The Effect Of Intestinal Microbes On Systemic Immunity

[Guest guest]

The Effect of Intestinal Microbes on Systemic Immunity

BY LEO GALLAND, M.D.

 

EXCERPTED FROM POWER HEALING (RANDOM HOUSE, 1998)

BY LEO GALLAND, M.D.

 

In keeping with its immense surface area and intense exposure to

foreign antigens, the intestinal tract is the largest organ of immune

surveillance and response in the human body (Targan et al., 1987). It

should not be surprising that events occurring in its lumen or on the

mucosal surface have systemic effects on immune function and disease

resistance. This chapter examines the contribution made by luminal

organisms commonly encountered in humans: bacteria, protozoa and

yeasts. Particular attention win be given to data concerning a role

for Giardia lamblia infestation and Candida albicans colonization in

the pathogenesis of chronic fatigue and immune dysfunction.

 

BACTERIAL MICROFLORA

 

Over 500 species of bacteria live in the healthy human alimentary

canal; in the average adult they weigh about one kilogram. The normal

colonic microflora ferment soluble fibre to yield short-chain fatty

acids which supply 5-10% of human energy requirements (McNeil, 1984).

Endogenous flora synthesize at least seven essential nutrients,

supplementing dietary intake: folic acid, biotin, pantothenic acid,

riboflavin, pyridoxine, cobalamin and vitamin K (Mackowiak, 1982).

They participate in the metabolism of drugs, hormones and

carcinogens, including digoxin (Lindenbaum et al., 1981),

sulphasalazine, and estrogens (Gorbach, 1982). By demethylating

methylmercury, gut flora protect mice from mercury toxicity (Rowland

et al., 1984). They prevent potential pathogens from establishing

infection by numerous mechanisms, which include: production of short-

chain fatty acids and bacteriocin, induction of a low oxidation-

reduction potential, competition for nutrients, deconjugation of the

bile acids (which renders them bacteriostatic), blockade of adherence

receptors and degradation of bacterial toxins (Savage, 1980).

 

Germ-free animals have mild to moderate defects in immune function

when compared to control animals. These include lower levels of

natural antibodies, hyporesponsive macrophages and neutrophiles,

defective production of colony-stimulating factors, leukopenia,

lymphoid hypoplasia, subnormal interferon levels and weak delayed

hypersensitivity (DHS) responses. They are more susceptible to

infection with intracellular parasites such as Listeria,

Mycobacterium and Nocardia, but are not more susceptible to viral

infection (Mackowiak, 1982). Adverse effects of endogenous bacteria

have also been described, indicating the complexity of the host-

saprophyte relationship. In diseases where host immune response is

the primary cause of pathology, such as lymphocytic choriomeningitis,

germ-free animals fare better than control animals (Mackowiak, 1982).

 

The immunologic effects of normal gut flora are in part due to

antigenic stimulation and in part to the bacterial origin of specific

immune activators, such as endotoxin lipopolysaccharicle (LPS) and

muramyl dipeptides Worrison and Ryan, 1979; Mackowiak, 1982; Stokes,

1984). An important role for these substances in normal immune

regulation has not been established, however (Mayrhofer, 1984).

 

The gut flora of healthy individuals is very stable (Sears et al.,

1950,19-%); this stability may in part be due to interbacterial

inhibition (Sprunt and Redman, 1968). Alteration in the level of

normal flora by antibiotics has long been known to allow secondary

infection by pathogenic bacteria and yeasts (Keefer, 1951; Seelig,

1966).

 

Occasional publications describe abnormal fecal flora in patients

with atopic eczema. Kuvaeva et al. (1984) studied 60 infants in

Moscow with IgE mediated food allergy and eczema. They reported a

decrease in anaerobic bacteria and lactic acid-producing aerobes and

an increase of Enterobacteriaceae. Severity of eczema was directly

proportional to severity of dysbiosis. No control data are given.

Ionescu et al. (1986) studied fecal flora in children and adults with

atopic eczema. Compared with healthy controls, there was a marked

reduction in Lactobacillus, Bifidobacterium and Enteroccoccus species

in the great majority of cases. This was associated with increased

concentrations of Candida species, Proteus, Klebsiella, and

Staphylococcus allreus, and appearance of atypical coliforms and

Clostridium innocutan. The high frequency of hypoalbumenernia,

indicanuria and steatorrhea in the eczema group suggested small bowel

bacterial overgrowth with secondary malabsorption. In neither of

these studies is it possible to determine whether abnormal bowel

flora caused allergy or whether food-allergic disease destabilized

gut flora.

 

Immunologic reactions to normal or abnormal components of the

bacterial gut flora are implicated in the etiology of some

inflammatory disorders. Reactive arthritis may occur after intestinal

infection with Salmonella typhimurium, Yersinia enterocolitica

serotype 3, Shigella flexneri, Campylobacter jejuni and Clostridium

difficile (Inman, 1988). Because arthritogenic potential is strain-

specific and because 60-80% of patients with reactive arthritis carry

the HLA-B27 gene, it is likely that genetically determined antigenic

crossreactivity plays a role (Yu et al., 1989).

