Reduction of in vitro genotoxicity of pollen extracts by removal of flavonoids

Abstract

The invention relates to grass pollen extracts containing reduced amount of flavonoid glycosides in order to minimize the risks of genotoxicity of the grass pollen extracts. The invention also relates to a method of preparing grass pollen extracts containing reduced amount of flavonoid glycosides by ultrafiltration. Flavonoid glycosides are naturally present in grass pollen extracts and they have been identified as being responsible for the formation of flavonoid aglycones, which are genotoxic in vitro, under the influence of enzymes contained in the grass pollen extracts.

Claims

1-4. (canceled)

5. A purified pollen extract containing pollen allergens wherein each flavonoid is contained in amount which is less than 0.002 g per 100 g of pollen derived fraction, as expressed in aglycone equivalent.

6. The purified pollen extract according to claim 5, wherein said pollen extract is a mixture of pollen extracts consisting of, or comprising, cocksfoot, meadow grass, rye-grass, sweet vernal-grass and timothy grass extracts.

7. A pharmaceutical composition which comprises a purified pollen extract containing pollen allergens wherein each flavonoid is contained in amount which is less than 0.002 g per 100 g of pollen derived fraction, as expressed in aglycone equivalent.

8. The pharmaceutical composition according to claim 7, wherein said pollen extract is a mixture of pollen extracts consisting of, or comprising, cocksfoot, meadow grass, rye-grass, sweet vernal-grass and timothy grass extracts.

9. A method of treating pollen allergy comprising administering to a subject in need thereof a purified pollen extract containing pollen allergens wherein each flavonoid is contained in amount which is less than 0.002 g per 100 g of pollen derived fraction, as expressed in aglycone equivalent.

10. The method according to claim 9, wherein said pollen extract is a mixture of pollen extracts consisting of, or comprising, cocksfoot, meadow grass, rye-grass, sweet vernal-grass and timothy grass extracts.

11. A purified allergen extract which is obtainable by a method comprising a step of ultrafiltration of an aqueous allergen extract on a 1-10 kDa membrane with at least 5 volumes of an ammonium bicarbonate solution, and a step of sterile filtration on a 0.22 μm filter.

12. A purified allergen extract according to claim 11, wherein ultrafiltration of said allergen extract in the method of preparing the purified allergen extract is performed on a 5-10 kDa membrane with 10 to 30 volumes of a 0.4 g/L to 0.8 g/L ammonium bicarbonate solution.

13. A purified allergen extract according to claim 11, wherein said allergen is selected from the group consisting of a tree pollen allergens, grass pollen allergens, herb pollen allergens, weed pollen allergens, mite allergens, venom allergens, animal hair and dandruff allergens and food allergens.

14. A purified allergen extract according to claim 11, wherein the method of preparing the allergen extract further comprises a step of formulating said allergen extract into a pharmaceutical composition.

15. A pharmaceutical composition obtainable by a method comprising: ultrafiltration of an aqueous allergen extract on a 1-10 kDa membrane with at least 5 volumes of an ammonium bicarbonate solution, and a step of sterile filtration on a 0.22 μm filter; and formulating said allergen extract into a pharmaceutical composition.

Description

FIGURES

[0094] FIG. 1 shows cumulated area of the peaks corresponding to the different flavonoids found in a grass pollen extracts before (0 h) and after a 3-h and a 24-h incubation at 37° C.

[0095] FIG. 2 illustrates the amount of flavonoids contained in a concentrated extract of 5 grass pollens before and after ultrafiltration with 2.5 volumes (as per the previous non-optimized manufacturing process) and 15 volumes (as per the optimized manufacturing process) of 120 ppm ammonium bicarbonate. Results are expressed as the sum of the integrated surfaces of peak 1 (RT 14.83 min) and peak 2 (RT 15.00 min).

EXAMPLES

Example 1: Raw Materials and Initial Process (i.e. Non Modified Process) of Preparing Pollen Extract

[0096] Raw materials consisted of defatted pollen of cocksfoot, meadow grass, rye-grass, sweet vernal-grass and timothy grass.

[0097] The process of purification used to prepare purified pollen extract, or purified mixture of pollen extracts typically comprise the steps of: [0098] optionally mixing pollens originating from different species, if a mixture of pollens is intended to be purified; [0099] extracting the pollens by contacting the pollens with extraction solution, typically an aqueous solution such as distilled water or a buffered distilled water solution; [0100] separating the aqueous phase from the solid phase, for instance by centrifugation, to recover the aqueous phase which contains the allergens extracted from the pollen (pollen extract); [0101] clarifying by filtration the pollen extract; [0102] concentrating the pollen extract by passage on a 1-kDa or 5-kDa membrane; [0103] submitting the retentate to ultrafiltration on a 1-kDa membrane with washing with 2.5 volumes of purified water; and [0104] filtrating through a 0.22-μm filter to sterilise the purified pollen extract.

[0105] The purified pollen extract is then typically freeze-dried and formulated into an appropriate pharmaceutical composition, e.g. a tablet.

[0106] Pollen raw materials and pollen extracts prepared according to the above process have been found to be genotoxic in a Mouse Lymphoma Assay (MLA)/TK test when using a continuous 24-h treatment without metabolic activation (“S9-”).

