Agaricus blazei Murill (AbM) of the family Basidiomycetes is a popular edible medicinal mushroom, originally native to a small village, Piedade, in the highland areas of Atlantic forest near São Paulo, Brazil. It has traditionally been used for the prevention of a range of diseases, including cancer, hepatitis, atherosclerosis, hypercholesterolemia, diabetes and dermatitis 12. Because of its alleged health effects, the mushroom was brought to Japan in the mid-60-ies and subjected to biomedical research. AbM was found to be rich in immuno-modulating substances such as β-glucans 34 and proteoglycans 5, and it had anti-infection 67 and anti-tumor 45 effects in mice.

Anti-tumor and anti-infection immunity are both due to Th1 responses, which also do promote autoimmune disease when overshooting. On the other hand, anti-helminth and anti-rejection immunity are due to Th2 responses, which may also induce IgE-mediated allergy, whereas delayed-type hypersensitivity is believed to involve Th1 cells. Since, according to the original Th1/Th2 dichotomy 8 there is an inverse relationship between Th1 and Th2 responses, we set out to look for substances that increased Th1 responses and thus, presumably, would reduce allergy. Moreover, we looked for substances with broad immunogenic specificity and hence a broad range of possible therapeutical activity. This criterion fits substances containing so-called pathogen-associated molecular patterns, which stimulate innate immunity via binding to a few different receptors with broad specificities like Toll-like receptors and dectin-1.

In order to test putative functional Th1-stimulating substances, a mouse model for systemic bacterial infection was chosen rather than a tumor model, because of the more rapid outcome of an anti-bacterial than an anti-tumor response. We tested different β-glucans, which are known stimulators of innate immunity with anti-tumor 9 and anti-infection 10 activities. We found that one β-1,3-glucan from Sclerotinia sclerotiorum was highly protective against sepsis in a mouse model for systemic S. pneumoniae, although only when given i.p. and not p.o. 11. However, surprisingly, we detected that s.c. administration of both this β-glucan and other β-glucans from barley and baker's yeast, in addition to moulds per se, also increased specific IgE levels in a mouse model for allergy 1213. This is in agreement with the finding of increased allergic responses of mold-derived β-1,3-glucan in an airway inhalation model in the mouse 14. Since AbM is another more recently discovered source of strong innate stimulatory properties 1516, with a high content of β-glucan and anti-tumor properties in the mouse 3, we tested whether extracts of AbM from different producers had anti-infection effects in the said mouse model for pneumococcal sepsis. We found that the current extract, AndoSan™, containing approximately 80% of AbM and 20% of two other Basidiomycetes mushrooms; Hericium erinaceum and Grifola frondosa, was the most effective: It was the only extract that decreased bacteremia statistically significantly and increased the survival rate of the exposed animals 17. Moreover, it had more profound anti-infection effect even when given p.o. via a gastric catheter than did any of the above β-glucans given i.p..

There are anecdotes about persons who have used AbM for other purposes than allergy, and who have experienced less allergic symptoms when ingesting the remedy. To our knowledge the very few papers on AbM or other Basidiomycetes mushrooms and allergy in English scientific literature rather report on induction of allergy; cheilitis and increased delayed-type sensitivity due to AbM 1819, hypersensitivity pneumonitis caused by Grifola frondosa 20, and allergic contact dermatitis from Hericeum erinaceum exposure 21. Based on preliminary anti-infection and anti-allergy results in our laboratory with the current extract of mainly AbM (AndoSan™), a patent application was filed in 2004 22. There are other publications on beneficial effects of the mushrooms in the patent literature, foremost of Japanese origin: One patent (A61K 35/84, 05.08.2002) claims that an essence extracted from mycelium of Basidiomycetes, including Hericium erinaceum, can prevent and cure allergic symptoms, especially atopic dermatitis. Another (WO 02/15917) claims the use of AbM in treatment of autoimmune and skin diseases, due to down-regulation of immune function. Yet another (WO 93/207923) describes the isolation from Agaricus hortensis of anti-allergic components, especially for dermatological usage. Extracts of AbM have also been found to have anti-allergic effect based on inhibition of basophilic leukocytes (US2003/0104006). Finally, EP0413053 describes a process for producing an anti-allergic substance from Basidiomycetes mycelium, including that of AbM and Grifola frondosa.

