Modulation of immune function by dietary lectins in rheumatoid arthritis
Loren Cordain*, L. Toohey, M. J. Smith and M. S. Hickey
Department of Health and Exercise Science, Colorado State University, Fort Collins, CO 80523, USA
(Received 2 March 1999 – Revised 5 July 1999 – Accepted 30 August 1999)
Despite the almost universal clinical observation that inﬂammation of the gut is frequently
associated with inﬂammation of the joints and vice versa, the nature of this relationship remains
elusive. In the present review, we provide evidence for how the interaction of dietary lectins with
enterocytes and lymphocytes may facilitate the translocation of both dietary and gut-derived
pathogenic antigens to peripheral tissues, which in turn causes persistent peripheral antigenic
stimulation. In genetically susceptible individuals, this antigenic stimulation may ultimately
result in the expression of overt rheumatoid arthritis (RA) via molecular mimicry, a process
whereby foreign peptides, similar in structure to endogenous peptides, may cause antibodies or T-
lymphocytes to cross-react with both foreign and endogenous peptides and thereby break
immunological tolerance. By eliminating dietary elements, particularly lectins, which adversely
inﬂuence both enterocyte and lymphocyte structure and function, it is proposed that the
peripheral antigenic stimulus (both pathogenic and dietary) will be reduced and thereby result
in a diminution of disease symptoms in certain patients with RA.
Rheumatoid arthritis: Lectins: Immune function
Rheumatoid arthritis (RA) is a complex autoimmune dis-
ease characterized by persistent inﬂammation of the syno-
vium, local destruction of bone and cartilage and a variety of
systemic manifestations which may ultimately result in
functional disability. Three major aspects of RA suggest a
fundamental autoimmune-mediated disease: (1) the pre-
sence of often massive lymphocytic inﬁltrates and activated
CD4+T-lymphocytes within the inﬂamed synovium;
(2) production of large amounts of rheumatoid factor (RF)
by B-lymphocytes and plasma cells in the synovium, and
(3) the observation that immuno-suppression inﬂuences the
course of RA (Williams, 1996). RA affects approximately 1%
of the adult population with females being two to four times
more susceptible than males (Grossman & Brahn, 1997).
Autoimmune diseases occur when the body loses the
ability to discriminate self proteins from non-self proteins.
This loss of tolerance ultimately results in the destruction
of self tissues by the immune system. Typically, auto-
immune diseases are characterized by the presence of
auto-antibodies and autoreactive T-lymphocytes acting
against speciﬁc self proteins (Dalton & Bennett, 1992).
Most autoimmune diseases are thought to develop via the
interaction of an environmental factor or factors with a
speciﬁc hereditary component. The genetic components
most closely associated with the expression of autoimmune
diseases are those genes which code for the human leuco-
cyte antigens (HLA). The mechanism (or mechanisms) by
which autoimmunity is manifested in genetically suscepti-
ble individualsvia environmental factors is not clearly deﬁned,
however, it is increasingly being recognized that the process of
molecular mimicry (Fig. 1), by which a speciﬁc foreign antigen
may induce an immune cross-reaction with self antigens, may
be involved in a variety of autoimmune diseases (Oldstone,
1987; von Herrath & Oldstone, 1995) including RA (Albani
& Carson, 1996; Baum et al. 1996; Wilson et al. 1997).
In the present review, we propose that the interaction of
dietary lectins with enterocytes and lymphocytes facilitates
the translocation of both dietary and gut-derived bacterial
antigens to peripheral tissues, which in turn causes persis-
tent peripheral antigenic stimulation. In genetically suscep-
tible individuals, this antigenic stimulation may ultimately
result in the expression of overt RA via molecular mimicry
in which foreign peptides are similar enough in structure to
cause T-lymphocytes and antibodies to cross-react with endo-
genous peptides and break immunological tolerance. By elim-
inating dietary elements, including lectins, which adversely
British Journal of Nutrition (2000), 83, 207–217 207
Abbreviations: BSA, bovine serum albumin; HLA, human leucocyte antigens; Ig, immunoglobulin; PHA, phytohaemagglutinin; RA, rheumatoid arthritis;
RF, rheumatoid factor.
*Corresponding author: Dr Loren Cordain, fax +1 970 491 0445, email firstname.lastname@example.org
inﬂuence both enterocyte and lymphocyte structure and func-
tion, it is proposed that the peripheral antigenic stimulus will be
reduced and thereby result in a diminution of disease symptoms.
Relationship of gut inﬂammation to rheumatoid arthritis
There is a strong relationship between gut inﬂammation and
joint inﬂammation which has been recognized for decades
in clinical practice and which has been substantiated in a
variety of animal models of arthritis. Approximately 20 %
of all patients with inﬂammatory bowel disease (Crohn’s
disease, ulcerative colitis) are complicated by joint inﬂam-
mation (Hazenberg et al. 1992). Conversely, occult intestinal
inﬂammation, which may be related to non-steroidal anti-
inﬂammatory drug therapy or may be disease associated, occurs
in approximately 67 % of patients with RA (Sartor, 1989).
Because gut inﬂammation is known to increase intestinal
permeability, it has been suggested that an increased uptake
of luminal bacterial components across the inﬂamed mucosa
leads to a systemic distribution of bacterially derived,
arthropathic products (Sartor et al. 1996). In support of
this notion is the observation that HLA-B27 transgenic rats
develop arthritis when raised in conventional environments,
but do not do so under germ-free conditions (Taurog et al.
1994; Rath et al. 1996). Further, treatment with metronida-
zole (an antibiotic preferentially acting on anaerobic bac-
teria) attenuates gastrointestinal inﬂammation and can
prevent reactivation of arthritis in animal models (Sartor
et al. 1996). It has been consistently shown that a single
intraperitoneal injection of cell-wall fragments of Eubacterium
aerofaciens, a main resident of the human intestinal ﬂora,
can elicit both acute and chronic arthritis in a rat model
(Hazenberg et al. 1992; Kool et al. 1992). The main
constituents of the Eubacterium aerofaciens cell wall are
peptidoglycan–polysaccharide complexes, of which a
65kDa heat-shock protein is the smallest bioactive unit
(Klasen et al. 1994). These data point to the critical role
that gut-derived bacteria may play in eliciting RA.
In human subjects, as previously mentioned, intestinal
inﬂammation frequently accompanies RA (Sartor, 1989).
Further, there is considerable evidence to suggest that
intestinal permeability may be increased in patients with
RA (Katz & Hollander, 1989; Mielants, 1990), particularly
when joint disease is active (Smith et al. 1985). Patients with
RA have also been shown to maintain a high frequency of
small-intestinal bacterial overgrowth (Henriksson et al. 1993),
particularly with anaerobic bacterial species (Benno et al.
1994; Eerola et al. 1994). Although the mechanism of action
is not entirely clear, there is convincing evidence to show that
antibiotic therapy has anti-rheumatic activity in many patients
with RA (Trentham & Dynesius-Trentham, 1995; O’dell et al.
1997). As in animal models, the human data are suggestive
that increased uptake of luminal bacterial components across
the inﬂamed mucosa leads to a systemic distribution of
bacterially derived, arthropathic products.
Translocation of intestinal antigens to the periphery
Common mucosal pathogens, particularly those found in the
gut, may play an important role in the aetiology of RA by
virtue of their ability to initiate an autoimmune response via
interaction with the immune system. Clearly implicit in this
model is the ability of intraepithelial pathogens and intact
proteins to escape enzymic digestion and to cross the
gastrointestinal barrier and enter peripheral circulation.
