The effects of exogenous xylanase supplementation on the in vivo generation of xylooligosaccharides and monosaccharides in broilers fed a wheat-based diet

ABSTRACT 1. This study quantified xylanase-induced changes in soluble monosaccharides, xylooligosaccharides (XOS) and volatile fatty acid (VFA) contents of the different sections of the gastrointestinal tract (GIT) and whether these were related to altered bird performance. 2. An in vitro digestion of the wheat-based diet was carried out with the xylanase (Econase XT at 16,000BXU/kg diet) to compare the in vitro and in vivo generation of these XOS and monosaccharides. For the in vivo study, 80 male Ross 508 b roiler chicks were split into two groups fed a wheat-based diet with or without Econase XT (16,000BXU/kg diet) for 21 days. 3. There were no effects of Econase XT inclusion on growth performance characteristics, likely a result of the high-quality wheat diet, the corresponding high performance of the control group (FCR average of 1.45 in controls) and the relatively young age of the birds (from four to 26 days of age). 4. Econase XT supplementation increased the xylotetraose (X4) content in the colon (P = 0.046, enzyme x GIT section interaction) and the xylose contents in the colon and caeca (P < 0.001, enzyme x GIT section interaction). 5. The trend for increased acetate production in the caeca of Econase XT treated birds (P = 0.062) suggested that the XOS generated were subsequently fermented in the caeca, potentially impacting upon the types of microbiota present. 6. The present study suggested that wheat arabinoxylan degradation was enhanced by xylanase supplementation, which may have increased the production of beneficial volatile fatty acids (VFA) in the caeca, and thereby potentially modulated the caecal microbiome, but without affecting bird performance at this early age.


Introduction
Arabinoxylan (AX) is the most abundant hemicellulose in the endospermic cell wall of wheat (Jmares and Stone 1973).Nonstarch polysaccharides (NSP), like AX, are anti-nutritive in monogastrics, which lack the endogenous enzymes to break down plant cell walls (Choct and Annison 1990).The result is an increase in digesta viscosity, which has negative effects on performance (Bedford 2000).The supplementation of endo-β1, 4-xylanases effectively hydrolyses the xylan backbone of AX, generating arabinoxylan-oligosaccharides (AXOS), which encompasses both arabino-xylooligosaccharides and xylooligosaccharides (XOS) (Jommuengbout et al. 2009).Monosaccharides are generated if any contaminating xylosidase or arabinofuranosidase is present in an enzyme product or within the diet itself, which cleave β-xylosidic glycosidic linkages into monomeric pentoses (e.g.xylose and arabinose) and hexoses (e.g.glucose and galactose).This results in a reduction in viscosity and improved growth rates and feed efficiency (Bedford and Classen 1992).It is hypothesised that the resulting short-chain AXOS from enzyme hydrolysis are utilised for fuel by the microbiota occupying the distal gastrointestinal tract (GIT), thereby having positive effects on the microbiome (Choct et al. 1999a).
NSP account for 10-12% of the dry matter (DM) in wheat (Knudsen 1997), of which 19-21% is soluble (Rodehutscord et al. 2016).The insoluble NSP in wheat mainly provides protection in the form of the cell wall (Simon et al. 2015).As per the cell wall degradation hypothesis, xylanase shows significant cell wall destruction in vivo (Bedford and Autio 1996), which enhances access to cell contents by endogenous enzymes, increasing their efficacy, thereby increasing amino acid retention, pancreatic amylase and mucin secretion to aid digestion (Cowieson and Bedford 2009).
There is evidence that feeding xylanase influences the caecal microbiome, where beneficial bacteria ferment the end products of enzyme hydrolysis (Mccracken et al. 2006;Masey-O'neill et al. 2014).This is presumably from the generation of AXOS during hydrolysis, as feeding near-pure AXOS results in similar performance benefits to xylanase inclusion itself (Morgan et al. 2019).AXOS are thus displaying prebiotic effects, in that they are fermented into beneficial volatile fatty acids in the microbiome, particularly acetate and butyrate in the caeca and colon (Choct et al. 1999b).Walker et al. (2005) suggested that AXOS substrate fermentation in the colon significantly lowered pH balance, which gave a boost to butyrate-producing bacteria, whilst hampering the proliferation of Bacteroides spp.which have the capacity to cause harm to their host.This subsequently improves the gut environment and could lead to better mineral absorption and immune response, along with increasing gizzard grinding, which all lead to more efficient nutrient absorption (Kim et al. 2011).
One theory is that there is an increased recovery of energy from the diet in the form of volatile fatty acids (VFAs).The issue with this theory is that the AXOS studies so far have used doses ranging from 0.1 g -10 g per kilogram of diet (Ribeiro et al. 2018;Suo et al. 2015;Eeckhaut et al. 2008).The lower levels of inclusion are unlikely to provide enough substrate to generate significant amounts of energy in the form of VFA.The other issue is that, in chickens, it is necessary to feed xylanase over an extended period for a response to be observed.If the enzyme was generating significant quantities of AXOS instantaneously, which were then fermented, there would not be a delay in the response seen after enzyme inclusion.However, studies show that performance benefits in wheat-based diets are often not seen until the birds reach 21 d of age (Mendes et al. 2013).
In a new hypothesis (Bedford 2018), suggested that the chicken microbiome can be 'trained' or adapted over time as a result of xylanase supplementation.Rather than non-starch polysaccharide enzymes (NSPases) producing more fermentable sugars, it may be that NSPases produce AXOS that signal to the microbiome, thereby increasing the capacity to degrade fibre.This theory is supported by a recent study which showed that chickens fed xylanase for 35 d had greater fermentation of pentoses and AXOS in their caecum than controls (Bedford and Apajalahti 2018).
Whether it is more efficacious to provide AX which is hydrolysed into AXOS in situ via endogenous enzyme supplementation or to supplement diets with AXOS that has been prepared in vitro remains to be established.
The aim of this study was to establish if the XOS and monosaccharides generated when a commercially available xylanase (Econase XT) was fed to broilers in a wheat-based diet between four and 25-26 d of age affected the nutrients generated and/or utilised along the GIT.In addition, in vitro digestion of the wheat-based diet was compared to the contents of the different sections of the GIT in vivo.Caecal VFA contents were determined as an indication of possible effects of Econase XT on the microbiome.