The aim of the present study was to examine whether the extract that was most effective against systemic pneumococcal infection, also could protect against allergy development when given to a mouse model for allergy. For this purpose the model allergen ovalbumin (OVA) was injected s.c. and AbM extract as adjuvant was given orally, and levels of specific IgE and IgG2a antibodies were determined in serum. In addition, Th1, Th2 and Treg cytokines were measured in supernatants of cultured spleen cells from the mice.

Our results are strengthened by the similar findings, observed in two different mouse strains after s.c. injection of OVA in the tail base, and in a third mouse strain (Balb/c) after s.c. injection of both mushroom extract and OVA in the foot pad. In the latter Th2-prone mice the so-called PLN assay was used, which was originally employed for toxicological screening of substances that would inflame the foot pad-draining PLN, but which is also convenient for examining systemic IgE response in serum to an allergen given with adjuvant 23. The lacking increase in PLN weight in mice injected AbM extract relative to PBS in the foot pad, agrees with the assumed anti-inflammatory anti-allergic effect of the AbM as seen from the tendency of generally lowering of Th2 cytokine levels in spleen cell cultures ex vivo.

Increased specific IgE levels are not equivalent with allergic disease, but a prerequisite for IgE-mediated allergy. Hence, our findings of decreased anti-OVA IgE levels secondary to AbM intake in animals that were otherwise sensitized to OVA, strongly indicates a protective effect of AbM against IgE-mediated allergy. We did not examine allergy signs in the mice. These would have been similar to egg allergy, as in food allergy. Possible skin rashes would have been difficult to assess in the mice, and nude mice could not have been used because they lack normal lymphocytes, which are a prerequisite for an allergic immune response. In possible follow-up studies, the allergen should be given via the natural route; e.g. p.o. if using ovalbumin, although this would be costly. Instead, a common food allergen like peanut could have been used, or if one wished to examine airways allergy in the case of aeroallergens, another cheap aeroallergen like birch pollen, although with novel ELISAs for these antigens. The finding of relatively far lower anti-OVA IgE levels in the repeated PBS controls in Figure 5, may be due to the stress invoked by such repeated intragastric procedure. In preliminary experiments, in which repeated pre-OVA treatment of mice with the mushroom extract or PBS was delivered intragastrically by the highly trained technicians to increase the dose, all mice looked sick and one animal died, presumably from stress, which is known to impair immunity.

Previously, we have used pure β-glucans from yeast and fungi together with ovalbumin s.c. in the very same PLN model and, contrary to the present observation, found increased specific anti-OVA IgE levels in serum 1213. Hence, either the administration route is critical, or the particular β-glucans of the current mushroom extract does either promote a different outcome than the other β-glucans, or other stronger anti-allergic immunomodulating substances in the mushroom extract do overcome a possible general "pro-allergic" effect of β-glucans. If the latter is true, we assume that the anti-allergy effects of the AbM extract in vivo is mediated via immunomodulating substances in the extract that are smaller and more readily absorbable than β-glucans.

As to possible side effects, there are conflicting reports regarding the effect of AbM on liver function. Whereas one report suggests that use of AbM for several weeks may have induced severe hepatic dysfunction in three cancer patients 24, another says that AbM extract normalized liver function in patients with chronic hepatitis B virus infection 25. Moreover, our studies on patients with chronic hepatitis C virus infection 26 and on AbM intake in healthy volunteers 27, revealed no pathological effect whatsoever on hematological parameters including those for liver-, pancreatic- and renal function, even when volumes equivalent by body weight to that given to the mice, were taken.

The generally observed AbM-induced all-over reduction in Th2 cytokines IL-4 and IL-5 relative to Th1 cytokines IFNγ and IL-2 production ex vivo in our present cultures of spleen cells, agrees with the original Th1/Th2 dichotomy 8. However, this theory has been modified towards suggesting that T regulatory cells are crucial for fine-tuning both Th1 and Th2 responses by the regulatory cytokines IL-10 and TGF-β. However, our measurement of suggestive reduced levels of the Treg cytokine IL-10 in the AbM groups, is difficult to interpret. In contrast, when the extract was given in vitro to cell cultures there was an increase in proinflammatory cytokines 15. This apparent discrepancy must be due to the fact that whereas cells in vitro are subjected to all substances in the extract including β-glucans with large m.w., which are abundant in AbM 3, mainly smaller substances are taken up from the digestive tract in humans and are active in the blood in vivo. Although, β-glucans in the intestines could stimulate Peyer's patches in jejunum, we have in fact observed that the genes in leukocytes predominantly affected by AbM in vitro and in vivo were quite different 2628. Whereas genes related to proinflammatory cytokines were strongly induced in vitro – presumably by β-glucan, genes involved in cell signalling and cycling and transcriptional regulation and thus foremost related to anti-tumor defence, were upregulated in vivo 26. Thus, the microarray analyses agree with the assumption that AbM extract especially promotes a Th1 anti-tumor and anti-infection response in the body and hence reciprocally inhibits a Th2 response. This is supported by the reported immuno-modulatory effects of AbM in mice 19.