Example 2: Quantification of External Genotoxic Contaminants

[0107] Four groups of genotoxic elements from the agroenvironment have been assessed in grass pollen raw materials, as summarized below in table 1:

TABLE-US-00001 TABLE 1 Genotoxic contaminants tested in raw materials tablets Class of potential Quantification contaminants method Samples tested Results Outcomes Artificial Gamma 2 batches (from 2 No activity due Complies with radioelements spectrometry different suppliers) for to the artificial the EEC .sup.134Cs and .sup.137Cs either cocksfoot, radioelements regulation meadow grass, rye- .sup.134Cs and .sup.137Cs (Council grass, sweet vernal- was detected in Regulation (EC) grass or timothy grass pollen No 616/2000 of pollen species = a raw materials 20 Mar. 2000) total of 10 batches Heavy metals Cadmium (Cd): a total of 17 batches Cd is Cd is largely (Cd, Ni and Cr) Inductively of cocksfoot, undetectable below (at least Coupled Plasma- meadow grass, rye- (<0.5 ppm) in all 250 fold) the Mass grass, sweet vernal- tested batches calculated limit Spectrometry grass or timothy dose (125 ppm) (ICP-MS) after pollen species from 3 mineralization suppliers Nickel (Ni) and 2 batches (from 2 Ni and Cr were <10 Ni and Cr are chromium (Cr): different suppliers) for ppm and <2 ppm, largely below Graphite Furnace either cocksfoot, respectively, in (more than 10 Atomic meadow grass, rye- the tested batches fold and 60 fold, Absorption grass, sweet vernal- respectively) the Spectrometry grass or timothy calculated limit (GFAAS) pollen species = a dose (125 ppm) total of 10 batches Mycotoxins HPLC 2 batches (from 2 undetectable or all mycotoxins (aflatoxins B1&2, suppliers, except for 1 just above the are largely G1&2; DON*/ species) for either detection limit below (1,250 vomitoxin; cocksfoot, meadow (aflatoxin B1 in fold or more) the fumonisins B1&2; grass, rye-grass, sweet 1 batch) in all calculated limit ochratoxin A; vernal-grass or tested batches doses (depending zearalenone) timothy pollen species = on the considered a total of 10 batches alfatoxin) Benzo[a]pyrene isotope dilution 2 batches (from 2 undetectable or benzo[a]-pyrene Gas different suppliers) for just above the is largely below Chromatography either cocksfoot, detection limit (at least 1.6 10.sup.5 followed by Mass meadow grass, rye- (in 1 batch) in fold) the Spectrometry grass, sweet vernal- all tested batches calculated limit (GC-MS) grass or timothy dose (125 ppm) pollen species = a total of 10 batches *Deoxynivalenol.

[0108] All four categories of contaminants which could lead to a genotoxic potential in vitro were undetectable and/or largely below the calculated limit doses according to the appropriate (i.e. health of food) guidelines, in at least 10 batches of raw materials used to make grass pollen tablets.

[0109] Thus, the genotoxic profile of grass pollen raw materials and extracts must be explained by grass pollen intrinsic substance(s) only.

Example 3: Search for Intrinsic Substances with a Genotoxic Potential In Vitro

[0110] Four substances or substance groups which might be grass pollen intrinsic substances explaining the observed in vitro genotoxic potential were selected for further analysis: [0111] chlorogenic and caffeic acids (Fung V A, Cameron T P, Hughes T J, Kirby P E, Dunkel V C. Mutat. Res. 1988; 204: 219-228.), [0112] coumarin (Kevekordes S, Spielberger J, Burghaus C M, Birkenkamp P, Zietz B, Paufler P, Diez M, Bolten C, Dunkelberg H. Anticancer Res. 2001; 21: 461-469; Möller M, Stopper H, Haring M, Schleger Y, Epe B, Adam W, Saha-Moller C R. Biochem. Biophys. Res. Commun. 1995; 216: 693-701), [0113] alkaloids (Liu S X, Cao J, Yuan J, Huang P, Shua P Q, Honma M. Zhongguo Zhong Yao Za Zhi. 2003; 28: 957-961), and [0114] flavonoids (Caria H, Chaveca T, Laires A, Rueff J. Mutat. Res. 1995; 343: 85-94; Müller L and Kasper P. Mutat. Res. 2000; 464: 19-34 (review); Antognoni F, Ovidi E, Taddei A R, Gambellini G, Speranza A. Altern. Lab. Anim. 2004; 32: 79-90; Meltz M L, McGregor J T. Mutat. Res. 1981; 88: 317-324; Snyder R D, Gillies P J. Environ. Mol. Mutagen. 2002; 40: 266-276; Nagao M, Morita N, Yahagi T, Shimizu M, Kuroyanagi M, Fukuoka M, Yoshihira K, Natori S, Fujino T, Sugimura T. Environ. Mutagen. 1981; 3: 401-419).

[0115] Importantly, two forms of flavonoids can be distinguished: [0116] flavonoid glycosides as exemplified by hyperoside:

##STR00001## [0117] and flavonoid aglycones, as exemplified by quercetin (the aglycone of hyperoside)

##STR00002##

[0118] The presence or absence of caffeic or chlorogenic acids, coumarin, alkaloids, glycosylated flavonoids or aglycones, in grass pollens, was assessed using Thin Layer Chromatography (TLC). Briefly, grass pollen raw materials or extracts were processed using organic solvents (hexane, methanol . . . ) and were allowed to migrate into a silica gel in a solvent mixture specific for the compound to be tested.