 

Ankylosing spondylitis (AS) occurs almost exclusively in HLA-

B27positive individuals. An increased rate of intestinal colonization

with Klebsiella pneumoniae has been described in this condition,

according to some but not all studies (Kinsella, 1988). Immunologic

cross-reactivity has been shown for HLA-B27 antigen expressed on the

host cell membrane and antigens present in K. pneumoniae, S. flexneri

and Y. enterocolitica, suggesting molecular mimicry in the

pathogenesis of this disease (Yu, 1988). Workers in Australia have

demonstrated bacteria with cross-reactive antigenic determinants in

bowel flora of B27-positive AS patients; these bacteria are almost

never found in B27-positive controls without AS (McGuignan et al.,

1986).

 

Endotoxemia has been described in patients with psoriasis (Rosenberg

and Belew, 1982a) and cystic acne (Juhlin and Michaelson, 1984).

Activation of the alternative complement pathway (APC) by gut-derived

endotoxin may play a role in the pathogenesis of these disorders.

Exposure of macrophages to endotoxin causes release of cytokines,

such as interleukin-1 (Il-1) and tumor necrosis factor (TNF). These

peptides have powerful effects on the immunologic and metabolic

response to infection. Whether gutderived endotoxins influence

cytokine production in vivo is unknown.

 

GIARDIA LAMBLIA

 

Although it was first described by van Leuwenhoek in 1681, it is only

in the past 25 years that G. lamblia has been acknowledged as an

important pathogen (Gillon, 1984). Giardiasis is the commonest cause

of parasitic disease in the United States (Myer and Jarroll, 1980)

with an overall prevalence estimated at 7.4%, which is about the same

as its average worldwide prevalence (Mahmoud and Warren, 1975).

Prevalence in Great Britain varies from 2% to 10% (Felman and

Nikitas, 1985). At least 27% of Giardia infections identified at the

University of Edinburgh Medical School had been acquired within the

UK and a diagnosis of giardiasis had not been suspected in two-thirds

of cases (Gibb, 1989).

 

Reports based on stool screening may underestimate the prevalence of

giardiasis. Comparison of stool examination with duodenal aspiration

has consistently shown that stool examination fails to identify

infected patients even at the height of acute infection. Single stool

specimens have a sensitivity of zero (Rosenthal and Liebman, 1980) to

50% (Kamath and Murugasu, 1974). Collecting multiple specimens over

several days increases the sensitivity to 85-90% (Gillon, 1984).

 

To overcome the limitations of stool analysis we developed a

diagnostic technique by which rectal mucus obtained at anoscopy is

stained with a monoclonal antibody to Giardia cysts and examined by

epifluorescence microscopy (Galland and Bueno, 1989). We recently

conducted a two-year retrospective study of 218 patients who

presented to our medical clinic with a chief complaint of chronic

fatigue (Galland et al., 1990). G. lamblia infection was identified

by rectal swab in 61 patients. The symptoms of patients with and

without giardiasis, are shown in Table 1.

 

All patients with giardiasis and 86% of patients without giardiasis

complained of digestive symptoms, but these were generally mild. The

most interesting difference between the two groups lies in the

positive association between giardiasis and symptoms such as myalgia,

muscle weakness, flu-like feelings, sweats and adenopathy. In fact,

61% of fatigued patients with giardiasis had been diagnosed elsewhere

as suffering from chronic fatigue syndrome (CFS) or ME, compared to

only 19% of fatigued patients without giardiasis. Cure of giardiasis

resulted in clearing of fatigue and related 'viral' symptoms

(myalgia, sweats, flu-like feelings) in 70% of cases, some palliation

of fatigue in 18%, and was of no benefit in 12%. This study shows

that giardiasis can present with fatigue as the major manifestation,

accompanied by minor gastrointestinal complaints and sometimes by

myalgia and other symptoms suggestive of ME. It indicates that G.

lamblia infection may be a common cause of CFS, at least in the

United States. It is noteworthy that tricyclic antidepressants, a

standard treatment for CFS, suppress the growth of Giardia in vitro

(Weinbach et al., 1985).

 

The mechanism by which G. lamblia causes disease is not known.

Experiments with human volunteers demonstrate that the ability of

Giardia to produce infection and to cause diarrhea depends upon the

strain of G. lamblia used (Nash et al., 1987), the inoculum dose

(Rentdorff, 1954; Rentdorff and Holt, 1954), and previous exposure to

the organism (Nash et al., 1987). Heterogeneity of Giardia isolates

from humans in the same city occurs and has been proposed as one

mechanism for variability of clinical response to infection (Korman

et al., 1986).

 

Table 1 Systemic symptoms of CFS patients

 

With giardiasis

(%) (N = 63) Without giardiasis

(%) (N = 157)

Depression 61 41

Muscle weakness 46 19

Headache 41 36

Sore throat 41 11

Lymphadenopathy 36 8

Arthralgia 36 27

Myalgia 34 18

Flu-like feelings 34 6

Poor exercise tolerance 30 10

 

The ease with which G. muris infections are established in mice

varies with the genetic background of the host (Belosevic et al.,

1984). Acute murine giardiasis suppresses the response of splenic and

mesenteric lymphocytes to sheep erythrocytes; susceptible mice

express a greater degree of immunosuppression than do resistant mice

(Belosevic et al., 1985a). Acute giardiasis in rodents is associated

with increased intestinal production of prostaglandins E and F

(Ganguly et al., 1984b) and cyclic AMP (Ganguly et al., 1984a),

probably a result of macrophage activation (Kanwar et al., 1987).