From a functional perspective, in healthy subjects the
contents of the gut lumen lie outside the body and contain a
toxic or antigenic load from which the body needs to be
protected. Protection is supplied by a number of mechan-
isms including: the intestinal mucosa, intestinal secretions
(primarily mucus and secretory immunoglobulin (Ig)A),
and intramural lymphocytes (Crissinger et al. 1990). The
primary intestinal barrier is supported by the liver, through
which all enterically derived substances must pass before
entering the peripheral circulation. Under normal circum-
stances in healthy subjects, the intestinal immune apparatus
208 L. Cordain et al.
amino acid sequence
of foreign antigen
amino acid sequence
of self antigen
Fig. 1. A schematic representation of simple two-way molecular mimicry in whichforeign
and self antigens may trigger T-cell cross-reactivity because of a shared sequence of
mounts rapid and potent effector responses to prevent
invasion by pathogenic viruses and bacteria (Mowat,
1987). Further, healthy, functionally intact epithelial
mucosa cells normally do not allow passage of more than
small amounts (approximately 2 %) of intact dietary pro-
teins (Mowat, 1987; Travie & Menzies, 1992). However,
translocation of viable bacteria from the gastrointestinal
tract to extra-intestinal sites (mesenteric lymph nodes, liver,
spleen, kidney and blood) has been shown to occur under
three circumstances: (1) disruption of ecological equili-
brium which allows intestinal bacterial overgrowth, (2)
deﬁciencies in host immune defences, and (3) increased
permeability of the intestinal barrier (Berg, 1992).
Undegraded dietary peptides have also been shown to enter
the peripheral circulation (Hurby et al. 1985), particularly
when intestinal permeability is increased by disease (Travis
& Menzies, 1992), non-steroidal anti-inﬂammatory drugs
(Travis & Menzies, 1992; Bjarnason & Peters, 1996), ethanol
(Bjarnason et al. 1984; Keshavarzian et al. 1994), acetic acid
(Fabia et al. 1993) and dietary lectins derived from legumes
and cereal grains (Liener, 1986; Pusztai, 1993).
Common dietary staples such as cereal grains and legumes
contain glycoproteins called lectins which have potent anti-
nutritional properties (Table 1) which inﬂuence the structure
and function of both enterocytes and lymphocytes (Liener,
1986; Pusztai, 1993). Wheat-germ agglutinin derived from
dietary wheat products is heat stable and resistant to
digestive proteolytic breakdown in both rats (Pusztai et al.
1993a) and human subjects (Brady et al. 1978) and has been
recovered intact and biologically active in human faeces
(Brady et al. 1978). Wheat-germ agglutinin and lectins in
general bind surface glycans on gut brush-border epithelial
cells causing damage to the base of the villi which includes
disarrangement of the cytoskeleton, increased endocytosis
and shortening of the microvilli (Liener, 1986; Sjolander
et al. 1986; Pusztai, 1993). The structural changes induced
by wheat-germ agglutinin on intestinal epithelial cells elicit
functional changes including increased permeability (Sjo-
lander et al. 1984) which may facilitate the passage of
undegraded dietary antigens into systemic circulation (Pusz-
tai, 1993). High-wheat-gluten diets have been shown to
induce jejunal mucosal architectural changes in normal
subjects (Doherty & Barry, 1981). In rats dietary wheat-
germ agglutinin is rapidly transported across the intestinal
wall into the systemic circulation where it is deposited in
blood and lymphatic vessel walls (Pusztai, 1993a).
Under normal circumstances, when the luminal concen-
tration of intact dietary proteins is low, absorbed proteins
generally elicit a minimal allergic response because of the
limiting inﬂuence of T-suppressor cells. Because of their
resistance to digestive proteolytic breakdown, the luminal
concentrations of lectins can be quite high, consequently
their transport through the gut wall can exceed that of other
dietary antigens by several orders of magnitude (Pusztai,
1989a), and absorbed dietary lectins can be presented by
macrophages to competent lymphocytes of the immune
system (Hurby et al. 1985; Pusztai, 1989a).
Not only do dietary lectins increase gut permeability
(Sjolander et al. 1984; Greer & Pusztai, 1985) thereby
allowing increased passage of dietary and gut-derived
bacterial antigens into the periphery (Liener, 1986; Pusztai,
1993), they may also cause a bacterial overgrowth which
facilitates the preferential growth of gut bacteria such as
Escherichia coli and Lactobacillus lactis (Banwell et al. 1988)
which are associated with the expression of RA because they
contain an amino acid sequence (Q(K/R)RAA) which is also
found in the gene products of the HLA system of a high
percentage of patients with RA (Auger & Roudier, 1997).
Phytohaemagglutinin (PHA), a dietary lectin derived from
kidney beans, causes accelerated enterocyte cell turnover
which leads to an increase in the proportion of juvenile cells
209Rheumatoid arthritis and diet
Table 1. A non-comprehensive list of edible plants containing lectins and the physico-chemical properties of their puriﬁed lectins (modiﬁed from
Genus and species Common name Peritoneal* Oral† mass (Da) Sugar speciﬁcity
Peanut, groundnut ? ? 110000 Galactose
Jack bean ++110000 Mannose, glucose
Horse gram −+113000
Hyacinth bean ++n.a. n.a.
Soyabean +−122000 Galactose,
Barley ? +40000
Lentil −? 52000 Mannose, glucose
Winged bean ? +120000 Fucose
Rice ? ? 10000
Mung bean −? n.a. n.a.
Scarlet runner bean ? ? 120000
Lima bean −+124000
Kidney bean ++120000
Garden pea, split pea −−53000 Mannose, glucose
Castor bean ++60000 Galactose,
Potato ? ? 46000 Diacetylchitobiose
Wheat ? +36000
Horse bean, broad bean ? ? 50000 Mannose, glucosamine
n.a., not available.
*Death results from the intraperitoneal injection of crude seed extract or puriﬁed lectin.
†Growth inhibition caused by adding puriﬁed lectin to the diet of experimental animals.
(containing high levels of membrane polymannosylated
receptor glycans) on the small-intestinal villi (Pusztai,
1993). Because Escherichia coli preferentially binds to man-
nose receptors on enterocytes, the increase in the number of
mannose receptors via increased juvenile enterocytes allows
Escherichia coli bacteria to successfully out-compete other
resident gut micro-organisms. It is likely that PHA and other
dietary lectins operate in the same manner to induce similar
changes in other gut ﬂoral species (Pusztai et al. 1991).
Legume and cereal lectins alter the microﬂora of the gut
(Liener, 1986; Banwell et al. 1988; Pusztai et al. 1993b),
causing both inﬂammation (Wilson et al. 1980; Liener,
1986; Pusztai et al. 1993b) and increased intestinal perme-
ability (Greer et al. 1985) which in turn facilitates the
translocation of gut pathogens to the periphery. Kidney-
bean lectin (PHA) is lethally toxic for conventional rats
when given in high doses (Wilson et al. 1980), but is non-
toxic for germ-free animals (Rattray et al. 1974). Thus,
PHA’s toxic effects could be directly attributed to its ability
to increase the translocation of gut-derived bacteria to the
periphery. In the case of RA, dietary lectins may operate in a
similar manner to indirectly increase the expression of the
disease by facilitating movement of bacterial antigens with
arthrogenic properties to the periphery.
Because dietary lectins are able to cross the gastrointestinal
barrier rapidly and enter the circulation intact (Pusztai et al.
1989), they may also be able to interact directly with synovial
tissues. Although not a characteristic model of RA with all of
its symptoms, a rabbit model of arthritis has shown that the
direct injection of legume-derived dietary lectins into the knee
joint induces the development of severe arthritis.Speciﬁcally,
single injections of Lens culinaris lectin (derived from lentils),
Pisum sativum lectin (derived from peas), or the lectin
concanavalin A derived from the jack bean (Canavalia
ensiformis) were able to induce severe arthritis characterized
by an ampliﬁcation of the initial inﬂammatory response due to
T-lymphocyte stimulation (Brauer et al. 1985).
Other legume-derived lectins such as soyabean aggluti-
nin, concanavalin A, and lectins derived from other
Phaseolus (bean) species have been demonstrated to inﬂu-
ence intestinal structure and function negatively (Liener,
1986), as have lectins derived from groundnuts (Ryder et al.
