Immunophenotyping in post-giardiasis functional gastrointestinal disease and chronic fatigue syndrome

In 2007, three years after a Giardia outbreak in Bergen, Norway, 1252 persons with stool-microscopy confirmed Giardia infection during the outbreak received a questionnaire regarding fatigue and abdominal complaints [3]. Five years after the outbreak, in 2009, 253 persons reporting chronic fatigue in the previous study received a mailed invitation, and 53 of these chose to participate in the present study, see Figure 1. These patients went through a clinical evaluation and were screened with a battery of routine blood tests. A total sample size of 50 was required for comparing two groups to obtain statistically significant results with a power of 80% with standardized effect size of 1 and alpha of 0.01. Due to the stratification of the patient population through clinical evaluation and questionnaires, we included the 53 patients willing to participate. The control group of 30 sex and age matched individuals included 22 Giardia exposed, non-fatigued questionnaire respondents and 10 healthy Giardia unexposed individuals. All were HIV negative and were not taking immunomodulatory or antibiotic medications. Written consent was obtained from the participants. The study was approved by the Regional Committee for Ethics in Medical Research and the Norwegian Social Society Data Services in Bergen, Norway.

Patient and control blood samples for exclusion of other diseases and for the purpose of this study were drawn once between 08am and 09am after overnight fast and analyzed in parallel during the same period. Fecal samples were screened by microscopy and 18S PCR [32] of feces to rule out chronic giardiasis.

On the day the blood samples were drawn, both patients and controls completed a Rome II questionnaire [5] enabling the classification of abdominal symptoms into IBS and FD or other FGID. An abdominal symptom scoring form was used to quantify abdominal symptoms [33]. Questions about abdominal complaints before the Giardia infection enabled identification of post-infectious FGIDs. To evaluate severity of fatigue, patients and controls also completed the Fatigue Questionnaire [30], a validated set of 11 questions addressing different aspects of fatigue.

Patients and controls were classified according to the Rome II questionnaire with regard to FGID. Patients were clinically evaluated once, but in separate consultations with specialists in internal medicine, psychiatry and neurology. They were classified as PI-CFS or PI-idiopathic chronic fatigue (ICF) according to the 1994 Fukuda criteria [34], or as “fatigue other causes” when most plausibly due to other co-morbidities, or as “fatigue recovered” for individuals who had fully recovered from the fatigue (Morch et al., submitted).

Three different categorizations of patients and controls were formed for the analyses of the relative importance of CFS and FGID co-morbidity on immunological variables; a fatigue categorization, a FGID categorization and a combined categorization (Table 1). In the fatigue categorization, exploratory analyses of lymphocyte subsets in the small group of patients with PI-ICF (n=5) showed similarities with the PI-CFS group. The two groups were therefore analyzed together unless otherwise clearly stated.

In the analyses we also set up a combined categorization with one group of participants who had both FGID and CFS/ICF, one group who had FGID only, and one group with fatigue other cause with or without FGID. These groups were compared to each other and to healthy controls with neither FGID nor fatigue.

Lymphocyte subpopulation quantification was performed using the BD Multitest 6-color TBNK kit with BD Trucount Tubes for relative and absolute concentration determination (BD Biosciences, San Jose, CA, USA). The samples were prepared according to the manufacturer’s instructions and analyzed on a BD Canto II flow cytometer (BD Biosciences).

