Formula with a specific beta-lactoglobulin peptide

Abstract

A formula comprising a whey protein hydrolysate comprising specific beta-lactoglobulin derived peptides and probiotics for inducing oral immune tolerance against milk protein, preventing or treating of oral immune intolerance against milk protein, and/or reducing the risk of developing oral immune intolerance against milk protein.

Claims

1.-30. (canceled)

31. A method for: inducing of oral immune tolerance against milk protein; and/or preventing or treating of oral immune intolerance against milk protein; and/or reducing the risk of developing oral immune intolerance against milk protein; and/or improving or enhancing of oral immune tolerance against milk protein, in a human subject, comprising administration to the subject a nutritional composition comprising a protein hydrolysate from mammalian milk, preferably milk from a species of the genus Bos, Bison, Bubalus or Capra, wherein the protein hydrolysate comprises at least one of the peptides having a sequence selected from the group consisting of SEQ ID NO: 2, SEQ ID NO: 3, SEQ ID NO: 4 and SEQ ID NO: 5, and wherein the composition comprises at least one strain of a lactic acid-producing, probiotic bacterium, and wherein the composition comprises less than 6 μg allergenic beta-lactoglobulin per g of total protein.

32. The method according to claim 31, wherein protein hydrolysate comprises at least one of the peptides having a sequence selected from the group consisting of SEQ ID NO: 2 and SEQ ID NO: 4, preferably at least one peptide having a sequence SEQ ID NO 4.

33. The method according to claim 31, wherein the composition comprises less than 3.5 μg allergenic beta-lactoglobulin, per g of total protein.

34. The method according to claim 31, wherein the composition comprises at least 95 wt %, hydrolysed whey protein based on total protein.

35. The method according to claim 31, wherein the composition comprises less than 10 wt % of peptides or proteins having a size of 5 kDa or above, based on total protein.

36. The method according to claim 31, wherein at least 1 wt % of peptides or proteins present in the composition has a size of 1 kDa or above, based on total protein.

37. The method according to claim 31, wherein the composition comprises an amount of allergenic beta-lactoglobulin of above 0.8 μg per g total protein.

38. The method according to claim 31, wherein the human subject is an infant.

39. The method according to claim 31, wherein the human subject is at risk of developing or suffering from milk protein allergy of milk from a species of the genus Bos, Bison, Bubalus or Capra, and wherein the human subject is an infant.

40. The method according to claim 31, wherein the composition comprises a strain of lactic acid-producing, probiotic bacterium.

41. The method according to claim 40, wherein the strain of lactic acid-producing, probiotic bacterium belongs to the genus Bifidobacterium.

42. The method according to claim 31, wherein the composition comprises one or more non-digestible oligosaccharide(s) selected from the group consisting of fructooligosaccharide, non-digestible dextrin, galactooligosaccharide, xylooligosaccharide, arabino-oligosaccharide, arabinogalacto-oligosaccharide, gluco-oligosaccharide, glucomanno-oligosaccharide, galactomanno-oligosaccharide, mannan-oligosaccharide, chito-oligosaccharide, uronic acid oligosaccharide, sialyloligosaccharide, and fucooligosaccharide, and mixtures thereof, preferably fructo-oligosaccharides.

43. The method according to claim 31, wherein the composition comprises long-chain polyunsaturated fatty acids.

44. The method according to claim 31, wherein the composition is an infant, follow-on formula or young child formula.

45. A nutritional composition comprising: a. a strain of lactic acid-producing, probiotic bacterium belonging to the genus Bifidobacterium; b. a milk protein hydrolysate from a species of the genus Bos, Bison, Bubalus or Capra, wherein the milk protein hydrolysate comprises at least one of the peptides having a sequence selected from the group consisting of SEQ ID NO: 2, SEQ ID NO: 3, SEQ ID NO: 4 and SEQ ID NO: 5; c. less than 6 μg allergenic beta-lactoglobulin, per g of total protein, d. less than 10 wt % of peptides or proteins having a size of 5 kDa or above, based on total protein, and e. at least 50 wt %, preferably at least 95 wt % of hydrolysed whey protein based on total protein, f. optionally one or more non-digestible oligosaccharide(s) selected from the group consisting of fructooligosaccharide, non-digestible dextrin, galactooligosaccharide, xylooligosaccharide, arabino-oligosaccharide, arabinogalacto-oligosaccharide, gluco-oligosaccharide, glucomanno-oligosaccharide, galactomanno-oligosaccharide, mannan-oligosaccharide, chito-oligosaccharide, uronic acid oligosaccharide, sialyloligosaccharide, and fucooligosaccharide, and mixtures thereof, and g. optionally long-chain polyunsaturated fatty acids, preferably docosahexaenoic acid (DHA), more preferably at least 0.35 wt. % DHA based on total fatty acids.

46. The nutritional composition according to claim 45, wherein protein hydrolysate comprises at least one of the peptides having a sequence selected from the group consisting of SEQ ID NO: 2 and SEQ ID NO: 4.