[0119] No chlorogenic, caffeic acid coumarin or alkaloid could be found in a 5 grass pollen mix.

[0120] As regards flavonoids, bands were observed in every single grass pollen species using a protocol aimed at detecting flavonoid glycosides. Conversely, flavonoid aglycones could not be found in grass pollen using the appropriate protocol.

[0121] Accordingly, among all the assayed substances or substance groups known to possess a genotoxic potential in vitro, only flavonoids were found in grass pollens. More specifically, flavonoids found in grass pollens were flavonoid glycosides only, as no flavonoid aglycones could be detected. This is consistent with published data since the most common flavonoids isolated from pollen are flavonoid glycosides whereas aglycones do not occur naturally in pollen (Mo Y, Nagel C, Taylor L P. Proc. Natl. Acad. Sci. USA. 1992; 89: 7213-7217; Campos M G, Webby R F, Markham K R. Z. Naturforsch. [C]. 2002; 57: 944-946).

Example 4: Explanation for the In Vitro Genotoxic Potential of Grass Pollen Extracts by the Intrinsic Flavonoids

[0122] 4.1. Working Hypothesis and Scientific Approach

[0123] In view of the above findings, it was concluded that the genotoxic potential of grass pollen extracts must be associated with intrinsic flavonoids. However, no in vitro genotoxicity was reported for flavonoid glycosides, which are contained in the pollen extracts, whereas flavonoid aglycones, which were undetectable in the extracts, are known to display a genotoxic potential in vitro (Antognoni F, Ovidi E, Taddei A R, Gambellini G, Speranza A. Altern. Lab. Anim. 2004; 32: 79-90; Nagao M, Morita N, Yahagi T, Shimizu M, Kuroyanagi M, Fukuoka M, Yoshihira K, Natori S, Fujino T, Sugimura T. Environ. Mutagen. 1981; 3: 401-419; Brown J P. Mutat. Res. 1980; 75: 243-277).

[0124] As an enzyme capable of deglycosylating a flavonoid glycoside into its aglycone counterpart has been described in a pollen model (Taylor L P, Strenge D, Miller K D. The role of glycosylation in flavonol-induced pollen germination. Adv. Exp. Med. Biol. 1998; 439: 35-44), the hypothesis: was made that the in vitro genotoxic potential observed with grass pollen extracts would be due to deglycosylation of non-genotoxic flavonoid glycosides into in vitro genotoxic flavonoid aglycones by specific enzyme(s), both the flavonoid glycosides and the enzyme(s) being extractable from the pollens.

[0125] To demonstrate this working hypothesis, it was investigated whether: [0126] grass pollen flavonoids are identifiable as flavonoid glycosides only present in quantifiable amounts, [0127] those flavonoid glycosides are deglycosylated into flavonoid aglycones under the conditions of the 24-h long term protocol but not of the 3-h short term protocols of the MLA/TK test, [0128] the amounts of flavonoid aglycones released under such conditions are consistent with the results of the MLA/TK test obtained with the freeze-dried extracts of grass pollens.

[0129] 4.2. Identification of Grass Pollen Flavonoids by Reverse Phase HPLC and HCl-Hydrolysis/Reverse Phase HPLC

[0130] Grass pollen flavonoids were first identified by reverse phase HPLC, according to their hydrophobicity. Reverse phase HPLC was performed on a Atlantis dC18 column, 4.6×250 mm, 5 μm, Waters, Milford, Mass., USA; with a gradient elution: 0.1% formic acid qsp water to 0.1% formic acid qsp acetonitrile in 30 min; flow rate: 1 mL/min; run time: 45 min; detection wavelength: 354 or 254 nm; injection volume 50 μL. Importantly, grass pollen extracts were diluted 5 times in methanol before analysis in order to solubilize both flavonoid glycosides and flavonoid aglycones, the latter being hardly soluble in water.

[0131] HCl-hydrolysis prior to reverse phase HPLC analysis was also used since hydrolysis of flavonoid glycosides leads to deglycosylation. HPLC analysis of the deglycosylated molecules then allows identifying their aglycone moieties. For HCl-hydrolysis, flavonoids of a grass pollen crude extract were first purified on Amberlite XAD-2, a resin currently used for flavonoid isolation (Gil M I, Ferreres F, Francisco A. Tomás-Barberán F A. J. Agric. Food Chem. 1998; 46: 2007-2012; D'Arcy B R. Antioxidants in Australian floral honeys: Identification of health-enhancing nutrient components: a report for the Rural Industries Research and Development Corporation. D'Arcy B R, Barton, A. C. T.: Rural Industries Research and Development Corporation, 2005, 84 p). Flavonoid aglycone and flavonoid glycoside standards were purchased from Extrasynthese (Genay, France).

[0132] Reverse phase HPLC indicated that an extract of 5 grass pollens contain: 1. a majority of highly polar flavonoids, appearing as peaks n° 1 and n° 22. small amounts of moderately polar flavonoids, appearing as peaks n° 3, 4 and 53. no detectable flavonoid aglycones. Retention times were as follows: peak n° 1 14.975, peak n° 2 15.151, peak n° 3 15.449, peak n° 4 16.022, peak n° 5 16.365, peak n° 6 17.117 (absorbance at 354 nm).