Peripheral blood mononuclear cells (PBMC) were isolated by density gradient separation from BD Vacutainer Na-citrate CPT tubes (BD, Franklin Lakes, NJ, USA). After harvesting, the PBMC were washed twice in PBS and cell suspensions (50μl) were stained 30 minutes in the dark using combinations of the following optimally titrated fluorescent dye-conjugated antibodies anti-CD8a-FITC, anti-CD27-FITC, anti-CD26-PE, anti-β7-PE, anti-CD45RO-PE, anti-CD4-PerCP-Cy5.5, anti-CD56-PerCP-Cy5.5, anti-CD19-PE/Cy-7, anti-CD45RA-PerCP-Cy5.5, anti-HLADR-PE/Cy7, anti-CD25-PE/Cy7 (BioLegend, SanDiego, CA, USA), anti-CD3-ECD (Beckman Coulter, Brea, CA, USA) and anti-CD127-PerCP-Cy5.5 (eBioScience, SanDiego, CA, USA). Appropriate isotype controls from the same manufacturer were used at equal concentrations. After staining, cells were washed once, resuspended in PBS-paraformaldehyde solution (1%) and analyzed the same day using a Beckman Coulter Cytomics FC 500 MPL flow cytometer. In a typical acquisition 7 × 10⁴ lymphocytes (min 2.3 × 10⁴, max 1.7 × 10⁵) were collected. The collected data were analyzed with FlowJo 7.6 software (Tree Star Inc, Ashland, OR, USA).

Unless otherwise stated, the data are presented as mean (standard deviation (SD)). Chi-squared tests were used for categorical comparisons between groups. Lymphocyte subset data and stimulation indices were analyzed using the Kruskal-Wallis test for all groups within each categorization and then Mann Whitney U test to compare two groups. The General Linear Model was used for multivariate analyses of possible interactions and confounding factors like sex and age on lymphocyte subset levels and for correlations between these and symptom scores. Due to multiple comparisons in lymphocyte subset analyses and a high number of variables, we chose a nominal significance level of 0.01. PASW 18 (SPSS Inc, Chicago, Ill, USA) was used for statistical analysis.

Peripheral blood lymphocyte subset results for the three analysis categorizations are shown in Table 2. We found the CD4:CD8 ratio to be reduced and the CD8 T-cell percentage to be increased in the FGID groups compared to the no-FGID control group. The same pattern was seen in the combined categorization, with elevated CD8 T-cell percentages and concentrations in both the group with FGID&CFS/ICF and the FGID, no-fatigue group.

As has been reported by others [35, 36] the NK-cell percentage measurements were significantly affected by sex, with men having higher levels than women (p=0.006). NK-cell levels also increased with increasing age (p=0.033). Correcting for this using multivariate analysis showed that PI-CFS/ICF patients had a lower percentage of NK-cells than the no-fatigue controls (p=0.005) when analyzing within the fatigue categorization. Also in the combined categorization the FGID&CFS/ICF patient group had lower percentage (p=0.006) and concentration (p=0.045) of NK-cells. The percentage and concentrations of CD3 and CD4 T-cells and CD19 B-cells were similar in all groups tested.

A significant correlation was found between the fatigue scores and abdominal symptom scores (R=0.421, p<0.001). NK-cell percentage correlated significantly with both fatigue scores (p=0.003) and with abdominal symptom scores (p=0.003) (Figure 2) and so did NK-cell concentrations with p=0.01 and p=0.009, respectively. Levels of CD8 T-cells did not show any correlation with symptoms.

CD4 and CD8 T-cells were analyzed for expression of activation markers HLA-DR, CD25, CD26, and CD27 and CD45RO. CD56⁺ NK-cells were analyzed for CD26. CD19⁺ B-cells were analyzed for activation markers CD25, CD26 and CD27. Additionally, we analyzed percentage of gut-homing integrin β7 positive T-cells, B-cells and CD4⁺CD25⁺⁺CD127^neg T-regulatory cells.

The percentage of CD26 positive CD3 T-cells in the “all FGID” group (78.5%(9.4)) was insignificantly decreased compared to the control group without FGID (83.4%(6.4)) in multivariate analysis (p=0.036). We did not find any significant differences along the FGID or CFS categorizations regarding the percentage of CD56^dim or CD56^bright NK-cells or for CD26 positive subpopulations of these NK-cell subsets. Nor did we find any significant differences between groups for activation markers and memory markers in T-cells, B-cells or in the percentage of gut-homing integrin β7 positive T-cells and B-cells or CD4⁺CD25⁺⁺CD127^- Treg cells (data not shown).