47. The nutritional composition according to claim 45, wherein the amount of allergenic beta-lactoglobulin is above 0.8 μg per g protein and/or wherein the composition comprises more than 1 wt % of peptides or proteins with a size of 1 kDa or above, based on total protein.

48. The nutritional composition according to claim 45, wherein the strain of lactic acid-producing, probiotic bacterium belongs to the species Bifidobacterium breve.

49. The nutritional composition according to claim 45, wherein the non-digestible oligosaccharides comprises at least two non-digestible oligosaccharide(s) selected from the group consisting of fructo-oligosaccharides and galacto-oligosaccharides.

50. The nutritional composition according to claim 45, which is an infant formula, follow-on formula or young child formula.

51. A method for providing nutrition to a human subject at risk of developing or suffering from allergy, more preferably milk protein allergy, comprising administering to the human subject the nutritional composition according to claim 45.

52. The method according to claim 51, wherein the human subject is an infant or young child.

Description

FIGURE LEGENDS

[0174] FIG. 1: Comparison of fragmentation spectra of a representative peptide (SEQ ID NO: 3; shown at the top) experimentally recorded (B) and its synthetic equivalent (A). The mass difference between the annotated fragment ions of the experimental and the synthetic peptide corresponds to the presence of a stable isotope labelled lysine in the synthetic peptide, italic in the sequence (+8 Da). The * indicates water loss fragments of the corresponding b-ions.

[0175] FIG. 2: T cell responses after stimulation with the identified peptides. Cow's milk-specific T cell lines (TCLs) of three different donors were stimulated with synthetic equivalents of the identified peptides (LIV . . . AAS=SEQ ID NO: 9; LIV . . . SLL=SEQ ID NO: 11; TMK . . . SLL=SEQ ID NO: 3; TMK . . . DAQ=SEQ ID NO: 4; DIQ . . . RVY=SEQ ID NO: 5). Cow's milk protein (CMP) was taken along as control. A stimulation index 2 was considered significant.

EXAMPLE 1: HYDROLYSED WHEY PROTEIN AND NEW METHOD FOR IDENTIFICATION OF BETALACTOGLOBULIN PEPTIDES

[0176] A mixture of acid whey protein and demineralised sweet whey protein (wt ratio protein 77:23) was dissolved in water (purified by reversed osmosis) and afterwards hydrolysed under specific conditions. Acid whey and sweet whey are commercially available. Sweet whey is the by-product of rennet-coagulated cheese and comprises caseinoglycomacropeptide (CGMP), and acid whey (also called sour whey) is the by-product of acid-coagulated cheese, and does not contain CGMP. Suitable sources for the acid whey protein are demineralised whey (Deminal, Friesland Campina, the Netherlands) and whey protein concentrate (WPC80, Friesland Campina, the Netherlands). A mixture of microbial endopeptidases and exopeptidases was used for hydrolysing these two protein sources using the method as described in example 1 of WO 2011/151059. Subsequently, the hydrolysed protein solution was spray dried. The size distribution of the peptides in this protein hydrolysate was determined by means of size exclusion high pressure liquid chromatography as known in the art. In short the total surface area of the chromatograms is integrated and separated into mass ranges expressed as percentage of the total surface area. The mass ranges are calibrated using peptides/proteins with a known molecular mass. The whey protein hydrolysate can be categorized as a partial or moderate protein hydrolysate.

[0177] The resulting hydrolysate powder was used as sole protein source in infant formulas. 10 different batches of infant fomulae were produced in a similar way. The amount of betalactoglobulin as determined by ELISA method as known in the art was between about 0.8 and 3.5 μg/g total protein.

[0178] In order to determine the presence of specific sequences in biological samples, mass spectrometry (MS) is the method of choice as opposed to the more traditional techniques that are currently used to characterize protein hydrolysates. A recent development in MS, termed peptidomics, allows for the characterization of peptide sequences with great sensitivity and specificity (Dallas et al. J Nutr 2015; 145:425-433). This technique is capable of closing the gap between understanding the impact of sequence specificity in relation to the biological activity a protein hydrolysate might have.

[0179] Samples were essentially prepared as described by Butré et al with the addition of a reduction and alkylation step (Butre et al. Anal Bioanal Chem 2014; 406):5827-5841). All chemicals were obtained from Sigma Aldrich. Briefly, the partial Hydrolysate protein (pHP) batches were diluted to 0.5% (v/v) using 50 mM ammonium bicarbonate followed by the reduction of peptides with 4 mM DTT and alkylation with 8 mM iodoacetamide. The mixture was cleared by centrifugation at 20,000×g for 10 min and diluted to 0.1% (v/v) using 0.1 M acetic acid.