β-glucans may stimulate macrophages and other cells of innate immunity after binding to cellular receptors like CD11b/18, Toll-like receptor and dectin-1 [reviewed in 17]. Stimulation by AbM of peripheral blood leukocytes resulted both in an upregulation of such receptors 262829, activation of NFKB via TLR2 stimulation 30, and mediation via them of increased release of proinflammatory cytokines 15 and Th1 cytokines IFNγ, IL-12, and IL-23α 162831. Although one report of reduced release of Th2 cytokine IL-4 after AbM stimulation in vitro also found reduced IL-2 and IFNγ levels 32, IL-12- and IFNγ-mediated NK cell activation by AbM p.o. has been documented in mice 16. Even though the present AbM extract should occasionally give reduced IFNγ levels, the increased expression of the IFN receptor gene after AbM extract intake in humans 26, may overcome a reduction in the concentration of the ligand and result in an increased Th1 response. When measuring different cytokines in serum from humans after 12 days intake of the current AndoSan™ extract mainly containing AbM, there was a significant reduction in both pro-inflammatory, Th1 and Th2 cytokines 31. This indicates a general anti-inflammatory effect of AndoSan™ in vivo, which agrees with its current anti-allergic effect.

For intragastric delivery of AbM extract a volume of 200 μl was chosen because this is, according to our veterinary, the maximal ventricular volume in a 5–6 weeks old mouse. In an initial experiment, we tried to give the AbM extract repeatedly on subsequent days via a gastric catheter in order to possibly inhibit the specific IgE response completely. However, this procedure was dropped because it was too stressful for the mice even in hands of our well-trained technicians. The unexpected result in the last two columns of Figure 5 may in fact reflect this concern. Also, we did not use a higher concentration of this extract than what was sold on the health food market. If translated to human intake, the equivalent of 200 μl to a 25 g mouse would be 560 ml to a 70 kg individual. In fact, a daily low intake in healthy volunteers of 60 ml AndoSan™ for 12 days gave a significant 50% reduction in levels of the allergy-promoting cytokine IL-4 in blood and left the other allergy-related cytokines IL-5, IL-7 and IL-13 at negligible levels 27. Addition of AbM extract to drink water for the mice in our set-up would have been more natural, but the intake of AbM is impossible to monitor as accurately as with intragastric delivery. In the current allergy model we did not test other extracts of AbM from other manufacturers that did not have a significant effect against pneumococcal infection in mice 17. Therefore, the question is not fully answered as to whether there is an absolute link between the anti-bacterial infection and anti-allergy effect of a substance or an extract like AndoSan™. Moreover, even though we have seen that Agaricus bM is the main TLR2 stimulating mushroom of AndoSan™ 30, it is likely that the former anti-bacterial and current anti-allergic effect of this mixed mushroom product may be partly due to possible synergistic effects of the other mushrooms, Hericium erinaceum and Grifola frondosa, and components thereof contained in the extract. The mouse model of allergic airways disease should be used in a follow-up study with OVA and the mushroom extract in order to confirm that also allergic symptoms like development of airways hyper responsiveness are reduced by AndoSan™ intake. Whether the AbM extract is effective against allergy in the human setting must be tested in a clinical trial, e.g. in persons with aeroallergy during the pollen season taking 60 ml a day for a few weeks.

From our results with mice we conclude that a mushroom extract, mainly containing AbM, may prevent the development of IgE-mediated allergy when given before allergen immunization. Even more interesting, the extract seemed to have a therapeutic effect when given together with or as late as 3 weeks after the allergen immunization. Three weeks in the mouse equals several months in a human, suggesting that also established allergy in patients can be reverted.

Possible conflict of interest: GH filed a patent application (WO2005/065063: "Use of the mushroom Agaricus blazei Murill for the production of medicaments suitable for treating infections and allergies") with priority Jan 2004, based on preliminary anti-infection and anti-allergy experiments in 2003. LKE has no competing interests.

LKE supervised the animal experiments and the Ig, cytokine and other measurements that were performed together with technicians. GH had the idea for the project and did most of the data analysing and writing, and both authors collaborated on the study design.