Diet and enzyme used
The diet was formulated by a commercial feed manufacturer in mash form (Target Feed Ltd, England) and the ingredients and nutritional values are shown in Table 1.Econase XT 25 is a commercially available xylanase with reported β-1,4 endoxylanase activity (160,000 BXU/g) provided by AB Vista (Marlborough, UK) and used at the recommended dose of 100 µg/g diet, which provided 16,000 BXU/kg, where each BXU represents the amount of enzyme required to release 1 μmol of reducing sugar per min from xylan under defined test conditions.

Total hydrolysis of non-cellulosic polysaccharides in the diet
Total non-cellulosic sugar contents in the diet were determined by total hydrolysis of non-cellulosic polysaccharides using Trifluoroacetic acid (TFA), as previously described (Fry 1988).Briefly, diet samples were ground to a fine powder (0.5 mm diameter) and suspended to 10 mg/ml in 2 M TFA in triplicate.Tubes were sealed and heated to 120°C for 1 h in an autoclave, then allowed to cool to room temperature before being centrifuged at 2236 × g for 10 min at room temperature.The supernatant was then diluted 1:100 with 10 mM NaOH and transferred into 2 ml clear vials for sugar analysis.The levels of the four monosaccharides (arabinose, galactose, glucose and xylose) were quantified, but there were no measurable XOS.

In vitro digestion of the wheat-based diet with Econase XT
The diet samples (four replicates) were individually ground to a fine powder (0.5 mm diameter) and 0.2 g was resuspended in 40 ml of 50 mM sodium citrate buffer (pH 5.2) either without (control -no enzyme) or with Econase XT 25 at 16,000 BXU/kg.The buffer was chosen to represent the average pH of the broiler digestive tract (Mabelebele et al. 2014).Digestion reactions were then placed in a shaking incubator at 150 RPM and a temperature of 41°C for up to 24 h.At each time point (0, 3, 6, 9, 12 or 24 h), 1 ml of each digest was removed and added to 9 ml of 10 mM NaOH at room temperature, mixed, centrifuged at 2236 × g for 10 min at room temperature and then frozen at −20°C prior to sugar analysis.

