Lymphocyte subset differences in patients with chronic fatigue syndrome, multiple sclerosis and major depression
Abstract
Chronic fatigue syndrome (CFS) is a heterogeneous disorder of unknown aetiology characterized by debilitating fatigue, along with other symptoms, for at least 6 months. Many studies demonstrate probable involvement of the central and autonomic nervous system, as well as a state of generalized immune activation and selective immune dysfunction in patients with CFS. The aim of this study was to compare the lymphocyte subsets of patients with chronic fatigue syndrome to those of patients with major depression and multiple sclerosis as well as those of healthy control subjects. No differences were found in total numbers of T cells, B cells or natural killer (NK) cells. However, differences were found in T, B and NK cell subsets. Patients with major depression had significantly fewer resting T (CD3+/CD25–) cells than the other groups. Patients with major depression also had significantly more CD20+/CD5+ B cells, a subset associated with the production of autoantibodies. Compared to patients with multiple sclerosis, patients with CFS had greater numbers of CD16+/CD3– NK cells. Further study will be required to determine whether these alterations in lymphocyte subsets are directly involved in the pathophysiology of these disorders, or are secondary effects of the causal agent(s).
Introduction
Chronic fatigue syndrome (CFS) is a heterogeneous disorder of unknown aetiology characterized by debilitating fatigue, myalgia, arthralgia, headache, sore throat, adenopathy, sleep disruption, difficulties with memory and concentration and post-exertional malaise [1,2]. Because of its clinical features and frequent onset after an acute febrile illness, one commonly held hypothesis is that CFS is due to infection with one or more viruses [3], leading to chronic, low-level activation of the immune system characterized by the overproduction of several cytokines. This immunological activation, in turn, is postulated to alter neurotransmitter function and to lead to the complex of symptoms seen in CFS. Although a number of studies have associated various infectious agents with CFS [4–10] most investigators do not believe that a single novel infectious agent will be shown to cause CFS. Indeed, no known agent has been conclusively shown to cause CFS.
Many studies have demonstrated central and autonomic nervous system abnormalities in patients with CFS, including neuroendocrine abnormalities reflecting dysfunction of the hypothalamus and pituitary gland [11–32]. Another commonly held hypothesis is that these neuroendocrine abnormalities are the primary disorder, and that they lead to secondary immune dysfunction. A variant of this hypothesis is that a chronic low-grade central nervous system infection triggers the neurological abnormalities seen in CFS.
Numerous immunological abnormalities have been described in patients with CFS. While a number of these putative abnormalities have been inconsistently reported, a few appear to be more robust: depressed natural killer (NK) cell function by in vitro testing, increased numbers of activated cytotoxic T cells, increased levels or production of proinflammatory cytokines, and deficiencies in immunoglobulin subclasses [33]. However, past studies have not included comparison groups with other fatiguing illnesses.
The aim of this study was to compare lymphocyte subsets in patients with CFS, major depression and multiple sclerosis (MS) and healthy control subjects. Patients with major depression and MS often experience symptoms of chronic fatigue in the course of their illness. We hypothesized that if the immunological abnormalities observed in CFS were specific and relevant to its pathogenesis, then immunological parameters in CFS should differ from those of disorders that share some of its symptoms but have a different pathophysiology.
Methods
Subject selection
Patients with CFS were recruited from those seen in the practice of A.L.K. and met the 1988 Centers for Disease Control and Prevention (CDC) case definition for CFS [34], which was the current definition at the time this study was initiated. Under this case definition there are 11 symptom criteria (fever, sore throat, painful lymph nodes, weakness, myalgia, post-exertional fatigue, headaches, arthralgia without redness or swelling, neuropsychological complaints, sleep disturbance and sudden onset) and three physical examination criteria (low-grade fever, non-exudative pharyngitis and cervical or axillary adenopathy). The physical examination criteria must be documented by a physician on two occasions at least 1 month apart. To establish a diagnosis of CFS under this definition, a patient must have a new onset of persistent or relapsing fatigue that reduces average daily activity by 50% for at least 6 months and exhibit either (a) six symptom and two physical examination criteria, or (b) eight symptom criteria. Other clinical conditions that could cause similar symptoms must be excluded. The patients with CFS that were selected for this study were not chosen because of the intensity of their illness symptoms: the only requirement was that they met the then-current CDC criteria. Patients were receiving symptomatic treatments at the time of the study.
Patients with major unipolar depression at the time of recruitment were enrolled from the practices of K.L.B. and S.N.W. All subjects were freely consenting out-patients, and were included if they met criteria for major depression as indicated in the Diagnostic and Statistical Manual for Psychiatry version III, revised (DSM-III-R). All were receiving standard treatments for depression, including pharmacological agents.
