WO1997030717A1

WO1997030717A1 - Adhesion inhibitors, preparation comprising them and method for producing them

Info

Publication number: WO1997030717A1
Application number: PCT/SE1997/000313
Authority: WO
Inventors: Patricia Lynne Conway; Arthur Ouwehand
Original assignee: Arla Ekonomisk Förening
Priority date: 1996-02-26
Filing date: 1997-02-25
Publication date: 1997-08-28
Also published as: AU2108997A; CA2243067A1; EP0904091A1; SE508045C2; SE9600716D0; SE9600716L

Abstract

Products for inhibiting the adhesion of pathogens to gastrointestinal epithelial mucosa in animals including humans are described. The products contain high molecular weight carbohydrates produced by Lactobacillus strains of enteric origin preferably having a composition of N-acetylglucose-amine:galactose:glucose in a ratio of 1:3:2. There is also described preparations for oral administration comprising as active ingredient the products above, when purified, partially purified, or modified chemically while retaining the biological activity, as well as subunits and complexes thereof which retain the biological activity, or Lactobacillus strains producing the products, or culture supernatant resulting from culturing Lactobacillus strains producing the products. Further, there is described a method of preparation of the products by culturing Lactobacillus strains producing the products.

Description

ADHESION INHIBITORS, PREPARATION COMPRISING THEM AND METHOD FOR PRODUCING THEM

The invention concerns products for inhibiting the adhesion of pathogens to gastrointestinal epithelial mucosa in animals including humans, as well as prepara¬ tion comprising the products as active ingredient and a method of producing the products.

Enteropathogenic Escherichia coli strains producing K88 fi briae are found to be one of the major causes of post-weaning diarrhoea in piglets (Jonsson & Conway, 1992) . It has been shown that K88 fimbriae allow the bac- teria to bind the brush border of small intestine epithe¬ lial cells (Gaastra & de Graaf, 1982) . In this way, the bacteria are able to resist the secretion of fluids into the lumen and the peristaltic movement of the intestine, which will increase especially after the production of toxins by the bacteria. Inhibition of this attachment might reduce the virulence of the pathogen. Work on the inhibition of enteropathogenic E. coli growth is rather common (rev. e.g. Lindgren & Dobrogosz, 1990; Cherrington et al, 1991), It has previously been found that spent culture fluids from Lactobacillus fermentum strain 104r reduced the adhesion of wild type E. coli K88ac strain 1107 to 35 days old piglet ileal mucus (Blomberg et al, 1993a) .

It has now been found that lactic acid bacteria (LAB) of enteric origin release products upon culturing which inhibit the adhesion of pathogens to gastrointes¬ tinal epithelial mucosa in animals including humans. The products are high molecular weight carbohydrates. One example of a product according to the invention has an estimated molecular weight of approximately 200-1700 kDa. The products contain N-acetylglucose-amine:galactose:glu¬ cose in a ratio of 1:3:2. They are products of cell growth and most likely they are initially cell bound (maybe associated with the cell wall or intracellular) . The products interfere with the mucus and inhibit adhe¬ sion, probably due to steric hindrance.

The products according to the invention also encom¬ passes derivatives thereof and preparations, both impure and pure, which contain the active component, as well as subunits and complexes thereof which retain the biologi¬ cal activity.

Preferably, the lactic acid bacteria originate from the gastrointestinal tract of the host to which the pro- ducts will be administered. In the following Table 1 examples are given of strains which produce carbohydrate products for use as adhesion inhibitors according to the invention.