Chicken trial
The trial was conducted at the University of Nottingham Bio-Support Unit using 80, one-day-old male Ross 308 broiler chickens (average body weight 42 g), obtained from P D Hook Hatcheries Limited (Cote, Bampton OX18 2EG).Birds were housed and cared for according to the UK Animals (Scientific Procedures) Act 1986 (ASPA) Code of Practice for the care and accommodation of animals (February 2013), approval reference number 197.Birds were wing tagged for identification, acclimatised in one group, and fed the control diet (without enzyme) between one to four days of age.On day four they were allocated to 20 pens in groups of four birds per pen, matched with birds of similar weights.Hence, there were 10 replicate pens per diet (control and enzyme treated) each containing four birds.
All birds were given ad libitum access to diet and water throughout the study and were raised under controlled conditions of light, temperature and humidity as recommended by the breeder.Temperature was maintained at 32°C on arrival and reduced by approximately 1°C per day until 21°C was reached, as per the Ross 308 Management Guidelines.Body weight for individual birds was monitored and recorded following the introduction to the experimental diets on four, 11 and 18 d of age and immediately following culling on 25 or 26 d of age (five pens per treatment were culled on two consecutive days).Diet consumption and body weight gain were measured between day four and cull at 25 or 26 d of age to calculate feed conversion ratio (FCR).Birds from pens 1-10 were culled at 25 d of age and pens 11-20 at 26 d of age by the Schedule 1 method (Animals (Scientific Procedures) Act 1986).Samples of intestinal digesta were collected from 5 cm segments of the mid-jejunum, proximal ileum and colon, as well as the caecal contents for each bird.Digesta samples were snap-frozen using liquid nitrogen immediately after collection and stored at −80°C prior to analysis.One representative bird closest to average pen weight from each pen was analysed for monosaccharide and XOS contents, whereas the first bird from each pen was selected for VFA analysis in the caecal contents and processed accordingly, before being sent to Alimetrics Ltd (Finland) for further analysis.

Identification and quantification of sugars using HPAEC-PAD
The sample contents of arabinose, galactose, glucose and xylose, as well as the XOS were determined using High-Performance Anion-Exchange Chromatography coupled with Pulsed Amperometric Detection (HPAEC-PAD) following the method of Xu et al. (2013).Analysis was carried out using a Dionex ICS-3000 with a Dionex CarboPac PA20 column (3 mm x 150 mm) and CarboPac PA20 Guard (3x30 mm) for the monosaccharide analysis.A CarboPac PA200 column (3 mm x 250 mm) and CarboPac PA200 guard (3 mm x 50 mm) were used for the oligosaccharide analysis.An injection volume of 10 μl was used throughout for all standards and samples.Monosaccharide standards (arabinose, galactose, glucose and xylose) were purchased from Sigma-Aldrich, UK and XOS standards (xylobiose, -triose and -tetraose) from Megazyme, Ireland.Serial dilutions for each standard (2.0, 1.0, 0.5 and 0.25 g/l for monosaccharides and 2, 1, 0.5 and 0.25 g/l for XOS) were made fresh for each batch of analyses, diluted 1:100.
For the monosaccharides, a single eluent, containing 10 mM NaOH solution, was used as the mobile phase at 0.5 ml/min for 14 min.For oligosaccharides, two eluents were used in a gradient for the mobile phase, 0.1 M sodium hydroxide (Solution A) and 0.1 M NaOH containing 0.5 M sodium acetate (solution B) in standard quadruple waveform, as described by Xu et al. (2013).The gradient program used for XOS determination was 100% solution A at 0 min, rising to 80% Solution A and 20% Solution B at 25 min, before returning to 100% Solution A after 25 min elapsed.Both eluents were stored in plastic-pressurised bottles with inert nitrogen gas at 6-9 psi.Data were collected with Dionex Chromeleon software (Version 6.7).Dry matter content of the digesta was determined by oven drying the digesta at 60° C for 36 h.Monosaccharide and XOS contents from in vivo digesta samples were then adjusted for dry matter content.

