Is Oral Vaccination against Escherichia coli Influenced by Zinc Oxide?

Simple Summary Zinc oxide remains a widely used compound for the control and prevention of post-weaning Escherichia coli (E. coli) diarrhoea in piglets. It is sometimes administered concomitantly with oral E. coli vaccines. In this research, we assessed the influence of the administration of zinc oxide in piglet feed in combination with a bivalent vaccine against E. coli. We studied the immune activation, intestinal integrity, production of secretory IgA, and excretion of E. coli via faecal samples at different times post-vaccination. Although the main difference observed was the excretion of the E. coli vaccine strain, the immune response determined in both vaccine groups was similar, irrespective of the presence of ZnO in the feed. Abstract Background: Although zinc oxide has been banned at therapeutic doses in the EU, its use is still legal in most countries with industrial pig farming. This compound has been shown to be very effective in preventing E. coli-related diseases. However, another strategy used to control this pathogen is vaccination, administered parenterally or orally. Oral vaccines contain live strains, with F4 and F18 binding factors. Since zinc oxide prevents E. coli adhesion, it is hypothesised that its presence at therapeutic doses (2500 ppm) may alter the immune response and the protection of intestinal integrity derived from the vaccination of animals. Methods: A group of piglets were orally vaccinated at weaning and divided into two subgroups; one group was fed a feed containing 2500 ppm zinc oxide (V + ZnO) for the first 15 days post-vaccination (dpv) and the other was not (V). Faeces were sampled from the animals at 6, 8, 11, 13, and 15 dpv. Unvaccinated animals without ZnO in their feed (Neg) were sampled simultaneously and, on day 15 post-vaccination, were also compared with a group of unvaccinated animals with ZnO in their feed (ZnO). Results: Differences were found in E. coli excretion, with less quantification in the V + ZnO group, and a significant increase in secretory IgA in the V group at 8 dpv, which later equalised with that of the V + ZnO group. There was also some difference in IFNα, IFNγ, IL1α, ILβ, and TNFα gene expression when comparing both vaccinated groups (p < 0.05). However, there was no difference in gene expression for the tight junction (TJ) proteins responsible for intestinal integrity. Conclusions: Although some differences in the excretion of the vaccine strain were found when comparing both vaccinated groups, there are no remarkable differences in immune stimulation or soluble IgA production when comparing animals orally vaccinated against E. coli in combination with the presence or absence of ZnO in their feed. We can conclude that the immune response produced is very similar in both groups.

with a capacity of 7 piglets per pen. One room per treatment was used, and each group included 42 animals housed in 6 pens. Feed was administered by means of a 4-hole hopper, with free disposal of feed, and water was supplied ad libitum by means of a nipple drinker.
Four groups were set up: a negative control group without vaccination or therapeutical levels of zinc oxide in their feed (Neg), a positive control non-vaccinated group with 2500 ppm zinc oxide (ZnO) in their feed, a vaccinated group without zinc oxide in their feed (V) and, finally, a vaccinated group with 2500 ppm zinc oxide in their feed (V + ZnO). Zinc oxide was administered during the first 15 days after weaning. A schematic of the experimental design is shown in Figure 1.
One hundred sixty eight pure Large White piglets from the farm of the University of Murcia were used. The farm is in Guadalupe de Maciascoque (Murcia, Spain; 38.00662156425343 N, −1.1749983147290364 W). Piglets were subjected to usual farm management, including treatment with toltrazuril, iron dextran, and tooth-filing on the third day.
The animals were uniquely identified with numbered ear tags and were weaned at an age of 24 ± 1.2 days. After weaning, the animals were housed in nursery rooms of 6 pens with a capacity of 7 piglets per pen. One room per treatment was used, and each group included 42 animals housed in 6 pens. Feed was administered by means of a 4-hole hopper, with free disposal of feed, and water was supplied ad libitum by means of a nipple drinker.
Four groups were set up: a negative control group without vaccination or therapeutical levels of zinc oxide in their feed (Neg), a positive control non-vaccinated group with 2500 ppm zinc oxide (ZnO) in their feed, a vaccinated group without zinc oxide in their feed (V) and, finally, a vaccinated group with 2500 ppm zinc oxide in their feed (V + ZnO). Zinc oxide was administered during the first 15 days after weaning. A schematic of the experimental design is shown in Figure 1.

Vaccine and Vaccination
The animals were orally vaccinated with the Coliprotect F4 + F18 vaccine (Elanco GmbH, Monheim am Rhein, Germany), which contains live non-pathogenic E. coli strain O8:K87 and live non-pathogenic E. coli strain O141:K94 (1.3 × 10 8 to 3.0 × 10 9 ), thus ensuring the presence of antigens F4ac and F18ac. The animals were vaccinated the day after weaning, when the piglets had adapted to the nursery. The vaccine was reconstituted with mineral water and 2 mL was administered to each piglet individually by drenching.
The animals were housed in four rooms separated by stalls. The two feeds (with or without ZnO) used for the vaccinated groups were randomly assigned to each room.

