In vitro digestion techniques for simulating physiological processes are widely extended in animal nutrition. Generally, enzymatic approaches have been applied for digestion gastric and small intestine in studies for monogastrics (Boisen & Fernandez 1995, 1997; Yegani et al., 2013), with results obtained being similar and highly correlated with in vivo feeding studies (Chiang et al., 2008). However, the enzyme approach does not fit so well for microbial fermentation studies, since used enzymes are commonly from aerobic fungi, and do not necessarily act over the substrate at the same level, and conditions for their optimal activity does not necessarily match with environmental conditions.

Previous efforts have been published to adapt the gas production technique to studies of hindgut fermentation in monogastric animals, such as pigs (Fondevila et al., 2002; Williams et al., 2005), horses (Lowman et al., 1999) or rabbits (Calabró et al., 1999; Rodríguez-Romero et al., 2011). Special consideration must be paid to the fact that incubated substrate must be pre-digested simulating gastric and small intestine digestion to accurately simulate characteristics of what arrives to the caecum (Fondevila et al., 2002).

As studied for ruminants, a major issue in the application of the gas production technique is the restricted viability of the microbial inoculum, that force to use it fresh, immediately after extraction (Mould et al., 2005). The immediate use of inoculum (fresh) in in vitro fermentation studies is recommended for most procedures (Menke & Steingass, 1988; Theodorou et al., 1994). However, this is often difficult to ensure in practical situations. Alternatively, refrigeration of microbial inoculum has been proposed, provided that inoculum is maintained at 4-6ºC for less than 6 h (Robinson et al., 1999; Hervás et al., 2005). Prates et al. (2010) demonstrated that the freezing process of microbial inoculum, either in liquid N (-196ºC) or even in a regular freezer (-20ºC) do not noticeably affect fermentation activity of the microbial community, either quantitative or qualitatively, respect to the conventional use of fresh inoculum. Thus, preservation of digesta contents for being used in subsequent in vitro studies is considered an important issue that improves the advantages of in vitro methods. The response to freezing microbial inoculum has been previously tested for ruminants, but it has not yet been validated for fermentation studies in monogastrics, as it is the objective of the present study. Methods were tested with a common substrate supplemented with different additives aiming to reach a range of in vitro microbial fermentative activity.

with the preservation method as the main plot (Mi, n=2) and the donor pig as a block (Pi, n=3) that were tested against the interaction PMij, and the dietary additive as the subplot (Ak, n=4) and the interaction MAjk (n=6) that were contrasted against the residual εl(ijk). The average of three bottles for each treatment incubated with the same inoculum (incubation run) was considered as the experimental unit. In case of detecting significant treatment effects, differences were contrasted by the Tukey t-test at p<0.05. A trend to significance was considered when p<0.10.

Addition of riboflavin to the substrate increased in a 0.09 proportion (14 to 18 mL gas/g OM) the volume of gas produced respect to control from 8 h onwards, and it was also higher that with CRINA-TEP from 4 to 10 h incubation (p<0.05). No effect of the interaction additive × preservation method was recorded (p>0.05), but since the major objective of this work was to compare among inocula, the gas production pattern of incubated substrates with fresh or frozen inocula is shown separately in Figure 1.

Comparison of gas production of fresh and frozen caecal inocula is shown in Figure 2. The gas production pattern from samples incubated with liquid N frozen inoculum did not differ from that with fresh inoculum (p>0.10), and only a trend (p=0.081) for a lower volume with the frozen inoculum was noticed at 4 h incubation. A significant effect of the additive included on the volume of gas produced was detected from 4 h onwards (p<0.05). In any case, the interaction inoculum x additive was also not significant (p>0.05) throughout the incubation period.

As it is showed by the ANOVA, the proportion of the effect of donor pigs to overall variation in gas production ranged from 0.32 to 0.58. Thus, the coefficient of variation of the donor pig × preservation in the split-plot analysis was relatively low (ranging 0.066-0.152 from 4 to 12 h), highlighting the homogeneity in the response to the inoculum preservation.

No effect of the dietary additive was recorded on methane production, ammonia or SCFA concentration, nor on the molar SCFA proportions along the incubation (p>0.10). Since the interaction additive × preservation method also resulted non-significant (p>0.10), Table 1 only shows average means for the preservation methods for these parameters at 6 and 12 h incubation. When comparing preservation methods, no differences (p>0.05) were recorded on CH₄ production along the 12 h incubation between fresh and frozen inocula, either in terms of volume produced or on its proportion of total gas. The ammonia concentration after 6 h incubation tended (p=0.062) to be higher with frozen than fresh inocula, but this trend disappeared at 12 h (p>0.10). Total SCFA concentration or their molar proportions were not affected by the preservation method of inoculum at 6 nor at 12 h incubation (p>0.05).

Considering the location of the fermentation site in the digestive tract of monogastric animals, as well as the fact that feeds for pigs and poultry are mostly compound of digestible carbohydrates and proteins that disappear in the first part of the digestive tract, pre-digestion with gastric and small intestine enzymes is a necessary step for the evaluation of hindgut fermentation in these animals (Fondevila et al., 2002). In our experiment, the pepsin-pancreatin treatment digested, on average, a proportion of 0.80 of initial feed DM, which should obviously affect the composition of incubated substrate (not analysed). As a further reference, simulation of polysaccharide digestion with enzymes (Viscozyme) as the last step of the digestion enzymatic process (Boisen & Fernández, 1997) digested 0.453 ± 0.115, but it accounted for only 0.051 ± 0.0138 of total feed DM digestion.

