The Effects of Agrimonia pilosa Ledeb, Anemone chinensis Bunge, and Smilax glabra Roxb on Broiler Performance, Nutrient Digestibility, and Gastrointestinal Tract Microorganisms

Simple Summary There is a global effort to reduce the use of antibiotics in broiler production, and the antibacterial activity of extracts from Agrimonia pilosa Ledeb, Anemone chinensis Bunge, and Smilax glabra Roxb has been previously identified in vitro. The present study determines the effects of the dietary inclusion of these extracts on broiler production performance, nutrient digestibility, and on the levels of Campylobacter spp. and Escherichia coli, and lactic acid bacteria in the caecum. The inclusion of S. glabra Roxb and A. chinensis Bunge in broiler feed increased the number of lactic acid bacteria in the caeca of broilers, in comparison to the inclusion of antibiotics in poultry feed. The inclusion of S. glabra Roxb also reduced the levels of E. coli and Campylobacter spp., and improved the feed efficiency and weight gain. The plant extract of A. chinensis Bunge also increased weight gain in broilers. This study highlights the benefits of using S. glabra Roxb and A. chinensis Bunge in broiler feed as alternatives to antibiotics. Abstract Poultry farming is growing globally, particularly in developing countries, to meet the demands of growing populations for poultry meat and eggs. This is likely to lead to an increase in the use of antibiotics in poultry feed, thus contributing to the development and spread of antibiotic resistance which, poses a serious threat to human and animal health worldwide. One way of reducing this threat is to reduce the use of antibiotics in poultry production by finding effective and sustainable antibiotic alternatives that can be used to support poultry health and productivity. Therefore, this study evaluates the incorporation of three medicinal plants, Anemone chinensis Bunge, Smilax glabra Roxb, and Agrimonia pilosa Ledeb, in poultry feed on production performance, nutrient digestibility, and bacteria in the chicken caecum in a 35-day performance trial with 420-day-old male Ross 308 broilers. Groups of randomly selected chicks received one of six dietary treatments. These included five experimental diets of reduced nutrient specifications as a negative control (NC); with amoxicillin as a positive antibiotic control (PC1); with A. pilosa Ledeb (NC1); with A. chinensis Bunge (NC2); and with S. glabra Roxb (NC3). One other positive control diet contained the recommended nutrient specification (PC2). Weight gain and feed intake were measured weekly and used to calculate the feed conversion ratio as performance parameters. Bacteria were enumerated from chicken caecum using a traditional plating method and selective agar. S. glabra Roxb and A. chinensis Bunge showed comparable effects to amoxicillin with significantly increased weight gain in birds offered these diets, compared to those offered the negative control from days 0 to 35 (p < 0.001). S. glabra Roxb exhibited effects similar to the amoxicillin control group with an improved feed conversion ratio (p < 0.001). In addition, S. glabra Roxb decreased numbers of E. coli and Campylobacter spp. on days 21 (p < 0.05) and 35 (p < 0.01) and increased numbers of lactic acid bacteria comparable to the antibiotic group on days 14 (p < 0.001) and 35 (p < 0.01). The findings of this in vivo trial highlight the potential of S. glabra Roxb and A. chinensis Bunge as beneficial feed material to promote poultry health and productivity in the absence of antibiotics.