Feeds
All feeds were produced using a basal diet, the formula for which is given in Table 1.
To the feeds of the ZnO and V + ZnO groups, 2500 ppm of zinc oxide was added, while the basal diet had 150 ppm of zinc oxide.

Vaccine and Vaccination
The animals were orally vaccinated with the Coliprotect F4 + F18 vaccine (Elanco GmbH, Monheim am Rhein, Germany), which contains live non-pathogenic E. coli strain O8:K87 and live non-pathogenic E. coli strain O141:K94 (1.3 × 10 8 to 3.0 × 10 9 ), thus ensuring the presence of antigens F4ac and F18ac. The animals were vaccinated the day after weaning, when the piglets had adapted to the nursery. The vaccine was reconstituted with mineral water and 2 mL was administered to each piglet individually by drenching.
The animals were housed in four rooms separated by stalls. The two feeds (with or without ZnO) used for the vaccinated groups were randomly assigned to each room.

Feeds
All feeds were produced using a basal diet, the formula for which is given in Table 1.
To the feeds of the ZnO and V + ZnO groups, 2500 ppm of zinc oxide was added, while the basal diet had 150 ppm of zinc oxide.

Weighing of Animals
Piglets were weighed at weaning, and at days 0, 8, 14, and 25 post-weaning, using a GRAM K2 scale (GRAM, Barcelona, Spain) suitable for weighing animals. It was decided not to weigh the animals every day of faecal sampling to avoid continuous stress to the animals. With these data, the average daily gain (ADG) was calculated over each of the periods (ADG 14 ; from day 0 to 14 and ADG 25 : from day 0 to 25). Personal protective equipment (PPE) and specific boots were used to enter each of the rooms to avoid transferring bacteria between rooms.

Faecal Sampling
Faecal samples were taken from 20 animals from each group and divided into two portions: one 100 mg portion was preserved by RNAlater (Invitrogen, Waltham, MA, USA) for gene expression quantification and the other for DNA isolation for E. coli quantification. The former was kept refrigerated for 24 h after sampling and then kept at −80 • C until analysis; the portion used to obtain DNA was immediately placed in a −80 • C environment. The animals were left for 6 days following vaccination to allow for multiplication of the vaccine strain and, thereafter, they were sampled on 6,8,11,13, and 15 days postvaccination (dpv). Sampling ceased on day 15, as this was the last day the animals ate feed containing zinc oxide. The ZnO group was only sampled 15 dpv, as an indicator at the end of the study.

Isolation of Nucleic Acids
Nucleic acids were isolated by means of Genejet RNA purification and Genejet PCR purification kits (Thermo Fisher, Waltham, MA, USA). The quantity of DNA and RNA was quantified using a Nanodrop 2000 (Thermo Fisher, Waltham, MA, USA). Regarding the RNA samples, cDNA was synthesised using oligo dT as a primer in order to utilise only the mRNA in the synthesis.

Secretory IgA Quantification
The secretory IgA content of the RNAlater used to preserve the faeces was quantified. The preservative was recovered before RNA isolation, centrifuged at 3900× g for 10 min, and preserved at −80 • C until analysis. IgA quantification was performed as described in the literature [12,31] using the Pig IgA ELISA kit (Bethyl Laboratories, Montgomery, TX, USA) according to the manufacturer's specifications.

Statistical Analysis
All data were analysed using the SPSS program (SPSS Inc., Chicago, IL, USA). The results of the production data were analysed by means of a one-way ANOVA. For the comparison of quantifications, a Kruskal-Wallis test was used, as well as a subsequent Mann-Whitney U test to constitute a two-samples-comparison analysis. Violin plots were produced for comparison of results with GraphPad Prism v. 8.2 (GraphPad, San Diego, CA, USA). In addition, correlations among parameters were calculated as partial correlations controlled for group and sampling. The significance level and tendency were set at p ≤ 0.05 and p ≤ 0.10, respectively.
For data reduction, Discriminant Function Analysis (DFA), which uses Wilks' lambda to determine the efficiency of analysis and generates as many functions as it deems necessary to explain the model, was used. A dot plot was generated when two or more functions were produced.

Weights and Growth
The mean weights found for each group in each sampling are shown in Figure 2.
to determine the efficiency of analysis and generates as many functions as it deems necessary to explain the model, was used. A dot plot was generated when two or more functions were produced.

Weights and Growth
The mean weights found for each group in each sampling are shown in Figure 2. Average daily gains from day 0 to 14 (ADG14) and from day 0 to 25 (ADG25) are shown in the table below (Table 3). Table 3. Average daily gain (ADG) at the 14th and 25th days of experiment, sorted by groups.