Patients with MS were recruited from the practice of G.A.M. from a large university MS speciality clinic. All patients had relapsing-remitting MS and met the criteria set forth by Poser and colleagues [35], including strongly confirmatory brain magnetic resonance imaging (MRI) scans. All enrolled patients had Kurtzke expanded disability status scores (EDSS) of 1·0 (no disability, minimal neurological signs) to 3·0 (moderate disability, fully ambulatory), and Hauser ambulation index scores of 1 (normal gait, significant fatigue) or 2 (noticeably abnormal gait or episodic imbalance). None had co-morbid conditions that could have contributed to their fatigue. In the MS patients selected for study, fatigue limited their activities by more than 50%, constituted their major MS-related limiting symptom, and was clearly disproportionate to their EDSS neurological impairments. None was clinically depressed or was taking antidepressant medication. None had ever received myelosuppressive or immunomodulatory treatment. None had had an MS attack or had received corticosteroids within 30 days prior to study entry and phlebotomy.
Control subjects were recruited from hospital employees who were asked to complete a health questionnaire to ensure that they were in good general health and had no history of depression, MS, CFS or any of the chronic symptoms that are required for a diagnosis of CFS. This study was approved by the Institutional Review Boards of Brigham and Women's Hospital, Boston, MA and the former New England Deaconess Hospital (now Beth Israel Deaconess Medical Center), Boston, MA. Written informed consent was obtained from all subjects prior to collecting blood samples.
At the time of enrolment, subjects in the depression, MS and control groups were matched for age (within 5 years) and sex to the patients with CFS. All blood samples were drawn between 8 and 11 o’clock in the morning to reduce variations caused by the normal daily fluctuations in lymphocyte subset levels.
Antibodies and reagents
Fluorochrome-conjugated murine monoclonal antibodies obtained from Coulter Immunology (Hialeah, FL, USA) included T11 (CD2), T3 (CD3), T4 (CD4), T1 (CD5), T8 (CD8), Mo1 (CD11b), 3G8 (CD16), B1 (CD20), B6 (CD23), IL-2R1 (CD25), 4B4 (CD29), NKH1 (CD56) and I-2 (HLA-DR). Fluorochrome-conjugated Leu-7 (CD57) was purchased from Becton-Dickinson (Mountain View, CA, USA). Fluoroscein isothiocyanate (FITC)-conjugated HB-7 (CD38) was a kind gift of Dr Thomas Tedder (Duke University Medical Center, Durham, NC, USA).
Isolation of peripheral blood mononuclear cells (PBMC)
Approximately 120 ml of fresh whole blood were collected in tubes containing ethylenediamine tetra-acetic acid (EDTA). Blood samples were coded prior to delivery to the laboratory, so that personnel performing the laboratory tests were blinded to the identity of the subjects. Peripheral blood mononuclear cells (PBMC) were separated by Ficoll-diatrizoate density gradient centrifugation, washed in medium containing 5% human antibody serum and cryopreserved in liquid nitrogen.
Immunofluorescence analysis and calculation of lymphocyte subset numbers
Cryopreserved PBMC were thawed, washed in medium containing 5% human AB serum, stained directly with FITC- and phycoerythrin (PE)-conjugated monoclonal antibodies, and incubated at 4°C in the dark for 20–30 min; the cells were then washed, fixed in 1% formaldehyde in phosphate-buffered saline (PBS), and stored in the dark at 4°C until flow cytometric analysis. Two-colour flow cytometry was performed using an EPICS 752 or EPICS Elite instrument from Coulter (Hialeah, FL, USA); 10 000 events were collected for each sample. During analysis, forward- and side-scattering properties were used to create a lymphocyte gate. Thresholds for discriminating levels of staining above background were established by analysis of PBMC stained with FITC- and PE-conjugated control monoclonal antibodies. The absolute number of various lymphocyte subsets was calculated by multiplying the total lymphocyte count (derived from a routine complete blood count performed on the same day a blood sample was obtained) by the percentage of cells in a sample expressing the relevant phenotype (derived from flow cytometric analysis).
The lymphocyte subsets analysed in this study are summarized in Tables 1 and and2.2. Data from other investigators on immunological abnormalities in patients with CFS were presented and/or published during the course of this study and surface antigens tested by immunophenotyping were modified accordingly. Therefore, not every blood sample was tested for every lymphocyte subset.