Introduction
Poultry farming represents one of the most rapidly growing global markets within the agri-food industry. The nature of poultry farming with its relatively short production cycle, efficient feed conversion, and comparatively low costs per unit output have contributed to its worldwide growth [1]. In developing countries, the annual growth for global poultry meat production is estimated to increase from 2005 to 2050 by 2.4% [2]. Globally, poultry meat production is estimated to increase from 137 million tonnes in 2020 to 145 million tonnes in 2029 [3]. It is also predicted that this will contribute to one third of the overall 60% increase in the global consumption of antibiotics in animals raised for food from around 63,000 tons in 2010 to 104 000 tons by 2030 [4]. Many countries continue to use antibiotics in poultry feed for therapeutic use, prophylactic use, and to support growth [5], therefore the number of antibiotics used is likely to increase with the increasing demand for poultry meat. However, the increased use and misuse of antibiotics in poultry feed is one of the main causes of the development and spread of antibiotic resistance [6,7]. This increment in antibiotic usage could potentially cause increased bacterial resistance, which is associated with the emergence and spread of new types of multi-resistant foodborne bacteria, such as Salmonella, Campylobacter spp., and Escherichia coli, [8,9]. At a global level, one of the main strategies to reduce the growing threat of antibiotic resistance is to reduce, and ultimately phase out, the unnecessary use of antibiotics in animal farming [8,10,11]. This has stimulated increased interest in exploring alternatives to antibiotics in poultry feed [12]. Therefore, finding an effective and sustainable supplement that can be used to improve performance and bird health could contribute to reducing the amount of antibiotics used in poultry farming. Some research suggests that plants might be a viable alternative to antibiotics in poultry diets [13,14]. Studies of plant extracts used to supplement poultry diets have shown that they can provide similar benefits to antibiotics, such as inhibiting pathogens, increasing beneficial bacteria, and improving digestibility, growth and performance [15]. Therefore, the aim of the present study is to evaluate the effect of plant supplementation on broiler chicken performance, nutrient digestibility, and bacteria in the gut of broiler chickens in an in vivo poultry trial. The plants used in this study are Agrimonia pilosa Ledeb, Anemone chinensis Bunge, and Smilax glabra Roxb, which were selected based on the results from previous in vitro assays [16]. These three plant extracts exhibited significant in vitro antibacterial activity against Listeria monocytogenes, Salmonella enteritidis, and Escherichia coli over 24 h. It is hypothesised that the inclusion of the selected plants will improve broiler performance and increase beneficial bacteria in the caecum to levels equivalent to an antibiotic supplementation.

The Determination of Antibacterial Activity In Vitro
A reference strain of Campylobacter jejuni was obtained from the National Collection of Type Cultures (NCTC 11322). Clinical isolates of C. jejuni were obtained as frozen stocks from the Agri-Food and Biosciences Institute from retail packs of raw chicken (RC152, RC104, and RC179). A reference strain of E. coli was obtained from the American Type Culture Collection (ATCC 25922). Clinical isolates of E. coli were obtained as frozen stocks from Queen's University Belfast initially from urinary tract infection sources (RC152, RC104, and RC179). Each bacterium was identified using 16S PCR, Sanger sequencing, and BLAST analysis as per manufacturer's instructions for using MyTaq™ Red Mix [17]. C. jejuni and E. coli were selected for antimicrobial susceptibility testing because they are pathogens that cause foodborne diseases and are commonly found in the intestinal tract of poultry [18,19].
Three dry plant samples were purchased from a pharmaceutical company (Tong Ren Tang, Beijing) and included the herb of Agrimonia pilosa Ledeb, the tuber of Smilax glabra Roxb, and the root of Anemone chinensis Bunge. The accepted names of these plants are in accordance with a list available online [20]. These plants were prepared as in the previous methods [16].

Diets, Experimental Design, and Husbandry
Day-old male Ross 308 (n = 420) broiler chicks were randomly allocated to 30 pens (size 1.5 × 2 m). The trial was approved by the Animal Welfare Ethical Review Body at the Agri-Food and Biosciences Institute (AFBI) and conducted under the Animals Scientific Act 1986. The chickens were assigned to one of six dietary treatments (five pen replicates; 14 chicks per pen) and had free access to drinking water via bell drinkers. The temperature was gradually decreased from 32 • C to 22 • C. An 18 h light:6 h dark lighting programme was followed according to a commercial manual [22].