TABLE 1

Strains used to produce products inhibiting the adhesion of E. coli strains expressing fimbriae Strain Origin

Lactobacillus fermentum 104r porcine, enteric Lactobacillus murinus C39 porcine, enteric Lactobacillus fermentum KLD human, enteric Lactobacillus sp. HBL8 human, enteric Lactobacillus sp. LMN9 human, enteric Lactobacillus fermentum 104s porcine, enteric Lactobacillus sp. LAB32 porcine, enteric

Target organisms are E. coli strain expressing fimbriae. These fimbriae allow the baceria to adhere to brush border of small intestine epithelial cells and to overlying mocus (25) . Examples of targets organisms are given in the following Table 2. TABLE 2 E. coli strains expressing different fimbriae used as target organisms

Escherichia coli strain Origin K12 PMK 002 (K88^~) porcine K12 K88ab (1) porcine K12 K88ab (2) porcine K12 K88ac porcine K12 PMK 005 (K88ac) porcine 1107 (K88ac) (wild type) porcine K12 K88ad porcine K88ad (wild type) porcine K99 (wild type) porcine 987P (wild type) porcine F41 (wild type) bovine CFA/I human CFA/II human CFA/IV human HB101 (pAZZ 50) (S) human The products according to the invention could be administered to animals including humans in different ways. Thus, preparations for oral dosage may comprise the viable Lactobacillus strain as such. It may, however, also comprise the culture supernatant resulting from culturing the Lactobacillus strain, or the purified, or partly purified high molecular weight carbohydrate pro¬ ducts produced by the Lactobacillus strain, as well as chemically modified products which still retain the acti¬ vity of the purified products. The carbohydrate products can be prepared by cultur¬ ing a strain of Lactobacillus of enteric origin in complex media under semi-anaerobic growth conditions. The optimal production occurs in the late log phase and early stationary phase of the growth. The products are recover- ed from the growth medium into which they are released during culturing. For use in preparations for oral dosage, the carbo¬ hydrate products are preferably mixed with pharmaceu¬ tically acceptable constituents.

The inhibitory activity was detectable in late log phase and early stationary phase. It was found that more of the active products were present after death of the cell, which can suggest that the active products origin¬ ate from lysed cells. This is also supported by the observation that cells did not die in the medium where acetate was omitted and no adhesion inhibiting activity could be observed. Despite the fact that cell wall frag¬ ments were not found to be inhibiting, treatment of retentate fractions with lysozyme removed the adhesion inhibiting activity. From this it could be deduced that the active components are soluble cell wall fragments coming from lysed cells.

The inhibitory activity was shown to be largest at 37°C. Although the activity is strongly reduced at 0°C, it is unlikely to be caused by inactivation of enzyme acti- vity, since diluted spent culture fluid was not able to increase inhibitory activity upon prolonged incubation.

Fractionation by gelfiltration shows that one form of the active component has an Mr of approximately 1700 kDa. After pronase treatment the Mr is reduced to approximately 1100 kDa, suggesting that the active com¬ ponent might be associated with protein. It is not clear whether this is a 600 kDa protein or a smaller protein binding several subunits together or giving the active component a confirmation with a larger Stoke' s radius. The additional fractions with adhesion inhibiting acti¬ vity that were found after pronase treatment, could probably contain pronase as it is heat sensitive and of about the same size.

From the above it is clear that the molecular weight of the products according to the invention is not to be limited to 1700 kDa, but products having both smaller and larger molecular weights, such as subunits and complexes ') 5 thereof, are encompassed as long as they have the same activity.

The active component is not likely to be protein- aceous since activity was still demonstrable after pronase treatment and after heat treatment at 121°C for 20 min.

Since lipase treatment interfered with the adhesion assay, it was not possible to exclude the fact that the inhibitory component contained lipid using the results from the lipase study. However, after solvent extraction the activity remained in the water phase and no activity could be observed in the chloroform phase. This indicates that the active component is not hydrophobic.

Chemical analysis showed the presence of glucose, N- acetylglucose-amine and galactose in a ratio of 1:3:2.