[0133] Based on the peaks area, peaks n° 1 and n° 2 represent more than 90% of the overall detected flavonoids within grass pollen. As a consequence identification was focused on the components of peaks n° 1 and 2.

[0134] According to their RT, the main flavonoids of grass pollen extract (peaks n° 1 and n° 2) were probably flavonoid diglycosides. However, because the observed RT did not correspond to any of the flavonoid standards we tested (flavone, rhamnetin, isorhamnetin, kaempferol, quercetin, quercetin-4′-O-glucoside, isorhamnetin-3-O-glucoside, kaempferol-3-O-glucoside, quercetin-3-O-galactoside, quercetin-3-O-glucoside, kaempferol-3-O-glucorhamnoside, quercetin-3-O-glucorhamnoside, kaempferol-3-O-robinoside-7-O-rhamnoside), the main flavonoids of grass pollen extracts could not be identified at this stage.

[0135] In order to determine the aglycone moieties of the grass pollen flavonoid glycosides, a deglycosylation was performed by HCl-hydrolysis of flavonoids purified from a grass pollen crude extract. HCl-hydrolysis released quercetin, kaempferol and isorhamnetin. Thus, the flavonoids of grass pollen extracts were glycosylated forms of quercetin, kaempferol and isorhamnetin.

[0136] Assuming that the peak areas obtained for quercetin, kaempferol and isorhamnetin are representative of their relative concentrations, the flavonoid aglycones produced by HCl-hydrolysis of a grass pollen extract are composed of 18% quercetin, 5% kaempferol and 77% isorhamnetin.

[0137] 4.3. Further Identification of Grass Pollen Flavonoids by Mass Spectrometry Analysis

[0138] Prior to mass spectrometry analysis, flavonoids of grass pollen extracts were separated using the HPLC method described above and manually collected from the HPLC column. Mass spectrometry analyses were performed using electrospray ionization-tandem mass spectrometry (ESI-MS/MS) in negative ion mode on a ThermoElectron LCQ Duo ion-trap mass spectrometer (San Jose, Calif., USA), after direct infusion of the samples. Fragmentation was obtained by collison-induced dissociation with helium.

[0139] According to data obtained from commercially available standards, mass spectra of flavonoid glycosides are easily interpretable: electrospray fragmentation leads mainly to the loss of part or all of their glycoside moiety(ies). Most particularly, all tested flavonoid glucosides (kaempferol-3-O-glucoside, quercetin-3-O-glucoside, isorhamnetin-3-O-glucoside, quercetin-3-O-glucorhamnoside, and kaempferol-3-O-robinoside-7-O-rhamnoside) loose a 120-Dalton part of the glucose or the whole 163-Dalton glucose moiety during fragmentation.

[0140] The two main peaks resolved in the HPLC method were analyzed by mass spectrometry analysis:

[0141] 1. Peak n° 2 with a Mean Retention Time of 15.00 Min:

[0142] One single ion appeared as a major parent ion, at m/z 639, therefore corresponding to a molecule of a 640-Da mass, referred to as “[M]”. A minor parent ion was observed at m/z 1279, most likely corresponding to a [2M-H]− negative dimer ion of the 640-Da molecule (“H” standing for “1 hydrogen atom”).

[0143] Cocksfoot and timothy grass pollens were described to contain a 640-Da flavonoid diglycoside, namely isorhamnetin-3, 4′-diglucoside (Inglett G E. Nature 1956; 178: 1346. Inglett G E. J. Org. Chem. 1957; 22: 189-192), also found in other pollens such as the one of Crocus (Kuhn R, Löw I. Chem. Ber. 1944; 77: 196-202). Since the fragmentation spectrum obtained for the 15.00 min-HPLC peak can be easily interpreted as a the one of isorhamnetin-3, 4′-diglucoside, it was concluded that the component of the major HPLC peak n° 2 corresponds to isorhamnetin-3, 4′-diglucoside.

[0144] 2. Peak n° 1 with a Mean Retention Time of 14.83 Min:

[0145] According to mass spectrometry analysis in negative ion mode, four ions appeared as major parent ions: at m/z 639, 625, 609 and 463, therefore corresponding to 640-Da, 626-Da, 610-Da and 464-Da molecules, respectively. According to fragmentation analysis, the 640-Da molecule was isorhamnetin-diglucoside of peak n° 2 that contaminated peak n° 1.

[0146] Fragmentation spectra of parent ions at m/z 625 and 609 can be easily interpreted as the ones of quercetin-diglucoside and kaempferol-diglucoside, respectively. Assuming that both flavonoid glycosides are produced through the same metabolism as isorhamnetin-3, 4′-diglucoside, and that they will be deglycosylated by the same specific enzymatic machinery, it was concluded that they are quercetin-3, 4′-diglucoside (m=626 Da) and kaempferol-3, 4′-diglucoside (m=610 Da), respectively.

