In 6-months-old Norwegian Red calves treated by intramuscular injection (IMI) of benzyl-penicillin, Grønvold et al. [51] observed the emergence in E. coli of resistance to benzyl-penicillin and to other classes of antibiotics, whereas no resistance was observed in non-treated calves. Considering that E. coli has low permeability to benzyl-penicillin, the mechanisms underlining the emergence of such resistance remain unclear.

Antibiotic therapy based on trimethoprim-sulfamethoxazole, a folic acid synthesis inhibitor, followed by ceftiofur (3rd GC), delayed diversification in species composition of calves GIM, whereas inter-individual variability, which usually decreases with maturation, remained overall elevated, suggesting that antibiotics delayed maturation of the GIM [52].

History of amoxicillin therapy has been associated to the rise of resistance in E. coli isolates from calves’ feces [53]. To our knowledge, longitudinal studies analyzing the effects of amoxicillin on calves’ GIM were lacking. We thus prospectively collected feces from calves (n = 16) aged from 5 to 26 days, belonging to five breeds (Charolais/Montbeliard, Montbeliard, Prim’ Holstein, Charolais/Prim’ Holstein, Limousin/Montbeliard), and resident in different farms (n = 7) in the Rhône-Alpes region (France), during the period October 2018–March 2020. Eleven out of 16 calves were suffering from omphalitis (umbilical cord infection) and were treated with IMI of amoxicillin (Longamox^®, 15 mg/kg) for a duration varying between 4 and 16 days. The remaining five calves did not receive antibiotic treatment, and their feces were sampled at the same pace of the treated ones. The abundance of 41 ARGs, intI1/2/3, and of 16S rDNA was analyzed by qPCR [54]. Seventeen out of 41 investigated ARGs were found in the feces of all calves before amoxicillin treatment. The bla_TEM gene and bacterial abundance were comparable between the treated and untreated group before treatment (ratio bla_TEM/16S rDNA: 0.013 in both groups). At the end of amoxicillin treatment (T1), the amount of bla_TEM increased in treated calves (bla_TEM/16S rDNA ratio: 0.040) along with other ARGs (tetA, strA and strB), and intI1, index of class 1 integrons. These data suggest co-selection by amoxicillin treatment of ARGs related to other antibiotic classes and potential multidrug development (Figure 1). A decrease of all ARGs was observed 1 week after amoxicillin withdrawal (T2) (bla_TEM/16S rDNA ratio: 0.008). The amount of bla_TEM constantly decreased in the untreated group (bla_TEM/16S rDNA ratio: T1, 0.005; T2: 0.002). The difference observed in the amount of ARGs at pretreatment and post-treatment, or between treated and untreated group, was not statistically significant (Wilcoxon Mann Whitney or Wilcoxon signed-rank test, p > 0.05). Several factors could confound the effect of the amoxicillin treatment on the bla_TEM amount, for instance variation of the bla_TEM gene amount among individuals, probably due to variation of calves’ age, which ranged from 5 to 26 days. Age is a determinant for GIM composition and ARGs amount at early life. In addition, calves were distributed in seven commercial farms probably contributing to the difference in ARGs content as well, because of different farm management. Environmental exposure of all calves to ARGs cannot be excluded, as calves of the untreated and treated group lived in the same farm, thus probably influencing the level of difference of bla_TEM amount between the two groups. For a better understanding of antibiotics action on the GIM and selection of ARGs, experiments in environmentally controlled set-up would be a benefit for avoiding confounding factors influencing GIM composition and ARGs variation further than antibiotic action. However, studies in commercial farms are necessary to model antibiotic therapies effects in a real-life environment and evaluate ARGs propagation to other hosts or in the farm environment.

Prophylactic subcutaneous injection (SCI) of tulathromycin caused a decrease of Bifidobacterium genus (Actinobacteria) in treated calves [55]. On the contrary, comparison of metaphylactic therapy based on enrofloxacin, a fluoroquinolone, or tulathromycin in Holstein calves did not evidence major changes in the GIM neither at the phylum level nor for gene function. Besides, Desulfovibrionales (Proteobacteria), which include species of potential pathogens for humans [56], had a higher relative abundance in the enrofloxacin-treated group 56 days post-withdrawal [57].

Metaphylactic SCI of tildipirosin did not alter the GIM of Holstein calves, at least at the phylum level [58]. Several antibiotic therapies caused a decrease of GIM diversity and E. coli amount during the treatment and until 15 days after withdrawal in Holstein calves observed in three different commercial farms [59].

Keijser et al. [60] analyzed the effects of a low and high dose of oxytetracycline in treated calves. The high-dose was administered for 5 days, whereas the low-dose was administered for 7 weeks. Major changes compared to a group of untreated animals were observed over time. Both oxytetracycline doses correlated with a decrease of Ruminococcus, Coprobacillus, and Lachnospiraceae, all belonging to the Firmicutes phylum, along with an increase of Prevotella (phylum Bacteroidetes), Faecalibacterium, and Blautia (phylum Firmicutes), compared to an untreated group. The selection of tetM gene and other ARGs, such as mel and floR, occurred only in high-dose treated calves and lasted for all the study period (42 days). Oultram et al. [61] analyzed the effects of oxytetracycline (IMI), tulathromycin (SCI), and florfenicol (SCI) used to treat pneumonia and otitis occurring in 7-week-old Holstein calves, hosted in a commercial farm. Considering the five most abundant detected genera, Lactobacillus, Faecalibacterium, Bacteroides, Parabacteroides, and Sharpea, a statistically significant decrease in the oxytetracycline-treated calves was observed for Lactobacillus genus compared to the control group. Overall, antibiotic treatment slightly decreased species richness in the calves’ microbiota 1 week after withdrawal. However, no statistical significance was observed compared with control group. Thames et al. [13] studied by qPCR the effects of neomycin and tetracycline orally administered on the abundance of selected ARGs (tetC/G/O/W/X, ermB/F, sul1/2; intI1), and found that only tetO was significantly more abundant in the treated group.

Lhermie et al. [62] analyzed the effect of fluoroquinolones at low (2 mg/L) and high (10 mg/L) doses administered by IMI in young bulls (7–10-months-old) and calves (2–5-weeks-old). The therapy moderately selected for resistant Enterobacterales compared to the untreated group, and with less detectable effects in young bulls, probably because of a more mature GIM compared to calves that was expected to contain less Enterobacterales and more species difficult to cultivate (the study was conducted by cultivation). However, calves were colonized by fluoroquinolone-resistant bacteria before the treatment. Dobrzanska et al. [6] analyzed the effect of thiamphenicol. At 7 days from thiamphenicol administration, Proteobacteria increased because of E. coli expansion, along with the emergence of mcr-2, a less prevalent gene than mcr-1 responsible for colistin resistance, and oqxB gene, encoding for antibiotic efflux pump. In the treated group, a rise of methanogenic Archaea and Prevotellaceae, typically associated to weight gain, was also observed.