Table 1
Absolute numbers of total lymphocytes and major lymphocyte subsets.a
| Subset | CFS | Depression | MS | Control | P-valueb |
|---|---|---|---|---|---|
| Total lymphocytes | 1900 (23) | 2095 (24) | 2130 (22) | 1760 (25) | 0·499 |
| CD2+ PBLc | 1172 (23) | 944 (24) | 1228 (22) | 1197 (25) | 0·156 |
| CD3+ T cells | 940 (23) | 816 (24) | 1097 (22) | 974 (25) | 0·352 |
| CD20+ B cells | 348 (23) | 432 (23) | 272 (21) | 245 (23) | 0·048 |
| CD56+ NK cells | 284 (23) | 230 (24) | 216 (22) | 201 (25) | 0·441 |
Table 2
| Subset | CFS | Depression | MS | Healthy controls | Four-way P-value | Pairwise statistically significant differences | Trends (not statistically significant) |
|---|---|---|---|---|---|---|---|
| T cell subsets | |||||||
| CD3+/CD25– | 892 (14) | 376 (11) | 809 (13) | 1009 (12) | 0·006* | Dep ↓versus All | |
| CD3+/CD16+ | 169 (14) | 172 (11) | 88 (13) | 72 (12) | 0·259 | CFS ↑versus controls | |
| Dep ↑versus controls | |||||||
| CD3+/CD25+ | 121 (14) | 155 (11) | 74 (13) | 81 (12) | 0·061 | Dep ↑versus MS | |
| CD8+/HLA-DR+ | 64 (23) | 34 (23) | 47 (22) | 26 (24) | 0·082 | CFS ↑versus controls | |
| CD8+/CD38+ | 196 (18) | 133 (24) | 90 (22) | 134 (23) | 0·057 | CFS ↑versus MS | |
| B cell subsets | |||||||
| CD20+/CD5+ | 130 (19) | 471 (11) | 100 (11) | 154 (14) | 0·004* | Dep ↑versus All | |
| CD20+/CD25+ | 62 (19) | 89 (11) | 29 (13) | 30 (14) | 0·044 | CFS ↑versus MS | |
| CFS ↑versus controls | |||||||
| Dep ↑versus MS | |||||||
| Dep ↑versus controls | |||||||
| CD20+/CD23+ | 17 (7) | 29 (23) | 49 (18) | 49 (16) | 0·179 | CFS ↓versus MS | |
| CFS ↓versus controls | |||||||
| NK cell subsets | |||||||
| CD16+/CD3– | 506 (14) | 333 (11) | 238 (13) | 309 (12) | 0·008* | CFS ↑versus MS | |
| CD56+/CD25+ | 20 (18) | 13 (23) | 0 (22) | 0 (23) | 0·179 | CFS ↑versus MS | |
| CFS ↑versus controls | |||||||
| Dep ↑versus MS | |||||||
| Dep ↑versus controls | |||||||
| CD56+/CD16– | 66 (19) | 116 (11) | 78 (13) | 58 (14) | 0·115 | Dep ↑versus controls | |
| CD56+/CD2− | 80 (20) | 52 (24) | 39 (22) | 42 (24) | 0·123 | CFS ↑versus MS | |
Values are median absolute number of cells per µl. Numbers in parentheses are the number of samples tested.
Statistical analysis
For each sample the number of CD2+ cells was derived from single-colour analysis. The absolute numbers of other major lymphocyte subsets (i.e. CD3+, CD8+, CD20+, CD56+) were calculated from the results of two-colour staining. For example, the number of CD20+ B cells were calculated for each sample by first calculating the following cell number sums: (CD20+/CD5– + CD20+/CD5+) (CD20+/CD29– + CD20+/CD29+) (CD20+/CD25– + CD20+/CD25+) and (CD20+/CD23– + CD20+/CD23+). For each sample, the average value of these four sums was the reported CD20+ cell number value. Similar analyses were conducted for the CD3+, CD8+ and CD56+ cells.
The gender make-up of the four comparison groups was tested for statistically significant differences using a χ2 test on the 2 × 4 contingency table. As the age data were not normally distributed, differences in age between the four comparison groups were tested for significance using a Kruskal–Wallis analysis of variance (anova) on rank test. Similarly, Kruskal–Wallis anova on rank tests were used to detect differences between the comparison groups for each lymphocyte subset, as the cell number data also were not normally distributed. If a difference was found, pairwise comparisons using Dunn's method were performed to identify specific differences among the four groups. The sigmastat (version 1·0, Jandel Scientific, San Rafael, CA, USA) software package was used for all statistical tests.