[0147] Kaempferol-3, 4′-diglucoside has already been described in pollen, namely in pollen of Trillium species (Yoshitama K, Tominaga T, Kanemaru Y, Yahara S. XVI Internationl Botanical Congress, Abstract n° 2539). Quercetin-3, 4′-diglucoside has also been described in plants, namely in onion (Bonaccorsi P, Caristi C, Gargiulli C, Leuzzi U. J. Agric. Food Chem. 2005; 53: 2733-2740; Mullen W, Crozier A. J. Oil Palm Res. 2006; Special Issue (April): 65-80). However, to our knowledge this is the first time quercetin-3, 4′-diglucoside is described in pollens.

[0148] According to its molecular mass, the 464-Da molecule might be quercetin-monoglycoside. This is confirmed by the fragment ion at m/z 301, corresponding to quercetin. The presence of quercetin-monoglycoside in peak n° 1 is surprising, as standard quercetin-monoglycosides display longer retention. Quercetin-monoglycoside of peak n° 1 might be complexed to other components of this peak, thus sharing the same retention time. It might also be produced by alteration of peak n° 1's quercetin-diglucoside during the purification steps. Anyhow, according to the mass spectrum this component of peak n° 1 is quantitatively less important than quercetin- and kaempferol-diglucoside.

[0149] Altogether, it was thus found that the extracts of 5 grass pollens contain the main following flavonoid glycosides, found in decreasing amounts: isorhamnetin-diglucoside, quercetin-diglucoside, kaempferol-diglucoside.

[0150] 4.4 Quantification of Isorhamnetin-, Quercetin- and Kaempferol-Glycosides in Freeze-Dried Extracts of 5 Grass Pollens by HPLC-DAD after HCl-Hydrolysis

[0151] Since no standard is available for isorhamnetin-, quercetin- and kaempferol-diglucosides, those flavonoids were quantified after HCl-hydrolysis. This induces deglycosylation into isorhamnetin, quercetin and kaempferol (aglycones) for which standard do exist. Knowing the molecular masses of both the aglycones and the diglucosides, the concentration of the latters can easily be deduced from the corresponding formers' concentrations.

[0152] Quantification of isorhamnetin, quercetin and kaempferol obtained after HCl-hydrolysis was performed by HPLC-DAD. In this method, the flavonoid aglycones are separated by HPLC and detection is performed using a diode-array detector (or DAD). This allows recording the absorption spectrum in the ultraviolet range. Given that two flavonoid aglycones do not share the same absorption spectra, HPLC-DAD allows the identification of a flavonoid aglycone on the basis of both its retention time and its absorption spectrum, by comparison with the corresponding standard molecule. Knowing the ratios between the molecular masses of the diglucosides and the aglycones, the concentration of isorhamnetin-, quercetin- and kaempferol-diglucosides can be easily deduced from the measured concentration of their aglycone counterparts.

[0153] Three batches of freeze-dried extracts of 5 grass pollens obtained using the non-optimized ultrafiltration step were quantified for isorhamnetin-, quercetin- and kaempferol-diglucosides after they were transformed into their aglycones by HCl-hydrolysis.

TABLE-US-00002 TABLE 2 Concentration of isorhamnetin, quercetin and kaempferol in three batches of sieved extract of 5 grass pollens, as measured by HPLC-DAD after HCl-hydrolysis, and deduced concentration of their respective diglucoside counterparts (in μg/mg) concentration of flavonoid aglycones deduced concentration of flavonoid diglucosides total isorhamnetin quercetin kaempferol total isorhamnetin Quercetin kaempferol flavonoid diglucoside diglucoside diglucoside flavonoid batch no (316 Da) (302 Da) (286 Da) aglycones (640 Da) (626 Da) (610 Da) diglucosides 50299 1.70 0.34 0.10 2.14 3.44 0.70 0.21 4.36 50300 1.00 0.26 0.08 1.34 2.03 0.54 0.17 2.73 50311 1.30 0.35 0.09 1.74 2.63 0.73 0.19 3.55 Mean 1.33 0.32 0.09 1.74 2.70 0.66 0.19 3.55 % 77 18 5 100 76 19 5 100

[0154] The mean relative amounts of isorhamnetin, quercetin and kaempferol found after HCl-hydrolysis of grass pollen freeze-dried extracts are exactly the same as the ones estimated after HCl-hydrolysis of purified flavonoids of a crude pollen extract that is: 77% isorhamnetin, 18% quercetin and 5% kaempferol. These results confirm that, on a quantitative basis, grass pollen flavonoids are in the following order, from most to less abundant: isorhamnetin diglucoside, quercetin diglucoside, kaempferol diglucoside.

[0155] Overall, the freeze-dried extracts of 5 grass pollens obtained using the non-optimized ultrafiltration step contain 0.36% (w/w) of flavonoid diglucosides.

[0156] 4.5 Deglycosylation of Grass Pollen Flavonoid Glycosides Under the Conditions of the 24-h Protocol but not of the 3-h Protocols of the MLA/TK Test

[0157] The short term and long term protocols of the MLA/TK test involves incubation for 3 h at 37° C. and incubation for 24 h at 37° C., respectively. To demonstrate that flavonoid glycosides of grass pollens are deglycosylated in the conditions of the long term protocol but not of the short terms protocols of the MLA/TK test, a crude extract of grass pollen was placed at 37° C. and then sampled after 3 h and 24 h incubations. Samples were kept frozen until analyzed for their contents in flavonoid glycosides and flavonoid aglycones.

