The chronic fatigue immune dysfunction syndrome (CFI DS) is a major subgroup of the chronic fatigue syndrome (C'FS), CFIDS is characterized by a protean group of signs and symptoms described frequently after an antecedent viral illness. Despite a plethora of pseudonyms used in the literature since the late nineteenth century [1, 2] and the recent classification of this syndrome by the Centers for Disease Control and Prevention (CDC; Atlanta) [3], controversy abounds over both the etiology and pathogenic mechanism(s) involved. Nonetheless, we [4] and other investigators [5, 6] have reported a strong association between immune dysfunction and a serological viral reactivation pattern among patients in this group. This finding appeared similar to that for a variety of conditions, such as chronic active hepatitis [7], juvenile rheumatoid arthritis, and systemic lupus erythematosus [8], in which a definite association between a particular HLA-DR/DQ haplotype and increased disease frequency has been reported. We thus elected to examine a cohort of patients with CFIDS, as defined by the CDC, with use of HLA-DR/DQ typing. The frequency of HLA class II antigens in this group was compared with that among a large group of closely matched controls.

Materials and Methods

Patients. One hundred ten patients were evaluated at our clinic over a 6-month period. The patients were all Cauca-sians (including l 7~/c who were Ashkenazi Jews). The fe-rnale-to-male ratio was 5.6:1, and all patients fit the CDC criteria established for this syndrome. When immune status and viral serologies were examined, viral reactivation patterns were observed for all patients to one or more of the following viruses: Epstein-Barr virus, cytomegalovirus, and human herpesvirus type 6. In addition, all patients demonstrated significantly decreased B cell mitogenic reactivity (when the pokeweed mitogen test was performed) and natural killer cell function as well as significantly increased frequency of activated T cells (CD3, HLA-DR') and levels of interleukin 2, as previously reported by our group [4].

Controls. Six hundred sixteen healthy Caucasian donors for the National Marrow Donor Program or for the Apheresis Program at the Community Blood Centers of South Florida were selected to provide samples that would yield unambiguous tissue-typing results.

HLA-DR/DQ Antigens. For this study, the HLA-DR and DQ types of all controls and patients were assigned to the "broad" types of HLA-DR and HLA-DQ antigens before statistical analysis. The following assignments were allowed: DR1 – 10 and DQ1 – 3. HLA "splits" were included in the broad specificity to which they belonged. Thus, DR15 [2] and DR16 [2] were included together within the broad type DR2.

HLA-DR/DQ Typing HLA-DR/DQ typing of patients and controls was performed serologically. B cells were isolated from heparin-treated peripheral blood samples <24 hours old. B cells were isolated as rosettes formed with Dyna-beads (HLA Class II monoclonal antibody – coated magnetic microspheres; HLA-Cell Prep II, Dynal, Great Neck, NY) using a Dynal Magnetic Particle Concentrator (neodymium-iron-boron permanent magnets). The rosettes were added directly to commercial tissue typing trays at a concentration of 200 rosettes per tissue-typing well (Genetic Testing Institute, Brookfield, WI; Biotest Diagnostics, Danville, NJ; and One Lambda, Canoga Park, CA) with use of an automatic Cell Dotter (Biotest Diagnostics). The cells were incubated with the serum panels for 30 minutes in a Sure-Temp incubator (Raleigh, NC) at 24ºC. Commercial rabbit complement (preselected for mediating lymphocytotoxicity by lot number; i.e., high titer and low background; Pel-Freez, Brown Deer, WI) was added directly to each well (5 pL per well, neat) with use of the Cell Dotter, and the cells were incubated for an additional 30 minutes at 24ºC. Complement-mediated lympholysis was measured by the addition to each well of a cocktail of acridine orange (0.1%), ethidium bromide (0.03%) and EDTA (5~Jo) in PBS (pH, 7.4; 2 pL per well). The reactions (percent killing in each well) werescored immediately on visual examination with use of a Leitz Flurovert microscope outfitted with ultraviolet and incident light (MicroOptics, Fort Lauderdale, FL: filter cube system [l, 2/3 Leitz 513 – 604 narrow-band blue-light excitation]; high-performance interference; short-pass filter [BP 450 – 490 nm]; beam-splitting mirror [RKP 510 nm]; log-pass suppression filter [LP 515 nm]). Microscopes were outfitted with computerized, semiautomatic stages and score-recording devices (Robbins Scientific, Mountain View, CA). Patterns of reactions with the panels of antisera were statistically analyzed, and HLA antigens were assigned for each individual tested.

Statistical analyses. Differences in the frequency of assignment to the broad types of HLA-DR and HLA-DQ antigens between groups were assessed using Pearson x2 statistics. Relative risk ratios and their associated 95% confidence intervals were computed for purposes of illustration. Fisher’s exact test was performed in each instance of probable statistical significance.