Measurement of volatile fatty acids
The VFAs in the caeca of one randomly selected bird from each pen were analysed as free acids by Alimetrics Ltd (Espoo, Finland), using gas chromatography, as described previously (González-Ortiz et al. 2019;Holben et al. 2002).In brief, 1 g caecal contents were vigorously mixed with 1 ml H 2 O for 5 min, before 1 ml of 0.8 M perchloric acid was added and the mix was shaken to extract the VFAs.The acids measured included acetic, butyric, lactic, propionic, valeric and total branched chain fatty acids.

Data and statistical analysis
For the in vitro digests, one source was used for the diet and analytical samples (digestions and Dionex analysis) which were carried out in four batches of control and Econase XT combinations.The data were then processed in Excel (Microsoft, 2013) and expressed as means and standard error of the mean (SEM).
Standards for the four monosaccharides or XOS were run at the start and end of each batch, standard curves were generated from the areas under the curve and presented as g/100 g of diet or digesta.Data were then analysed by one-(enzyme) or two-way (enzyme x time or enzyme x GIT section) ANOVA, as appropriate, using Genstat statistical software (19th Edition), with blocking for batch and tube.Bonferroni post-hoc tests were used to identify significant differences between groups following a significant ANOVA.Significance was set at P < 0.05 with P < 0.10 indicating a strong trend.
For average daily feed intake (AFDI), average daily gain (ADG) and FCR, data were analysed by one-way ANOVA (Genstat statistical software, 19 th Edition).For body weight gain, data were analysed by two-way (treatment x time) ANOVA with repeated measures (Genstat statistical software, 19 th Edition).P < 0.05 denoted statistical significance in both instances.
For the correlations, data were analysed by Pearson correlation coefficients, assuming gaussian distribution, using GraphPad Prism (Version 8.1.2).A two-tailed test was undertaken to determine statistical significance with a 95% confidence interval, hence P < 0.05 was denoted statistical significance, and P < 0.10 a strong trend.

Total hydrolysis of non-cellulosic polysaccharides in trial diet
The total sugar contents were determined by TFA hydrolysis of the control diet (Table 2).As expected, the highest monosaccharide in the diet was glucose, while xylose and arabinose contents were similar (Table 2).It was assumed that the vast majority of these two monosaccharides were present as arabinoxylan (AX), and, as such, the total AX content and arabinose to xylose ratio were calculated.Galactose was present in fairly high amounts, presumably due to the use of soyabean meal in the diet, which is known to contain high levels of galactose (Irish and Balnave 1993).The total hydrolysis values for each monosaccharide were subsequently used to calculate the proportion that was released during the in vitro digestion and in the in vivo experiment, with and without Econase XT supplementation.

In vitro digestion of the wheat-based diet with or without Econase XT -release of xylooligosaccharides (XOS) and monosaccharides
There were significant enzyme x time interactions (P < 0.001) for the release of xylotetraose (X 4 ), xylotriose (X 3 ) and xylobiose (X 2 ).The X 4 was released from the diet in the absence of Econase XT but post hoc Bonferroni tests revealed that the release was higher with Econase XT at 3, 9, 12 and 24 h (P < 0.001, Figure 1a).Similarly, X 3 was released from the diet in the absence of Econase XT, peaking at 3 h (Figure 1b) and then either declining (in the control) or remaining flat (Econase XT).Post hoc Bonferroni tests indicated higher concentrations with Econase XT only at 12 and 24 h, with no differences at earlier timepoints.In contrast, there was no release of X 2 from the diet in the absence of Econase XT (Figure 1c), with significant release of X 2 only observed after 24 h incubation with Econase XT.There were significant enzyme x time interactions for the release of xylose (P = 0.034) and arabinose (P < 0.001).As observed for X 3 , xylose was released in the absence of Econase XT and both groups peaked at 3 h, before declining (Figure 1d).Post hoc Bonferroni tests indicated higher levels with Econase XT only at 24 h, with no differences at earlier timepoints.In contrast, arabinose was released linearly from 0 to 9 h, both with and without Econase XT (Figure 1e).Post-hoc Bonferroni tests showed significantly more was released in the presence of Econase XT at 12 h and 24 h.Unlike the release of xylose, there were no significant enzyme x time interactions for the release of galactose or glucose (Figure 1f and g, respectively), nor were there any effects of enzyme, but the release of both increased with time (P < 0.001).