Parameter
Neg A significant difference in ADG14 was found when comparing the Neg group with all other groups, while there was no difference between the ZnO and V groups, and there was a significant difference between the ZnO and V + ZnO groups and between the V and V + ZnO groups ( Table 2). For ADG25, the main differences were between the ZnO groups with the V group and the V + ZnO group and between the V and V + ZnO groups. In this case, the highest ADG corresponded to the ZnO group, followed by Neg and the two vaccinated groups, which obtained the lowest ADG25. In the first part of the experiment, the vaccinated groups demonstrated significantly greater growth than the two control groups, but in the second half of the study, there was compensatory growth, especially in the ZnO Average daily gains from day 0 to 14 (ADG 14 ) and from day 0 to 25 (ADG 25 ) are shown in the table below (Table 3). A significant difference in ADG 14 was found when comparing the Neg group with all other groups, while there was no difference between the ZnO and V groups, and there was a significant difference between the ZnO and V + ZnO groups and between the V and V + ZnO groups (Table 2). For ADG 25 , the main differences were between the ZnO groups with the V group and the V + ZnO group and between the V and V + ZnO groups. In this case, the highest ADG corresponded to the ZnO group, followed by Neg and the two vaccinated groups, which obtained the lowest ADG 25 . In the first part of the experiment, the vaccinated groups demonstrated significantly greater growth than the two control groups, but in the second half of the study, there was compensatory growth, especially in the ZnO group. This shows that even in the absence of vaccination or E. coli prophylaxis, animals eventually adapt and recover homeostasis and compensatory growth.

Frequency of Positive Samples
The frequencies of positive samples for each E. coli-related gene tested, sorted by sampling day, are shown in the table below ( Table 4).
As expected, none of the Neg samples showed amplification for CRISPR, as this gene is only present in the vaccine strain. This indicates that there was no transfer from the vaccine strain to the Neg group throughout the experiment. On the other hand, none of the samples from the Neg group were positive for adhesion factor F4. These two findings suggest that all samples with F4 amplification were amplified as a consequence of vaccination, since this fimbria is present in the vaccine strain. The main differences were found in the 6 and 8 dpv samples, for which group V showed higher frequencies for F4 and CRISPR. On day 8 postvaccination, both vaccinated groups had an increased frequency of F4, the V + ZnO group had a slight increase in CRISPR frequency, and both vaccinated groups had an increased Animals 2023, 13, 1754 7 of 21 frequency of O8, while the Neg group had reduced frequencies of positive samples. On 11, 13, and 15 dpv, there were differences in the frequency of O8-positive samples when comparing the vaccinated groups and the V + ZnO group with the Neg group. In the vaccinated groups, the frequency of this gene decreased throughout the experimental period while, in the Neg group, it increased. Table 4. Frequency of ECOTotal-, F4-, H7-, CRISPR-, and O8-positive samples. The p-value indicates differences between expected and observed frequencies and AR = adjusted residual for positive samples. The p-values indicating statistically significant differences are marked in bold. The frequency of CRISPR-positive samples showed a bimodal distribution over the duration of the experiment for the V group, but not for the V + ZnO group; in the V group, 55% of the pigs excreted E. coli with this gene in the first sampling. This percentage was maintained on day 8, decreased to 15% on day 11, and increased again to 35% on day 13 before decreasing slightly on day 15. In the V + ZnO animals, the frequency of CRISPR-positive samples started from 40% 6 dpv, increased to 55% 8 dpv, dropped to 15% on day 11, but did not increase as in the V group, instead remaining at 15% until the end of the experimental period ( Figure 3). duration of the experiment for the V group, but not for the V + ZnO group; in the V group, 55% of the pigs excreted E. coli with this gene in the first sampling. This percentage was maintained on day 8, decreased to 15% on day 11, and increased again to 35% on day 13 before decreasing slightly on day 15. In the V + ZnO animals, the frequency of CRISPRpositive samples started from 40% 6 dpv, increased to 55% 8 dpv, dropped to 15% on day 11, but did not increase as in the V group, instead remaining at 15% until the end of the experimental period ( Figure 3). The Ct for ECOTotal for each group and for each sampling date is shown in Figure  4. The Ct for ECOTotal for each group and for each sampling date is shown in Figure 4.

Quantification of E. coli Genes
The quantification of E. coli (ECOTotal) showed significant differences between the vaccinated and control groups at all sampling dates except 11 dpv, with a higher Ct in the V + ZnO group. The difference between the Neg and V + ZnO groups increased linearly (r = 0.924, p = 0.025) from 3.7 Ct (6 dpv) to 8.8 Ct (15 dpv). In the case of the comparisons between Neg and V groups, there were also differences in Ct except for the 6 dpv and 11 dpv samples, ranging from 0.2 cycles (6 dpv) to 4.84 more cycles at 15 dpv; this decrease was also linear, although not as marked or significant as in the comparison with the V + ZnO group (r = 0.794, p = 0.108) ( Figure 5).
After analysing the H7 gene, a significant difference was observed 15 dpv when comparing the Neg group with the V group (p = 0.001). However, when comparing the Neg group with the V + ZnO group, there were differences at 11 dpv (p = 0.002), 13 dpv (p = 0.033), and 15 dpv (p = 0.002). Differences between the V and V + ZnO groups were observed on days 11 (p < 0.001) and 13 (p = 0.016).
The CRISPR gene was not amplified in any of the samples from the Neg group, as shown above. Comparing the V and V + ZnO groups, no significant differences were