In order to investigate in which way the active com¬ ponent interacts with the mucus, E. coli 1107 was treated with retentate and washed. This treatment did not affect the adhesion. If E. coli 1107 was allowed to adhere to immobilized mucus prior to treatment with retentate, cells were not removed. This indicates that the component acts either on another place than the K88 receptor or has a lower affinity for the receptor than K88 fimbriae. The latter possibility is unlikely since simultaneous expo- sure of immobilized mucus to both retentate and E. coli 1107 gives the same inhibition of adhesion as with pre- treatment of the mucus with retentate. Without being bound to any theory, it is assumed that the active com¬ ponent has a higher affinity for its site of action than the K88 fimbriae for their receptor site. It might there¬ fore be postulated that E. coli 1107 adhesion is inhibit¬ ed due to steric hindrance by the active component.

Adhesion to mucus fractions containing detectable proteins can be inhibited by treatment with spent culture retentate. Since the adhesion to the neutral lipids from the mucus could not be inhibited by retentate, it might be that the active component only affects the protein receptor. Alternatively, the component may require non- -lipid substances to influence adhesion to the neutral lipids and the lack of inhibition by retentate to neutral lipids could be attributable to the assay used. The invention will be described further in detail by way of the following non-limiting examples. Materials and methods Bacteria and culture conditions

L. fermentum strain 104r was isolated from porcine gastric squamous epithelium (Henriksson et al, 1991) .

Primary cultures (from stocks stored at -80°C in 40% gly¬ cerol) were grown overnight in broth (Mann, Rogosa and Sharpe; MRS Oxoid) using a 1% inoculum, 37°C and a candle jar. Cultures were then grown in a lactobacillus defined medium (LDM) , prepared according to Kotarski & Savage (1979) except that Tween 80 was omitted, using an 1% inoculum from the primary culture. Cultures were also grown in LDM where both Tween 80 and sodium acetate had been omitted. Spent culture fluids were collected by centrifuging the 24 hour cultures at 10 000 x g for 20 min and dialysing (molecular weight cut off 12 to 14 kDa) at least 3 times at 4°C for approximately 6 h against 5 liter Milli-Q water, purified by Milli-Q plus (Millipore Corp.). To determine the relation between adhesion by spent culture fluid and L. fermentum 104r growth, 10 ml samples were taken from a 250 ml culture at different time intervals and centrifuged and dialysed as described above. Spent culture fluids and retentates were stored at -20°C prior to use. As a control uninoculated medium was dialysed as for the spent culture fluid.

Primary cultures of E. coli K88ac strain 1107 (here after referred to as E. coli 1107) were grown overnight in tryptone soya broth (TSB; Oxoid) at 37°C using an 1% inoculum from stocks stored at -80°C in 40% glycerol. E. coli 1107 was radioactively labelled by inoculating 5 ml TSB, containing 1 μCi.ml^"1 of methyl-1,2-³H-thymidine.ml^-1 (120 Ci.mmol^"1; Amersham International), with 1% from the primary culture and growing at 37°C to an absorbance (600 nm) of 0.5±0.01. Cells were harvested by centrifuga¬ tion (approx. 3000 x g) , washed in phosphate buffered saline (PBS; 10 mM phisphate; pH 7.2) and resuspended in PBS and the absorbance at 600 nm adjusted to 0.5. The effect of the presence of retentate on adhesion and the effect of retentate on E. coli 1107 was studied by re- suspending cells in retentate (60 min 37°C) and subse¬ quently using this suspension directly in the adhesion assay or by washing and resuspending in PBS. In vitro adhesion assay

