefore, and despite the decreased IL-10 production, we conclude to a shift from a Th1 to a Th2 type of cytokine production.

Immunophenoptyping   The percentage and absolute number of CD3+CD45RA+CD45RO+ cells were significantly higher in patients reporting an acute onset of their symptoms [23.1% (20.6–31.5) vs 17.4% (14.2–20.4), P = 0.01 and 505.2 (422.8–606.0) vs 321.4 (286.5–444.0), P = 0.04] (Fig. 5). No differences were observed in the amount of lymphocytes, B cells, NK cells, nor CD4+ or CD8+ cells.

Figure 5.

 Number of CD3+CD45RA+CD45RO+ cells in acute vs unspecified onset FGID. *P = 0.04.


In this pilot study, we evaluated the characteristics of immune function in a group of consecutive FGID patients. As compared to healthy subjects, the stimulated expression of IL-5 and IL-13 by lymphocytes was enhanced in patients suffering from IBS, FD and NCCP. Conversely, the stimulated monocytic IL-12 and lymphocytic IL-10 expressions were reduced in IBS and FD, while IFN-γ expression was also reduced in FD patients. Except for an increase in the numbers of CD3+CD45RA+CD45RO+ cells, no distinct cellular profile was detected. Patients with a presumed acute onset of their symptoms had higher serum IL-10 levels and more CD3+CD45RA+CD45RO+ cells, while TNF-α levels following stimulation with LPS are higher in FD patients reporting an acute onset.

The possible pathophysiological involvement of cytokines in FGID has been addressed by only a few papers, and these studies were limited to IBS. Dinan et al.22 described increased plasma levels of IL-6, while plasma TNF-α and IL-10 concentration remained identical between IBS patients and healthy subjects. Liebregts et al.,17 on the other hand, found an increase in both unstimulated and stimulated IL-6 production and a decrease in unstimulated TNF-α production in IBS, which were attributable to D-IBS and PI-IBS. The increased IL-6 production in IBS has been attributed to the changes in the hypothalamic-pituitary-adrenal axis (HPA).22 It is conceivable that similar interactions play a role, but studying an association with HPA axis activation was beyond the scope of this study. On the other hand, we did not detect a difference in plasma IL-6 concentration in any FGID in this study. Anxiety and depression have also been implicated in altered immune function in several disease states, including FGID.17,23 Interestingly, we found significant correlations between both anxiety and depression scores and stimulated IL5, IL13 and TNF levels. However, these correlations were only significant when grouping controls and patients, and larger studies will need to address whether this reflects the differences between groups, or whether this is also found within patient (sub)groups. The correlation between stimulated TNF production and HADS scores is in accordance with findings from Liebregts et al.17 in IBS. There is also an extensive literature on the interaction between the immune system and psychological disorders.24,25 Unexpectedly, we also noted a positive correlation between depression scores and basal serum IL-10 levels. Given the relatively low numbers of subjects in this study, confirmation will be required in future studies in specific subgroups.

Our results indicate lower LPS-stimulated IL-10 in FGID as compared to controls. Further analysis revealed that this was attributable to a significant decrease in FD and a borderline significant reduction in IBS, which is in accordance with the previous report of a low-producer genotype of IL-10 in IBS.18 In acute onset FGID, increased plasma levels of IL-10 were present, while enhanced stimulated TNF-α production was detected in acute onset FD only. An increase in baseline TNF-α has been described previously by Liebregts et al.17 in IBS. Furthermore, a high-producer genotype of TNF-α has been described previously in IBS.19 In our study, however, no differences were detected between acute and unspecified onset IBS, probably owing to the limited number of patients.