1992). Legumes are almost always consumed in the cooked
state, and it is often assumed that cooking eliminates lectin
activity. However, Grant et al. (1982) have demonstrated
that residual lectin activity is present in kidney beans
(Phaseolus vulgaris) even when cooked at 858for 6 h or
at 908for 3h. Lectin activity has been demonstrated in
wheat, rye, barley, oats, maize (Liener, 1986) and rice
(Tsuda, 1979). Maize, like wheat, can alter intestinal
epithelial structure and function (Mehta et al. 1972). The
biological activities of cereal lectins are similar because
they are closely related to one another both structurally and
immunologically (Peumans & Cammue, 1986).
Clinical and experimental evidence implicating diet in
rheumatoid arthritis aetiology
The control of RA by dietary manipulation has been
infrequently tested and has not always yielded convincing
results (Shatin, 1964; Ziff, 1983; Darlington et al. 1986;
Buchanan et al. 1991), probably because the gastrointestinal
tract may not play a pathogenic role in all cases of RA and
because most previous clinical trials have only controlled for
single, rather than multiple dietary elements which may
simultaneously inﬂuence disease expression and progression.
Van de Laar & van der Korst (1992) demonstrated
symptomatic improvement in a subset of patients with RA
who were seropositive for RF when they were placed on
elemental diets (protein-free diets consisting of essential
amino acids, glucose, trace elements and vitamins). Twice
as many of the food-sensitive patients showed improvement
during a milk-free leg of the trial, and all food-sensitive
patients showed marked disease exacerbation during food
re-challenge. The authors concluded: ‘The existence of a
subgroup of patients in whom food intolerance inﬂuences
the activity of rheumatoid factor seropositive rheumatoid
arthritis deserves serious consideration’. In support of this
conclusion is a more recent experiment which showed a
signiﬁcant (P=0⋅04) improvement in the number of sore
joints in a group (n10) of patients with RA who followed an
elemental diet for 3 weeks (Haugen et al. 1994). Further, in
the only controlled study of elemental diets in the treatment
of RA, patients experienced improvements in grip strength
(P=0⋅008) and Ritchie score (P=0⋅006) that relapsed
following food re-introduction (Kavanaghi et al. 1995). In
Crohn’s disease approximately 20 % of the patients experi-
ence joint inﬂammation together with gut inﬂammation
(Hazenberg et al. 1992). Elemental diets have been shown
to be as effective as corticosteroids in treating the disease,
and most subjects (84 %) achieve disease remission with
elemental diets (Riordan et al. 1993). The most frequent
food intolerances were to cereals, dairy products and yeast
(Riordan et al. 1993).
In coeliac disease there is a characteristic T-cell-mediated
destruction of the intestinal villi which results in malabsorp-
tion and increased intestinal permeability (Hamilton et al.
1982). All symptoms of coeliac disease are eliminated
following removal of gluten-containing cereals (wheat,
rye, barley and oats). RA has frequently been demonstrated
to occur concurrently with coeliac disease (Collins & Maki,
1994; Lepore et al. 1996). Multiple studies of arthritic
patients have demonstrated elevated antibody levels for
gliadin (O’Farrelly et al. 1988; Lepore et al. 1993), and
gluten-free diets have been shown to be effective in redu-
cing arthritic symptoms in coeliac patients (Bourne et al.
1985; Charkravarty & Scott, 1992; Lepore et al. 1993).
These studies support the concept that wheat-containing
diets can increase intestinal permeability and thereby allow
gut-derived antigens access to the periphery. Because
removal of gluten-containing grains not only eliminates
coeliac disease, but also symptoms of arthritis, such diets
may be of beneﬁt for some patients with RA. No large clinical
trials have been undertaken speciﬁcally to examine the effec-
tiveness of gluten-free diets in the treatment of RA, however,
there are numerous case studies reporting alleviation of RA
symptoms with grain-free diets (Shatin, 1964; Williams,
1981; Beri et al. 1988; Lunardi et al. 1988). Additionally,
complete withdrawal of food during fasting reduces objective
and subjective indices of the disease (Kjeldsen-Kragh et al.
1991). Collectively, these studies suggest that modulation of
intestinal physiology by dietary substances may allow both
210 L. Cordain et al.
dietary and pathogenic antigens access to the periphery,
thereby causing persistent immune system stimulation.
Milk and dairy products have frequently been implicated
in the aetiology of RA. O’Farrelly et al. (1989) demon-
strated that ﬁfty-three of ninety-three patients with RA had
elevated circulating IgG antibodies to milk, wheat or both
dietary proteins. Bovine serum albumin (BSA), a milk
protein, contains an amino acid sequence homologous
with human collagen type I, C1q, and sera from RA patients
displayed reactivity to synthetic peptides containing the
BSA residues responsible for the homology (Perez-
Maceda et al. 1991). Additionally, exogenous BSA peptides
have been found to be bound to RA HLA-DR susceptibility
alleles (Chicz et al. 1993). Case studies have shown that
elimination of milk and dairy products from the diets of
patients with RA improved symptoms, and the disease was
markedly exacerbated on re-challenge (Parke & Hughes,
1981; Panush et al. 1986). No large-scale controlled trial
testing the effect of dairy products on RA development and
progression has been undertaken. In animal models of RA,
disease symptoms are routinely induced in dogs (Ohashi et
al. 1996), rats (Grifﬁths, 1992) and rabbits (Thomsen et al.
1985) by injecting the synovium with BSA. Further, milk
drinking is known to induce rheumatoid-like joint lesions in
rabbits drinking cows’ milk (Welsh et al. 1985).
Immunological and molecular mechanisms of
Inherited susceptibility to RA is associated with the genes
found on the short arm of chromosome 6 which code for
the HLA system. On chromosome 6, the HLA system is
sub-divided into class I (HLA-A, HLA-B, HLA-C) and class
II segments (HLA-DR, HLA-DQ, HL-DP). Within the class
II segment, the HLA-DRB1 genes, which encode the HLA-
DR4 and HLA-DR1 molecules, convey enhanced sus-
ceptibility to RA. The speciﬁc function of HLA molecules
is to bind internally processed antigens (both exogenous
and endogenous in nature) and to present them to
T-lymphocytes. Thus, the relationship between HLA sus-
ceptibility haplotypes and RA indicates an antigen-driven
response (Weyland & Goronzy, 1997). More speciﬁcally, it
has been observed that most (76%) of the HLA alleles
associated with RA contain, in the third hypervariable
region of their bchains, an amino acid sequence composed
of the amino acid motif Q(K/R)RAA (glutamine-(lysine/
arginine)-arginine-alanine-alanine) (Rowley et al. 1997).
Not only does the Q(K/R)RAA motif increase the prob-
ability of RA, it increases the severity of destruction (Larsen
score.1⋅62) and the likelihood of developing early erosive
disease (Wagner et al. 1997). How the Q(K/R)RAA motif
increases RA susceptibility or severity is still controversial.
However, a multistep molecular mimicry model (Fig. 2) has
been proposed where the Q(K/R)RAA sequence is an
antigenic epitope exhibited on several gut microbial pro-
teins (Escherichia coli,Lactobacillus lactis,Brucella ovis,
Proteus mirabilis), and patients with RA respond more
strongly to these antigens than healthy subjects (Albani
& Carson, 1996; La Cava et al. 1997; Auger & Roudier,
HLA molecules themselves are frequently processed and
presented by antigen-presenting cells and these HLA-
derived peptides sometimes represent the majority of
211Rheumatoid arthritis and diet
Homologous amino acid motif
Fig. 2. Rheumatoid arthritis may arise from three-way molecular mimicry in which the Q(K/
R)RAA (glutamine-lysine/arginine-arginine-alanine-alanine)amino acid susceptibility motif
is shared by gut bacterial proteins, by self human leucocyte antigen (HLA) proteins and by
a putative tissue auto-antigen. Bacterial peptides containing the Q(K/R)RAA sequence
and entering the periphery (facilitated by dietary lectins) may stimulate T-cells causing
them to cross-react with the putative auto-antigen.
peptides presented by the antigen-presenting cells (Cao et al.