The adhesion of radioactively labelled E. coli 1107 to i..eal mucus from a 35 days old pig was studied using a modification of the method of Laux et al (1984) as out- lined in Conway et al (1990) . In brief, mucus was col¬ lected from the ileum of a 35 days old piglet in Hepes Hanks buffer (HH; pH 7.4) and stored at -20°C or -80°C. It was immobilised (0.5 mg protein.ml^"1) on polystyrene microtiterwells (Nunc; polysorp) by overnight incubation at 4°C. Excess mucus was removed by washing twice with 0.25 ml HH. Spent culture fluid dialysis retentates or fractions (0.10 ml per well) were added to the immobi¬ lised mucus and the microtiterplates were then incubated at 37°C for 1 h. Wells similarly treated with uninoculat- ed dialysed medium served as controls. The mucus coated wells were then washed twice with HH (0.25 ml) and the adhesion of the radioactively labelled cells studied as previously described (Conway et al, 1990) . The inhibition of adhesion was assessed as the procentage of radioacti- vity in the test relative to the control in which mucus was treated with unincoluated medium. All assays were performed in quadruplet. Kinetics of the adhesion inhibiting component

In order to determine the kinetics of the inhibitory component, the adhesion assay was performed on ice (0°C) and at 37°C, and incubated with spent culture fluid for different times. To determine whether enzymatic activity was involved the spent culture fluid was diluted 10 times with Milli-Q water and also incubated for different time intervals. The rest of the experiment was performed as outlined above. Fractionation of spent culture fluid retentate

Spent culture fluid from L. fermentum 104r after growth in LDM-medium was fractionated by gel filtration of culture supernatant retentate using Superose 6 prepa¬ rative grade (Pharmacia) . Aliquots of spent culture fluid (100 ml) were dialysed against Milli-Q water and then concentrated approx. 25 times by ultra filtration using a 10 kDa cut off membrane (Millipore) . This concentrate was freeze dried and resuspended in 2 ml Milli-Q, thus con¬ centrating the material 50 times. A 1 ml sample of this concentrate was applied to an XK 16/70 column packed with Superose 6 preparative grade (Pharmacia) and eluted with PBS at a flow rate of 13.4 ml.cm^~2.h^_1. The fractionation was carried out at 4°C and the eluate collected as 2 ml fractions. The absorbance (280 nm) was recorded and carbohydrate content determined with glucose as standard (Duboiε et al, 1956) immediately after fractionation. Concentrated supernatant was also treated with pronase (Calbiochem) 0.1 mg.ml^"1 and incubated for 30 min at 37°C and fractionated as described above. To a Con A-Sepharose 4B (Pharmacia) column, bed volume 18.5 ml, 10 ml retentate was applied, flow rate 13 ml.cm^~2.h^_1. The column was then washed with 20 ml buffer (Tris-HCl, 0.02 M; NaCl 0.5 M, pH 7.4) . Finally the column was eluted with buffer (as mentioned above) containing 0.5 M α-D-methylmannose or with a borate buffer (0.1 M, pH 6.5). The fractionation was carried out at 4°C and the eluate collected as 2 ml fractions. Characterisation of the adhesion inhibiting component

Prior to the addition of spent culture fluid reten- tate or the medium control to the immobilised mucus to test the presence of adhesion inhibiting activity, reten¬ tate was pretreated in different ways, (i) Retentate and 9 medium were autoclaved at 121^CC for 20 min before testing in the adhesion assay, (ii) Treatment with lipase (Calbiochem) was performed by using 4 mg.ml^-1 and incuba¬ tion for 1 h at 37°C, then the sample was boiled for 10 min and centrifuged at about 13 000 x g for 5 min.

(iii) A lipid extraction was performed according to Bligh & Deyer (1959) . A 5 ml sample of dialysed spent culture fluid was added dropwise to a mixture of 5 ml chloroform and 10 ml of methanol. The mixture was left stirring for 30 min. After this period, 5 ml chloroform and 5 ml

Milli-Q water were added and allowed to stir for 5 min. The mixture was centrifuged for 5 min at approx. 3000 x c. After separation the chloroform was evaporated under a stream of nitrogen and the water phase was freeze dried. Both fractions were redissolved in 5 ml Milli-Q water. The lipid fraction was sonicated to enhance dis¬ solving, (iv) Lysozyme (Serva) treatment was performed with 2 mg.ml^"1 and incubation for 1 h at 37°C, then the samples were boiled and centrifuged as described above for lipase. (v) Cell wall fragments were obtained using