Quantification of E. coli Genes
The quantification of E. coli (ECOTotal) showed significant differences between the vaccinated and control groups at all sampling dates except 11 dpv, with a higher Ct in the V + ZnO group. The difference between the Neg and V + ZnO groups increased linearly (r = 0.924, p = 0.025) from 3.7 Ct (6 dpv) to 8.8 Ct (15 dpv). In the case of the comparisons between Neg and V groups, there were also differences in Ct except for the 6 dpv and 11 dpv samples, ranging from 0.2 cycles (6 dpv) to 4.84 more cycles at 15 dpv; this decrease was also linear, although not as marked or significant as in the comparison with the V + ZnO group (r = 0.794, p = 0.108) ( Figure 5). When analysing the Ct for the H7 gene for each sampling date, differences were observed in group V when comparing the 8 dpv sample with the 15 dpv sample (p = 0.006), and when comparing the 11 dpv sample with 13 dpv (p = 0.003) and 15 dpv (p = 0.010) samples. In the V + ZnO group, differences were found when comparing the Ct at 6 dpv with those at 11 dpv (p = 0.016), 13 dpv (p = 0.004), and 15 dpv (p = 0.05), and when comparing the 8 dpv sample with the 13 dpv sample (p = 0.001). Regarding CRISPR gene quantification, differences were only found in the V + ZnO group when quantification at 6 dpv was compared with that at 8 dpv (p = 0.008).
No differences were found between samples in any of the experimental groups for F4 and O8.
Positive correlations were found between all genes used to quantify E. coli excretion ( Table 5). After analysing the H7 gene, a significant difference was observed 15 dpv when comparing the Neg group with the V group (p = 0.001). However, when comparing the Neg group with the V + ZnO group, there were differences at 11 dpv (p = 0.002), 13 dpv (p = 0.033), and 15 dpv (p = 0.002). Differences between the V and V + ZnO groups were observed on days 11 (p < 0.001) and 13 (p = 0.016).
The CRISPR gene was not amplified in any of the samples from the Neg group, as shown above. Comparing the V and V + ZnO groups, no significant differences were found, although, in all cases, the Ct was higher in the V + ZnO group. Comparing the Ct, there was a linearly increasing difference (r = 0.924, p = 0.076) between 6 dpv (1) and 13 dpv (6.33). Finally, on day 15, this difference was limited to one cycle.
There were no differences between the Ct obtained for the Neg, V, and V + ZnO groups for the O8 and F4 genes in any of the analyses. None of the samples analysed from the Neg group were positive for F4.
When conducting a paired samples analysis, there were differences between the Ct for ECOTotal in group V when comparing 6 dpv quantification with that of 13 dpv (p < 0.001) and with that of 15 dpv (p < 0.001), as well as when comparing 11 dpv with 13 dpv samples (p = 0.01). In the V + ZnO group, however, differences were found when comparing quantification at 6 dpv with those at 11 dpv (p < 0.001), 13 dpv (p = 0.001), and 15 dpv (p = 0.011), and when comparing Ct at 8 dpv with those at days 11 (p < 0.0001), 13 (p = 0.001), and 15 (p = 0.001). The Neg group had no difference in Ct when comparing any of the samples.
When analysing the Ct for the H7 gene for each sampling date, differences were observed in group V when comparing the 8 dpv sample with the 15 dpv sample (p = 0.006), and when comparing the 11 dpv sample with 13 dpv (p = 0.003) and 15 dpv (p = 0.010) samples. In the V + ZnO group, differences were found when comparing the Ct at 6 dpv with those at 11 dpv (p = 0.016), 13 dpv (p = 0.004), and 15 dpv (p = 0.05), and when comparing the 8 dpv sample with the 13 dpv sample (p = 0.001). Regarding CRISPR gene quantification, differences were only found in the V + ZnO group when quantification at 6 dpv was compared with that at 8 dpv (p = 0.008).
No differences were found between samples in any of the experimental groups for F4 and O8.
Positive correlations were found between all genes used to quantify E. coli excretion (Table 5). Regarding total E. coli quantification, the highest correlation was for the H7 sample, with a similar correlation for F4 and CRISPR, again suggesting that F4 is acquired from vaccination, supported by the absence of amplification in the Neg group samples (Table 4).

Intestinal Integrity
Quantifications for the three TJ proteins are shown in Figure 6. For occludin, a fourfold increase in gene expression was found for the V group when compared to the Neg group (p = 0.016) at 6 dpv. At 8 dpv, a 19-and 30-fold increase in gene expression was observed in the vaccinated groups (p < 0.001), a difference that was not maintained at 11 dpv but was again 24-and 30-fold higher at 13 dpv (p = 0.001). At 15 dpv, there was a 9-fold and 25-fold reduction in the vaccinated groups compared to the ZnO group (p = 0.027), but differences were found when compared to the Neg group or in comparisons of the Neg group with the ZnO group.
A similar pattern was observed for zonulin, with significant 330-and 500-fold increases in the vaccinated groups at 6 dpv (p < 0.001), 16,980-and 21,671-fold increases at 8 dpv (p < 0.001), 623-and 500-fold increases at 11 dpv (p < 0.001), and 850-and 1300-fold increases at 13 dpv (p < 0.001). However, at 15 dpv, the vaccinated groups had a 1.55-and 1.3-fold decrease compared to the Neg group (p = 0.033), but there was no difference between the vaccinated groups and the ZnO group.
Regarding total E. coli quantification, the highest correlation was for the H7 sample, with a similar correlation for F4 and CRISPR, again suggesting that F4 is acquired from vaccination, supported by the absence of amplification in the Neg group samples (Table  4).