The effect on Broiler performance of Econase XT in a wheat-based diet
There were no differences in average daily feed intake (ADFI), average daily gain or FCR in broiler chickens fed the diets for three weeks (Table 3; Figure 2).There were no mortalities.

The effect on release of xylooligosaccharides (XOS) and Monosaccharides at different sections of the broiler gastro-intestinal tract (GIT) of Econase XT in a wheatbased diet
There was a significant enzyme x GIT section interaction for X 4 release (P = 0.038, Figure 3a), such that X 4 concentration was highest in the colon of the birds supplemented with Econase XT, but there was no X 4 observed in the caeca.
There was no enzyme x GIT section interaction for X 3 release (Figure 3b) and no effect of Econase XT, but there was a highly significant effect of GIT section (P < 0.001).The X 3 content was highest in the ileum followed by the jejunum and colon, with the lowest concentrations observed in the caeca.As for X 3 , there was no enzyme x GIT section interaction nor any effect of Econase XT for X 2 release (Figure 3c), but there was a highly significant effect of GIT section (P < 0.001).The concentration of X 2 declined down the GIT, although the concentrations in the ileum and colon were not significantly different, nor were those in the colon and caeca.
As observed for X 4 , there was a significant enzyme x GIT section interaction for xylose (P < 0.01, Figure 4a).Econase XT increased xylose content in the colon and caeca, with no detectable xylose in the jejunum or ileum of any of the birds, irrespective of whether the enzyme was present.Unlike xylose release, there were no significant enzyme x GIT section interactions for arabinose, galactose or glucose (Figure 4b, c and d, respectively).Econase XT significantly increased the release of arabinose (P = 0.012) and tended to increase Galactose (P = 0.071) in all sections of the GIT.There were significant differences between the different sections of the GIT in the release of arabinose, galactose and glucose (P < 0.001).Arabinose concentrations were higher in the colon than in the jejunum and ileum, with the caeca being intermediate (Figure 4b), whereas galactose was highest in the colon (Figure 4c).In contrast, glucose concentrations declined down the GIT but with no difference between the colon and caeca (Figure 4d).
As xylanase increased X 4 in the colon (but not in the caeca) and xylose in the caeca, correlations were performed to investigate relationships between colonic X 4 , X 3 or X 2 contents and caeca xylose or other monosaccharides.The correlations for X 3 or X 2 were not significant (Supplemental Table 1).There appeared to be positive relationships between X 4 and both xylose and arabinose contents, but negative relationships were seen for galactose and glucose contents.However, only colonic X 4 and caecal arabinose contents showed a trend for a positive correlation (P = 0.089, Figure 5b), whereas none of the other correlations were significant (Figure 5).

Proportions of volatile fatty acids in the caeca
Acetic acid was the most abundant VFA in the caeca (Figure 5), but its total concentration was not statistically different between samples from the control birds (91.7 mM) and Econase XT birds (97.8 mM, P = 0.338).However, samples from the control birds had a slightly lower (P = 0.062) proportion of acetic acid (74%) than birds supplemented with Econase XT (78%).The proportions of other VFA were in the order butyric acid>lactic acid>propionic acid>branchedchain fatty acids>valeric acid (Figure 6), but there were no significant differences in proportions in the caeca of birds fed the control or Econase XT supplemented diets.