Searching the literature regarding CD4 and CD8 T-cell subsets in FGID we found two studies which did not find differences in these T-cell subsets [15, 16] and two studies with lowered CD4/CD8 ratio in IBS patients; one due to high total CD8 cells [37] and one due to low CD4 cells [38]. While CD4 and CD8 T-cell numbers and percentages may fluctuate considerably within an individual over time, the CD4:CD8 ratio is found to be relatively stable [39]. We found the CD4:CD8 ratio to be low due to increased CD8 T-cell numbers in our PI-FGID study population. Similar peripheral blood T-cell patterns are reported in patients suffering from a number of viral diseases like mononucleosis, dengue, RSV and cytomegalovirus infection and herpes simplex recrudescence as well as in chronic toxoplasmosis infection [39, 40]. A low grade ongoing immune response against reactivated viruses is a hypothesis of CFS morbidity [41].

CD8-T-cells have been shown to be responsible for the mucosal injury and temporary loss of disaccharidase activity in acute giardiasis [28]. Our finding of elevated CD8 T-cell levels in post-giardiasis FGID patients may indicate an association between the mucosal injury during the acute infection and the observed prolonged syndromes.

NK-cells do not seem to play a role in clearing acute Giardia infection [29]. No previous study has evaluated their possible role in post-giardiasis IBS or CFS. Two studies in general IBS populations found peripheral blood CD56⁺ NK-cells levels to be similar to healthy controls [16, 42]. A post-prandial decrease in NK-cells in IBS patients relative to controls has also been described [43].

The majority of studies of NK-cells in heterogeneous CFS populations have not found differences in NK-cell levels [11, 12]. However, two studies have found significantly reduced NK-cell levels in both PI-CFS and non-PI-CFS patients [19, 44]. One small study has shown NK-cell levels to return to normal after successful recovery in the PI-CFS group [45].

In our study, fatigue scores correlated with NK-cell levels, and the patients who had recovered from chronic fatigue showed NK-cell levels comparable to the non-fatigue control group. These data suggest that decreased NK-cell levels are associated with ongoing fatigue in post-giardiasis PI-CFS. As we also found NK-cell levels to be correlated with abdominal symptom scores, an alternative explanation might be that NK-cell levels play a role in sensitization or interpretation of sensory stimuli.

We could not replicate the elevated levels of CD26 expression in T and NK-cells as a robust marker of CFS reported by Fletcher et al. [13]. Instead, there was a tendency towards lower CD26⁺ T-cells in FGID patients.

A study in CFS patients recently reported higher levels of CD4⁺CD25⁺⁺FoxP3⁺ cells [12]. We found no difference in regulatory T-cells with regard to fatigue categorization as measured by CD4⁺CD25⁺⁺CD127^- cells, and also not with regard to FGID categorization. Also a previous study regarding regulatory T-cells in FGID patients did not find any such difference [14].

NK-cell levels are influenced by sleep and depression [46]. We did not control for poor sleep, but no patients in the CFS/ICF group had clinical depression. NK-cells are also known to fluctuate considerably with exercise, but the resting immune function evaluated in this study, is not very different in athletes compared to non-athletes [47]. The cytotoxic activity of NK-cell has been more consistently found decreased in CFS studies, but may be a bystander effect as it has been shown to be decreased in chronic stress like unemployment, while NK-cell concentration was not [48].

Out of the 253 invited persons, 53 consented to participate. We believe the main reason for the low participation rate was that the basis for invitation to participate was fatigue score at two years previously and that many had recovered during this time span. Another reason may be that the study logistics with five appointments for clinical evaluations and testing was too demanding for some. However, a chance for bias regarding increased health seeking behavior in the participants cannot be excluded. Also, we cannot fully exclude the contribution of an additional viral trigger in the development or perpetuation of fatigue symptoms.