Intestinal Integrity
Quantifications for the three TJ proteins are shown in Figure 6. For occludin, a fourfold increase in gene expression was found for the V group when compared to the Neg group (p = 0.016) at 6 dpv. At 8 dpv, a 19-and 30-fold increase in gene expression was observed in the vaccinated groups (p < 0.001), a difference that was not maintained at 11 dpv but was again 24-and 30-fold higher at 13 dpv (p = 0.001). At 15 dpv, there was a 9-fold and 25-fold reduction in the vaccinated groups compared to the ZnO group (p = 0.027), but differences were found when compared to the Neg group or in comparisons of the Neg group with the ZnO group.

Immune Stimulation
The changes in gene expression by groups and sampling dates are shown in Figure 7. Regarding calprotectin, there were only differences in the vaccinated groups versus the Neg group with higher quantification in the two vaccinated groups, and with an increase of 52and 4-fold more than the Neg group (p = 0.004 and p < 0.001). At 15 dpv, the V group had a trend towards significance with 32 times more mRNA than the ZnO group (p = 0.062). At 6 dpv, the V + ZnO group had a 2.33-fold increase in quantification over the V group (p = 0.046).
IFNα had differences between the three groups at 6 dpv, with higher quantification in the Neg group and a 20-fold and 5-fold difference with V and V + ZnO groups (p < 0.001). At 8 dpv, there was a 48-fold increase in the V + ZnO group over the V group (p < 0.001) and at 13 dpv, the V group had26-and 1024-fold differences with the V + ZnO and Neg groups (p = 0.002). At 11 dpv, there were no differences between groups and at 15 dpv, there were differences between all groups with the highest quantification in the Neg and ZnO groups and the lowest in the two vaccinated groups with differences of 300-and 32-fold less in the V group (p < 0.001) and 91-and 9-fold less in the V + ZnO group (p < 0.001).
For IFNγ, significant differences in quantification were found in the two vaccinated groups compared to the Neg group at 6 dpv with 64-fold increases (p < 0.001), at 8 dpv with more than 4000-fold increases (p < 0.001), at 11dpv with a 16-fold and 8-fold increase, respectively (p < 0.001), and at 13 dpv with a 256-fold increase in mRNA quantification (p < 0.001). At 15 dpv, the V + ZnO group showed a 5-fold increase over the Neg group (p = 0.022) and among the vaccinated groups with a 5-fold increase in the V group compared to the V + ZnO group (p = 0.030). No difference was observed when comparing the ZnO group with all others. At 11 dpv, a 3-fold increase was also determined in the V group compared to the V + ZnO group (p = 0.046).
For IL1α, a 9-fold increase was found in the two vaccinated groups compared to the Neg group (p < 0.001) at 8 dpv and a 12-and 8-fold increase at 13 dpv (p < 0.001 and p = 0.002, respectively). At 15 dpv, the highest quantification was observed in the ZnO group with a 13-fold increase over the Neg group (p = 0.01) and 4500-and 14000-fold increases compared to V and V + ZnO groups (p < 0.001), and there was a trend towards a difference between the V and Neg groups with 1.8-fold more mRNA in the vaccinated groups (p = 0.065). With respect to IL1β, there were differences between the vaccinated and Neg groups at 6 dpv with a 21.5-and 7-fold reduction (p < 0.001), at 13 dpv with a 20-and 8-fold reduction (p < 0.001), and at 15 dpv with a 315-and 69-fold reduction with respect to the Neg group (p < 0.001). Differences were also found between the V and V + ZnO groups at 8 dpv with a 5-fold increase for the V + ZnO group compared with the V group. In all samplings, the V group had the lowest quantification and the Neg group had the highest with the V + ZnO group's quantification falling between them.   IL6 showed lower quantification in the vaccinated groups compared to the Neg group at 11 dpv (19-and 8.7-fold decrease; p = 0.001) and at 15 dpv (2280-and 67-fold decrease; p < 0.001), but higher quantification at 8 dpv (13-and 27-fold increase over Neg; p < 0.001). At 15 dpv, both vaccinated groups showed a lower quantification than the ZnO group (3.7and 2.8-fold increase; p < 0.001).
IL8 gene expression only showed a fourfold decrease in the vaccinated groups compared to the Neg group at 11 dpv (p = 0.028). IL10 showed an 8.44-and 2.6-fold reduction in V and V + ZnO groups compared to the Neg group at 6 dpv (p = 0.004 and p = 0.049, respectively) and a 34-and 67-fold reduction at 13 dpv (p = 0.049) with no further significant differences observed.
IL12p35 had a 30-fold lower quantification in the V + ZnO group compared to the Neg group (p = 0.015) at 6 dpv and both vaccinated groups showed an 88-and 42-fold reduction over the Neg group at 11 dpv (p < 0.001). However, both vaccinated groups had 29-and 31-fold higher gene expression than Neg at 8 dpv (p < 0.001) and 1.4-and 7-fold higher gene expression at 13 dpv (p < 0.001). At 15 dpv, the V group showed a 32-fold reduction over the V + ZnO group (p = 0.002) and a 19-fold reduction over the Neg group (p = 0.002). IL12p40 showed no difference at 6 dpv, but there was a 52-and 29-fold increase in the V and V + ZnO groups compared to the Neg group (p < 0.001) at 8 dpv and an 8-and 15-fold increase at 13 dpv (p = 0.005). However, at 11 dpv, the two vaccinated groups showed a 5-and 4-fold decrease compared to the Neg group (p = 0.004). Finally, at 15 dpv, mRNA quantification was 3.1, 2, and 5 times higher in the ZnO group than in the V, V + ZnO, and Neg groups (p = 0.001).
TGFβ expression increased 121-and 153-fold in the V and V + ZnO groups compared to the Neg group (p < 0.001) at 8 dpv and 30-and 33-fold at 13 dpv (p < 0.001), while there were no differences in the other samplings.
Finally, a 4-and 2.6-fold decrease in TNFα expression was observed in the vaccinated groups compared to the control (p < 0.001) at 6 dpv, and a 23-and 24-fold increase at 8 dpv (p < 0.001). The V + ZnO group had a 1.5-and 4-fold increase compared to Neg and ZnO groups at 13 dpv (p < 0.001). At 15 dpv, there was a 78-and 50-fold reduction in the vaccinated groups compared to the Neg group (p < 0.001) and a 5.5-and 3.7-fold reduction compared to the ZnO group (p < 0.001).