[0158] To demonstrate that deglycosylation was due to an active enzyme-dependent process, grass pollen flavonoids were separated from grass pollen proteins by purification on Amberlite XAD-2 resin (Gil M I, Ferreres F, Francisco A. Tomás-Barberán F A. J. Agric. Food Chem. 1998; 46: 2007-2012; D'Arcy B R. D'Arcy B R, Barton, A. C. T.: Rural Industries Research and Development Corporation, 2005, 84 p) and then analyzed before and after a 24-h incubation at 37° C.

[0159] Analysis of samples was performed using the HPLC method described above for the detection of flavonoids.

[0160] Incubation of a grass pollen extract at 37° C. induces a slight but detectable decrease in levels of flavonoid diglucosides (peaks with RT 14.98-min and 15.16 min-min retention times) as soon as after 3 h. After 3 h, however, appearance of the aglycones quercetin (RT=20.2 min) and isorhamnetin (RT=22.2 min) remains negligible, whereas no kaempferol (RT=22.0) could be detected. In fact, the decrease in flavonoid diglucosides mainly corresponds to a ˜2-fold increase of two peaks' surfaces, with 16.4-min and 17.1-min retention times, respectively. A 16.4-min RT corresponds to quercetin monoglucoside and a 17.1-min RT corresponds to both kaempferol- and isorhamnetin-monoglucosides. Therefore, the increase of the two peaks is most likely a consequence of a partial deglycosylation of flavonoid-diglucosides into flavonoid-monoglucosides.

[0161] Extending the 37° C.-incubation to 24 h leads to a marked decrease of flavonoid diglucosides, a marked increase of flavonoid monoglucosides, a dramatic increase of the flavonoid aglycones quercetin, kaempferol and isorhamnetin (FIG. 1).

[0162] In return, when flavonoid glycosides isolated from a grass pollen extract were incubated for 24-h at 37° C., no change in the HPLC profile was observed. Since the isolated flavonoids did not contain any detectable protein according to an SDS-PAGE experiment, this confirmed that the phenomenon observed with a whole pollen extract is an enzyme-driven deglycosylation of flavonoid glycosides.

[0163] As (a) the sum of all peaks areas was rather well conserved and (b) the decrease of a flavonoid diglucoside-corresponding peak was related to the increase or appearance of flavonoid monoglycoside and/or aglycone-corresponding peak(s), it could be considered that the ratio between the two peaks areas is equivalent to the molar ratio of the corresponding flavonoids. On this basis, the decrease in quercetin-diglucoside leaded to an equal increase in molecules of quercetin-monoglucoside and in molecules of quercetin aglycone.

[0164] Based on the peaks areas, the decrease in flavonoid diglucosides after a 24-h incubation at 37° C. was in the order of ˜40%. Since about half of the quercetin-diglucoside was transformed into quercetin aglycone, this means that ˜20% of quercetin-diglucoside was transformed into quercetin (aglycone) under the conditions of the 24 h-long term protocol of the MLA/TK test.

[0165] Altogether, those results indicate that: [0166] deglycosylation of grass pollen flavonoid diglucosides into flavonoid aglycones occurs under the conditions of the 24-h long term protocol of the MLA/TK test, [0167] such a complete deglycosylation hardly occurs in the conditions of the 3-h short term protocols of the MLA/TK test, [0168] the deglycosylation of flavonoid glycosides is driven by grass pollen extractible enzymes.

[0169] Most particularly, ˜20% of grass pollen quercetin-diglucoside is transformed into quercetin (aglycone).

[0170] 4.6. Consistency of the Amounts of Flavonoid Aglycones Released in the Conditions of the 24-h Long Term Protocol of the MLA/TK Test with Results from this Test

[0171] To our knowledge, quercetin is the only flavonoid aglycone that was studied in the MLA/TK test, namely in a work by Meltz and MacGregor who used a 4-h short term protocol (Mutat. Res. 1981; 88: 317-324). On the other hand, it was proved to be the most genotoxic flavonoid aglycone by other in vitro genotoxic tests (Nagao M, Morita N, Yahagi T, Shimizu M, Kuroyanagi M, Fukuoka M, Yoshihira K, Natori S, Fujino T, Sugimura T. Environ. Mutagen. 1981; 3: 401-419; Brown J P. Mutat. Res. 1980; 75: 243-277; Czeczot H, Tudek B, Kusztelak J, Szymczyk T, Dobrowolska B, Glinkowska G, Malinowski J, Strzelecka H. Mutat. Res. 1990; 240: 209-216 MacGregor J T, Jurd L. Mutat. Res. 1978; 54: 297-309).

[0172] Therefore, on the basis of the study by Meltz and MacGregor, it was determined whether the deglycosylation of grass pollen quercetin-diglucoside into quercetin aglycone could by itself explain the results of the MLA/TK test obtained with the freeze-dried extracts of grass pollen.

[0173] The amounts of quercetin released during the 24-h/S9- MLA/TK test was calculated on the basis of: the mean concentration of quercetin-diglucoside in freeze-dried extracts, as determined above; and the level of deglycosylation of this molecule into quercetin aglycone in the conditions of this protocol, as also determined above.

[0174] The induction ratios that should result from those amounts was deduced from the data of Meltz and MacGregor. Such deduced induction ratios were then compared to the induction ratio actually observed.