The freezing at -70ºC has been applied as routine procedure for preserving pure bacterial cultures in microbiological studies, either with or without cryoprotectants (Gibson & Khoury, 1986; Malik, 1991; Perry, 1998), despite Fonseca et al. (2001) suggested some selective effect depending on the microbial membrane structure. in vitro fermentation studies in which freezing failed to maintain microbial fermentative capacity (Robinson et al., 1999; Hervás et al., 2005; Denek et al., 2010) used large batches of inoculum, and did not consider the time of freezing and thawing in the preparation process. In this regard, a fast processing is mandatory for ensuring the maintenance of cell viability and fermentative capacity (Malik, 1991; Perry, 1998). Dispensing in tubes with high surface to volume ratio should ensure a fast processing, and therefore the viability of the microbial community of the inoculum.

Other more recent approaches in ruminant fermentation studies have assayed the ability of frozen inoculum, either with or without cryoprotectants, to maintain similar fermentation characteristics than fresh inoculum. Chaudhry & Mohamed (2012) obtained a high correlation among fresh and frozen (-20ºC) inoculum in in vitro DM degradation pattern but results were of 0.20 lower magnitude. Belanche et al. (2019) froze small volumes of inoculum (35 mL) at -80ºC using DMSO as cryoprotectant, and Spanghero et al. (2019) froze 250-300 mL aliquots at -80ºC without cryoprotectant, but both works detected a lower gas production than the fresh inoculum. Despite the differences in freezing temperatures, it can be suspected that there is a common methodological aspect in all these efforts: either the freezing or the thawing periods, or both, were carried out relatively slowly. This might be due because of large volumes may take delay freezing and thawing, thus maintaining exposure of microbial community to low temperatures for an extended period of time, thus reducing its viability, as it could be the case of Robinson et al. (1999), Hervás et al. (2005), Chaudhry & Mohamed (2012) or Spanghero et al. (2019). In other sense, the thawing process lasted for less than one minute in Denek et al. (2010), 5 minutes in Belanche et al. (2019), 2 h in Spanghero et al. (2019) and 12 h in Chaudhry & Mohamed (2012). Further, in the case of Denek et al. (2010) and Belanche et al. (2019), where small amounts of inoculum were frozen in liquid N, DMSO was used as cryoprotectant, but this was not removed from the medium and its toxicity at culture temperatures (Hubálek, 2003) might probably affect microbiota when became stable at optimum temperature. This reinforces the work from Prates et al. (2010) who hypothesised that the length of the freezing and thawing processes is critical, and that a cryopreservant is not really necessary for studies of microbial fermentation. Further, the use of glycerol as cryoprotectant enhanced later fermentation as it was used as energy source for microorganisms. Besides, recovery of frozen culturable rumen bacteria is in general adequate (Zehavi et al., 2018) and activity of other microbial types such as protozoa is minimal under conventional in vitro media (Dehority, 2008). In other way, the freezing temperature does not seem to be determinant: Prates et al. (2010) did not observe major differences between freezing at -20ºC or in liquid N, provided that inoculum was dispensed in small tubes for maximising the surface to volume ratio. In the present experiment, the same protocol of Prates et al. (2010) for preserving and activating the inoculum was followed for pig caecal contents, and the obtained results confirm the positive responses obtained by those authors.

Result from our study showed that both types of inoculum behaved similarly in terms of rate and magnitude of substrate fermentation. Further, individual responses for each donor animal also showed a similar fermentative pattern. Values of fermentation parameters did not differ among inocula, as well as for ammonia concentration at 12 h incubation. It is worth considering that no differences on molar volatile fatty acids proportions were recorded by Prates et al. (2010) despite the processing method of the inocula (fresh, refrigerated at 6ºC for 4 h, frozen at -20ºC or frozen in liquid N) after 6 or 24 h incubation.

In a recent study (Ma et al., 2022) it was observed that, despite overall microbial diversity was not affected by freezing rumen inoculum, it affected the abundance of certain genera of bacteria, and Ruminococcus increasing and Christensenellaceae R7 tending (p=0.09) to decrease, the latter being involved in butyrate production (Morotomi et al., 2012). However, results from Ma et al. (2022) do not support a selective positive effect on protein degradation bacteria that should justify the higher ammonia concentration observed in the medium with frozen inoculum. In line with the above discussion, it is worth considering that, in their work inoculum was frozen in liquid N, but thawing was carried out at 37ºC for 10 min. In any case, further microbiological analyses are necessary to determine these effects on caecal microbial fermentation.

In conclusion, despite the mentioned approaches in in vitro studies for ruminants, to our knowledge this has not been boarded with inoculum from monogastric animals. Our results confirm that microbial inoculum can be dispensed in tubes, frozen in liquid N and stored at -80ºC without compromising its in vitro fermentative activity, either quantitatively or qualitatively.