Just as production of PGE by host macrophages may contribute to

diarrhea, activation of a population of suppressor macrophages in

mesenteric lymph nodes contributes to immunosuppression in acute

giardiasis (Belosevic et al., 1985b). Disaccharidase deficiency, a

frequent complication of giardiasis, may also be immunologically

mediated. Sensitized gerbils develop a depression of disaccharidase

activity when exposed to Giardia antigens; live organisms are not

required (denHollander et al., 1988).

 

In chronic infection, the immune response to the host also appears to

be a critical determinant of outcome (denHollander et al., 1988).

Chronic giardiasis in humans has been associated with deficiency of

secretary IgA (Vinayak et al., 1987) and with impaired macrophage

cytotoxicity (Smith et al., 1982). In animals infected with G. muris,

both T helper/inducer lymphocytes and mast cells are critical for

clearance of the parasite (Heyworth et al., 1987; denHollander et

al., 1988), whereas cytotoxic T cells and natural killer cells are

not (Heyworth et al., 1986). It is noteworthy that athymic mice with

chronic giardiasis do not develop mucosal damage (Roberts-Thompson

and Mitchell, 1978). Gillon et al. (1982) have proposed that the

release of enteropathic lymphokines by intraepithelial T cells is the

cause of the intestinal injury in chronic giardiasis. In humans, the

severity of malabsorption observed with chronic giardiasis is more

closely related to the presence of intraepithelial lymphocytes and

the antibody titer to Giardia cyst antigen than to the estimated

parasite burden (Solomons, 1982).

 

In rodents and humans, therefore, acute Giardia infection elicits a

protective response from mast cells and T helper lymphocytes

(responsible for stimulating sIGA secretion) which is essential for

clearance of parasites. A macrophage response occurs as well; this is

both protective (Smith et al., 1982) and immunosuppressive (Belosevic

et al., 1985b), depending perhaps on the activity of different

macrophage populations. Chronic giardiasis is a disease of immune

dysregulation in which effector lymphocytes mediate tissue damage.

Defective control of macrophage-lymphocyte communication in Giardia

infection is likely, and appears to be genetically determined.

Defective macrophage-lymphocyte communication is also a feature of

human atopic disease (Galland, 1986) and the relationship between

human giardiasis and allergy is therefore of interest.

 

Immunologic hypersensitivity to G. lamblia has been reported; the

result may be asthma (Fossati, 1971; Lopez-Brea et al., 1979),

urticaria (Harris and Mitchell, 1949; Wilhelm, 1958; Webster, 1958;

Dellamonica et al., 1976; Weisman, 1979; Kennou, 1980; Farthing et

al., 1983), arthritis (Goobar, 1977; Farthing et al., 1983; Woo and

Panayi, 1984; Shaw and Stevens, 1987; Galland, 1989) and uveitis

(Carroll et al., 1961). Hypersensitivity reactions may occur in the

absence of digestive complaints (Wilhelm, 1957; Kennou, 1980;

Galland, 1989). In none of these cases was the mechanism of

hypersensitivity known; eosinophilia. was a feature in only two cases

(Kennou, 1980; Farthing et al., 1983). A high frequency of pre-

existing atopic disease occurs in patients with chronic giardiasis

(Chester et al., 1985; Galland et al., 1990) and may be a factor in

susceptibility to infection. We have observed that when several

members of a family are infected with Giardia, symptoms tend to be

more prominent among those with allergy. We suspect that the immune

dysregulation which underlies atopy allows the immunologic response

to Giardia infection to favor chronic disease.

 

Two other features of chronic giardiasis are relevant to an

understanding of CFS: the effect of G. lamblia infection on

nutritional status and its interaction with other organisms,

specifically viruses, bacteria and fungi.

 

G. lamblia can cause intestinal protein loss without producing

diarrhea (Sherman and Lieberman, 1980). Specific micronutrient

deficiencies have also been described in chronic giardiasis. Low

levels of carotene and folate (Brasitus, 1983) and abnormal vitamin A

and folic acid absorption curves (Solomons, 1982) occur in a large

minority of patients with chronic symptoms. Serum vitamin B12 may be

low (Cowan and Campbell, 1973), and abnormal Schilling tests occur in

a substantial number of patients (Solomons, 1982). Direct competition

between parasite and host for vitamin B,2, as suggested by Cowan and

Campbell (1973), seems unlikely, as Giardia selectively damages the

duodenum and upper jejunum, and cobalamin is absorbed in the distal

ileum. Bacterial overgrowth of the small bowel has been described in

giardiasis (Yardley et al., 1965; Tandon et al., 1977; Tompkins et

fil., 1978; Rogers, 1979) and is associated with severity of

malabsorption (Tompkins et al., 1978; Tandon et al., 1977). Solomons

(1982) has proposed bacterial overgrowth as a possible cause of

abnormal Schilling tests in giardiasis. Bacteroides fragilis produces

a substance which binds the B12-intrinsic factor complex (Mackowiak,

1982) and may cause malabsorption.

 

Colonization of the jejunum with Candida albicans was reported in 30%

of patients with giardiasis and was absent in controls (Naik et al.,

1978). The implications of intestinal candidiasis for CFS are

described later in this chapter. Some strains of G. lamblia contain

double-stranded RNA viruses (denHollander et al., 1988). The role of

Giardia as a vector for viral infection requires further study.