Discussion
According to the available published papers, this is the first in vivo study to quantify the generation of XOS and release of monosaccharides through the GIT of broilers and compare these findings to an in vitro model of digestion for the  A:X ratio calculated as Arabinose (g/100 g cereal) divided by Xylose (g/100 g cereal) ± SD of four replicates.
same enzyme and diet combination.The main findings of this study are that Econase XT supplementation in a wheatbased diet increased X 4 content in the colon, arabinose content throughout the GIT and xylose content in the colon and caeca of male broilers.Although other XOS and monosaccharides were detected in the GIT, they were unaffected by Econase XT supplementation.
There was a higher quantity of X 4 in the colon of broilers fed Econase XT, but this was not seen in the caeca, where X 4 was undetectable in both treatment groups.Econase XT increased the generation of X 4 along the GIT from the jejunum to the ileum before peaking in the colon and being negligible in the caeca.There are currently no reports of X 4 transporters, and AXOS are considered to be resistant to digestion via saliva, gastric juices, pancreatin and enzymes secreted from the intestinal mucosa, limiting their absorption (Fujikawa et al. 1991).Therefore, the reduced X 4 content between the colon and caeca was likely a result of factors acting on AXOS in the GIT, potentially the microbiome.These factors may be responsible for the increased xylose .Two-way ANOVA indicated significant enzyme x time interactions for X 4 , X 3 , X 2 (all P < 0.001), xylose (P = 0.034) and arabinose (P < 0.001).There were significant effects of time for galactose (P < 0.001) and glucose (P < 0.001), but no effects of enzyme.contents seen in the colon and caeca.AXOS are known to be readily fermented by the microbial populations that inhabit the caeca (Kiriyama et al. 1992).Their fermentation, along with that of xylose, can lead to greater acetate production, so the increased proportion of acetate observed with Econase XT potentially indicated greater fermentation of AX products in the caeca (Johnson et al. 2006).Interestingly, there were increases in X 4 generation associated with Econase XT, but not X 2 or X 3 .This suggested that the endogenous xylanase enzymes present in both the wheat source and in the GIT were able to hydrolyse bonds in the AX chain associated with X 2 -X 3, but the addition of Econase XT was necessary to create AXOS products with more than three degrees of polymerisation.Since X 2, X 3 and X 4 are not considered to be absorbed due to a lack of transporter system, concentrations of each XOS might be expected to increase down the GIT, before reaching the caeca where they may be metabolised by the microbiome.This was true for X 4 , but X 2 declined and X 3 remained constant along the GIT, suggesting that X 2 may be more easily degraded to xylose in the lower small intestine or that there are differential specificities for oligosaccharide fermentation with different degrees of polymerisation.It is important to remember that the concentrations along the GIT reflect a single time point, the balance between generation of XOS and monosaccharides by enzymatic action and disappearance by absorption/utilisation in vivo, whereas there was no disappearance mechanism in the in vitro model.
Comparing the data, the X 4 generation in vitro was most like the in vivo findings.At the 9 h time point in vitro, there was a clear increase in X 4 associated with Econase XT addition to the diet, which may reflect the digesta reaching the colon in vivo, where X4 contents were increased by Econase XT.There was no effect of Econase XT on the in vitro levels of X 3 , which was reflected in vivo.In contrast, there was an increase in X 2 in vitro, but only at the 24 h time point, which did not reflect in vivo digests.This could be because broiler GIT transit time was more rapid, with the marker first appearing in the faeces 90 min after consumption with a peak at 220 min (Summers and Leeson 1986).
The monosaccharides, arabinose and galactose, both increased linearly over 24 h during the in vitro digests, with arabinose release being greater in the presence of Econase XT during the final 12 h of digestion compared to the control.This contrasted with the in vivo profiles, where arabinose and galactose were either very low or undetectable in the jejunum or ileum, yet markedly higher in the colon, before decreasing again in the caeca.This could be because enzyme hydrolysis of arabinose and galactose did not occur before reaching the colon, or because arabinose and galactose were rapidly removed from the GIT by their corresponding transporter proteins in the jejunum and ileum.Arabinose and galactose appear to be well utilised in the early stages of GIT, based on data from other species (Schutte 1990;Csáaky and Ho 1965;Wagh and Waibel 1966).
Glucose release increased linearly throughout the 24 h time period in vitro, contrasting with in vivo, where glucose levels decreased down the GIT, with no effect of Econase XT either in vitro or in vivo.The gradual decrease down the GIT was likely due to glucose being released from the starch component of the diet, followed by rapid absorption in the jejunum and ileum (Klasing 1998), via both glucose transporter (GLUT) and sodium-glucose cotransporter (SGLT) proteins (Braun and Sweazea 2008), resulting in little getting through to the colon and caeca.