L. fermentum cells grown in LDM. The cells were harvested in mid log-phase by centrifugation (20 min at 4°C, 10 000 x g) and washed twice with HH. Cells were homo¬ genised for one hour at 2°C in a Vibrogen-Zellmϋhle. The homogenate was centrifuged at 25 000 x g, 20 min at 4°C. The pellet was washed twice and resuspended in HH to one eighth of the original culture volume and then frozen (-20°C) until use. (vi) For carbohydrate analysis spent culture fluid retentate was concentrated 100 x and 900 μl was applied on a Sepharose CL 4B column after elution with PBS and active fractions were pooled. Monosaccharide analysis was performed on pooled fractions as described by Hardy et al (1988) . (vii) Glucose-oxidase (Boehringer Mannheim) treatment was performed with 5.0 mg.ml^-1 and incubation for 1 h at 37°C, then the samples were treated as with lipase. (viii) Endo-β-galactosidase (Boehringer Mannheim) treatment was performed with 0.05 mg.ml^"1 and incubation for 1 h at 37°C, then the samples were treated as described for lipase. Characterisation of the mechanism of inhibition

Ellipsometry measurements were performed in a Rudolph Research model 436 with a vertical sample orien¬ tation, in order to detect alterations of the mucus by treatment with retentate. This technique measures the changes in polarisation of light when reflected on a surface. These changes are strongly influenced by the presence of thin films adsorbing to the surface. The principles of ellipsometry are described in Welin (1992) . From the ellipsometric angles, polariser and analyser, the thickness of the adsorbed film is calculated. Mucus was immobilised on a hydrophobic silicon surface and treated with retentate or medium as in the adhesion assay. Retentate was also radioactively labelled by reductive methylation using the method of Jentoff & Dearborn (1979) . An aliquot of retentate (6 ml) was mixed with 1 μl ³H-formaldehyde (37%), NaCNBH₃(7.6 mg) and Hepes (15.6 g), pH 7.5 and incubated overnight at 22°C. The reaction was stopped by dialysis against Milli-Q water. The labelled retentate was tested for adhesion inhibiting activity. Immobilised mucus was incubated with the label¬ led retentate. After 1 hour at 37°C, retentate was trans- ferred to scintillation vials and activity was determin¬ ed. The wells were washed twice with 250 μl HH and the activity in the washings determined by scintillation. The wells were then incubated with 250 μl 1% SDS at 60°C. After 1 h the contents were transferred to scintillation vials and activity was determined. Complementary contact angles of immobilised mucus were determined with a NRL contact angle goniometer. Mucus was immobilised as in the adhesion assay and the supporting surface was immersed in HH with the mucus coated surface facing down. Air bubbles were released under this surface and their complementary contact angles determined. Adhesion and inhibition of E. coli 1107 adhesion to neutral lipids was studied in order to determine which receptor is affected by retentate. Neutral lipids (0.5 mg.ml^"1 in methanol) from mucus, pre¬ pared as described in Blomberg et al (1993b), were immo¬ bilised by evaporation of the solvent. After evaporation BSA was added (0.5 mg.ml^"1) and the rest of the experiment was performed as described above for mucus.

A 2 ml sample (4.5 mg.ml^"1 protein) of freshly pre¬ pared porcine ileal mucus was fractionated using a XK 16/70 column packed with Sepharose CL 4B (Pharmacia) and eluated with HH at a flow rate of 15 ml.cm^"2.h^-1. The eluate was collected as fractions (2 ml each) and moni¬ tored by measuring absorbance (280 nm) . Fractions were immobilised as described above for the in vitro adhesion assay, in order to block any non-covered spaces, the wells were incubated 1 h at 37°C with 0.5 mg.ml^-1 BSA after removing the fractions. The rest of the adhesion assay was performed as outlined above.