[0175] It has been shown above that: [0176] freeze-dried extracts obtained using the non-optimized ultrafiltration step contain on average of 0.32 μg/mg of quercetin in the form of flavonoid glycosides, mostly quercetin diglucoside; [0177] ˜20% of quercetin diglucoside is deglycosylated into quercetin aglycone under the conditions of the 24-h long term of the MLA/TK test.

[0178] The concentrations of freeze-dried extracts tested in the 24-h/S9- protocol of MLA/TK were 13.5 mg/mL or below. The corresponding concentrations of released quercetin were then 0.864 μg/mL (20%×0.32 μg/mg×13.5 mg/mL) or below.

[0179] The lowest concentration of quercetin tested by Meltz and MacGregor was 10 μg/mL. Therefore, to determine the ratio that should be obtained with the lowest concentration of 0.864 μg/mL or below, the relationship between (a) the concentrations of quercetin tested by Meltz and MacGregor and (b) the corresponding induction ratios they observed was mathematically modelled. Since those ratios started reaching a plateau at the first concentration tested, we choose to use a logarithmic function for such a modelization. Given that the ratio is necessarily of 1 for 0 μg/mL of mutagen (control), the logarithmic function will be of the form:

y=a.Math.ln(x+1)+1  (1)

[0180] where y is the induction ratio, x is the concentration of quercetin and a is a constant number.

[0181] Indeed, according to equation (1), the induction ratio y is of 1 for a quercetin concentration x of 0.

[0182] Using the ratios obtained by Meltz and MacGregor for 10, 20 and 30 μg/mL of quercetin allows a mathematical modelization by the following equation with an excellent correlation coefficient (r=0.99):

y=3.0822×ln(x+1)+1  (2).

[0183] Two freeze-dried extracts obtained using the non-optimized ultrafiltration step were studied in the 24-h/S9- MLA/TK test, namely: batch n° 40244 and batch n° 52494, which displayed a genotoxic potential at concentrations of 13.5 and 12.9 mg/mL, respectively. The corresponding concentration of quercetin released during the test is of 0.83-0.86 μg/mL (see above for calculation). According to equation (2) the induction ratio that should be obtained at this concentration is of 2.9 (2.86-2.91). The induction ratios obtained experimentally for batch n° 40244 and 52494 were of 2.1 and 3.5, respectively, that is, a mean of 2.8, which is virtually identical to the interpolated induction ratio of 2.9.

[0184] Since the induction ratio actually observed is identical to the induction ratio interpolated from the published data on the genotoxic potential of quercetin in MLA/TK test, we conclude that most, if not all, of the genotoxic potential of freeze-dried extracts of grass pollen obtained using the non-optimized ultrafiltration step can be explained by the production of quercetin through deglycosylation of quercetin-diglucoside during the test.

[0185] However, contribution of the deglycosylation of kaempferol- and isorhamnetin-diglucosides, albeit theoretically negligible, cannot be excluded.

[0186] 4.7. Optimization of Ultrafiltration Step

[0187] The manufacturing process of grass pollen extracts involves ultrafiltration step with 2.5 volumes of washing with purified water on a 1 kDa-membrane.

[0188] Ultrafiltration step on a 1 kDa-membrane was compared with ultrafiltration step on a 5 kDa-membrane. Washings were performed with from 1 to 15 volumes of purified water. It was found that the cut-off value of the membrane had no impact of the immunoreactivity and protein content of the pollen allergen preparation. The volume of purified water for washing did neither influence the quality of the allergen preparation. It was concluded that ultrafiltration may be performed on a 5 kDa membrane with up to 15 volumes of water without altering the quality of the pollen extract.

[0189] However, clotting was observed afterwards upon filtration with a 0.22 μm filter.

[0190] To avoid this clotting, ultrafiltration with a solution containing 120 ppm ammoniac, equivalent to 0.56 g/L ammonium bicarbonate, was used for washing instead of purified water. No filtration clotting could be observed upon filtration with a 0.22 μm filter.

[0191] It was further checked that activity and protein content of the pollen extracts were unaltered by replacement of purified water with a 0.56 g/L ammonium bicarbonate solution for washings. No difference between the two washing solutions could be seen up to 15 volumes of washings. However, above 15 volumes of purified water, a decreased allergenic activity was detected whereas the ammonium bicarbonate solution was assayed up to 30 volumes of washing without detrimental effect of the allergenic activity of the pollen extract.

[0192] Flavonoid dosages indicated that washing with 2.5 volumes of purified water or ammonium bicarbonate 0.56 g/L enabled to remove about 80% of flavonoids. Flavonoids were completely removed by 15 volumes of purified water or ammonium bicarbonate 0.56 g/L. Washing with 30 volumes of ammonium bicarbonate 0.56 g/L did not improved any further flavonoid removal.

[0193] Ultrafiltration with 15 volumes of washing on a 5 kDa-membrane was selected for further characterisation.

[0194] 4.8. Demonstration of the Elimination of the Grass Pollen Flavonoids by an Optimized Ultrafiltration Step

[0195] The manufacturing process of grass pollen extracts has been optimized at the ultrafiltration step, involving 15 volumes of washing on a 5 kDa-membrane instead of the previously used 2.5 volumes of washing on a 1 kDa-membrane.