There were only increases in xylose release due to Econase XT supplementation at 24 h in vitro, whereas xylose contents were significantly higher in the colon and caeca of Econase XT supplemented birds.The lack of xylose seen in the jejunum and ileum was probably due to rapid absorption in the small intestine, as previously described (Schutte et al. 1991;Longstaff et al. 1988).The increase in xylose in the colon and caeca may have resulted from colonic and caecal bacteria producing their own xylanases after being activated by Econase XT, which more aggressively attacked the fibre component of the diet.This suggested that Econase XT supplementation may be inducing adaptations to the microbiome in the colon and caeca over time, rather than providing an acute effect of increasing AX hydrolysis.This was supported by the ratio of free arabinose and xylose being closer to one, which was much higher than would be susceptible to Econase XT attack, suggesting that activity other than the added Econase XT might have been responsible for the release of these monomers.
One question was why X 3 and X 2 concentrations along the GIT were not different between controls and Econase XT treated birds?It has previously been shown that wheat can contain endogenous xylanase activity (Dornez et al. 2008), suggesting that the diet used contained endogenous activity that generated XOS without the inclusion of exogenous Econase XT.The generation of XOS without any exogenous enzyme may be due to the birds being fed a mash rather than a pelleted diet.Pelleting requires high temperature and pressure, which would be expected to denature the proteins Xylotriose (X 3 ) c: Xylobiose (X 2 ).Two-way ANOVA indicated a significant enzyme x gut section interaction for X 4 (p = 0.038) and significant effects of gut section for both X 3 and X 2 (both P < 0.001).a,b,c,d Columns with different superscript letters were significantly different (P < 0.05, Bonferroni post hoc test).In b and c post hoc Bonferroni tests were for effects of gut section.present, thereby reducing any endogenous xylanase activity.The results from the in vitro digests for the same diet would appear to confirm this, with a similar release of xylose, galactose and glucose observed with or without Econase XT addition.This suggested that endogenous xylanase causes the release of certain monosaccharides in vitro, but this is not  enhanced further by exogenous xylanase.It was therefore suggested that there might be greater effects of Econase XT in broilers fed a pelleted diet or a mash diet where the wheat contained either less endogenous xylanase activity or more xylanase inhibitors.Despite this, there were still significant effects of Econase XT on the generation of X 4, xylose and arabinose, particularly in the colon and caeca.It was proposed that the Econase XT produced more X 4 in the colon, thereby activating the gut microbiome through the provision of substrate, leading to further enzyme activity to produce the higher levels of xylose and arabinose seen in the colonic digesta and caeca.This appeared to be supported by the correlation analysis, where there tended (P = 0.089) to be more arabinose in the caeca of broilers, with the highest X 4 contents being in the colon.There was a similar positive relationship for caecal xylose and colonic X 4 , although this was not significant (P = 0.112).The fact that similar relationships were not seen for X 3 and X 2 suggested that undigested X 4 produced in the colon was being utilised in the caeca, potentially resulting in the increased xylose and proportion of acetate seen in the caeca of Econase XT supplemented broilers.This increase in the proportion of acetate in the caeca with Econase XT supplementation agreed with previous studies (Kabel et al. 2002;Ravn et al. 2018), showing that acetate and butyrate were produced by the fermentation of pure AXOS or via xylanase activity.However, there was no effect of Econase XT on butyrate in the present study.Whether the observed increase in acetate represented a change in the microbiome is unclear, but did support the hypothesis that the chicken microbiome can be 'trained' or adapted over time as a potential mechanism for the effects of xylanase supplementation.
The lack of effect on bird performance could be due to the limited time frame, and late exposure to Econase XT, as most studies supplemented xylanase from day of hatch, when the microbiome is rapidly developing.Birds were given Econase XT supplementation from four days old, for a total of 21d.
This exposure time may be too short or may be before the birds' microbiome had established or had completely adapted.A recent study (Figueiredo et al. 2012) showed that xylanase supplementation from one day of age only had significant effects on FCR at 28 days of age.
The present study showed that there was degradation by Econase XT of AX in the wheat-based diet, as indicated by higher xylose, arabinose and X 4 levels present in digesta, particularly in the colon and caeca of male broilers.This may have been responsible for the observed increase in acetate production by the caecal microbiome and indicative of the microbiome adapting.Although these effects did not result in improved performance of the birds, this could be explained by (a) the presence of endogenous xylanase activity in the diet, particularly as it was not pelleted, and (b) the relatively short exposure time to Econase XT at such an early stage of broiler development.