 

AMOEBAE

 

Entamoeba histolytica infects 10% of the world's population (Walsh,

1986a,b). Cysts can be found in stool samples of 2% to over 40% of

individuals, depending on the area and level of hygiene and

sanitation (Guerrant, 1986). Amoebic antibodies, indicative of past

or present invasive infection, were found in 1% of general hospital

patients, 2% of random serum specimens, and 4% of healthy military

recruits in the United States (Walsh, 1986b). Amoebic infection is

found in about one-third of homosexual men attending clinics for

sexually transmitted diseases in the United States (Petri and Ravdin,

1986). Over 90% of individuals infected with Entamoeba histolytica

are asymptomatic.

 

The clinical response to amoebic infection is better understood than

the clinical response to Giardia infection. Pathogenic strains of

Entamoeba histolytica are able to evade lysis by both classical and

alternative pathways of complement (Reed et al., 1986). Intestinal

bacteria, E. coli in particular, are necessary for this complement

resistance and for amoebic virulence (Wittner and Rosenbaum, 1970).

It is suggested that ingested bacteria lower the redox potential and

allow the amoebae to escape destruction by oxidative enzymes (Gitler

and Mirelman, 1986). Whereas amoebae of low virulence are killed by

granulocytes, highly virulent amoebae resist phagocytosis and instead

kill the attacking leukocytes (Guerrant et al., 1981; Chadee et al.,

1985). Mirelman (Mirelman, 1987; Mirelman et al., 1986) has reported

that one can reversibly change the zvmodeme patterns of Entamoeba

histolytica isolates from non-pathogenic to invasive by culturing

amoebae with the gut flora of patients who have either invasive

disease or no symptoms. His work, which is controversial, suggests

that pathogenicity may actually be determined by the bacterial milieu.

 

The immunologic effects of amoebic infection have been the focus of a

recent trans-Atlantic controversy. Workers in London observed that

Entamoeba histolytica infestation of HIV-infected homosexual men

involved only non-pathogenic amoebic zymodemes; amoebic antibodies

were absent and there was no association with diarrhea or increased

morbiditv (AllasonJones et al., 1986). Several North American groups,

on the other 6nd, have found that AIDS patients with diarrhea are

often infected with 'nonpathogenic' amoebae. Treatment with

metronidazole or paromomyci Produces relief of diarrhea in parallel

with the disappearance of these amoebae from feces (Rolsten et al.,

1986; Sullam et al., 1986; Pearce and Abrams, 1987). Several

researchers in the United States have advanced the notion that

infection with E. histolytica and other parasites may promote the

development of AIDS in HIV-infected individuals (Pearce, 1983; Pearce

and Abrams, 1984, 1986; Archer and Glinsman, 1985; Krogstad, 1986;

Petri and Ravdin, 1986; Croxson et al., 1988). Entamoeba histolytica

contains a soluble lectin which is mitogenic for T lymphocytes (Chen

et al., 1985; Petri and Ravdin, 1986). T helper cell activation by

this lectin may induce HIV replication in vivo. A soluble Entamoeba

histolytica protein, although not mitogenic itself, induced HIV

replication in tissue culture of lymphocytes obtained from three out

of seven men with chronic HIV infection (Croxson et al., 1988).

 

Synergism between intestinal parasites and lymphotrophic retroviruses

has also been advanced as an explanation for the pathogenesis of

Burkitt's lymphoma (Burkitt, 1983) and adult T cell leukemia/lymphoma

(Tajima. et al., 1981). It seems likely that the clinical importance

of amoebic infection is related as much to the characteristics of the

host as of the parasite.

 

Chronic Entamoeba histolytica infection of humans has been associated

with autoimmune phenomena, including the appearance of antibodies to

colonic epithelial cells (Salem et al., 1973) and development of

symmetrical polyarthritis very similar to rheumatoid arthritis (RA)

(Zinneman, 1950; Rappaport et al., 1951; Kasliwal, 1970). Singh et

al. (1985) measured amoebic antibody levels in 41 Indian patients

with a primary diagnosis of RA, 35 age- and sex-matched healthy

volunteers, 162 hospital inpatients and 26 patients with other

arthritides. Amoebic antibodies were elevated in 39% of RA patients

and 0-11% of the various control groups. Only two patients with RA

had experienced recent diarrheal disease. These authors suggest that

an excessive and prolonged antibody response to Entamoeba histolytica

or other enteric organisms may contribute to joint inflammation in RA.

 

Galland (1989) described a patient with rheumatoid-like arthritis and

antinuclear antibodies whose arthritis went into rapid and complete

remission upon treatment of G. lamblia infection with metronidazole.

Relapse occurred when the patient acquired Entamoeba histolytica

during a trip to Egypt; remission occurred slowly following treatment

of amoebiasis. Diarrhea, polyarthritis and circulating antinuclear

antibodies developed in a United States serviceman heavily infested

with Endolimax nana, allegedly a non-pathogen (Burnstein and Liakos,

1983). Metronidazole rapidly reversed all abnormalities. The reported

cases of amoebic arthritis may represent a variant of parasitic

rheumatism, an inflammatory polyarthropathy produced by circulating

antigen-antibody complexes (Bocanegra, 1988). The presence of

autoantibodies, however, is not characteristic of parasitic

rheumatism, and suggests other mechanisms of immune dysfunction:

either a pre-existing disease is exacerbated by intercurrent amoebic

infection or amoebic infection itself provokes autoirnmunity, perhaps

mediated by the action of immune response genes (Singh et al., 1985).