Results

Relationship between growth phase of L. fermentum 104r and inhibition of E. coli 1107 adhesion

The adhesion inhibiting effect is first detectable in late log-phase and early stationary-phase of L. fer¬ mentum 104r. When a sample taken after 24 h of growth was diluted 10 times, no adhesion inhibiting effect could be observed (94.5% SD 15.21, n=4), however a sample taken after 86 h of growth was diluted 10 times, an adhesion of 54.2% SD 2.3 (n=4) was observed. When L. fermentum 104r was grown in LDM without Tween 80 and sodium acetate, no adhesion inhibiting activity could be detected in spent culture fluid, nor did the cells die as fast as in LDM where only Tween 80 had been omitted, results not shown. Kinetics of the adhesion inhibiting activity

The maximum inhibitory activity was obtained after approx. 15 min, when the assay was performed on ice or at 37°C, however the adhesion of E. coli 1107 was reduced to approx. 70% and 15% for 0° and 37°C incubations respecti¬ vely. When spent culture fluid was diluted 10 times, the adhesion was reduced to approx. 90% and remained at this level with prolonged incubation. Fractionation of culture supernatant retentate

When concentrated retentates of cultures grown in LDM broth were fractionated by gel filtration, the adhesion inhibitory activity was found in the fractions corresponding to have a relative molecular weight (Mr) of approx. 1700 kDa. When retentate was treated with pronase, the activity was detected in the fractions con- sistent with an Mr of approx. 1100 kDa. These fractions contained virtually no protein but considerable amounts of carbohydrate. After pronase treatment additional frac¬ tions with adhesion inhibiting activity could be observ¬ ed. This activity was sensitive to heat treatment, and has an Mr consistent with that of pronase.

Upon affinity chromatography with Concanavaline-A no activity could be recovered from the column, neither upon eluation with 0.5 M α-D-methylmannose nor with elution with 0.1 M borate buffer. Characterisation of the adhesion inhibiting component Autoclaving the spent culture fluid retentate did not affect the adhesion inhibitory effect (autoclaved 6.78% SD 0.04 adhesion v.s. untreated 7.23% SD 1.46, n=3) . Results from treatments with lipase could not be interpreted because these enzyme treatments interfered with the adhesion assay, since both treated medium and retentate caused a 1.5 to 3 fold increase in adhesion as compared to wells which has been exposed to untreated medium. After lipid extraction, very little inhibitory activity was found in the chloroform phase (94.50% SD 10.35 adhesion, n=3) with most of the activity being present in the water-methanol phase (34.76% SD 24.02 adhesion, n=3) . Treatment with lysozyme totally removed the activity (100.46% SD 1.42, n=3) . In contrast cell wall fragments had a stimulatory effect on adhesion in concentrations higher than the original culture concent¬ ration and no effect when diluted to the original con- centration. Carbohydrate analysis of spent culture reten¬ tate fractions revealed the presence of N-acetylglucose- -amine, galactose and glucose in a ratio of 1:3:2 in the most active fractions. Based on these results, retentate was treated with glucose-oxidase and endo-β-galactosid- ase. Glucose-oxidase removed most of the activity (81.90% SD 9.37, n=4), while endo-β-galactosidase did not affect the adhesion inhibiting activity (27.51% SD 9.38, n=3) . Characterisation of the mechanism of inhibition Using ellipsometry no change was detectable in thickness of the absorbed mucus film (4.3 ng.mm^-2 immobi¬ lised material) . Retentate was radioactively labelled in order to ascertain whether material in the retentate was retained on the mucus after washing. Of the total radio- activity in the retentate added to the mucus, 92.61% SD