Results

The frequency of HLA-DR and DQ antigens in our control population (n = 616) was compared with the frequency of the same antigens reported for Caucasians in the United States (n = 254) who were studied at the 11th International Histocompatibility Workshop held in Yokohama, Japan [9], and no significant differences were observed.

Table I depicts the data for our patient group as compared with those for the control population. A significant association between CFIDS and the presence of HLA-DQ3 was noted (P < .01). The association with HLA-DR4 and HLA-DR5 may represent the slightly increased percentage of Ash-kenazi Jews (in whom these antigens are more prevalent) in our patient population as compared with that among controls. In addition. the association with HLA-DQ3 could represent an additive effect for patients who also have HLA-DR4 and/or HLA-DR5 since most DR4-, DR5-, and DR9-positive individuals are also DQ3-positive by virtue ot the well-known phenomenon of linkage disequilibrium. Results of Fisher’s exact test, however, suggest that the association between CFIDS and the presence of HLA-DQ3 is biologically significant.

Discussion

There are few published reports of the possible assciciation of genetic loci in the immune response in this syndrome despite the findings of an association between >S00 diseases and HLA class I and/or class II antigens in the last 10 years [10]. Alpart and colleagues [8] compared 59 Caucasian patients living in the United Kingdom with 247 controls without finding significant correlates. In their study, however, the patient population was selected only if they evidenced persistent enteroviral infection. Another study, albeit small, examined 12 patients with CFIDS who had concurrent sleep disorders and found that 11 of 12 were DQw1-positive [9]. The current series. therefore, is the largest group (n = 110) ot patients with CFIDS who were selected in the ahsence of subcategorization by symptoms or viral serology, and the results presented are intriguing. DQ3 (relative risk for CFIDS, 1.8) was significantly more prevalent in the patients than either the South Florida or North American Caucasian control groups. It is possible that DR4 (relative risk for C'FIDS, 1.6) and DR5 (relative risk for CFIDS, 1.8) are also associated with an increased risk of developing CFIDS. although this association will require further study of a larger cohort patients because of the number of Ashkenazi Jews in our patient population.

The relative risk determined in this study is less striking than that found for the development of ankylosing spondylitis (Reiter's syndrome) and the presence of HLA-B27 (relative risk, 69 for Caucasians) or narcolepsy and the presence of DR2 (relative risk, 129 for Caucasians); however, all of these relative risks are similar to those associated with the presence of HLA-DR antigens and the development of chronic active hepatitis [7], systemic lupus erythematosus, and juvenile rheumatoid arthritis [8]. These findings thus strongly suggest that further evaluation of persons with CFIDS, including an investigation of an HLA Class I linkage dysequilibrium and use of the molecular techniques of oligonucleotide sequencing of DR and DQ alleles, are warranted.

The data presented herein suggest that CFIDS, together with a variety of immune-mediated diseases, such as chronic active hepatitis [7] and juvenile rheumatoid arthritis [8], may share similar sequences of pathogenetic mechanisms. Therefore, it may be speculated that in a particular subpopulation a genetic predisposition may be triggered immunologically by any of a number of potential stimuli, resulting in a state of chronic immune dysequilibrium (figure 1). This model could easily explain the recent findings with regard to acute viral infections, chronic active viral infection, allergies, or other as yet unreported mechanisms that are obscured by the process of chronic immune activation and viral reactivation.

References

I. Gantz NM, Holmes GP. Treatment of patients with chronic fatigue syndrome. Drugs 1989;38:855 – 62.
2. Levine PH, Krueger GRF, Straus S. The postviral chronic fatigue syndrome: a roundtahle [letter]. J Infect Dis 1989;160:722 – 4.
3. Holmes GP, Kaplan JE, Gantz NM, et al. C'hronic fatigue syndrome: a working case definition. Ann Intern Med 1988; 108:387 – 9.
4. Klimas NG, Morgan R, Salvado F, Fletcher MAP. Immunologic abnormalities of chronic fatigue syndrome. J Flin Microbiol 1990;28: 1403 – 10.
5. Gupta S, Vayuvegula B. A comprehensive immunological analysis in chronic fatigue syndrome. Scand J Immunol 1991;33:319 – 27.
6. Landay AL, Jessop C, Lennette ET, Levy JA. Chronic fatigue syndrome: clinical condition associated with immune activation. Lancet 1991;338:707 – 12.
7. Talal N. Autoimmunity. New York: Academic Press, 1977:24 – 5.
8. Alpart S, koide J, Takada S. T cell regulatory disturbances in the rheumatic diseases. Rheum Dis Clin North Am 1987;13:431 – 47.
9. The Data Book of the l 1th International Histocompatibility Workshop. Tokyo: International Histocompatibility Society, 1991;2:807 – 14.
l0. Tiwari J, Terasaki P. HLA and disease associations. New York: SpringerVerlag, 1985:383-401.