The Measurements and Sample Collection
Body weight was recorded weekly on a per pen, and individual basis and feed intake was recorded weekly on a per pen basis. Weight gain and the feed conversion ratio (FCR) were calculated. Mortality and morbidity rates were recorded throughout [24].
Litter samples were obtained to assess the litter quality at 20, 27, and 34 days from 6 evenly distributed positions within each pen, and mixed to create a composite sample per pen. The percentage moisture content was determined by weight loss [25] after sufficient oven drying (N175CF Genlab MINO/175/F) at 80 • C for 48h. Litter pH was determined as per manufacturer's instructions and using a pH meter (Hanna Instruments, Fischer Scientific, Hampton, NH, USA). Ammonia concentrations from each pen were measured on days 20, 27, and 34. One 15 litre clear plastic container (470 mm × 300 mm × 170 mm) was placed in each pen. One Gastec Passive Dosi-tube No. 3D for ammonia (SKC, Dorset, UK), with a measurement range of 2.5 to 1000 ppm, was held in place in each plastic container with two level Diall M4 × 46 mm hollow wall anchor cup hooks for 4 h to obtain an on-the-spot time-weighted average. This is a modification by Devenish Nutrition Ltd. personnel to a previous method [26]. The ammonia concentrations were calculated as per manufacturer's instructions. One NH 3 monitor, the multi-gas detector eco (International Gas Detectors ltd, Stockport, UK) with a measurement range of 0-100 ppm, was also placed in each plastic container for an on-the-spot reading at 3 separate times at days 27 and 34.
On days 14 and 35, 2 broilers per pen were randomly selected, anaesthetised, then killed by cervical dislocation. The ileum was dissected from each bird [27], then immediately frozen. Ileal contents were dried in an oven (N175CF Genlab MINO/175/F) at 80 • C for 72 h. Ileal samples were powdered through a 1 mm screen using a hammer mill (Christie and Norris, AFBI), according to the manufacturer's instructions. Ileal samples were combined on a per pen basis to produce six pen replicates of ileal contents per treatment.
On days 7, 14, 21, and 35, 5 birds from each treatment group were euthanised as above to determine the average bacterial count from the caecum. A 10 −1 dilution was prepared by adding 1 g of the caecum contents to 9 mL of maximum recovery diluent (MRD) (Oxoid Code CM0733B Thermo Scientific, Waltham, MA, USA). A 10 −1 to 10 −6 dilution series was prepared in MRD. Lactic acid bacteria (LAB), E. coli, and Campylobacter spp. were enumerated by conventional microbiological techniques using a suitable selective agar. For LAB, a 1 mL portion of each dilution was pour-plated, in duplicate, using De Man, Rogosa and Sharpe (MRS ISO) agar (Oxoid Code CM1153B). Once set, the plates were overlaid with another layer of MRS agar to restrict oxygen and ensure an anaerobic incubation. To enumerate E. coli, 100 µL of each suitable dilution was spread-plated onto duplicate TBX agar plates (Oxoid Code CM0945B). For Campylobacter spp., 100 µL of each suitable dilution was spread plated onto mCCDA agar plates (Oxoid Code CM0739 supplemented with SR0155E). In addition, to reduce the limit of detection, 1 mL of the 10 −1 dilution was spread over 3 mCCDA plates [28] and the plates were allowed to dry before incubation. The MRS agar plates were incubated at 30 • C for 72 h. The TBX plates were incubated at 37 • C for 24 h. The mCCDA plates were incubated in a microaerophilic workstation (Don Whitley Scientific, Bingley, UK) at 41.5 • C for 48 h, in an atmosphere of 5% oxygen, 10% carbon dioxide, and 85% Nitrogen.
The agar plates of an appropriate dilution level with 30 to 300 bacteria colonies were selected and colonies counted using a plate counter (Stuart Scientific, Bath, UK).