[0180] All samples were analysed by nanoflow liquid chromatography using an Agilent 1200 HPLC system (Agilent Technologies) coupled on-line to a LTQ Velos mass spectrometer (Thermo Fisher Scientific). The liquid chromatography part of the system was operated in a setup essentially as described previously (14). Peptides were trapped at 5 μl/min in 100% solvent A (0.1 M acetic acid in water) on a 2-cm trap column (100-μm inner diameter, packed in-house using Aqua C18, 5-μm resin (Phenomenex)) and eluted to a 20-cm IntegraFrit column (50-μm inner diameter, ReproSil-Pur C18-AQ 3-μm, New Objective) at ˜100 nl/min in a 90-min gradient from 10 to 40% solvent B (0.1 M acetic acid in 8:2 (v/v) acetonitrile/water). The eluent was sprayed via standard coated emitter tips (New Objective) butt-connected to the analytical column. The mass spectrometer was operated in data-dependent mode, automatically switching between MS and MS/MS. Full scan mass spectra (from m/z 300 to 1,200) were acquired at zoom scan rate after accumulation to a target value of 3,000. The five most intense ions at a threshold above 500 were selected for collision-induced at normalized collision energy of 35% after accumulation to a target value of 10,000.

[0181] All MS data was processed by Proteome Discoverer (version 2.1, Thermo Scientific). Peak lists were generated using a standard workflow. Peptide identification was performed by searching individual peak lists of CID fragmentation spectra against a database containing selected bovine whey and casein proteins using Mascot (version 2.4.1, Matrix Science). No enzyme was specified and no missed cleavages were allowed. Precursor ion mass tolerance was set to 0.2 Da and product ion mass tolerance to 0.5 Da. Carbamidomethylation (C) was set as fixed modification.

[0182] Of the 314 peptides identified by this approach, 101 could be assigned to betalactoglobulin (BLG) leading to a sequence coverage of BLG of 90%. Most of the remaining peptides originated from other abundant whey proteins such as α-lactalbumin and serum albumin. In total 13 betalactoglobulin peptides of minimal 9 AAs were identified in the region of interest (AA #13-48) (See Table 1 and FIG. 1). Six of them were identified consistently in all 10 batches. Unambiguous peptide identifications were obtained by comparing characteristics (retention time, peptide mass and fragmentation spectrum) of the experimental peptide with its stable-isotope labelled synthetic equivalent. An example is shown in FIG. 1 where the fragmentation spectra of the experimental and synthetic peptide show excellent correlation. In this way the identity of the six peptides found in all 10 batches was confirmed.

TABLE-US-00002 TABLE 1 Sequences of the beta-lactoglobulin peptides identified in 10 different batched of a partially hydrolysed whey-based infant formula Number of batches where the Amino peptide acid was number identified SEQ in BLG to be ID sequence Sequence present NO AA 1-22 LIVTQTMKGLDIQKVAGTWYSL 7 6 AA 1-25 LIVTQTMKGLDIQKVAGTWYSL 9 7 AMA AA 1-26 LIVTQTMKGLDIQKVAGTWYSL 10 8 AMAA AA 1-27 LIVTQTMKGLDIQKVAGTWYSL 10 9 AMAAS AA 1-30 LIVTQTMKGLDIQKVAGTWYSL 2 10 AMAASDIS AA 1-32 LIVTQTMKGLDIQKVAGTWYSL 10 11 AMAASDISLL AA 1-35 LIVTQTMKGLDIQKVAGTWYSL 5 12 AMAASDISLLDAQ AA 1-39 LIVTQTMKGLDIQKVAGTWYSL 2 13 AMAASDISLLDAQSAPL AA 1-42 LIVTQTMKGLDIQKVAGTWYSL 2 14 AMAASDISLLDAQSAPLRVY AA 5-35 QTMKGLDIQKVAGTWYSLAMAA 2 15 SDISLLDAQ AA 6-32 TMKGLDIQKVAGTWYSLAMAAS 10 3 DISLL AA 6-35 TMKGLDIQKVAGTWYSLAMAAS 10 4 DISLLDAQ AA 11-42 DIQKVAGTWYSLAMAASDISLL 10 5 DAQSAPLRVY

[0183] A similar analysis was performed with batches of Nutrilon pepti, an extensively hydrolysed whey protein containing infant formula marketed for the dietary management of cow's milk allergy. The BLG peptides of Table 1 were not present in Nutrilon pepti.

EXAMPLE 2 IN VITRO IDENTIFICATION OF HLA-DR-RESTRICTED PEPTIDES

[0184] Identification of BLG-derived peptides presented by human DCs was performed by ProImmune, as described (Lamberth et al, Sci Transl Med 2017 11; 9(372):10.1126/scitranslmed.aag1286.). Briefly, peripheral blood mononuclear cell samples from 12 HLA-DR-typed healthy adult donors were obtained. The donors were selected based on common 11 HLA-DRB1 alleles (Table 2). The HLA-DRB1 was prioritized in this study due to the facts that the HLA-DR molecule is the most predominant human MHC class II isotypes (>90%) (Sturniolo et al. Nat Biotechnol 1999; 17:555-561.) and that HLA-DRB1 gene locus is polymorphic while HLA-DRA1 gene locus is monomorphic (i.e., HLA-DRB1 genotype determines the whole HLA-DR molecule) (Marsh et al, 2010. Tissue Antigens 75:291-455.). Furthermore, HLA-DRB1 allele is expressed five times higher than its paralogs (HLA-DRB3, -DRB4 or -DRB5) and is present in all individuals (O'Leary et al. Nucleic Acids Res 2016 4; 44(D1):D733-45.).