In this study we tested 81 different lymphocyte subsets for statistically significant differences. Because 81 different null hypotheses are being tested, it would be appropriate to lower the threshold P-value for statistical significance for each test, so that the overall P-value for all tests combined is 0·05. However, due to the exploratory nature of this study and the fact that many of the lymphocyte subsets are not independent of each other, a P-value correction for multiple comparisons is not recommended [36,37]. Because no correction was made for the total number of statistical tests, we chose to use a more conservative P-value threshold of 0·01 to determine statistical significance. We also identified group trends in the lymphocyte subsets where no significant differences were found. We report as trends only those groups that have cell numbers at least twice that of another group.
Results
Composition of study subject groups
Study subjects included 23 patients with CFS, 22 patients with MS, 24 patients with major depression, and 25 healthy control subjects. Members of each group were matched for age (within 5 years) and gender (Table 3). Duration of illness was also compared among the patient groups. Patients with CFS had a significantly longer duration of illness than patients with depression (P= 0·003), but only six of 24 patients with depression provided the month and year of illness onset. From this we infer that most of the patients with depression could not recall specifically their illness onset. In contrast, all the patients with CFS remembered the onset of their illness because it was sudden.
Table 3
| No. of subjects | Age (years)a | Gender (% female) | Illness duration (years)a | |
|---|---|---|---|---|
| CFS | 23 | 38·6 | 73·9 | 3·9b |
| Depression | 24 | 35·5 | 75·0 | 0·8b,c |
| MS | 22 | 39·9 | 77·3 | 2·6d |
| Healthy controls | 25 | 37·7 | 76·0 | — |
| P-value | — | 0·843 | 0·995 | 0·003 |
Total lymphocyte counts and major lymphocyte subsets
Comparing the four study subject groups, there were no statistically significant differences in absolute numbers of total lymphocytes, CD2+ (T and most NK) cells, CD3+ T cells, CD20+ B cells or CD56+ NK cells (Table 1). No remarkable trends (absolute numbers of cells in one group at least twice the number in a comparison group) were detected among the groups for any of these major lymphocyte subsets.
T cell subsets
Comparing the four study groups, there was no trend towards any difference in total numbers of the CD4 helper T cell or CD8 suppressor/cytotoxic T cell subsets (data not shown). Patients with CFS tended to have increased numbers of activated cytotoxic T cells, as assessed by co-expression of CD8 with CD38 and HLA-DR (Table 2). Numbers of activated CD8+/CD38+ cells tended to be increased in patients with CFS compared to patients with MS, and numbers of CD8+/HLA-DR+ cells tended to be higher in patients with CFS than in healthy controls. However, these differences did not achieve statistical significance. We did not detect any significant differences in CD8+/CD11b+ cells comparing the four study groups (data not shown). Patients with depression had a significantly lower number of resting T (CD3+/CD25–) cells (P= 0·006) in the peripheral blood compared to all other study groups.
B cell subsets
CD20+/CD5+ B cells were significantly elevated in patients with depression compared to all other groups (P= 0·004). Activated CD20+/CD25+ B cells tended to be increased in patients with CFS and depression compared to patients with MS and healthy control subjects, but these differences were not statistically significant.
NK cell subsets
Of the various immunologic changes described in CFS patients, abnormalities in NK cell number and/or function have been reported most frequently. We therefore undertook an extensive analysis of NK cell subsets in CFS patients and the other study groups. As shown in Table 1, there were no statistically significant differences in absolute numbers of total CD56+ NK cells. We also detected no significant differences or trends for the following NK cell subsets: CD56+/CD8+, CD56+/CD11b+, CD56+/CD38+ or CD56+/CD57+. However, the number of CD16+/CD3– NK cells was significantly increased in patients with CFS compared to patients with MS (P= 0·008). Both patients with CFS and depression tended to have increased numbers of CD25+ NK cells compared to patients with MS and healthy control subjects, although the differences did not achieve statistical significance.
Discussion
We compared lymphocyte subsets of patients with CFS to those of patients with major depression and MS as well as those of healthy control subjects, with the goal of identifying immunophenotypic differences in patients with CFS compared to two patient groups with fatiguing illnesses − major depression and MS − and to healthy control subjects.
We found a trend (P= 0·057–0·082) towards increased numbers of CD8+/HLA-DR+ and CD8+/CD38+ in CFS patients compared to patients with MS or control subjects. It is possible that inclusion of larger numbers of study subjects would have allowed this trend to achieve statistical significance. This is in keeping with the reports of Klimas et al. [38], Landay et al. [39], Hassan et al. [40] and Peakman et al. [41], but not with several other reports [42–44].