[0196] To determine whether this optimized process is able to completely eliminate flavonoids, the latter were assayed at different steps of the optimized ultrafiltration step, namely: just before ultrafiltration, that is, after the concentration step, after ultrafiltration with 2.5 volumes of washing, after ultrafiltration with 15 volumes of washing. This was performed using the HPLC method described above for this purpose.

[0197] Flavonoids were also assayed in three batches of sieved freeze-dried extracts obtained using the optimized ultrafiltration step, as compared to three batches obtained using the previous non-optimized process. The same HPLC method was used for this assay.

[0198] Finally, flavonoids were accurately quantified in three batches of sieved freeze-dried extract obtained using the optimized ultrafiltration step. Quantification was performed by HPLC-DAD after HCl-hydrolysis.

[0199] As opposed to the previous non-optimized manufacturing process of 5 grass pollens extracts, the optimized process leads to complete elimination of grass pollen flavonoids (FIG. 2).

[0200] This was confirmed with freeze-dried extracts obtained using the optimized ultrafiltration step, which contained no detectable flavonoids, as opposed to the freeze-dried extracts obtained through the previous non-optimized process.

[0201] Using HPLC-DAD after HCl-hydrolysis, flavonoids, as expressed in aglycone equivalents, were below the limit of quantification in sieved freeze-dried extracts of grass pollens, that is, below 0.003% (or g/100 g), leading to a total concentration of flavonoid diglucosides below 0.15 μg per 300 IR tablet (Table 3).

TABLE-US-00003 TABLE 3 Concentration of isorhamnetin, quercetin and kaempferol in three batches of sieved freeze-dried extract of 5 grass pollens (active substance) obtained using the optimized ultrafiltration step as measured by HPLC-DAD after HCl-hydrolysis, deduced concentration of their respective diglucoside counterparts and deduced concentration of all flavonoid diglucosides in a corresponding 300 IR tablet deduced concentration concentration of flavonoid aglycones deduced concentration of flavonoid diglucosides of flavonoid diglucosides (in % or g/100 g) (in % or g/100 g)* in a 300 IR tablet (in μg)** total total total batch flavonoid isorhamnetin quercetin kaempferol flavonoid flavonoid no isorhamnetin quercetin kaempferol aglycones diglucoside diglucoside diglucoside diglucosides diglucosides 60099 <0.001 <0.001 <0.001 <0.003 <0.002 <0.002 <0.002 <0.006 <0.15 60106 <0.001 <0.001 <0.001 <0.003 <0.002 <0.002 <0.002 <0.006 <0.15 60113 <0.001 <0.001 <0.001 <0.003 <0.002 <0.002 <0.002 <0.006 <0.15

[0202] Therefore, the optimized ultrafiltration step in the optimized manufacturing process resulted in a thorough removal of flavonoids from the extracts.

CONCLUSIONS

[0203] In our efforts to determine the intrinsic causes of the in vitro genotoxic potential of grass pollen extracts, the followings were demonstrated: [0204] grass pollen extracts of 5 grass pollen contain flavonoid glycosides, mostly identified as quercetin-diglucoside, kaempferol-diglucoside and isorhamnetin diglucoside, [0205] under the conditions of the 24-h long term protocol of the MLA/TK test, grass pollen flavonoid glycosides are deglycosylated into their aglycone counterparts, namely: quercetin, kaempferol and isorhamnetin, [0206] most particularly, ˜20% of quercetin diglucoside are transformed into quercetin (aglycone) in those conditions, [0207] on the basis of published data, the corresponding amounts of produced quercetin can totally explain the level of genotoxicity observed with the sieved freeze-dried extracts of 5 grass pollens obtained using the non-optimized ultrafiltration step, [0208] almost no flavonoid aglycones are obtained under the conditions of the 3-h short term protocols of the MLA/TK test, consistent with the absence of genotoxic potential under those conditions, [0209] extending the ultrafiltration step, as per the optimized ultrafiltration step, results in removal of flavonoids from the extracts to undetectable amounts.

[0210] Since external genotoxic contaminants were undetectable and/or largely below the calculated limit doses in grass pollen raw materials, the in vitro genotoxic potential of grass pollen extracts observed in the 24 h/S9- protocol of the MLA/TK test must be explained by grass pollen intrinsic substances. In this respect, it was demonstrated that the following mechanism occurs under the conditions of the assay: nongenotoxic flavonoid glycosides from grass pollens, namely isorhamnetin-diglucoside, quercetin-diglucoside and kaempferol-diglucoside, are deglycosylated by pollen-derived enzymes into their aglycone counterparts, namely isorhamnetin, quercetin and kaempferol, respectively, which are well-known to display a genotoxic potential in vitro, although they have never been proved genotoxic in vivo.

[0211] On the basis of published data on quercetin genotoxicity in the MLA/TK test, the amounts of quercetin released by grass pollen can totally explain the grass pollen in vitro genotoxic potential of grass pollen.

[0212] The theoretical risk associated with the presence of flavonoids in grass pollen extracts has been definitely eliminated by extending the ultrafiltration step of the manufacturing process, resulting in complete removal of pollen-derived flavonoids from the extracts. It was calculated that a daily administration of 300 IR tablets of grass pollen extract obtained using the optimized ultfiltration step would lead to a theoretical daily intake below 0.15 μg of pollen-derived flavonoids, well below the daily intake of flavonoids through common diet which is of 20 mg to 1 g.