1995). It has been suggested that thymically selected T-cells
with weak afﬁnity for self HLA peptides may subsequently
be stimulated by peripheral exposure to microbial peptides
which mimic the HLA amino acid sequence (Baum & Staines,
1997). Speciﬁcally, the T-cell repertoire which is positively
selected during embryonic development, with weak afﬁnity
for the Q(K/R)RAA motif expressed on HLA-DR molecules,
may subsequently be stimulated by peripheral exposure to
Escherichia coli,Lactobacillus lactis,Brucella ovis and
Proteus mirabilis containing the mimicking epitope. The
Q(K/R)RAA motif, whether expressed on HLA-DR mole-
cules, synthetic peptides, bacterial or viral proteins represents
a strong epitope for B-lymphocytesresponsible for serological
cross-reactivity among Q(K/R)RAA-containing peptides
(Roudier et al. 1989; Albani et al. 1992).
The Q(K/R)RAA motif is also the basis for T-lympho-
cytes involved in the positive and negative selection of the
T-cell repertoire in the thymus during embryonic develop-
ment. The T-lymphocyte repertoire in most cases is entirely
deﬁned before exposure to environmental pathogens (Bevan
et al. 1994), and immature T-lymphocytes are positively
selected in the thymus by virtue of their low-afﬁnity binding
with HLA peptides presented by cells of the thymic epithe-
lium. These positively selected T-lymphocytes undergo
maturation and become part of the pool of mature, naive
T-lymphocytes that, after birth, are recruited in speciﬁc
immune responses (Jameson et al. 1995). Immature T-
lymphocytes in the thymus that bind with high afﬁnity to
HLA self peptide complexes have a high self-reactive
potential and are eliminated (negative selection) (Nossal,
1994). Therefore, thymically selected T-lymphocytes which
have been positively selected by virtue of their low afﬁnity
interactions with HLA-DR1B alleles containing the Q(K/
R)RAA motif, may be later triggered in the periphery on
exposure to foreign peptides with homologous amino acid
sequences (Albani & Carson, 1996).
Approximately 70–80 % of patients with RA have RF
present in their blood and synovial ﬂuid (Grossman &
Brahn, 1997). RF is an autoantibody since it has speciﬁcity
for the Fc (C g3 and C g2 domains) receptor of IgG
(Williams, 1992). Recent work utilizing crystal structure
analysis of the RF–IgG Fc, antibody–antigen complex
suggests that RF may have another entirely different speci-
ﬁcity separate from IgG Fc and that the reactivity with IgG
Fc probably occurs because of similarities with an unin-
dentiﬁed antigen (Sutton et al. 1998). Consistent with this
notion is earlier work indicating that viral and bacterial
proteins also bind IgG Fc receptors via microbial receptors
which resemble the IgG Fc receptor (Nardella et al. 1985,
1988). Further, RF Ig genes show clear evidence of somatic
mutation, indicating that RF production by B-lymphocytes
is a T-lymphocyte-dependent, antigen-driven process (Gor-
onzy & Weyland, 1993). More recent studies utilizing
computer modelling and crystallographic studies suggest
that the mechanisms that operate on RF selection in RA
synovia are similar to immune responses to exogenous
antigens (Mageed et al. 1997). Collectively, these data
suggest that RF production may occur principally in
response to foreign proteins and secondarily in response to
There is some evidence to suggest that RF production
may be inﬂuenced by dietary proteins. O’Farrelly et al.
(1989) showed that ﬁfty-three of ninety-three patients with
RA had raised levels of IgG antibodies to milk and/or wheat
proteins. Of the ﬁfty-three patients positive for dietary
proteins, forty-eight (90 %) had raised levels of IgA RF
whereas only seven (17 %) of the remaining forty non-diet-
sensitive RA patients had detectable levels of IgA RF. These
data are suggestive of a breakdown in gastrointestinal
tolerance to dietary antigens in this group of patients, and
indicate that RF production may occur in response to
gastrointestinally related antigens.
Interaction of dietary lectins with immune function
It is apparent that dietary lectins from both cereal grains and
legumes increase the translocation of gut-derived bacterial
and dietary antigens to the periphery by: (1) causing an
intestinal bacterial overgrowth and (2) increasing intestinal
permeability. Additionally, dietary lectins have the ability to
interact with components of the immune system which may
facilitate the autoimmune process. Table 2 lists these
212 L. Cordain et al.
Table 2. Inﬂuence of dietary lectins on gastrointestinal and immunological function
1. Facilitate preferential growth of bacteria such as
and Liener, 1986; Banwell
. 1988; Pusztai
which contain the Q(K/R)RAA susceptibility motif
2. Bind surface glycans on gut brush-border epithelial cells causing Liener, 1986; Sjolander
. 1986; Pusztai, 1993
disarrangement of the cytoskeleton, increasedendocytosis and
shortening of the microvilli
3. Increase gut permeability allowing increased passage of both dietary and Sjolander
. 1984; Greer & Pusztai, 1985; Liener, 1986
bacterial antigens to the periphery
4. Amplify HLA class II expression in intestinal epithelial cell lines Weetman
5. Stimulate T-cell proliferation Uder
. 1980; Clevers
6. Stimulate IFN-g, causing HLA class II expression Piccinini
. 1987; Lowes
7. Cause abnormal expression of ICAM in T-cells Koch
. 1994; Shingu
8. Stimulate the production of inﬂammatory cytokines (IL-1, TNFa) Firestein
. 1990; van den Bourne
HLA, human leucocyte antigens, IFN-g, interferon-g; ICAM, intracellular adhesion molecules; IL-1, interleukin-1; TNFa, tumour necrosis factor a.
multiple effects, and Fig. 3 shows schematically how dietary
lectins may facilitate the expression of RA via their
interaction with the gut and the immune system.
Dietary lectins including both wheat-germ agglutinin
(Pusztai et al. 1993a) and PHA (Pusztai et al. 1989) have
been shown to cross the gastrointestinal barrier rapidly and
enter the peripheral circulation. In rats dosed with PHA, up
to 10% of the lectin is found in circulation 3h after feeding
(Pusztai et al. 1989). Thus, PHA is a powerful oral immuno-
gen which produces a high titre of IgG anti-PHA antibodies
in animals and probably man (Pusztai, 1993). Antibody
development to PHA becomes measurable 10 d after the ﬁrst
dose and further feeding or re-introduction results in
booster effects (Pusztai, 1993). Thus, the gut anti-lectin
IgA system is ineffective against PHA since it cannot
prevent its absorption after its re-introduction (Pusztai,
1989b). This abrogation of the gut IgA response to PHA
and possibly to other lectins has important consequences in
autoimmune disease because it would allow dietary lectins
continuous access to T-lymphocytes at the gut mucosal
Further, because dietary lectins increase intestinal perme-
ability, they would promote increased activity between the
immune system and gut bacteria. The interaction of the
systemic immune system with bacterial and dietary antigens
at the gut mucosal surface could lead to activation of
previously quiescent Q(K/R)RAA-speciﬁc T-lymphocytes
which react with the Q(K/R)RAA amino acid motif of gut
pathogens (Albani & Carson, 1996). Normally, these acti-
vated T-lymphocytes would return to the intestinal mucosa
and would not travel to peripheral sites, such as the
synovium, without abnormal expression of intracellular
adhesion molecules (Albani & Carson, 1996). Numerous
in vitro experiments have demonstrated that PHA is a potent
stimulator of intracellular adhesion molecule expression in
RA (Koch et al. 1994; Shingu et al. 1994). Thus, lectin-
induced intracellular adhesion molecule expression in auto-
reactive T-lymphocytes would allow them to travel to
peripheral sites in the joint and to persist in the synovial
Immunogenic foreign antigen fragments could be brought
to the inﬂammatory sites by synovial type A macrophages
and by B-cells recruited and activated by the inﬂammatory
stimuli (Albani & Carson, 1996). Therefore RF-producing
B-cells, because of their ability to bind and ingest antigens
trapped in immune complexes, would represent powerful
213Rheumatoid arthritis and diet
Immune system activation
HLA molecule loaded with putative auto-antigen
Target tissue (joint connective tissue)
Fig. 3. A diagrammatic illustration of how dietary lectins may hypothetically interactwith the gut and immune system to inﬂuence the expression of
rheumatoid arthritis. Dietary lectins may: (1) facilitatepreferential growth of bacteriasuch as
the Q(K/R)RAA susceptibility motif, (2) increase gut permeability allowing increased passage of both dietary and bacterial antigens to the
periphery, (3) amplify human leucocyte antigen (HLA) class II expression in intestinal epithelial cell lines, (4) stimulate T-cell proliferation,
(5) stimulate interferon-g, (6) cause abnormal expression of intracellular adhesion molecules in T-cells, and (7) stimulate the production of
inﬂammatory cytokines (interleukin-1,tumour necrosis factor a).