TABLE-US-00003 TABLE 2 Donors with HLA-DRB1 typing information Donor ID DRB1_1 DRB1_2 P1 *03:01 *03:01 P2 *01:01 *04:01 P3 *01:01 *07:01 P4 *01:01 *04:05 P5 *04:01 *07:01 P6 *13:02 *14:01 P7 *03:01 *09:01 P8 *11:01 *15:01 P9 *03:01 *11:01 P10 *07:01 *15:01 P11 *03:01 *15:01 P12 *01:01 *04:04

[0185] Immature monocyte-derived DCs were generated in vitro and matured in the presence of tested hydrolysed protein (HP). DCs were harvested and lysed in order to obtain HLA-DR complexes by using a specific immunoaffinity method. Peptides were eluted from the HLA-DR complexes and subsequently analyzed by high resolution sequencing LC-MS/MS. The presence of six endogenous relevant proteins (i.e., ITGAM, ApoB, CLIP, TFRC, FcER2/FcGR2 and LAMP-1/3) was assessed as a control for this assay. Each donor sample had to express a minimum of 3 relevant proteins to be qualified for subsequent analysis. The resulting data of HLA-DR-restricted peptides were compiled and assessed using sequence analysis software referencing the Swiss-Prot Human Proteome Database with the incorporated test item sequences. The likelihood of peptides to be real findings is described by their expect value 0.05. The false discovery rate was determined to be <1%.

[0186] Using the ProPresent® antigen presentation assay by focusing on BLG-derived sequences, 15 relevant peptides with an expect value 0.05 were identified within 5 donor samples (Table 3). These peptides could be further grouped into 2 unique sequence groups, i.e., DIQ . . . DIS (AA #11-30) and AMA . . . APL (AA #23-39) (Table 4). Importantly, both sequence groups overlapped with the region of interest (i.e., AA #13-48 of mature BLG). This finding demonstrated that BLG-derived peptides of interest can be presented by HLA-DR molecules on human DCs when incubated with a specific HP.

TABLE-US-00004 TABLE 3 Significant fragments identified from peptide-HLA-DR complexes with expect value <0.05. Donor Sequence of Fragment Amino Acid Expect ID DRB1_1 DRB1_2 (Mature BLG) [SEQ ID NO] Start/End Value P2 *01:01 *04:01 AMAASDISLLDAQSAPL-[16] 23-39 0.043 -MAASDISLLDAQSAPLR [17] 24-40 0.0013 P3 *01:01 *07:01 -MAASDISLLDAQSAPL-[18] 24-39 0.009 P4 *01:01 *04:05 DIQKVAGTWYSLAMAASDI-[19] 11-29 0.000052 ----VAGTWYSLAMAASDIS [20] 15-30 0.00021 ------GTWYSLAMAASDIS [21] 17-30 0.00037 P7 *03:01 *09:01 DIQKVAGTWYSLAMAASDI--[19] 11-29 0.00019 DIQKVAGTWYSLAMAASDIS-[22] 11-30 0.00053 ----VAGTWYSLAMAASDI--[23] 15-29 0.002 ----VAGTWYSLAMAASDIS-[20] 15-30 0.000019 -----AGTWYSLAMAASDIS-[24] 16-30 0.0014 ------GTWYSLAMAASDI--[25] 17-29 0.00023 ------GTWYSLAMAASDIS-[21] 17-30 0.000015 P12 *01:01 *04:04 AMAASDISLLDAQSAPL-[16] 23-39 0.028 -MAASDISLLDAQSAPL-[18] 24-39 0.042

TABLE-US-00005 TABLE 4 HLA-DR allele association on unique region of mature BLG with significant fragments were defined by having expect value <0.05. Number of Significant Unique Region Fragments and Donor (Mature BLG) Amino acid individual ID DRB1_1 DRB1_2 [SEQ ID NO] start-end fragment P2 *01:01 *04:01 AMAASDISLLDAQSAPLR [39] 23-40 2 (23-39 and 24-40) P3 *01:01 *07:01 MAASDISLLDAQSAPL [18] 24-39 1 (24-39) P4 *01:01 *04:05 DIQKVAGTWYSLAMAASDIS [22] 11-30 3 VAGTWYSLAMAASDIS [20] (11-29, 15-30, GTWYSLAMAASDIS [21] and 17-30) P7 *03:01 *09:01 DIQKVAGTWYSLAMAASDIS [22] 11-30 7 (11-29, 11-30, 15-29, 15-30, 16-30, 17-29, 17-30) P12 *01:01 *04:04 AMAASDISLLDAQSAPL [16] 23-39 2 (23-39, 24-39)

[0187] Not all donor antigen presenting cells bound to peptides. This concerned HLA-DRB1*03:01 in homozygous donor P1. Similar arguments could be applied for donors P6 and P11.

EXAMPLE 3 MHC CLASS II BINDING IN SILICO ASSESSMENT

[0188] A limitation of the ProPresent® antigen presentation assay is the usage of general anti-HLA-DR antibody, and not a specific antibody against a particular HLA-DRB1 allele (e.g., anti-DRB1*01:01 antibody), to isolate peptide-HLA-DR complexes of interest. Hence, with a donor with heterozygous genotypes of HLA-DRB1, e.g., *01:01 and *04:01, it is uncertain which allele will present the identified peptide.