Laboratory Analysis
The dried samples of plants, diets, and ileal digesta were analysed for various chemical constituents, according to the methods outlined in AOAC [29]. All the results are reported on a DM basis. The gross energy (GE) content was measured using the isoperibol bomb calorimeter (Parr Instruments Co., Moline, IL, USA). Titanium dioxide content was measured Leone's method [30] with some modifications [31]. Amino acid content was

Statistical Analysis and Calculations
One-way analysis of variance (ANOVA) was used to analyse the data using GraphPad Prism (Version 5.0). A value of p < 0.05 was set to determine the statistical significance. Pen was taken as the experimental unit. If a significant difference was found (p < 0.05), a Fisher's Least Significant Difference test was run to test the pairwise differences between the treatment means. The feed conversion ratio was calculated as the ratio of weight of feed intake per pen to the average weekly weight gain per pen [33]. The FCR was adjusted for mortality by calculating the ratio of the weight of feed intake per pen to the average weekly weight gain of the survivors per pen + average weekly weight gain of the mortalities per pen [34]. The moisture content values were calculated using the following equation: percentage moisture content = [(Wet litter weight (g) − dry litter weight (g))/wet litter weight(g)] × 100 [35]. Ileal digestibility was calculated through the use of titanium dioxide as an indigestible marker [36] 3. Results

The Nutritional Composition of Plants
As shown in Table 3, A. chinensis Bunge contained the highest amino acid levels, ash, and oil levels, followed by A. pilosa Ledeb, and S. glabra Roxb with the exception of arginine, for which S. glabra Roxb contained the highest amount. Crude fibre content was highest for A. pilosa Ledeb (349 g/kg), followed by A. chinensis Bunge (280 g/kg), then S. glabra Roxb (258 g/kg).

Performance, Nutrient Digestibility, Litter Quality, and Mortality
From 0 to 35 days, diets containing A. chinensis Bunge (NC2), S. glabra Roxb (NC3), and amoxicillin (PC1) led to significantly increased weight gain (p < 0.001), compared to the weight gain of birds administered the other diets. This was reflected in the higher live weights at day 14 and day 35 for birds receiving diets containing amoxicillin (PC1) and A. chinensis Bunge (NC2). From 0 to 35 days, the feed conversion ratios were most efficient for the birds offered diets containing amoxicillin (PC1), A. pilosa Ledeb (NC1), and S. glabra Roxb (NC3), and least efficient in birds offered diets containing the reduced nutrient specification (NC) (p < 0.001; Table 4. From 0 to 35 days, the feed conversion ratios were more efficient in the birds fed diets containing the recommended nutrient specification (PC2), compared to the birds offered diets containing the reduced nutrient specification (NC) (p < 0.001; Table 5). Means without common superscripts ( abc ) indicate the significant differences between the treatments (rows). PC1 = positive control 1 (NC + 40 mg/kg amoxicillin); PC2 = positive control 2 (recommended nutrient specification); NC = negative control (reduced nutrient specification); NC1 = NC + herb of A. pilosa Ledeb; NC2 = NC + herb of A. chinensis Bunge; and NC3 = NC + tuber of S. glabra Roxb. NS = non-significant (p > 0.05). Overall, the diets containing A. chinensis Bunge (NC2) and S. glabra Roxb (NC3) resulted in similar improved growth rates and 35-day liveweights as birds offered the diets containing amoxicillin PC1), compared to the birds offered the reduced nutrient specification negative control diet (NC). In terms of the FCR, the diet containing S. glabra Roxb (NC3) improved the feed efficiency to a level equivalent to that achieved by birds offered the diet containing amoxicillin (PC1).
The ileal digestibility of dry matter was higher in the recommended nutrient specification (PC2), compared to all other treatments (p < 0.01). The ileal digestibility of ash ranged from 0.24 for birds receiving the antibiotic (PC1) to the highest levels of 0.53 and 0.57 for birds offered the recommended nutrient specification (PC2) and the A. chinensis Bunge (NC2) supplemented treatment group (Table 6). Means without common superscripts ( abc ) indicate the significant differences between the treatments (rows). PC1 = positive control 1 (NC + 40 mg/kg amoxicillin); PC2 = positive control 2 (recommended nutrient specification); NC = negative control (reduced nutrient specification); NC1 = NC + herb of A. pilosa Ledeb; NC2 = NC + herb of A. chinensis Bunge; and NC3 = NC + tuber of S. glabra Roxb. NS = non-significant (p > 0.05).
There was a total of 3.78% mortality. There were no significant differences in mortality between treatments and none of the mortality was due to treatments. Ten chickens died and seven chickens were euthanised as they were unable to stand or had breathing difficulties and associated with common causes of mortality.