Disclosure statement
No potential conflict of interest was reported by the authors.Figure 6.Effects of Econase XT inclusion in the diet on the proportions of Volatile Fatty Acids (VFA) in the caeca of male broilers.Data are means of 10 randomly selected birds (one from each pen) expressed as % of the total VFA content.There was a trend for Econase XT supplementation to increase the proportion of acetic acid in the caeca (P = 0.062). 3

Figure 1 .
Figure1.Time-dependent release of Xylo-oligosaccharides and monosaccharides from the wheat-based diet during an in vitro incubation at 41°C in the presence or absence of Econase XT for 24 hours.a: Xylotetraose (X 4 ) b: Xylotriose (X 3 ) c: Xylobiose (X 2 ) d: Xylose, e: Arabinose, f: Galactose, g: Glucose.Data show mean values ± standard error of the mean (SEM).Two-way ANOVA indicated significant enzyme x time interactions for X 4 , X 3 , X 2 (all P < 0.001), xylose (P = 0.034) and arabinose (P < 0.001).There were significant effects of time for galactose (P < 0.001) and glucose (P < 0.001), but no effects of enzyme.

Figure 2 .
Figure2.Effect of inclusion of Econase XT in diet on broiler growth.Data show mean values ± standard error of the mean (SEM).Data represent ten replicate pens per treatment, with four birds per pen.There was no enzyme x time interaction (P = 0.716), nor any effect of enzyme (P = 0.686), but there was a significant effect of time (P < 0.001).

Figure 3 .
Figure3.Effects of Econase XT inclusion in the diet on release of Xylooligosaccharides at different sections of the broiler gastro-intestinal tract (GIT).Data show mean g/100 g of digesta dry matter ± standard error of the mean (SEM) for 10 replicates per treatment (1 bird per pen per treatment) for a: Xylotetraose (X 4 ) b: Xylotriose (X 3 ) c: Xylobiose (X 2 ).Two-way ANOVA indicated a significant enzyme x gut section interaction for X 4 (p = 0.038) and significant effects of gut section for both X 3 and X 2 (both P < 0.001).a,b,c,d Columns with different superscript letters were significantly different (P < 0.05, Bonferroni post hoc test).In b and c post hoc Bonferroni tests were for effects of gut section.

Figure 4 .
Figure 4. Effects of Econase XT inclusion in the diet on release of Monosaccharides at different sections of the broiler gastro-intestinal tract (GIT).Data show mean g/100 g of digesta dry matter ± standard error of the mean (SEM) for 10 replicates per treatment (1 bird per pen per treatment) for a) Xylose, b) Arabinose,c) Galactose and d) Glucose.Two-way ANOVA indicated a significant enzyme x gut section interaction for xylose (P < 0.001) and significant effects of gut section for arabinose, galactose and glucose (all P < 0.001).There was a significant effect of enzyme for Arabinose (P = 0.012) and a trend for galactose (P = 0.071).a,b,c,d Columns with different superscript letters were significantly different (P < 0.05, Bonferroni post hoc test).In B, c and d post hoc Bonferroni tests were for effects of gut section.

Table 2 .
Monosaccharide composition, total Arabinoxylan (AX) content and Arabinose: Xylose (A:X) ratio of the wheat-based diet after Trifluoroacetic acid hydrolysis of non-cellulosic polysaccharides.

Table 3 .
Effect of inclusion of Econase XT in diet on broiler performance.