antigen-presenting cells, and their presence would partially
regulate the ampliﬁcation of the inﬂammatory process
(Albani & Carson, 1996).
In addition to maintaining elevated levels of intracellular
adhesion molecules, RA is also characterized by elevated
levels of the inﬂammatory cytokines, interleukin 1, and
tumour necrosis factor a(Odeh, 1997). Numerous in vitro
experiments show that PHA is a potent stimulator of both
cytokines in peripheral blood mononuclear cells (Firestein
et al. 1990; van den Bourne et al. 1997). To date no trials
have been conducted to determine if in vivo administration
of dietary lectins in human subjects is able to elicit similar
responses. However, such responses seem likely since intact
PHA is present in the peripheral circulation following its
ingestion (Pusztai et al. 1989).
Molecular mimicry of dietary antigens with self proteins
We have outlined the hypothesis that the homologous amino
acid motifs among bacterial antigens, the HLA-DRB1 genes
and putative auto-antigens in joint tissue may induce RA in
genetically susceptible individuals by virtue of immuno-
logical cross-reactivity in a three-way model of molecular
mimicry. In addition to bacterial antigens, viral antigens,
including the Epstein-Barr virus, may also induce cross-
reactivity in RA via three-way molecular mimicry (Albani
& Carson, 1996; Baum et al. 1996). Less well appreciated
are the homologous amino acid motifs which may occur
between dietary peptides and self and which have been
implicated in the aetiology of RA.
Perez-Maceda et al. (1991) showed that the sera from
patients with RA recognized BSA from cows’ milk and that
a sequence of BSA (residues 141–157) was highly homo-
logous with human collagen type I and the plasma com-
plement protein, C1q. In support of the immunogenicity of
this BSA fragment, sera from patients with RA displayed a
speciﬁc reactivity for a synthetic peptide containing the
BSA residues responsible for the homology (Perez-Maceda
et al. 1991). Chicz et al. (1993) have demonstrated that
exogenous BSA proteins can be bound to various HLA
alleles, including DR4 molecules, suggesting that BSA
proteins may have a signiﬁcant role in the development of
autoimmunity via molecular mimicry.
Ostenstad et al. (1995) have demonstrated that glycine-
rich cell-wall protein (GRP 1.8), which is a ubiquitous
storage protein found in virtually all cereal grains and
legumes, contains signiﬁcant amino acid homologies with
both ﬁbrillar collagen, procollagen and Epstein-Barr virus
nuclear antigen-1. A synthetic ﬁfteen amino acid sequence
derived from GRP 1.8 stimulated T-lymphocytes taken from
the synovial ﬂuid of patients with RA and caused a pre-
ferential expansion of synovial T-cells bearing V a2.1/V b
5.5 gene products, thereby indicating the involvement of
HLA complex-restricted auto-antigen recognition (Osten-
stad et al. 1995). The T-cell expansion caused by the GRP
1.8 analogue could be blocked by the addition of anti-DR
antibodies (Ostenstad et al. 1995), providing further evi-
dence of the potentiating role of molecular mimicry in the
aetiology of RA.
A third dietary antigen which may also induce RA via
molecular mimicry is the a-gliadin component of wheat
which shares signiﬁcant amino acid sequences with calre-
ticulin, an endoplasmic reticulin chaperone protein (Karska
et al. 1995). Anti-calreticulin antibodies have been found in
patients with RA (Routsias et al. 1993), and HLA-DR4
molecules from arthritic patients are known to present a
peptide fragment derived from calreticulin (Verreck et al.
In summary, dietary peptide fragments, derived from
both milk proteins and cereal grain and legume proteins,
maintain signiﬁcant amino acid homologies with collage-
nous tissues found in the synovium and are capable of
stimulating T-cells in an HLA restricted manner. Because
of the inherent lectin-induced permeability and ﬂoral changes
induced by cereal and legume consumption on the intestinal
epithelial cells, dietary antigens (with molecular mimicking
potential) which normally would not enter into the systemic
circulation, are rendered capable of doing so.
We have provided extensive evidence linking dietary sub-
stances to the development of RA. Dietary glycoproteins, as
well as other elements, can inﬂuence intestinal structure
and function so as to allow increased translocation of both
pathogenic and dietary antigens to the periphery causing
persistent immunological stimulation. Because of shared
amino acid motifs among exogenous peptides, HLA-derived
peptides and self tissue, cross reactivity may occur thereby
breaking immunological tolerance and resulting in the
expression of RA. It is proposed that by eliminating certain
dietary elements, including lectins, which adversely inﬂu-
ence both enterocyte and lymphocyte structure and function,
the peripheral antigenic stimulus will be reduced and
thereby result in a diminution of disease symptoms in
some but not all patients with RA.
Albani S & Carson DA (1996) A multistep molecular mimicry
hypothesis for the pathogenesis of rheumatoid arthritis.
Immunology Today 17, 466–470.
Albani S, Tuckwell J, Esparza L, Carson DA & Roudier J (1992)
The susceptibility sequence to rheumatoid arthritis is a cross-
reactive B epitope shared by the Escherichia coli heat shock
protein DnaJ and the HLA DRB1*0401 molecule. Journal of
Clinical Investigation 89, 327–331.
Auger I & Roudier J (1997) A function for the QKRAA amino acid
motif: mediating binding of DnaJ to DnaK. Implications for the
association of rheumatoid arthritis with HLA-DR4. Journal of
Clinical Investigation 99, 1818–1822.
Banwell JG, Howard R, Kabir I & Costerton JW (1988) Bacterial
overgrowth by indigenous microﬂora in the phytohemaggluti-
nin-fed rat. Canadian Journal of Microbiology 34, 1009–1013.
Baum H, Davies H & Peakman M (1996) Molecular mimicry in the
MHC: hidden clues to autoimmunity. Immunology Today 17,
Baum H & Staines NA (1997) MHC-derived peptides and the
T-cell repertoire: implications for autoimmune disease.
Cytokines, Cellular and Molecular Therapy 3, 115–126.
Benno P, Alam M, Henriksson K, Norin E, Uribe A & Midtvedt T
(1994) Abnormal colonic microbial function in patients with
rheumatoid arthritis. Scandinavian Journal of Rheumatology 23,
214 L. Cordain et al.
Berg RD (1992) Bacterial translocation from the gastrointestinal
tract. Journal of Medicine 23, 217–244.
Beri D, Malaviya AN, Shandilya R & Singh RR (1988) Effect of
dietary restrictions on disease activity in rheumatoid arthritis.
Annals of the Rheumatic Diseases 47, 69–77.
Bevan MJ, Hogquist KA & Jameson SC (1994) Selecting the T-
cell repertoire. Science 264, 796–797.