Secretory IgA Quantification
The quantifications of secretory IgA in samples from groups V and V + ZnO are shown in Figure 8. Significant differences between groups were found only at 8 dpv (p = 0.047). There was a much more rapid increase in the amount of secretory IgA in the V group than in the V + ZnO group. In fact, in the V group, there is a 159% increase in the amount of secretory IgA, while in the V + ZnO group, in the same period, there is a 7% increase in the amount of antibody. Between 8 dpv and 11 dpv, the V + ZnO group experiences an increase of 108% in the amount of IgA. The increase in secretory IgA in the V + ZnO group was slower, and this produces the difference observed at 8 dpv. in the V and V + ZnO groups compared to the Neg group (p < 0.001) at 8 dpv and an 8and 15-fold increase at 13 dpv (p = 0.005). However, at 11 dpv, the two vaccinated groups showed a 5-and 4-fold decrease compared to the Neg group (p = 0.004). Finally, at 15 dpv, mRNA quantification was 3.1, 2, and 5 times higher in the ZnO group than in the V, V + ZnO, and Neg groups (p = 0.001).
TGFβ expression increased 121-and 153-fold in the V and V + ZnO groups compared to the Neg group (p < 0.001) at 8 dpv and 30-and 33-fold at 13 dpv (p < 0.001), while there were no differences in the other samplings.
Finally, a 4-and 2.6-fold decrease in TNFα expression was observed in the vaccinated groups compared to the control (p < 0.001) at 6 dpv, and a 23-and 24-fold increase at 8 dpv (p < 0.001). The V + ZnO group had a 1.5-and 4-fold increase compared to Neg and ZnO groups at 13 dpv (p < 0.001). At 15 dpv, there was a 78-and 50-fold reduction in the vaccinated groups compared to the Neg group (p < 0.001) and a 5.5-and 3.7-fold reduction compared to the ZnO group (p < 0.001).

Secretory IgA Quantification
The quantifications of secretory IgA in samples from groups V and V + ZnO are shown in Figure 8. Significant differences between groups were found only at 8 dpv (p = 0.047). There was a much more rapid increase in the amount of secretory IgA in the V group than in the V + ZnO group. In fact, in the V group, there is a 159% increase in the amount of secretory IgA, while in the V + ZnO group, in the same period, there is a 7% increase in the amount of antibody. Between 8 dpv and 11 dpv, the V + ZnO group experiences an increase of 108% in the amount of IgA. The increase in secretory IgA in the V + ZnO group was slower, and this produces the difference observed at 8 dpv There was no correlation among performances and E. coli gene quantification. There were also no differences in the different weights recorded when comparing animals testing positive or negative for any of the E. coli genes studied (p > 0.05).

Correlations among E. coli Gene Quantification and Performances
There was no correlation among performances and E. coli gene quantification. There were also no differences in the different weights recorded when comparing animals testing positive or negative for any of the E. coli genes studied (p > 0.05).

Correlations among E. coli Gene Quantification and Immune Stimulation Gene Expression
The correlations between the quantifications for cytokine gene expression and Ct for E. coli genes are shown in Table 6. A positive correlation was found between the quantification of IFNα (p = 0.044) and IL1β (p = 0.043) with the observed CRISPR Ct. The correlations marked in bold were significant at 0.05 level.