[0189] Hence, the identified HLA-DR-restricted peptides were computed into the IEDB MHC Class II Binding Prediction software (http://tools.iedb.org/mhcii/) in order to assess binding of discovered peptides to the selected HLA-DRB1 alleles, as described (Wang P, et al. PLoS Comput Biol 2008 4; 4(4):e1000048.; Wang et al, BMC Bioinformatics 2010, 11:568-2105-11-568.) The default IEDB recommended prediction method was selected. For each peptide sequence (15 mers long), a percentile rank was generated by comparing the peptide's score against the scores of five million random 15 mers selected from the Swiss-Prot database. A lower percentile rank indicates a higher affinity of peptide binding to a particular MHC class II allele, in which the IEDB recommends to make a selection based on a consensus percentile rank of the top 10%. In order to be more stringent with the in silico assessment, a selection based on a consensus percentile rank of the top 3% was made.

[0190] The MHC class II prediction software was utilized to assess whether fragments of the identified two unique sequence groups of BLG would have high affinity to bind to the selected HLA-DRB1 alleles of the characterized donors from the ProPresent® assay, i.e., HLA-DRB1 *01:01, *03:01, *04:01, *04:04, *04:05, *07:01 and *09:01. As shown in Table 5, with an arbitrary threshold of percentile rank set at <3% (i.e., top 3% high binders), fragments of DIQ . . . DIS (AA #11-30) were predicted to have high affinity to bind to 5 HLA-DRB1 alleles, i.e., DRB1*01:01, *04:01, *04:04, *04:05 and *09:01. In contrast, it was estimated that only one fragment of AMA . . . APL (AA #23-39) bound to HLA-DRB1*07:01 with high affinity. Taken together, these findings suggest that fragments from AA #11-30 had a higher likelihood to be presented as T-cell epitopes than the ones from AA #23-39. Moreover, the data suggest that several common HLA-DRB1 alleles could present fragments derived from the identified two BLG-derived unique sequence groups.

[0191] These unique sequence groups of BLG (AA #11-30 and #23-39) overlapped and could be joined into one sequence (i.e., AA #11-39). This joined sequence was highly correlated with one constantly identified peptide in the tested HP (AA #11-42; DIQ . . . RVY). Therefore the sequence of AA #11-42 was subjected into the MHC class II prediction software. Importantly, the predicted complexes of AA #11-42 reconfirmed the predicted results of two in vitro identified unique sequences (Table 6), suggesting that both unique sequences of BLG could be derived from the AA #11-42 peptide that exists in the tested formula. Furthermore, parts of AA #11-42 were predicted to have high affinity to the same sets of identified HLA-DRB1 alleles, i.e., DRB1*01:01, *04:01, *04:04, *04:05, *07:01 and *09:01. In conclusion, the in silico assessment confirms the in vitro findings that BLG-derived peptides can bind to common HLA-DRB1 alleles.

TABLE-US-00006 TABLE 5 In silico assessment on potential peptide-HLA-DRB1 complexes of two identified BLG-derived peptides with percentile rank <3%. Predicted 15-mer Fragment of DIQKVAGTWYSLAMAASDIS (SEQ ID NO: 25) or AMAASDISLLDAQSAPL HLA-DRB1 (SEQ ID NO: 19) Percentile Allele [SEQ ID NO] Rank HLA-DRB1*09:01 AGTWYSLAMAASDIS [24] 0.86 HLA-DRB1*04:05 AGTWYSLAMAASDIS [24] 1.38 HLA-DRB1*04:05 VAGTWYSLAMAASDI [23] 1.40 HLA-DRB1*09:01 VAGTWYSLAMAASDI [23] 1.56 HLA-DRB1*04:05 KVAGTWYSLAMAASD [26] 2.26 HLA-DRB1*01:01 AGTWYSLAMAASDIS [24] 2.37 HLA-DRB1*04:04 AGTWYSLAMAASDIS [24] 2.44 HLA-DRB1*09:01 KVAGTWYSLAMAASD [26] 2.53 HLA-DRB1*04:01 AGTWYSLAMAASDIS [24] 2.84 HLA-DRB1*07:01 AASDISLLDAQSAPL [27] 2.39