 

CANDIDA ALBICANS

 

Candida species are part of the normal flora of the lower intestinal

tract of adult humans, being cultured from stool and rectal mucus of

23.2-82.4% of healthy subjects (Odds, 1988). Serious infection with

Candida albicans has increased dramatically over the past 40 years;

this increase is largely iatrogenic and may be attributed to

widespread use of antibiotics and immunosuppressive drugs (Seelig,

1966; Kirkpatrick, 1984). Candida albicans is an opportunist par

excellence and its ability to exploit pre-existing immune deficiency

in a host animal is well known, although the precise mechanisms

involved in the switch from commensalism to parasitism remain

uncertain (Odds, 1988). In contrast, little scientific attention has

been focused on the effect of Candida infection or colonization on

immune responses of the host.

 

That C. albicans is a potential allergen has been known for years.

Over 90% of a healthy adult population has delayed skin test

hypersensitivity (type IV) to antigenic extracts of C. albicans

(Dwyer, 1984). There are numerous reports of atopic diseases,

primarily asthma and allergic rhinitis, associated with type I

Candida hypersensitivity. Positive immediate hypersensitivity

reactions to intradermal or prick tests with C. albicans antigen are

more prevalent among asthmatics than among non-atopics (Itkin and

Dennis, 1966; Pepys et al., 1968; Kurimoto, 1975; Kabe et al., 1971).

One study found no difference (Gordon and Klaustermeyer, 1986) but

observed that `strong skin test reactivity' to Candida was associated

with atopy. El-Hefny (1968a) found Candida reactivity to vary

directly with severity of asthma. When challenged with inhaled

Candida antigen, asthmatics with immediate skin test hypersensitivity

develop acute bronchoconstriction (Itkin and Dennis, 1966; Pepys et

al., 1968; Kabe et al., 1971; Kurimoto, 1975; Edy and Pepys, 1980;

Akiyama et al., 1981). Pretreatment with inhaled cromolyn sodium

prevents experimental bronchoconstriction under these conditions

(Gordon and Klaustermeyer, 1986). Kurimoto (1975) concluded that type

I hypersensitivity, is involved in both the early and late phase

responses to Candida antigen but that late bronchial responses may

also involve type III hypersensitivity, as a transient drop in C3 and

C4 levels occurred.

 

Candida infections can induce an Arthus reaction in guinea pigs (Kabe

et al. 1971); Arthus-type reactivity to C. albicans was demonstrated

in 26% of asthmatics, being positively associated with severity and

duration of asthma (EI-Hefny, 1968b). Kurimoto (1975) frequently

provoked systemic reactions when administering Candida antigen by

inhalation to his Candida-allergic subjects and attributed this to

type III allergy. There are few published reports on the value of

hyposensitization with Candida extract in asthma treatment. El-Hefny

(1968a,b), who used an antigen she prepared herself, demonstrated in

a controlled study that Candida-sensitive asthmatics undergoing

hyposensitization with multiple antigens had a significantly better

outcome if C. albicans extract was included in the antigen mixture.

Other reports of improvement in asthma with Candida hyposensitization

are uncontrolled or anecdotal (Sclafer, 1957; Charpin, 1958; Kabe et

al., 1971; Gumowski et al., 1987).

 

Eczema and urticaria are also mentioned in the literature on Candida

allergy and are also reported to respond to immunotherapy (Sclafer,

1957; Charpin, 1958; Hold, 1966; Planes et al., 1972). James and

Warin (1971) found positive prick tests to C. albicans in 36 of 100

consecutive patients with chronic urticaria; they induced hives by

blind oral challenge with Candida extract in 25 of 33 patients.

Candida allergy was associated with immediate skin test reactivity to

inhalant molds and with positive responses to blind oral challenge

with Saccharomyces cerevisiae. Oral antifungal therapy with nystatin

tablets and amphotericin t:roches was combined with a yeast-free diet

in treatment of all Candida-allergic patients and 18 patients with

negative Candida prick tests. Clearing of urticaria occurred for 81%

of Candida prick-test-reactive patients and 39% of prick-test-

negative patients W < 0.01).

 

Gastrointestinal manifestations of Candida allergy have been reported

by Sclafer (1951), Liebeskind (1962), Holti (1966) and Alexander

(1975). Holti studied 65 patients with irritable bowel syndrome and

symptoms of explosive diarrhea and colicky abdominal pain; they had

been sick for an average of five years. All 56 patients with positive

skin wheals to C. albicans also had positive stool cultures for

yeasts. C. albicans was isolated from none of the nine patients with

a negative skin test to C. albicans and from 24% of a healthy control

group. Sixty-one per cent of Candida-allergic patients also reacted

to Saccharomyces cerevisiae. Treatment with oral nystatin was

associated with permanent disappearance of symptoms in 17 of 57

patients. Thirty-two additional patients were placed on yeast-free

diets and, within three days, nine were totally symptom-free and 14

were much improved. A double-blind controlled study of the effects of

administering C. albicans extract by mouth was conducted using five

patients with mucous colitis who had been free of symptoms for at

least four weeks. C. albicans extract, but not placebo, produced

diarrhea and borborygmi within 20 minutes in all five. In five

control subjects with positive Candida skin tests but no digestive

complaints, oral Candida extract produced no symptoms.