Lactic Acid Bacteria, Campylobacter spp., and E. Coli
The inclusion of S. glabra Roxb in broiler diets, from days 14 to 21, reduced Campylobacter spp., caecal colonisation, at day 14, from 3.3 Log 10 CFU g −1 to 2.2 Log 10 CFU g −1 Animals 2022, 12, 1110 9 of 15 (p < 0.01; Figure 1). Similarly, S. glabra Roxb in the broiler diets decreased the detection of E. coli from 8.2 Log 10 CFU g −1 to 7.4 Log 10 CFU g −1 (p < 0.05; Figure 2).  Although the other treatment groups did not show significant differences between the numbers of pathogens, the trend in the results over a period of 0 to 35 days suggests that the diet containing the antibiotic (PC1) resulted in higher levels of Campylobacter spp. and E. coli than the other diets containing A. pilosa Ledeb (NC1), A. chinensis Bunge (NC2), and S. glabra Roxb (NC3) (Figures 1 and 2).
Our results also show that the recommended nutrient specification diet (PC2), the reduced nutrient specification diet (NC), A. pilosa Ledeb diet (NC1), A. chinensis Bunge diet (NC2), and the S. glabra Roxb diet (NC3) resulted in significantly higher levels of lactic acid bacteria compared to amoxicillin (PC1) at day 14 (p < 0.001; Table 7). The negative control diet (NC) and the S. glabra Roxb (NC3) diet also resulted in significantly higher levels of lactic acid bacteria, in comparison to the levels detected when amoxicillin diet (PC1) was included in the diet at day 35 (p < 0.01; Table 7). These results suggest that, over the period of 0 to 35 days, the diets that did not contain the antibiotic resulted in significantly higher levels of lactic acid bacteria; in particular, the A. chinensis Bunge diet (NC2), and the S. glabra Roxb diet (NC3) resulted in the highest levels of lactic acid bacteria (Table 7). from 7 to 35 days of age. Means without common superscripts ( ab ) indicate the significant diffe (*, p < 0.05)between the treatments. PC1 = positive control 1 (NC+40mg/kg amoxicillin); PC2 = tive control 2 (recommended nutrient specification); NC = negative control (reduced nutrient ification); NC1 = NC + herb of A. pilosa Ledeb; NC2 = NC + herb of A. chinensis Bunge; and NC + tuber of S. glabra Roxb.

Discussion
From the in vitro studies [16], the most effective antibacterial activity against C. and E. coli was exhibited by A. pilosa Ledeb. Following this, A. chinensis Bunge and S. Roxb demonstrated strong antibacterial activity against C. jejuni and E. coli. A revi  Means without common superscripts ( abc ) indicate the significant differences between the treatments (rows). PC1 = positive control 1 (NC + 40 mg/kg amoxicillin); PC2 = positive control 2 (recommended nutrient specification); NC = negative control (reduced nutrient specification); NC1 = NC + herb of A. pilosa Ledeb; NC2 = NC + herb of A. chinensis Bunge; and NC3 = NC + tuber of S. glabra Roxb.