Like Straus et al. [42] and Natelson et al. [43], but unlike Landay et al. [39], we did not detect any trend towards depressed numbers of CD8+/CD11b+ suppressor T cells in patients with CFS compared to control subjects. Of note, in the study of Landay et al. [39] the increased percentages of CD8+/HLA-DR+ and CD8+/CD38+ cells and depressed proportion of CD8+/CD11b+ cells were seen only in the most severely affected patients with CFS (those who were functioning at less than 25% of their normal daily activity).
While reduced NK cell function has been found repeatedly in patients with CFS [45,46], phenotypic findings have been inconsistent: low [45,47–49], normal [39,42,43] or even increased [38,41,50] percentages or absolute numbers of NK cells have been found in the peripheral blood of these patients. Although we did not detect any significant difference in the number of total NK cells in CFS patients, patients with CFS had significantly increased numbers of the CD16+/CD3– NK cell subset in the peripheral blood compared to patients with MS. In contrast to Morrison et al. [50], we did not find an increased number of CD16– (CD56bright) NK cells or a decreased proportion of CD16+ NK cells in patients with CFS. Resting CD56dim CD16+ NK cells do not express detectable levels of CD25 on the cell surface [51,52] but can express this antigen after in vitro activation [53,54]. CD25+ NK cells were observed in patients with CFS and depression, but were not detected in the blood of patients with MS or healthy subjects. Analysis of larger numbers of patients and control subjects will be required to determine if this trend is significant. Moreover, additional studies are needed to determine whether NK cell function is activated in vivo in CFS patients with higher numbers of CD25+ NK cells.
In contrast to Klimas et al. [38] and Tirelli et al. [48], but in accordance with other investigators [39,42,47,49], we did not find increased numbers of either total B cells or of CD5+ B cells in patients with CFS compared to healthy subjects. Neither Klimas [38] nor Tirelli [48] evaluated the patients with CFS carefully for concomitant major depression. We have found that major depression may be characterized by higher numbers of CD5+ B cells (see below). We did find that patients with CFS tended to have increased numbers of activated B (CD20+/CD25+) cells compared to patients with MS and healthy control subjects, although this trend was not statistically significant (P= 0·044).
In comparison to the other groups, patients with major depression had significantly higher numbers of CD20+/CD5+ B cells. We are not aware that this has been reported previously. CD5+ (or B-1) B cells are a distinct subset of mature B cells that constitutes approximately 1–7% of peripheral blood leucocytes (PBL) in healthy adult humans [55,56]. The immunoglobulin genes of CD5+ B cells usually do not show evidence of somatic hypermutation and CD5+ B cells tend to produce low affinity IgM autoantibodies. Maes et al. [57,58] found increased levels of total B cells in depressed patients, whereas others have reported no difference in total B cell numbers compared to control subjects [59–63]. However, none of these studies describes an analysis of CD5+ B cells.
We observed a highly statistically significant decrease in resting (CD25–) T cells in patients with depression, compared to all other subject groups; this was accompanied by a non-significant trend (P= 0·061) toward higher numbers of activated CD25+ T cells. Maes et al. [58,64] have reported increased numbers of CD25+ and HLA-DR+ lymphocytes in patients with depression. They subsequently identified these CD2+/HLA-DR+ and CD7+/CD25+ lymphocytes as activated ‘T cells’[65]. However, because both T cells and NK cells express CD2 and CD7, it is possible that these CD2+/CD7+/CD25+/HLA-DR+ lymphocytes comprise CD16– NK cells as well as activated T cells. If so, the results of Maes et al. would be largely in agreement with ours with respect to depressed patients. In contrast to some previous studies [43,66–68], we did not observe any trend toward increased numbers of CD25+ T cells and CD8+/CD11b+ T cells in patients with MS compared to control subjects, nor any trend toward decreased numbers of CD8+/HLA-DR+ T cells. However, the most consistent abnormalities that have been reported in MS involve the CD4+/CD45RA+ T cell subset.
We have identified several differences in T, B and NK cell subsets in patients with CFS and depression that were not present in patients with MS. Some of these changes were shared by both patients with CFS and patients with depression, whereas others were different. We cannot determine whether the alterations in lymphocyte subsets that we observed are directly involved in the pathophysiology of these disorders, or whether they are secondary effects. Future studies will be required to distinguish between these possibilities.
Acknowledgments
We thank David A. Hafler MD, J. Stephen Heisel MD and Marika J. Hohol MD for patient recruitment and Thomas Manley, Christine Cameron and Keith Cochran for technical assistance. This work was supported in part by grant numbers R01-A127314 and U01-A132246 from the National Institute of Allergy and Infectious Diseases and by a gift from the DeYoung Foundation. M. J. R. is supported in part by NIH grant number MO1 RR0075–27S3.