 

These studies are described in some detail because they indicate that

Candida allergy is not a rare disease with limited symptoms, as

maintained by some authorities (American Medical Association Council

on Scientific Affairs, 1987), but a relatively common disorder with

protean manifestations.

 

A relationship between Candida allergy and Candida infection is

suggested by clinical research in vaginitis. Candida allergy has been

well described in patients with chronic vaginitis (Tomsikova et al.,

1980). Mathur et al. (1977) found that total IgE was elevated in sera

and cervicovaginal secretions of women with recurrent Candida

vaginitis and that most of this IgE reacted with Candida antigens.

Witkin et al. (1988, 1989) found anti-Candida IgE in vaginal

secretions of 18.8% and 27.8% of women with chronic vaginitis.

Vaginal specimens with IgE antibodies also contained detectable

levels of prostaglandin E2 (PGE2), an important mediator of

inflammation. Witkin et al. (1986, 1988) suggest that production of

PGE2, stimulated by vaginal allergy to Candida and other substances,

inhibits lymphocyte responses to Candida in the vagina, permitting

Candida infection to flourish. They found that macrophages of women

with recurrent vaginal candidiasis inhibit response of control

lymphocytes to Candida antigen: this inhibition is reversed by PG-

synthesis inhibitors and by exogenous 11-2. This group has recently

shown (Witkin et al., 1989) that cervical infection with human

papilloma virus (HFV) is strongly correlated with the presence of

anti-Candida IgE; 47.4% of 19 women with HPV and only 5.9% of 17

women without HIPV were positive for anti-Candida IgE (p < 0.025).

Conversely, nine out of 10 women with anti-Candida IgE compared to 10

out of 26 women without anti-Candida IgE harbored HPV. They speculate

that the immunosuppressive effects of Candida allergy permit chronic

viral infection of the uterine cervix. In small, uncontrolled studies

Palacios (1976) and Rosedale and Browne (1979) had demonstrated

reduction in episodes of vaginal thrush by hyposensitizing injections

of C. albicans extract, suggesting clinical utility for Witkin's

findings.

 

There are several clinical case reports of immunosuppression

occurring in vivo as an apparent side effect of Candida infection

(Cuff et al., 1986). Paterson et al. (1971) described a 20-year-old

female patient with a 15-year history of chronic mucocutaneous

candidiasis (CMC) who was anergic and whose plasma contained a factor

capable of extinguishing the blastogenic response of normal

lymphocytes to Candida and mumps. Treatment with intravenous

amphotericin B cleared the Candida infection and simultaneously

eliminated the circulating plasma inhibitor, slowly restoring normal

cell-mediated immunity (CMD. The patient remained free of yeast

infection and immunologically normal for at least seven months after

discontinuation of amphotericin B. The authors speculated that the

circulating inhibitor was yeast-derived.

 

Circulating immunosuppressive factors have been described in other

cases of CMC (Valdimarsson et al., 1973). The immunosuppressive

factors are thought to be soluble polysaccharides, such as mannan,

contained in the yeast membrane and released into the circulation

(Fischer et al., 1978). Mannan at high dose inhibits mitogen- and

antigen-stimulated proliferation of human lymphocytes in vitro. Lower

concentrations specifically inhibit the lymphocyte blastogenic

response to C. albicans, probably by competing for polysaccharide-

antigen binding sites (Nelson et al., 1984). In that macrophages

normally remove mannan from the circulation, the immunosuppressive

effects of mannan in vivo probably depend upon defective macrophage

function, which may be a factor in some cases of CMC (Fischer et al.,

1982).

 

A Candida albicans cell wall glycoprotein rich in mannan causes

histamine release from rat mast cells in vitro (Nosal et al., 1974;

Svec, 1974). An experiment in mice suggested that some

immunosuppressive effects of the glycoprotein reside in the protein

moiety and are mannan-independent (Carrow and Domer, 1985). There is

a case reported of refractory esophagitis caused by C. tropicalis in

a 28-year-old nurse that was associated with cutaneous anergy and a

circulating inhibitor that was not mannan but a lowmolecular-weight

protein derived from the yeast itself (Lee et al., 1986).

 

Additional mechanisms of Candida-induced immunosuppression exist.

Mouse lymphocytes incubated with formalin-killed C. albicans induce a

suppressor B lymphocyte, the appearance of which may explain the

increased susceptibility of mice treated with Candida extracts to

infection by a number of micro-organisms (Cuff et al., 1986). Human T

lymphocytes incubated with C. albicans polysaccharide produce a

nonspecific inhibitor of macrophage function which decreases

macrophage production of 11-1 and hence lymphocyte production of 11-

2, inhibiting lymphocyte proliferation (Lombardi et al., 1985). On

the other hand, injection of heat-killed C. albicans augments natural

killer (NK) cell activity in mice (Marconi et al., 1985; Wojdani and

Ghoneum, 1987). Glucan, another yeast-derived polysaccharide, may

initiate this effect by stimulating macrophages to release TNF, which

raises NK levels (Reynolds et al., 1980). This effect is similar to

the immunoenhancing effect of bacterial endotoxin.