Discussion
From the in vitro studies [16], the most effective antibacterial activity against C. jejuni and E. coli was exhibited by A. pilosa Ledeb. Following this, A. chinensis Bunge and S. glabra Roxb demonstrated strong antibacterial activity against C. jejuni and E. coli. A review of the literature suggested that crude extracts of A. pilosa Ledeb, S. glabra Roxb, and A. chinensis Bunge did not appear to have been previously screened for antibacterial activity against C. jejuni. These new findings highlighted the antibacterial activity of these medicinal plants against C. jejuni and justified their use in an in vivo poultry trial. The analysis of the nutritional chemical composition indicated that A. chinensis Bunge has the highest amino acid content, followed by A. pilosa Ledeb and S. glabra Roxb. Accurately formulating to amino acid requirements is known to improve performance [37]. There appears to be no previous research that demonstrates the nutritional chemical composition of these plants, and knowledge of the composition enables a more accurate formulation of the diets, as the inclusion rates of other ingredients can be modified accordingly to balance the diets for total amino acids and other essential nutrients. S. glabra Roxb contained the lowest crude protein content (4.6%) and amino acid content, compared to the other plants analysed in this study. The findings indicate substantial differences in plant nutritional value. They also highlight the need for analysis prior to dietary inclusion. It is recognised that, at the level of inclusion in this study, the nutritional contribution of these plants is minimal, but it is still of importance to accurately profile the nutritional profile of all ingredients to accurately formulate diets to the requirements.
The findings presented herein highlight that birds fed with diets containing A. chinensis Bunge and S. glabra Roxb have similar growth rates and live weights at day-35 to broilers fed diets that include amoxicillin. Other studies have also shown that the addition of antibiotics, such as salinomycin and zinc bacitracin [33,38], flavophospholipol antibiotic [39], and penicillin [40], to poultry diets has led to increased weight gain. However, this study has demonstrated that A. chinensis Bunge and S. glabra Roxb could be used in place of amoxicillin to result in similar growth weights, and therefore be an effective antibiotic alternative.
The addition of A. pilosa Ledeb, A. chinensis Bunge, and S. glabra Roxb to poultry diets resulted in improved bird growth rates. In terms of the FCR, the poultry diet that contained S. glabra Roxb resulted in comparable effects to amoxicillin and improved the feed efficiency of the birds, when compared to the birds in the negative control group. Previous studies have shown that birds fed diets supplemented with plants or plant extracts, or with antibiotics, can achieve an improved performance by the modification of the gastrointestinal microbiota [41]. The bird caecum has a functional role in nutrient absorption, which affects the overall performance and health [42]. Any increase in pathogens, such as Campylobacter spp. or E. coli, could have detrimental impacts on nutrient absorption, feed conversion ratios, and performance, whereas increases in lactic acid bacteria may have beneficial effects by improving growth performance, feed conversion efficiency, and immune responses, and help to combat enteric pathogens [43]. Our study shows that the addition of amoxicillin in broiler diets decreased the lactic acid bacteria in birds, highlighting that antibiotics can have a negative effect on beneficial bacteria, as well as harmful bacteria. Birds receiving diets containing amoxicillin showed fluctuations in the levels of lactic acid bacteria, Campylobacter spp., and E. coli. Birds on diets that contained amoxicillin initially showed reduced levels of lactic acid bacteria after starting the diet, and then they gradually increased until day 35. The gradual increase in lactic acid bacteria in birds on the diet that contained amoxicillin was lower than the levels of lactic acid bacteria found in birds on the other poultry diets. The findings in this study suggest that amoxicillin did not increase performance through selective effects towards beneficial bacteria, but rather through decreasing pathogens.
The changes in the levels of lactic acid bacteria in this study are similar to those reported in other poultry trials, which included antibiotics, monensin [44], zinc bacitracin [35], chlorotetracycline, virginiamycin, and amoxicillin [45]. These studies reported a decrease in lactic acid bacteria in the intestinal tract of poultry that were administered feed with antibiotics, as in this study. This study shows that diets containing S. glabra Roxb are more effective than diets containing amoxicillin in stimulating the development of lactic acid bacteria populations in the caecum by day 35. Other studies have also shown that one of the beneficial effects of including plants in the poultry diet is the occurrence of increased levels of lactic acid bacteria in the caecum [33,46]. This modulation of the caecum may have resulted in differences in the litter quality, but the results suggest that the litter moisture, pH, and ammonia output were not affected by the treatment.