Correlations among E. coli Genes and Intestinal Integrity
The correlations between the quantifications for TJ protein gene expression and Ct for E. coli genes are shown in Table 7. Only a partial correlation controlled for group and sampling date was found between the Ct obtained for O8 and the relative mRNA quantification for zonulin (Table 7). The correlations marked in bold were significant at 0.05 level.

Data Reduction including Immune System and Intestinal Integrity Biomarkers
The Wilks' lambda showed significance for all samples (p < 0.001) and two or more functions were produced, except for the sampling at 15 dpv, where only one function was produced and explained 100% of the variance (p = 0.003) (Figure 9).
The presentation of the DFA data showed that the Neg group was always far away from the vaccinated groups while, between the vaccinated groups, there was no significant distance after applying the functions. This reinforces the idea that vaccination in the presence or absence of ZnO produces a clear effect compared to non-vaccinated animals.
The correlations marked in bold were significant at 0.05 level

Data Reduction including Immune System and Intestinal Integrity Biomarkers
The Wilks' lambda showed significance for all samples (p < 0.001) and two or more functions were produced, except for the sampling at 15 dpv, where only one function was produced and explained 100% of the variance (p = 0.003) (Figure 9). The presentation of the DFA data showed that the Neg group was always far away from the vaccinated groups while, between the vaccinated groups, there was no significant distance after applying the functions. This reinforces the idea that vaccination in the presence or absence of ZnO produces a clear effect compared to non-vaccinated animals.
The group assignment is shown in the following table (Table 8). The group assignment is shown in the following table (Table 8). Table 8. Percentage of group membership assertion using DFA-calculated functions. The cases correctly grouped according to the original assignment were 66% at 6 dpv, 86% at 8 dpv, 72% at 11 dpv, 83.7% at 13 dpv, and 86.13% at 15 dpv.