TABLE-US-00007 TABLE 6 In silico assessment on potential peptide-HLA-DRB1 complexes of a BLG-derived long peptide with percentile rank <3 (top 3% binders). Predicted 15-mer Fragment of DIQKVAGTWYSLAM HLA-DRB1 AASDISLLDAQSAPLRVY Percentile Allele (SEQ ID NO: 5) [SEQ ID NO] Rank HLA-DRB1*09:01 GTWYSLAMAASDISL [28] 0.65 HLA-DRB1*09:01 AGTWYSLAMAASDIS [24] 0.86 HLA-DRB1*04:05 GTWYSLAMAASDISL [28] 1.30 HLA-DRB1*09:01 TWYSLAMAASDISLL [29] 1.32 HLA-DRB1*04:05 AGTWYSLAMAASDIS [24] 1.38 HLA-DRB1*04:05 VAGTWYSLAMAASDI [23] 1.40 HLA-DRB1*04:05 TWYSLAMAASDISLL [29] 1.40 HLA-DRB1*09:01 VAGTWYSLAMAASDI [23] 1.56 HLA-DRB1*09:01 WYSLAMAASDISLLD [30] 2.12 HLA-DRB1*09:01 YSLAMAASDISLLDA [31] 2.14 HLA-DRB1*01:01 GTWYSLAMAASDISL [28] 2.18 HLA-DRB1*04:05 KVAGTWYSLAMAASD [26] 2.26 HLA-DRB1*01:01 AGTWYSLAMAASDIS [24] 2.37 HLA-DRB1*07:01 AASDISLLDAQSAPL [27] 2.39 HLA-DRB1*07:01 SDISLLDAQSAPLRV [32] 2.44 HLA-DRB1*04:04 AGTWYSLAMAASDIS [24] 2.44 HLA-DRB1*04:04 GTWYSLAMAASDISL [28] 2.44 HLA-DRB1*04:04 TWYSLAMAASDISLL [29] 2.44 HLA-DRB1*01:01 TWYSLAMAASDISLL [29] 2.46 HLA-DRB1*09:01 KVAGTWYSLAMAASD [26] 2.53 HLA-DRB1*04:04 YSLAMAASDISLLDA [31] 2.57 HLA-DRB1*04:01 GTWYSLAMAASDISL [28] 2.83 HLA-DRB1*04:01 AGTWYSLAMAASDIS [24] 2.84 HLA-DRB1*04:04 WYSLAMAASDISLLD [30] 2.91

[0192] Several in silico assessments confirmed the in vitro ProPresent® findings on the peptide-HLA-DR complexes. However, we observed that some in silico predicted complexes did not correlate to the ones found in the ProPresent® assay, e.g., DRB1*04:04 with AGT . . . DIS. This discrepancy can be explained either by a small size of tested cohort in the ProPresent® assay (n=12), by the fact that each donor had a different combination of HLA-DRB1 allele types (i.e., interindividual variation) and/or by a tendency of in silico assessment to be over predictive (thus requires a stringent cut-off). Therefore, it is prudent to combine both in vitro and in silico approaches in order to identify peptide-MHC class II complexes.

EXAMPLE 4 T-CELL PROLIFERATION ASSAY

[0193] Synthetic peptides were obtained from JPT technologies. Only peptides containing at least 9 AAs were taken into account, as this is the minimal size for binding to MHC class II molecules and subsequent T-cell recognition (Holland et al, Front Immunol 2013 Jul. 1; 4:172.) Synthetic peptides (JPT technologies) identical to the peptides identified in all batches of the HP of example 1 were dissolved in dimethylsulfoxide (Sigma Aldrich) at a concentration of 5.3 mM. The peptides were tested on cow's milk-specific T-cell lines (TCLs) from three different donors. The peptides were tested on cow's milk protein-specific T-cell lines (TCLs) from three infant donors, diagnosed with cow's milk protein allergy, aged <1 year, 7.5 months and 6 years old. These TCLs were generated previously and have been shown to recognize epitopes in the region of interest (AA #13-48 of mature BLG). Proliferation was determined as described previously (Ruiter et al, Clin Exp Allergy 2006, 36(3):303-310). Stimulation indexes (Sls, ratio between proliferation of allergen/peptide-stimulated and non-stimulated T cells) were calculated and a SI≥2 was considered positive.

[0194] To confirm that the peptides identified in HP were recognized by T cells, synthetic peptides identical to these identified peptides were tested on cow's milk-specific human TCLs. As there was a considered overlap between the identified peptides, five peptides were tested (Table 7) that were identified in all batches and differed with more than 9A from each other. All tested peptides were able to induce proliferation (FIG. 2). However, each donor showed a different recognition pattern. This was also seen with the in silico prediction data (data not shown). All five peptides induced proliferation of TCL B suggesting that this TCL recognizes either the overlapping part of the peptides (AA #11-27) or multiple T-cell epitopes in this region. Also TCL A showed a proliferative response after stimulation with several peptides. As peptide LIV . . . AAS (AA #1-27) was not able to induce proliferation in TCL A while peptide LIV . . . SLL (AA #1-32) did, the region containing AA #28-32 (DISLL) was essential for the response of TCL A. Unexpectedly, TCL C recognized peptide LIV . . . AAS (AA #1-27), while the longer peptide containing the same sequence with 5 additional AAs, LIV . . . SLL (AA #1-32), was not able to induce a proliferative response indicating that longer peptides are not always presented better.