The discrepancy between lower stimulated IL-10 production in FGID vs healthy controls and increased plasma levels of IL-10 in acute onset FGID vs unspecified onset FGID raises a few questions. Considering the inhibition of Th1-cytokine production by IL-10,26 one would expect higher IL-10 levels in FGID in general. However, this was only present in acute onset FGID. Nevertheless, IL-10 is a very versatile cytokine exhibiting pleiotropic effects on different pathways involved in asthma and allergy. It even demonstrated therapeutic potential in these diseases.26 Based on this evidence, we hypothesize that IL-10 production by PBMC in FGID in general is insufficient to terminate the aberrant immunological response. On the other hand, the increased plasma levels in acute onset FGID could result from a persisting inappropriate cellular immune response despite effective removal of the initiating pathogen. Several studies provide the support to the infectious hypothesis in FGID. Community-based studies described a higher prevalence of IBS as well as FD following an outbreak of Salmonella gastroenteritis.5,8 Others focused mainly on PI-IBS.2,5–7,9 In an attempt to control this low-grade inflammation, IL-10 production may become enhanced with consequent inhibition of Th1-type responses and shift towards a Th2 type of humoral response.

In addition, we analysed the production of cytokines by PBMC and identified a decrease in IFN-γ and an increase in IL-5 and IL-13 production after stimulation with PHA in FD, IBS and NCCP as opposed to control subjects. These changes, together with the decreased stimulated IL-12 expression, are congruent with a shift from a Th1- to a Th2-immune profile. Based on unstimulated monocytic IL-10 and IL-12 production, a previous study concluded to a Th1 profile.16 However, in that study, stimulated cytokine release was not assessed. Furthermore, the slightly higher prevalence of asthma observed in IBS tends to support a shift towards Th2-immune profile in this disorder.27 Additionally, it has been demonstrated in IBS patients that airway responsiveness to metacholine was increased in the absence of a manifest allergic presentation.28 Food allergy has also been suspected to play a role in this disorder, although the success of allergen exclusion diets has been very unpredictable.29 In the small patient groups included in this study, the prevalence of allergy remained comparable between groups.

As a Th2 type of cytokines favours a humoral response, important increases in the number of certain immune cell subtypes are not to be expected. Indeed, only the number of combined CD3+CD45RA+CD45RO cells was increased in this study. It remains doubtful whether this subtype has any significance except indicating a transition of naïve T cells towards memory T cells. One in vitro study demonstrated a shift from the CD3+CD45RA+ to a CD3+CD45RO+ phenotype induced by IL-2.30 Interestingly, this phenotypic shift was associated with the expression of adhesion molecules, which could home these cells to specific tissues, such as the gastrointestinal tract. No changes in the number of B cells, T cells or NK cells were detected by immunophenotyping.

There are some limitations to this study. Firstly, although the subject’s age was not significantly associated with the subject’s group (healthy volunteer vs patients) in multivariate analysis, this variable should be adequately dealt with in confirmatory studies. Secondly, patients were recruited in a tertiary care centre, thus limiting the generalizability of our findings. Thirdly, the number of patients included in each group remained limited in this pilot study, and a larger scale study is needed to verify our results. This is especially important to verify the possible relationship between cytokine concentration and psychological variables within a specific patient group. Fourthly, in many subjects, the assay for circulating IL-6 and IL-10 did not reach the detection limit, which could induce bias in the results. A more sensitive test is required to clarify this issue. However, this was not the case for stimulated cytokine expression. Moreover, our data do not clearly indicate whether the observed cytokine profiles in FGID are part of the innate immune response of the patients or are a response of the adaptive immune system to a prior triggering agent, although the different cytokine profile observed in patients reporting an acute onset tends to favour the first hypothesis. As lymphocytes and monocytes were isolated from peripheral blood and as immunophenotyping was performed only in peripheral blood, any extrapolation to what is happening on the mucosal level remains hypothetical. Combining mucosal biopsies and plasma cytokine measurement would be interesting.

In conclusion, the immune response of patients suffering from FGID in general, and IBS and FD in particular, is shifted towards a Th2-cytokine profile without associated changes in circulating immune cells. We hypothesize that this change may play an aetiopathophysiological role in these disorders and could be therapeutically relevant.


Sebastien Kindt and Lukas Van Oudenhove are research fellows of the Fonds voor Wetenschappelijk Onderzoek–Vlaanderen, Belgium.