Bjarnason I & Peters TJ (1996) Inﬂuence of anti-rheumatic drugs
on gut permeability and on the gut associated lymphoid tissue.
`res Clinical Rheumatology 10, 165–176.
Bjarnason I, Peters TJ & Wise RJ (1984) The leaky gut of
alcoholism: possible route of entry for toxic compounds.
Lancet i, 179–182.
Bourne JT, Kumar P, Huskison EC, Maged R, Unsworth DJ &
Wojtulewski JA (1985) Arthritis and coeliac disease. Annals of
the Rheumatic Diseases 44, 592–594.
Brady PG, Vannier AM & Banwell JG (1978) Identiﬁcation of the
dietary lectin, wheat germ agglutinin, in human intestinal con-
tents. Gastroenterology 75, 236–239.
Brauer R, Thoss K, Henzgen S & Waldman G (1985) Lectin-
induced arthritis of rabbit as a model of rheumatoid arthritis. In
Lectins, vol. IV, pp. 29–38 [TC Bog-Hansen and E van
Driessche, editors]. Berlin: Walter de Gruyter & Company.
Buchanan HM, Preston SJ, Brooks PM & Buchanan WW (1991) Is
diet important in rheumatoid arthritis? British Journal of Rheu-
matology 30, 125–134.
Cao W, Tykodi SS, Esser MT, Braciale VL & Braciale TJ (1995)
Partial activation of CD8
T cells by a self-derived peptide.
Nature 378, 295–298.
Charkravarty K & Scott DGI (1992) Oligoarthritis – a presenting
feature of occult coeliac disease. British Journal of Rheumatol-
ogy 31, 349–350.
Chicz RM, Urban RG, Gorga JC, Vignali DA, Lane WS &
Strominger JL (1993) Speciﬁcity and promiscuity among natu-
rally processed peptides bound to HLA-DR alleles. Journal of
Experimental Medicine 178, 27–47.
Clevers HC, De Bresser A, Kleinveld H, Gmelig-Meyling FHJ &
Ballieux RE (1986) Wheat germ agglutinin activates human T
lymphocytes by stimulation of phosphoinositide hydrolysis.
Journal of Immunology 136, 3180–3183.
Collin P & Maki M (1994) Associated disorders in coeliac disease:
clinical aspects. Scandinavian Journal of Gastroenterology 29,
Crissinger KD, Kvietys PR & Granger DN (1990) Pathophysiology
of gastrointestinal mucosal permeability. Journal of Internal
Medicine 732, (Suppl.), 145–154.
Dalton TA & Bennett JC (1992) Autoimmune disease and the
major histocompatibility complex: therapeutic implications.
American Journal of Medicine 92, 183–188.
Darlington LG, Ramsey NW & Mansﬁeld JR (1986) Placebo-
controlled, blind study of dietary manipulation therapy in
rheumatoid arthritis. Lancet i, 236–238.
Doherty M & Barry RE (1981) Gluten-induced mucosal changes in
subjects without overt small-bowel disease. Lancet i, 517–520.
Eerola E, Mottonen T, Hannonen P, Luukkainen R, Kantola I,
Vuori K, Tuominen J & Toivanen P (1994) Intestinal ﬂora in
early rheumatoid arthritis. British Journal of Rheumatology 33,
Fabia R, Ar’Rajab A, Johansson ML, Willen R, Andersson R,
Molin G & Bengmark S (1993) The effect of exogenous
administration of Lactobacillus reuteri R2LC and oat ﬁber on
acetic acid-induced colitis in the rat. Scandinavian Journal of
Gastroenterology 28, 155–162.
Firestein GS, Alvaro-Gracia JM & Maki R (1990) Quantitative
analysis of cytokine gene expression in rheumatoid arthritis.
Journal of Immunology 144, 33347–33353.
Goronzy JJ & Weyland CM (1993) Interplay of T lymphocytes and
HLA-DR molecules in rheumatoid arthritis. Current Opinion in
Rheumatology 5, 169–177.
Grant G, More LJ, McKenzie NH & Pusztai A (1982) The effect of
heating on the haemagglutinating activity and nutritional prop-
erties of bean (Phaseolus vulgaris) seeds. Journal of the Science
of Food and Agriculture 33, 1324–1326.
Greer F & Pusztai A (1985) Toxicity of kidney bean (Phaseolus
vulgaris) in rats: changes in intestinal permeability. Digestion
Grifﬁths RJ (1992) Characterisation and pharmacological sensi-
tivity of antigen arthritis induced by methylated bovine serum
albumin in the rat. Agents and Actions 35, 88–95.
Grossman JM & Brahn E (1997) Rheumatoid arthritis: current
clinical and research directions. Journal of Womens Health 6,
Hamilton I, Cobden I, Rothwell J & Axon ATR (1982) Intestinal
permeability in coeliac disease: the response to gluten with-
drawal and single-dose gluten challenge. Gut 23, 202–210.
Haugen MA, Kjeldsen-Dragh J & Forre O (1994) A pilot study of
the effect of an elemental diet in the management of rheumatoid
arthritis. Clinical and Experimental Rheumatology 12, 275–279.
Hazenberg MP, Klassen IS, Kool J, Ruseler-van Embden JG &
Severijnen AJ (1992) Are intestinal bacteria involved in the
etiology of rheumatoid arthritis? APMIS 100,1–9.
Henriksson AE, Blomquist L, Nord CE, Midtvedt T & Uribe A (1993)
Small intestinal bacterial overgrowth in patients with rheumatoid
arthritis. Annals of the Rheumatic Diseases 52, 503–510.
Husby S, Jensenius JC & Svehag SE (1985) Passage of undegraded
dietary antigen into the blood of healthy adults. Scandinavian
Journal of Immunology 22, 83–92.
Jameson SC, Hogquist KA & Bevan MJ (1995) Positive selection
and thymocytes. Annual Review of Immunology 13, 93–126.
Karska K, Tuckova L, Steiner L, Tlaskalova-Hogenova H &
Michalak M (1995) Calreticulin – the potential autoantigen in
celiac disease. Biochemical and Biophysical Research Commu-
nications 209, 597–605.
Katz KD & Hollander D (1989) Intestinal mucosal permeability
and rheumatological diseases. Baillie
`res Clinical Rheumatology
Kavanaghi R, Workman E, Nash P, Smith M, Hazleman BL &
Hunter JO (1995) The effects of elemental diet and subsequent
food reintroduction on rheumatoid arthritis. British Journal of
Rheumatology 34, 270–273.
Keshavarzian A, Fields JZ, Vaeth J & Holmes EW (1994) The
differing effects of acute and chronic alcohol on gastric and
intestinal permeability. American Journal of Gastroenterology
Kjeldsen-Kragh J, Haugen M, Borchgrevink CF, Laerum E, Eek
M, Mowinkel P, Hovi K & Forre O (1991) Controlled trial of
fasting and one-year vegetarian diet in rheumatoid arthritis.
Lancet 338, 899–902.
Klasen IS, Melief MJ, van Halteren AG, Schouten WR, van
Blankenstein M, Hoke G, de Visser H, Hooijkaas H & Hazen-
berg MP (1994) The presence of peptidoglycan-polysaccharide
complexes in the bowel wall and the cellular responses to these
complexes in Crohn’s disease. Clinical Immunology and Immu-
nopathology 71, 303–308.
Koch AE, Shah MR, Harlow LA, Lovis RM & Pope RM (1994)
Soluble intercellular adhesion molecule-1 in arthritis. Clinical
Immunology and Immunopathology 71, 208–215.
Kool J, Gerits-Boeye MY, Severijnen AJ & Hazenberg MP (1992)
Immunohistology of joint inﬂammation induced in rats by cell
wall fragments of Eubacterium aerofaciens.Scandinavian Jour-
nal of Immunology 36, 497–506.