TABLE-US-00008 TABLE 7 Sequences of the beta-lactoglobulin peptides identified in 10 different batched of a partially hydrolysed whey-based infant formula and tested for T cell proliferation activity Amino acid number SEQ in BLG ID sequence Sequence NO AA 1-27 LIVTQTMKGLDIQKVAGTWYSLAMAAS 9 AA 1-32 LIVTQTMKGLDIQKVAGTWYSLAMAASDISLL 11 AA 6-32 TMKGLDIQKVAGTWYSLAMAASDISLL 3 AA 6-35 TMKGLDIQKVAGTWYSLAMAASDISLLDAQ 4 AA 11-42 DIQKVAGTWYSLAMAASDISLLDAQSAPLRVY 5

[0195] Not all peptides were recognized by each donor, which is an indicative for the variation between donors. Each donor expressed a different panel of MHC class II molecules. To determine if and which HLA-DRB1 allele expressed by the donors presented the peptides the same algorithm as described above was used. The in silico prediction confirmed that HLA-DRB1 alleles (*11:01 & *04:04) expressed by donor B were able to present all peptides. Based on the algorithm data TCL A should only recognize peptide DIQ . . . RVY (AA #11-42), however, this TCL showed a proliferative response after stimulation with four of the five peptides. A possible explanation might be that the other three peptides were not presented by HLA-DRB1 but by other MHC class II molecules.

[0196] In conclusion, example 1 to 4 demonstrate that a specific whey protein hydrolysate contains functional T-cell epitopes. Specific, beta-lactoglobulin peptides were identified to be present, that are HLA-DRB1 restricted peptides. These are presented to and recognized by T cells from different donors, including healthy donors and cow's milk protein allergic infants. This interaction was confirmed by in silico analysis. The HP may therefore stimulate the development of oral immunotolerance to protein.

EXAMPLE 5: SYNBIOTIC MIXTURE SHOWS A FURTHER IMPROVED EFFECT OF TOLERANCE INDUCTION COMPARED TO PROBIOTICS OR PREBIOTICS ALONE

[0197] Several published studies support a notion that dietary antigens (including peptides) within the intestinal lumen can be absorbed and subsequently captured and presented by intestinal antigen-presenting cells, i.e., antigen sampling mechanisms, without the need of a disrupted intestinal barrier. There are at least two distinct mechanisms that continuously work to sample dietary antigens at the healthy state, i.e., microfold/M cell-mediated transcytosis and goblet-cell-associated antigen passage. Both pathways provide dietary antigens to intestinal lamina propria CD103+ DCs, which in return could imprint gut-homing molecules on T and B cells, could promote differentiation of intestinal IgA-producing plasma cells as well as could generate intestinal Tregs, a key player in the state of hyporesponsiveness to fed antigen known as oral tolerance. This suggests that peptides within the tested HP can be orally absorbed, followed by antigen capture and presentation by intestinal CD103+ DCs and subsequently culminated in generation of functional Tregs.

[0198] For the development of oral tolerance, T-cell epitopes within the tested HP need to be presented under the right circumstances. In addition to TCR activation, co-stimulation and cytokine signalling play an important role in the generation of Tregs. Pro- or prebiotics play an important role in creating the right environment for this predisposition towards Tregs. The preventive effect of synthetic BLG peptides was strengthened in combination with a probiotic or prebiotic diet.

[0199] To demonstrate the effect of pro-, pre- and synbiotics an experiment was performed similar as described in example 4 of WO 2016/148572, with the same protocol, and the same concentration and mixture of synthetic peptides (‘Pepmix’), the pepmix identified in Table 8.

TABLE-US-00009 TABLE 8 pepmix according to WO 2016/148572 Peptide Sequence 1 (18AA) SEQ ID NO: 33 QKVAGTWYSLAMAASDIS 2 (18AA) SEQ ID NO: 34 WYSLAMAASDISLLDAQS 3 (18AA) SEQ ID NO: 35 AASDISLLDAQSAPLRVY 4 (18AA) SEQ ID NO: 36 LLDAQSAPLRVYVEELKP

[0200] A number of diets were compared: [0201] (a) synbiotic diet of example 1 of WO 2016/14472 (1 wt % scFOS/lcFOS in a wt ratio 9:1+2 wt % 2×10.sup.9 cfu/g Bifidobacterium breve M-16V, [0202] (b) probiotic diet of example 4 of WO 2016/148572 with 2 wt % 2×10.sup.9 cfu/g Bifidobacterium breve M-16V, and [0203] (c) prebiotic diet containing 1 wt % of scFOS/lcFOS wt ratio 9/1.

[0204] The short-chain (Sc-) and long-chain (lc-) fructo-oligosaccharides (FOS) were commercially available as Raftilose P95 (Orafti) and Raftiline HP, respectively.

[0205] All groups were pretreated with the pepmix and one of the above diets, were sensitized with whey+cholera toxin and challenged with whey.

[0206] After challenge with whey, the delta ear swelling of the mice having consumed (c) scFOS/lcFOS diet was about 174 μm, of the mice having consumed (b) the probiotic diet was about 158 μm, and of the mice having consumed (a) the synbiotic diet was about 112 μm. The ear swelling in the synbiotic group (a) was statistically significantly lower than in the corresponding scFOS/lcFOS group (c), and there was a trend of a lower response when compared with probiotic group (b) (p=0.08). These results are indicative for a further improved effect on oral immunotolerance induction: [0207] when using a lactic acid-producing bacterium, in particular a Bifidobacterium strain from the species B. breve, and [0208] particularly when using a lactic-acid producing bacterium in combination with NDOs.