0.38 (n=3) was left in the retentate when it was recover¬ ed from the wells, 4.90% SD 0.37 (n=3) of the activity was in the washings. Only 2.40% SD 0.15 (n=3) of the app¬ lied radioactivity was directly associated with the mucus layer and was recovered after solubilisation with 1% SDS. As a control, it was checked that labelling of the reten¬ tate did not affect the adhesion inhibiting activity (labelled retentate permitted 54.26% SD 4.35 adhesion compared to 45.37% SD 5.98 for unlabelled retentate, n=2) . There was no demonstrable difference in the contact angle of mucus treated with retentate relative to the untreated control. Immobilised mucus had a complementary contact angle of 20.22° SD 5,71 (n=68), while mucus treated with retentate had a complementary contact angle of 15,83° SD 5,52 (n=52) . E. coli 1107 cells that adhered to mucus untreated with retentate were not removed by subsequent addition of retentate (adhesion 99.15% SD 0.78, n=2) . Adhesion was inhibited, however, when E. coli 1107 cells were resuspended in retentate and applied to the immobilised mucus (15.00% SD 2.22, n=2) . This inhibi¬ tion was removed by washing the retentate from the cells prior to the adhesion assay (85.01% SD 8.51, n=2) . Adhe- 14 sion to neutral mucus lipids could not be inhibited by retentate (99.1% SD 6.1, n=2) . The number of cells adher¬ ing to neutral lipids was approx. 58% of that adhering to mucus. E. coli 1107 cells adhere to most of the mucus fractions.

Summary of properties of one example of the adhesion inhibitor

Size approx. 1 700 000 Da

(gelfiltration)

Heat stability 20 min 121°C Chemical composition N-acetylglucose-amine:galac- tose:glucose (1:3:2)

No inhibition 104r cell walls periodate inconclusive results resistance enzymatic activity; lipase unknown, interferes with adhesion assay pronase yes lysozyme no endo-β-galactosidase yes glucose oxidase reduced activity mucus destroys activity

Concanavalin A no activity to be rocovered

Anion-exchanger no activity to be recovered absorption to activated little activity recovered carbon absorption to 104r at low pH no solubilisation with detergents detergents (SDS, Na-desoxycholate, Brij 58) interfere with adhesion assay

Precipitation; ammoniumsulfate no activity precipitated

PEG-6000 no activity precipitated

PEG-400 no activity precipitated cold aceton no activity precipitated Lipid extraction activity in water phase Production late log/early stationary phase several LAB enteric origin media high acetate concentration required no production when grown in mucus production when grown on several C-sources (glucose, arabinose, galactose) dependant on biomass formation

Mechanism no inhibition adhesion to neutral mucus lipids (lipid receptor) no binding radioactively labelled supernatant no binding HRP labelled mucus to immobilised supernatant no reduction HPR activity from HPR labelled mucus no change in immobilised layer thickness

(ellipsometry) affects only specific protein peak in fractionated mucus active at 37°C, not 0°C no release of adhered

E. coli 1107 cells adhesion inhibition with both immobilised and soluble

Activity mucus against K88ab/ac/ad and S fimbriae References:

Bligh, E. G. & Dyer, W. J. (1959) A rapid method of total lipid extraction and purification, Canadian Journal of Biochemistry and Physiology, 37:911-917. Blomberg, L., Henriksson,A. & Conway, P.L. (1993a)

Inhibition of adhesion of Escherichia coli K88 to piglet ileal mucus by Lactobacillus spp., Applied and Environ¬ mental Microbiology, 59:34-39. Blomberg, L., Krivan, H. C, Cohen, P. S. & Conway, P. L. (1993b) Piglet ileal mucus contains protein and glyco- lipid (galactosylceramide) receptors specific for Escherichia coli K88 fimbriae, Infection and Immunity, 61:2526-2531. Conway, P. L., Welin, A. & Cohen, P. S. (1990) Presence of K88 specific receptors in porcine ileal mucus is age dependent, Infection and Immunity, 58:3178-3182. Dubois, M., Gilles, K. A., Hamilton, J.K., Rebers, P. A. & Smith, F. (1956) Colorimetric method for determination of sugars and related substances, Analytical Chemistry, 28:350-356.