La Cava A, Nelson JL, Ollier WER, MacGregor A, Keystone EC,
Thorne JC, Scavulli JF, Berry CC, Carson DA & Albani S
(1997) Genetic bias in immune responses to a cassette shared by
215Rheumatoid arthritis and diet
different microorganisms in patients with rheumatoid arthritis.
Journal of Clinical Investigation 100, 658–663.
Lepore L, Martelossi S, Pennesi M, Falcini F, Ermini ML, Ferrari
R, Perticarari S, Presani G, Lucchesi A, Lapini M & Ventura A
(1996) Prevalence of celiac disease in patients with juvenile
arthritis. Journal of Pediatrics 129, 311–313.
Lepore L, Pennesi M, Ventura A, Torre G, Falcini F, Lucchesi A &
Perticarari S (1993) Anti-alpha-gliadin antibodies are not pre-
dictive of coeliac disease in juvenile chronic arthritis. Acta
Paediatrica 82, 569–573.
Liener IE (1986) Nutritional signiﬁcance of lectins in the diet. In
The Lectins: Properties, Functions and Applications in Biology
and Medicine, pp. 527–552 [IE Liener, N Sharon and IJ Gold-
stein, editors]. Orlando, FL: Academic Press.
Lowes JR, Radwan P, Priddle JD & Jewell DP (1992) Character-
isation and quantiﬁcation of mucosal cytokine that induces
epithelial histocompatibility locus antigen-DR expression in
inﬂammatory bowel disease. Gut 33, 315–319.
Lunardi C, Bambara LM, Biasi D, Venturini G, Nicholis F, Pachor
ML & DeSandre G (1988) Food allergy and rheumatoid arthritis.
Clinical and Experimental Rheumatology 6, 423–426.
Mageed RA, Borretzen M, Moyes SP, Thompson KM & Natvig JB
(1997) Rheumatoid factor autoantibodies in health and disease.
Annals of the New York Academy of Sciences 815, 296–311.
Mehta SK, Kaur S, Avasthi G, Wig NN & Chhuttani PN (1972)
Small intestinal deﬁcit in pellagra. American Journal of Clinical
Nutrition 25, 545–549.
Mielants H (1990) Reﬂections on the link between intestinal
permeability and inﬂammatory joint disease. Clinical and
Experimental Rheumatology 8, 523–524.
Mowat AM (1987) The regulation of immune responses to dietary
protein antigens. Immunology Today 8, 93–98.
Nardella FA, Oppliger IR, Stone GC, Sasso EH, Mannik M,
Sjoquist J, Schroder AK, Christensen P, Johansson PJ & Bjork
L (1998) Fc epitopes for human rheumatoid factors and the
relationships of rheumatoid factors to the Fc binding proteins of
microorganisms. Scandinavian Journal of Rheumatology 75,
Nardella FA, Teller DC, Barber CV & Mannik M (1985) IgG
rheumatoid factors and staphylococcal protein A bind to a
common molecular site on IgG. Journal of Experimental Med-
icine 162, 1811–1824.
Nossal GJV (1994) Negative selection of lymphocytes. Cell 76,
Odeh M (1997) New insights into the pathogenesis and treatment
of rheumatoid arthritis. Clinical Immunology and Immuno-
pathology 83, 103–116.
O’dell JR, Haire CE, Palmer W, Drymalski W, Wees S, Blakely K,
Churchill M, Eckhoff PJ, Weaver A, Doud D, Erikson N, Dietz
F, Olson R, Maloley P, Klassen LW & Moore GF (1997)
Treatment of early rheumatoid arthritis with minocylcline or
placebo: results of a randomized, double-blind, placebo-con-
trolled trial. Arthritis and Rheumatism 40, 842–848.
O’Farrelly C, Marten D, Melcher D, McDougall B, Price R,
Goldstein AJ, Sherwood R & Fernandes L (1988) Association
between villous atrophy in rheumatoid arthritis and a rheuma-
toid factor and gliadin-speciﬁc IgG. Lancet ii, 819–822.
O’Farrelly C, Price R, McGillivray AJ & Fernandes L (1989) IgA
rheumatoid factor and IgG dietary protein antibodies are asso-
ciated in rheumatoid arthritis. Immunology Investigations 18,
Ohashi F, Shimada T, Sakaurai M, Ishihara S, Kuwamura M,
Yamate J & Ota M (1996) The production of arthritis in beagles
by an immunological reaction to bovine serum albumin.
Experimental Animals 45, 299–307.
Oldstone MBA (1987) Molecular mimicry and autoimmune dis-
ease. Cell 50, 819–820.
Ostenstad B, Dybwad A, Lea T, Forre O, Vinje O & Sioud M
(1995) Evidence for monoclonal expansion of synovial T cells
bearing V alpha 2.1/V beta 5.5 gene segments and recognizing a
synthetic peptide that shares homology with a number of
putative autoantigens. Immunology 86, 168–175.
Panush RS, Stroud RM & Webster EM (1986) Food-induced
(allergic) arthritis. Inﬂammatory arthritis exacerbated by milk.
Arthritis and Rheumatism 29, 220–226.
Parke AL & Hughes GRV (1981) Rheumatoid arthritis and food: a
case study. British Medical Journal 282, 2027–2029.
Perez-Maceda B, Lopez-Bote JP, Langa C & Bernabeu C (1991)
Antibodies to dietary antigens in rheumatoid arthritis – possible
molecular mimicry mechanism. Clinica Chimica Acta 203,
Peumans WJ & Cammue BPA (1986) Gramineae lectins: a special
class of plant lectins. In Lectins – Biology, Biochemistry,
Clinical Biochemistry, vol. 5, pp. 31–37 [TC Bog-Hansen and
E van Driessche, editors]. Berlin: Walter de Gruyter &
Piccinini LA, Mackenzie WA, Platzer M & Davis TF (1987)
Lymphokine regulation of HLA-DR gene expression in human
thyroid cell monolayers. Journal of Clinical Endocrinology and
Metabolism 64, 543–548.
Pusztai A (1989a) Transport of proteins through the membranes of
the adult gastrointestinal tract – a potential for drug delivery?
Advances in Drug Delivery Reviews 3, 215–228.
Pusztai A (1989b) Lectins. In Toxicants of Plant Origin, vol. III,
pp. 29–71 [PR Cheeke, editor]. Boca Raton, FL: CRC Press.
Pusztai A (1993) Dietary lectins are metabolic signals for the gut
and modulate immune and hormone functions. European Jour-
nal of Clinical Nutrition 47, 691–699.
Pusztai A, Ewen SWB, Grant G, Brown DS, Stewart JC, Peumans
WJ, Van Damme EJM & Bardocz S (1993a) Antinutritive
effects of wheat-germ agglutinin and other N-acetylglucosa-
mine-speciﬁc lectins. British Journal of Nutrition 70, 313–321.
Pusztai A, Ewen SWB, Grant G, Peumans WJ, Van Damme EJM,
Rubio LA & Bardocz S (1991) Plant (food) lectins as signal
molecules: effects on the morphology and bacterial ecology of
the small intestine. In Lectin Reviews, vol. 1, pp. 1–15 [DC
Kilpatrick, E Van Driessche and TC Bog-Hansen, editors]. St
Louis, MO: Sigma.
Pusztai A, Grant G, Spencer RJ, Duguid TJ, Brown DS, Ewen
SWB, Peumans WJ, Van Damme EJM & Bardocz S (1993b)
Kidney bean lectin-induced Escherichia coli overgrowth in the
small intestine is blocked by GNA, a mannose-speciﬁc lectin.
Journal of Applied Bacteriology 75, 360–368.
Pusztai A, Greer F & Grant G (1989) Speciﬁc uptake of dietary
lectins into the systemic circulation of rats. Biochemical Society
Transcations 17, 527–528.