EXAMPLE 6: MHC BINDING DOMAINS IN THE NATURAL, IDENTIFIED PEPTIDES OF THE HYDROLYSATE VS SYNTHETIC BETA-LACTOGLOBULIN PEPTIDES

[0209] Method: The sequences of the beta-lactoglobulin peptides that were identified in hydrolyzed infant formula (ID_PEP) and the sequences of the synthetic beta-lactoglobulin peptides that have been published in WO2016/148572 (SYN_PEP, Table 9) were tested in silico with the IEDB MHC Class II Binding Prediction Software (http://tools.iedb.org/mhcii/) to determine the number of MHC Class II binding domains in each peptide. In the software the IEDB recommended prediction method and the full HLA reference set were selected. This HLA reference set provides a population coverage of >99%. The software calculates a percentile rank by comparing the binding affinity of the predicted binding domain to a large set of peptides. The higher affinity, the lower the percentile rank. In this experiment, the top 2% binders were selected.

[0210] Results

[0211] The number of MHC class II binding domains in the identified peptides is higher than in synthetic peptides (see Table 9). The software predicted at least one binding domain for all identified peptides and in total six different domains. For the synthetic peptides four different domains were predicted. However, these domains were all derived from one sequence, namely SYN_PEP_1. No MHC Class II binding domains were predicted for the other two synthetic peptides.

TABLE-US-00010 TABLE 9 Number of MHC class II binding domains per betalactoglobulin peptide predicted IEBD Number of MHC class II binding Peptide Sequence domains  ID_PEP_1 LIVTQTMKGLDIQKVAGTWYSLAMAAS 1 ID_PEP_2 LIVTQTMKGLDIQKVAGTWYSLAMAASDIS 6 LL ID_PEP_3 TMKGLDIQKVAGTWYSLAMAASDISLL 6 ID_PEP_4 TMKGLDIQKVAGTWYSLAMAASDISLLDAQ 6 ID_PEP_5 DIQKVAGTWYSLAMAASDISLLDAQSAPLR 6 VY SYN_PEP_1 QKVAGTWYSLAMAASDIS 4 SYN_PEP_2 WYSLAMAASDISLLDAQS 0 SYN_PEP_3 AASDISLLDAQSAPLRVY 0

[0212] The number of MHC Class II binding domains is higher in the identified peptides than in the synthetic peptides. In other words, more T cell epitopes can be derived from the identified peptides than from the synthetic peptides. Therefore, the likelihood that the identified peptides will be presented to T cells is higher. This presentation to T cells is a requirement for oral immune tolerance induction.

EXAMPLE 7: HLA-DR EXPRESSION ON DENDRITIC CELLS AFTER MATURATION WITH A LACTIC ACID PRODUCING BACTERIUM

[0213] Method: Monocytes (CD14+ cells) were cultured for 7 days with GM-CSF and IL-4 to generate immature dendritic cells (DCs). After 7 days, the immature DCs were washed and incubated with either medium, LPS (100 ng/ml) or Bifidobacterium breve M-16V (Morinaga) in a 10:1 bacteria:DC ratio. After 48 hrs, the mature DCs were harvested and stained with APC-Cyanine7 labelled anti-human HLA-DR antibody to determine HLA-DR expression on the surface of the cells. The samples were analyzed on a BD FACS Canto flow cytometer with FACS DIVA software. Expression of HLA-DR is presented in mean fluorescence intensity (MFI).

[0214] Results: Maturation with Bifidobacterium breve M-16 v (M16 v-DCs) increases the expression of HLA-DR to a similar level as LPS (LPS-DCs, see Table 10).

TABLE-US-00011 TABLE 10 The expression of HLA-DR on different subtypes of dendritic cells DCs HLA-DR expression MFI (SD) Immature DCs 11338 (1919, 1) LPS-DCs 24676 (9694, 4) M16v-DCs 21714 (1516) DCs = dendritic cells, MFI = mean fluorescence intensity, SD = standard deviation

[0215] Bifidobacterium breve M-16V increases the expression of HLA-DR molecules on the surface of the dendritic cells. An increased expression of HLA-DR is indicative for an increased presentation of peptides to T cells.

EXAMPLE 8: INFANT FORMULA

[0216] A powdered infant formula packed with instructions that it should be reconstituted with water of 40° C. with 3 scoops of powder (13.74 g) added to 90 ml water, giving a final volume of 100 ml. The product is suitable from birth up to 6 months.

[0217] Per 100 ml the infant formula contains 66 kcal, 1.5 g protein (the hydrolysed whey protein of example 1), 3.3 g fat (vegetable oils, single cell oil and fish oil, comprising 0.4 wt % DHA and 0.35 wt % ARA based on total fatty acids), 7.2 g digestible carbohydrates (mainly lactose), 0.8 g non-digestible oligosaccharides (a 9/1 w/w mix of scGOS and lcFOS), Bifidobacterium breve M-16V in an amount of about 3*10.sup.7 cfu/g, and vitamins, minerals, trace elements and other micronutrients according to international directives for infant formulae.