Discussion
The influence of the presence of zinc oxide in feed, both in native high-dose and microencapsulated low-dose forms, has been extensively demonstrated both scientifically and empirically in field trials [32][33][34]. However, in this study, we found significantly higher growth during the first 14 days post-weaning in the groups vaccinated with or without zinc oxide in their feed. However, thereafter, compensatory growth occurred and the overall growth at 25 days was higher in the ZnO group than in the vaccinated group alone. It should be noted that the vaccinated animals did not present any health problems related to E. coli and no sample containing E. coli F4 was found in the control group. Therefore, in this case, it may be that the vaccine gave an initial advantage immediately after weaning, but at around 25 days post-weaning, the energy consumption due to the immune response, in the absence of pathogen-related problems, slightly detracts from growth efficiency. Some cases have been documented where the addition of ZnO to feed did not result in better growth compared to untreated groups [33] or even groups vaccinated orally against E. coli or with a symbiotic in their feed [35]. It has also been observed in different experiments with oral vaccination against E. coli F4 + F18 that there were no differences in ADG when comparing vaccinated animals with unvaccinated and untreated control animals [10]. Thus, Vangroenweghe's study recorded ADG values very similar to those observed in the current research. Compensatory growth comparing the first two weeks of rearing with the following two weeks has already been described with other feed additives such as symbiotics, similar to the effect observed in this study in the ZnO group [35]. In other cases, the lower ADG observed in vaccinated vs. control piglets, under commercial conditions, has been attributed to a reduction in antibiotic consumption in vaccinated vs. unvaccinated control animals [36], but no antibiotics were used in either group in our study, so we can rule out this type of effect.
Despite its almost universal use for the prevention of E. coli-related diarrhoea, it has also been shown that the addition of zinc oxide at 2500 ppm has a moderate effect on F4 strains [37] and there is no difference in the presence of faecal E. coli compared to animals without zinc oxide in their feed [38]. However, it has also been shown that in the presence of zinc oxide in feed, there is a reduction in the excretion of E. coli and a reduction in the frequency of diarrhoea [39].
Studies show that the presence of ZnO in E. coli F4 infections reduces the expression of pro-inflammatory cytokines such as IFNγ, TNFα, and IL6 [32,34] and increases TGFβ gene expression [40]. It has been shown that such activation of pro-inflammatory cytokines has a negative effect on the proteins of the immune system, especially IFNγ, TNFα, and IL1β [1,[41][42][43]. However, we should note that the animals have been vaccinated in this study, which produced variations in the gene expression of cytokines, key elements for the activation of the immune system. The presence of calprotectin mRNA in faeces following oral E. coli vaccination has been demonstrated previously [35,44] and results from the recruitment of antigen-presenting cells such as macrophages and neutrophils necessary to produce the immune response to the vaccine. It has previously been determined that oral vaccination with the vaccine used in this study resulted in increased IFNγ and TGFβ gene expression in the ileum and faeces, detectable at 6, 8, 14, and even 21 dpv in tissues compared to unvaccinated animals [12,35,44]. In the current study, vaccinated groups showed higher gene expression for both biomarkers than Neg or ZnO groups, significant at 8 and 13 dpv. Interestingly, TNFα is not above that of the Neg group until 8 dpv, and again at 13 dpv, while at 15 dpv, both vaccinated group values are below those of the Neg and ZnO groups. In the case of IFNγ, there was more mRNA in all quantifications except at 15 dpv, which may be related to the immune activation produced by vaccination. The fact that, at 15 dpv, both vaccinated group values were lower than those of the control groups may be the result of the increase in TGFβ observed at 13 dpv.
Interestingly, many of the immune system biomarkers showed a bimodal response, with an initial increase comparing 6 and 8 dpv, a subsequent decrease at 11 dpv, a further increase at 13 dpv, and a final decrease. This type of response is also observed in the frequency of CRISPR-positive samples, at least in the V group and in the Ct for ECOTotal. The cyclic increase in E. coli, including the vaccine strain, could be responsible for this bimodal response.
However, zinc does not only exert actions on the immune system or intestinal integrity or innate immunity. It is also capable of preventing bacteria from adhering to the intestinal epithelium, which could be one of the pillars of zinc's action against E. coli. This nonadhesion may be key to preserving the integrity of the TJ [40]. We have seen how TJ proteins were increased in the two vaccinated groups compared to the Neg group at 6, 8, and 11 dpv in the case of occludin, and on days 6, 8, 11, and 13 dpv in the case of zonulin. This would indicate that vaccination is producing a preservation of intestinal integrity even though immune activation is occurring. The effect of oral vaccination with the same vaccine on zonulin and occludin at 6 and 8 dpv, with increased gene expression compared to unvaccinated animals, has been previously described [44].
The efficacy of vaccination with bivalent F4 + F18 vaccines can lead to competition and the displacement of wild-type strains, as vaccine strains bind to recipients, thereby displacing wild-type strains and producing an IgM-and IgA-based immune response [4,9]. Precisely, the presence of zinc oxide could reduce the number of bacteria that bind to the receptors and, thus, could reduce the multiplication of bacteria and the effect on the immune system. When comparing the quantification of the studied E. coli genes, a linear decrease in the Ct for ECOTotal was observed in the two vaccinated groups, in contrast to the Neg group. However, this decrease was much more pronounced in the V + ZnO group, with a difference of more than 8 cycles compared to the Neg group, which meant almost 2.7 exponentials less. However, the V group had, at most, 4.8 cycles of difference at 15 dpv when compared with the Neg group, which was 1.5 exponentials less; the reduction had a much steeper slope than that observed for the V + ZnO group. When comparing the paired samples for CRISPR, a significant difference was observed when comparing the first quantification on day 6 with the second quantification at 8 dpv in the V + ZnO group with a difference of 3.7 cycles (more than one exponential), while the comparison in the V group showed a difference of only 1 cycle, meaning that half the number of bacteria were excreted. All these data indicated that the V + ZnO group excreted less E. coli in all quantifications and that there was a significant linear decrease in the amount of faecal E. coli excreted. In group V, the same effect occurred but was less significant. Our findings of decreased E. coli excretion following vaccination are consistent with those previously reported by Nadeau et al. [9]. However, a challenge was made in that work with a pathogenic strain and not in ours.
Our team previously determined the difference in the amount of secretory IgA in orally vaccinated animals vs. unvaccinated control animals or even parenterally vaccinated 21 dpv animals [12], findings consistent with other research [4,9]. However, while Nadeau et al. [9] found an increase in IgA around day 21, in this study, we observed an increase as early as day 8, although this increase remained at the same level until day 15 and was lower than that documented by us previously in similar animals on the same farm. It should be noted that, in the previous study [12], secretory IgA was studied in tissue samples, in contrast to the current study, which investigated secretory IgA in faeces, as it is difficult to compare the levels observed in tissue samples and faeces.

Conclusions
Although zinc oxide is, empirically, very effective in the control of E. coli-related diarrhoea, the results obtained in this study do not allow us to affirm that a clearly different response is produced after administering a bivalent oral vaccine in the presence or absence of zinc oxide in feed fed to the piglets immediately after vaccination. Differences were observed in the excretion of E. coli, including the vaccine strain, especially at the end of the study period, as well as differences in the presence of secretory IgA in the second sampling and in the gene expression of TNFα, IL1β, IL1α, IFNγ, or IFN α, but only in some of the samples and not consistently.
Although the discriminant capacity of FDA was high, there was a wide percentage of samples that the functions could not correctly assign to a vaccine group. We can conclude that there were no relevant differences in immune activation, IgA production, or intestinal integrity between groups of vaccinated animals depending on whether they took zinc oxide in their feed or not. It must be emphasised that this study was not an efficacy study, there was no experimental infection, and the animals did not suffer any pathological events related to E. coli. In fact, the Neg group did not have any positive samples for F4, one of the adhesion factors present in the vaccine and, therefore, there was no effect derived from the prevention of the action of these strains. Institutional Review Board Statement: Not applicable, as no procedures were carried out that caused pain equal to or greater than the insertion of a hypodermic needle or any kind of suffering to the animals.