Rath HC, Herfarth H, Ikeda JS, Grenther WB, Hamm TE, Balish E,
Taurog JD, Hammer RE, Wilson KH & Sartor B (1996) Normal
luminal bacteria, especially bacteroides species, mediate chronic
colitis, gastritis, and arthritis in HLA-B27/Human B2 micro-
globulin transgenic rats. Journal of Clinical Investigation 98,
Rattray EAS, Palmer R & Pusztai A (1974) Toxicity of kidney
beans (Phaseolus vulgaris L.) to conventional and gnotobiotic
rats. Journal of the Science of Food and Agriculture 25, 1035–
Riordan AM, Hunter JO, Cowan RE, Crampton JR, Davidson AR,
Dickinson RJ, Dronﬁeld MW, Fellows IW, Hishon S & Kerrigan
GN (1993) Treatment of active Crohn’s disease by exclusion
diet: East Anglian multicentre controlled trial. Lancet 342,
Roudier J, Petersen J, Rhodes G, Luka J & Carson DA (1989)
Susceptibility to rheumatoid arthritis maps to a T cell epitope
shared by the HLA Dw4 DR beta 1 chain and the Epstein Barr
216 L. Cordain et al.
virus glucoprotein gp110. Proceedings of the National Academy
of Sciences USA 86, 5104–5108.
Routsias JG, Tzioufas AG, Sakarellos-Daitsiotis M, Sakarellos C
& Moutsopoulos HM (1993) Calreticulin synthetic peptide
analogues: anti-peptide antibodies in autoimmune rheumatic
diseases. Clinical and Experimental Immunology 91, 437–441.
Rowley MJ, Stockman A, Brand CA, Tait BD, Rowley GL,
Sherritt MA, Mackay IR, Muirden KD & Bernard CC (1997)
The effect of HLA-DRB1 disease susceptibility markers on the
expression of RA. Scandinavian Journal of Rheumatology 26,
Ryder SD, Smith JA & Rhodes JM (1992) Peanut lectin: a mitogen
for normal human colonic epithelium and human HT29 color-
ectal cancer cells. Journal of the National Cancer Institute 84,
Sartor RB (1989) Importance of intestinal mucosal immunity and
luminal bacterial cell wall polymers in the aetiology of inﬂam-
matory joint disease. Baillie
`res Clinical Rheumatology 3, 223–
Sartor RB, Rath HC, Lichtman SN & van Tol EA (1996) Animal
models of intestinal and joint inﬂammation. Baillie
Rheumatology 10, 55–76.
Shatin R (1964) Preliminary report of the treatment of rheumatoid
arthritis with high protein gluten free diet and supplementation.
Medical Journal of Australia 2, 169–172.
´M, Hashimoto M, Nobunaga M, Isayama T, Yasutake C &
Naono T (1994) Production of soluble ICAM-1 by mononuclear
cells from patients with rheumatoid arthritis. Inﬂammation 18,
Sjolander A, Magnusson KE & Latkovic S (1984) The effect of
concanavalin A and wheat germ agglutinin on the ultrastructure
and permeability of rat intestine. International Archives of
Allergy and Applied Immunology 75, 230–236.
Sjolander A, Magnusson KE & Latkovic S (1986) Morphological
changes of rat small intestine after short-time exposure to
concanavalin A or wheat germ agglutinin. Cell Structure and
Function 11, 285–293.
Smith MD, Gibson RA & Brooks PM (1985) Abnormal bowel
permeability in ankylosing spondylitis and rheumatoid arthritis.
Journal of Rheumatology 12, 299–305.
Sutton BJ, Corper AL, Sohi MK, Jefferis R, Beale D & Taussig MJ
(1998) The structure of a human rheumatoid factor bound to IgG
Fc. Advances in Experimental Medicine and Biology 435,41–
Taurog JD, Richardson JA, Croft JT, Simmons WA, Zhou M,
Fernandez-Sueiro JL, Balish E & Hammer RE (1994) The
germfree state prevents development of gut and joint inﬂamma-
tory disease in HLA-B27 transgenic rats. Journal of Experi-
mental Medicine 180, 2359–2364.
Thomsen P, Bjursten LM, Ahlstedt S, Bagge U & Ericson LE
(1985) Proliferative synovitis in rabbit knee joints induced by
antigen and preformed immune complexes. Scandinavian Jour-
nal of Rheumatology 14, 239–251.
Travis S & Menzies I (1992) Intestinal permeability: functional
assessment and signiﬁcance. Clinical Science 82, 471–488.
Trentham DE & Dynesius-Trentham RA (1995) Antibiotic therapy
for rheumatoid arthritis. Scientiﬁc and anecdotal appraisals.
Rheumatic Disease Clinics in North America 21, 817–834.
Tsuda M (1979) Puriﬁcation and characterization of a lectin from
rice bran. Journal of Biochemistry 86, 1451–1461.
Udey MC, Chaplin DD, Wedner HJ & Parker CW (1980) Early
activation events in lectin stimulated human lymphocytes:
evidence that wheat germ agglutinin and mitogenic lectins
cause similar early changes in lymphocyte metabolism.
Journal of Immunology 125, 1544–1550.
van de Laar MA & van der Korst JK (1992) Food intolerance in
rheumatoid arthritis. I. A double blind, controlled trial of the
clinical effects of elimination of milk allergens and azo dyes.
Annals of the Rheumatoid Diseases 51, 298–302.
van den Bourne BE, Kijkmans BA, de Rooij HH, le Cessie S &
Verweij CL (1997) Chloroquine and hydroxychloroquine
equally affect tumor necrosis factor-alpha, interleukin-6, and
interferon-gamma production by peripheral blood mononuclear
cells. Journal of Rheumatology 24, 55–60.
Verreck FA, Elferink D, Vermeulen CJ, Amons R, Breedveld F, de
Vries RR & Koning F (1995) DR4Dw4/DR53 molecules contain
a peptide from the autoantigen calreticulin. Tissue Antigens 45,
von Herrath MG & Oldstone MBA (1995) Role of viruses in the
loss of tolerance to self-antigens and in autoimmune diseases.
Trends in Microbiology 3, 424–430.
Wagner U, Kaltenhauser S, Sauer H, Arnold S, Seidel W,
Hantzschel H, Kalden JR & Wassmuth R (1997) HLA markers
and prediction of clinical course and outcome in rheumatoid
arthritis. Arthritis and Rheumatism 40, 341–351.
Weetman AP, Volkman DJ, Burman KD, Gerrard TL & Fauci AS
(1985) The in vitro regulation of human thyrocyte HLA-DR
antigen expression. Journal of Clinical Endocrinology and
Metabolism 61, 817–824.
Welsh CJ, Hanglow AC, Conn P, Barker TH & Coombs RR (1985)
Early rheumatoid-like synovial lesions in rabbits drinking cow’s
milk. Joint pathology. International Archives of Allergy and
Applied Immunology 78, 145–151.
Weyland CM & Goronzy JJ (1997) Pathogenesis of rheumatoid
arthritis. Medical Clinics of North America 81, 29–55.
Williams R (1981) Rheumatoid arthritis and food: a case study.
British Medical Journal 283, 563.
Williams RC (1992) Rheumatoid factors: historical perspective,
origins and possible role in disease. Journal of Rheumatology 32
Williams RC (1996) Autoimmune mechanisms involved in the
pathogenesis of rheumatoid arthritis. Advances in Dental
Research 10, 47–51.
Wilson AB, King TP, Clarke EMW & Pusztai A (1980) Kidney
bean (Phaseolus vulgaris) lectin-induced lesions in the small
intestine. II. Microbiological studies. Journal of Comparative
Pathology 90, 597–602.
Wilson C, Tiwana H & Ebringer A (1997) HLA-DR4 restriction,
molecular mimicry and rheumatoid arthritis. Immunology Today
Ziff M (1983) Diet in the treatment of rheumatoid arthritis.
Arthritis and Rheumatism 26, 457–461.
qNutrition Society 2000
217Rheumatoid arthritis and diet