 

Zymosan, an insoluble yeast membrane polysaccharide, activates the

APC in vitro (Ray and Wuepper 1976); in vivo, the inflammation seen

in patients with CMC may be mediated by APC. Rosenberg and his

colleagues have proposed that psoriasis and Crohn's disease both

involve excessive and unregulated activation of the APC and state

that various microbial products, including zymosan from C. albicans,

may be stimulating the APC in vivo, causing the appearance of disease

in genetically susceptible individuals (Rosenberg et al., 1982,

1983). Having successfully treated scalp psoriasis with ketaconazole

(Rosenberg and Belew, 1982b), they proceeded to use oral nystatin for

treating psoriasis, with positive results (Crutcher et al., 1984).

They postulate that psoriasis is a systemic disease, which can be

triggered by Candida in the intestine, as well as by other infectious

agents.

 

Iwata et al. (1966) first described a high-molecular-weight protein,

isolated from a strain of C. albicans, with inflammatory and nerve-

growth-stimulating effects; they called it canditoxin. They have

since discovered several other substances, of low and high molecular

weight, which may serve as endo- or exotoxins by activating the

classical complement pathway Uwata, 1977a, 1977b). These substances

are not thought to be widely distributed, but rather confined to just

a few strains; their role in the pathogenesis of candidiasis is

unclear.

 

In summary, there are numerous and complex immunologic responses to

Candida constituents, both antigenic and non-antigenic, which may

follow colonization or infection; these cause release of inflammatory

mediators and alterations in CNU. Local infections may have systemic

effects.

 

CMC is known to occur in association with endocrine dysfunction and

circulating autoantibodies (Wuepper and Fudenberg, 1967). Although

chronic infection and autoimmune disease may result from defective T

lymphocyte function, some workers have speculated that polyclonal B

cell activation induced by Candida components may trigger

autoantibody formation (Zouali et al., 1983/1984). Mathur et al.

(1980) studied 40 women with chronic vaginal candidiasis (CVC). Anti-

ovarian and anti-thymocyte antibodies were present at a titer of 1 :

64 or greater in the sera of 27 and 19 patients, respectively. Both

autoantibodies were found at significant titers in 16. Mean

autoantibody levels of CVC patients were much higher than those of

controls. Autoantibody levels were strongly and positively associated

with Candida antibody levels. Absorption of sera from CVC patients

with thymocytes, ovarian follicles or C. albicans, significantly

lowered antibody titers to all three, suggesting antigenic cross-

reactivity. The authors speculated that a high level of multispecific

Candida antibody produced by chronic yeast infection in these

patients cross-reacted with ovarian and T cell antigens, producing

autoimmune phenomena.

 

In 1978, Truss first published six case reports of patients with

baffling neurological, psychiatric and inflammatory diseases who were

cured following treatment with oral nystatin and injections of

Candida antigen. A series of papers followed in which he d,-schbed

several cases of multiple sclerosis and one of Crohn's disease which

responded to the same therapy (Truss, 1981); he presented a

comprehensive treatment program for this condition which included

avoidance of dietary carbohydrates, limitation of exposure to food-

borne or air-bome fungi, administration of oral antifungal drugs and

immunotherapy with Candida extract (Truss, 1980a, K He initially

proposed that chronic Candida antigenemia may be responsibl~ for this

polymorphic illness (Truss, 1981), and later published a theory that

acetaldehyde production by intestinal yeast was the cause of

metabolic, immunologic and neuroendocrine abnormalities seen in his

patients (Truss, 1984). Although direct evidence for Truss'

hypopthesis has not been forthcoming, investigations by others since

his last paper have demonstrated that C. albicans can produce ethanol

in infant food formulas (Bivin and Heinen, 1985) and in the human

stomach (Bode et al., 1984), and that acetaldehyde, the principal

metabolite of ethanol, by conjugating with mammalian protein, induces

formation of polyclonal antibodies to acetaldehyde-protein adducts

that may mediate tissue damage (Israel et al., 1986).

 

Truss's concept that systemic illness may be provoked by mucosal

Candida infection has itself provoked considerable controversy

(Crook, 1984; Turner, 1985; Blonz, 1986; American Academy of Allergy

and Immunology, 1986). Although Truss's original case reports

primarily described patients with autoimmune or neurological

diseases, most patients diagnosed with this Candida-related complex

(CRQ have symptoms of fatigue, depression, allergy, food intolerance,

and a variety of gastrointestinal, gynecologic and musculoskeletal

complaints which are generally regarded as 'functional' (Truss, 1982;

Zwerling et al., 1984; Crook, 1986; Kroker, 1987; Mabray, 1988).

 

The first published controlled treatment study to test the CRC

hypothesis was done by Schinfeld (1987). He studied 30 patients with

premenstrual syndrome (PMS) that had not responded to treatment which

included psychotherapy, high-dose vitamin B6 and, in some cases,

psychotropic drugs. Twelve patients had no evidence of vaginal yeast

infection, symptomatic or asymptomatic, and the remainder suffered

from recent Candida, vaginitis. Both Candida-free and Candida-

infected groups were subdivided into active treatment (oral nystatin

and yeast-free diet) and contact-only treatment groups. All patients

remained symptomatic, with premenstrual depression as their major

complaint, but there were significant differences between treatment

and non-treatment groups. The worst outcome occurred in Candida-

infected patients who received no active treatment, and the best

outcome occurred in Candida-infected patients treated with oral

nystatin and a yeast-free diet. The small number of patients limits

the number of comparisons that can be made and the interpretation of

data.