Gaastra, W. & de Graaf, F. W. (1982) Host-specific fimbrial adhesins of noninvasive enterotoxigenic Escherichia coli strains, Microbiological Reviews, 46:129-161. Henriksson, A., Szewzyk, R. & Conway, P. L. (1991)

Characteristics of the adhesive determinants of Lacto¬ bacillus fermentum 104, Applied and Environmental Microbiology, 57:499-502. Jentoft, N. & Dearborn, D. G. (1979) Labelling of pro- teins by reductive methylation using sodium cyanoboro- hydride, Journal of Biological Chemistry, 11:4359-4365. Jonsson, E. & Conway, P. L. (1992) Development of probiotics for pigs, p. 260-316. In R. Fuller (ed.), probiotics. Chapman & Hall, Ltd., London. Kotarski, S. F. & Savage D. C. (1979) Models for study of the specificity by which indigenous lactobacilli adhere to murine gastric epithelia, Infection and Immunity, 26:966-975.

Laux, D.C., McSweegan, E. F., Williams, T. J. , Wadolkowski, E. A. & Cohen, P. S. (1986) Identification and characterization of mouse small intestine mucosal receptors for Escherichia coli K-12(K88), Infection and Immunity, 52:18-25.

Lindgren, S. E. & Dobrogosz, W. J. (1990) Antagonistic activities of lactic acid bacteria in food and feed fermentations, FEMS Microbiology Reviews, 87:149-164. Welin Klintstrδm, Ξ. (1992) PhD. thesis. Linkδping University, Linkδping, Sweden.

Claims

1. Products for inhibiting the adhesion of pathogens to gastrointestinal epithelial mucosa in animals includ¬ ing humans, c h a r a c t e r i s e d in that they con¬ tain high molecular weight carbohydrates produced by Lactobacillus strains of enteric origin and derivatives thereof.

2. Products according to claim 1, c h a r a c t e ¬ r i s e d in that they are produced by Lactobacillus strains originating from the gastrointestinal tract of the host to which the products will be administered.

3. Products according to claim l or 2, c h a r a c - t e r i s e d in that they have a molecular weight of approximately 200-1700 kDa and a composition of N-acetyl¬ glucose-amine:galactose:glucose in a ratio of 1:3:2.

4. Preparation for inhibiting the adhesion of patho¬ gens to gastrointestinal epithelial mucosa in animals including humans, c h a r a c t e r i s e d in that the active ingredient is chosen from the group consisting of viable Lactobacillus strains of enteric origin which produce high molecular weight carbohydrate products, the culture supernatant resulting from culturing of Lacto- bacillus strains of enteric origin which produce high molecular weight carbohydrate products, and purified, partly purified or chemically modified forms of, or sub¬ units or complexes with retained biological activity of, high molecular weight carbohydrate products produced by Lactobacillus strains of enteric origin.

5. Preparation according to claim 4, c h a r a c ¬ t e r i s e d in that the Lactobacillus strains are originating from the gastrointestinal tract of the host to which the preparation will be administered.

6. Preparation according to claim 4 or 5, c h a r a c t e r i s e d in that the high molecular weight carbohydrate products have a molecular wight of approximately 200-1700 kDa and comprises N-acetylglucose- amine:galactose:glucose in a ratio of 1:3:2.

7. A method of producing products for inhibiting the adhesion of pathogens to gastrointestinal epithelial mucosa in animals including humans, which products are high molecular weight carbohydrates, c h a r a c ¬ t e r i s e d in that a strain of Lactobacillus i cul- tured in complex media under semi-anaerobic growth condi¬ tions, whereby the optimal production of the products occurs in the late log phase and early stationary phase of the growth, and the products are recovered from the growth medium into which they are released during the culturing.