 

Another study (Dismukes et al, 1990) reported no significant

difference in the effect of oral nystatin or placebo on systemic

systems of 42 women with chronic vaginitis. In contrast to

Schinfield's study, however, these authors lumped together women with

Candida vaginitis and women with culturenegative vaginitis, possibly

explaining the high inter-patient variability they observed. They

also used a cross-over treatment design with no washout period and

failed to control for changes in diet and other self-care activities,

although all subjects were aware of Truss' treatment regimen, having

been selected from his waiting list. Despite these flaws, and the

authors' failure to provide any raw data in their report, a careful

reading of this highly publicized study indicates some differences in

the effect of oral nystatin and placebo that do not support their

main conclusion. Oral nystatin significantly reduced somatization

scores Nistress arising from perception of bodily functions') (p =

0.04). Analysis of other systemic symptoms showed benefits from oral

nystatin under conditions where placebo responses appeared to have

stabilized. Truss (personal communication) has performed a separate

statistical analysis and has prepared a critique of this study which

demonstrates a strong effect of nystatin on generalized symptoms,

when compared to placebo.

 

The largest study of CRC reported is that of Jessup, who presented a

retrospective analysis of 1100 patients to the First International

Conference on Chronic Fatigue Syndrome, San Francisco, California, 15

April 1989. These patients had been treated over a nine-year period

for fatigue, myalgia, headache, dizziness, depression, arthralgias,

night sweats, morning stiffness and post-strain malaise. Posterior

cervical adenopathy occurred in 35% and neurological examination was

abnormal (serial sevens, tandem gait) in 30%. Although 80% reported

the sudden onset of their disease following an acute flue-like

illness, pre-morbid characteristics of the group revealed a high

frequency of chronic or recurrent health problems. About 80% had

repeated antibiotic exposures for acne or respiratory or urinary

tract infection; 60% of these had developed sensitivity to

antibiotics. Alcohol intolerance, irritable bowel syndrome, recurrent

vaginitis, migraine headaches, urticaria and premenstrual tension

were very frequently encountered. Almost all patients had experienced

addiction to sugar or alcohol prior to the onset of chronic fatigue.

Patients treated between 1980 and 1987 showed little improvement. In

September, 1987, 685 patients were unemployed and receiving

disability payments. At that point Jessup began treating these

patients with ketaconazole 200 mg a day, combined with a diet free of

alcohol, added sugar, fruit or fruit juice. The average length of

treatment was five months (range three to 12 months). By April, 1989,

& 4% of these patients had recovered and only 12 patients remained on

disability. Jessup concluded that Candida infection was the major

cause of disease for those patients who responded to ketaconozole and

speculated that intestinal colonization with yeast produced a

systemic toxin.

 

The treatment results do not necessarily support Jessup's conclusion.

Although ketaconazole is an effective antifungal agent, it has

powerful and complex effects on the function of lymphocytes (Schutt

et al., 1987a) and monocytes (Claus et al., 1988) which are

consistent with inhibition of certain cytokine effects (Schutt et

al., 1987b, 1988). Ketaconazole eliminates the spontaneous lymphocyte

proliferation of patients with dermatophytoses (Schutt et al., 1988).

Elevated levels of 11-2 (Cheney et al., 1989), increased production

of alpha-interferon (Lever et al., 1988) and spontaneous lymphocyte

blastogenic activity (Olson et al., 1986) have been described in some

patients with CFS. It is possible that Jessup's success resulted from

alteration of aberrant immune responses by ketaconazole rather than

an antifungal effect. The large number of patients treated and the

extraordinarily good outcome in this report, however, mandate a

prospective controlled study of imidazole therapy for CFS.

 

SUMMARY AND CONCLUSIONS

 

This chapter has reviewed the systemic effects of immune responses to

organisms inhabiting the gut lumen and adhering to the intestinal

mucosal surface. Type I and type III allergic responses to yeasts and

protozoa occur: the former may precipitate asthma, urticaria,

vaginitis and irritable bowel syndrome; the latter may cause

arthritis. Type I allergic reactions may also produce local

immunosuppression. Cross-reactivity between human and microbial

antigens occurs for several strains of Enterobacteriaciae and for C.

albicans. Enterobacteriaceae are implicated in the pathogenesis of

some spondyloarthropathies; C. albicans and Entamoeba histolytica

infections may contribute to autoimmune phenomena. Nonspecific

activation of immune responses by microbial components is common and

may be necessary for normal maturation of the immune system.

Bacterial enclotoxin, yeast zymosan and protozoan lectins express non-

antigenic immune stimulatory activity, which may be undesirable,

eliciting inflammatory reactions, such as psoriasis, in susceptible

individuals, or inducing replication of lymphotrophic viruses. Yeast-

derived glyoproteins can cause immune suppression in vitro and in

vivo. Chronic G. lamblia infection can produce fatigue, myalgia,

asthenia and malnutrition without serious gastrointestinal symptoms.

 

Our results demonstrate that some patients with CFS have Giardia

lamblia infection as their primary, and unexpected, diagnosis. The

excellent response to ketoconazole reported by Jessup suggests that

fungal infection may play an important etiologic role in many CFS

patients. Mowbray's group has found chronic enteroviral antigenernia

in the majority of patients with post-viral fatigue syndrome (Yousef

et al., 1988). It is possible that bacterial clysbiosis, enteric

protozoan or yeast infection, or intestinal allergy may alter normal

immune responses of the gut, allowing persistence of viral

replication. The probable importance of enteric factors in the

pathogenesis of ME should guide diagnostic and treatment strategies.