Evaluation of dietary probiotic (Bacillus subtilis KMP-BCP-1 and Bacillus licheniformis KMP-9) supplementation and their effects on broiler chickens in a tropical region

ABSTRACT

The experiment aimed to investigate the effects of probiotic (Bacillus subtilis KMP-BCP-1 and Bacillus licheniformis KMP-9) supplementation on growth performance, ileal digestibility, jejunal histomorphology, colonic ammonia-nitrogen (NH₃-N) and blood parameters in broiler chickens. A total of 288, one-day-old female Ross-308 broilers were randomly allotted into 4 treatments with 6 replicates of 12 birds. Treatments were (T1) basal diet (BD), (T2) BD with 200 ppm amoxicillin, (T3) and (T4) BD with probiotics at the levels of 2.5 × 10⁷ and 5.0 × 10⁷ cfu/kg feed, respectively. Supplementation of amoxicillin or probiotics improved (P < 0.05) average daily gain (ADG) and feed conversion ratio (FCR) on d 42. Compared to those fed with T1, broilers fed with T4 increased (P < 0.05) the ratio of jejunal villus height to crypt depth, and improved (P < 0.05) ileal digestibility of dry matter and protein. They also had the highest numbers of lactic acid bacteria (LAB) and Bacillus spp. as well as the ratio of LAB to Escherichia coli (P < 0.05). Probiotics at the level of 5.0 × 10⁷ cfu/kg feed improved the growth performance of broilers in a tropical region by enhancing the balance of caecal bacteria, jejunal histomorphology and ileal digestibility.

KEYWORDS:

• Bacillus subtilis KMP-BCP-1
• Bacillus licheniformis KMP-9
• broiler
• caecal bacteria
• probiotic

1. Introduction

Poultry industry has played an important role in Thai economy including export and domestic consumption. Antibiotics were used to improve growth performance and disease prevention. However, there was concern about several deleterious effects of antibiotic growth promoters (AGP), for instance, the resistance of pathogen to antibiotics, accumulation of antibiotic residue in animal products and environment, imbalance of normal flora, and reduction in beneficial intestinal flora (Ahmad 2006). These results turned into severe restriction or total ban on AGP in animal industry in many countries. Therefore, it is necessary to look for alternative measures to replace AGP.

Probiotic has been a potential choice due to its beneficial effects on nutrient digestibility (Mountzouris et al. 2010), immune response (Yang et al. 2012) as well as growth performance and modulation of gut microflora (Mountzouris et al. 2007). However, efficiency of probiotic is likely dependent on strain selection, administrative level, application method, diet composition, ages of bird and its ability to survive in host and during storage (Hong et al. 2005; Zhang et al. 2013). Bacillus spp. has been a promising probiotic owing to a resistance of its spores to harsh environmental stress and transition during storage and handling (Cartman et al. 2008).

Heat stress has a negative influence on poultry production and health in a tropical climate. In addition to heat stress, ammonia in animal houses induces adverse effects on chicken health status. However, no information was reported on two local strains of probiotics (Bacillus subtilis KMP-BCP-1 and Bacillus licheniformis KMP-9) in broiler chickens raised in Thailand. The authors hypothesized that mixed Bacillus local strains might provide a variety of positive influences on broiler chickens in a hot climate. Therefore, the present research was conducted to evaluate the probiotic utilization in broiler chickens in terms of growth performance, blood parameters, nutrient digestibility, caecal bacteria, jejunal histomorphology and ammonia concentration.

2. Materials and methods

The experimental protocol was approved by the Chulalongkorn University Animal Care and Use Committee (permission No: 1431080).

2.1. Animals, diets and management

A total of 288 one-day-old female Ross-308 broilers were randomly allotted into 4 groups in a completely randomized design. Each group comprised 6 replicate pens with 12 chicks each. The initial body weight (BW) ranged from 41.67 to 42.09 g. All birds were raised in the open house with a 10 cm height of rice hull litter. Each pen with 1.5 m² space was composed of a self-feeder and a bell drinker. Birds could access feed and water ad libitum. The brooding temperature during the first week was maintained at 35 ± 2°C. Average temperature and relative humidity during the entire experiment from November to December were 28.7 ± 1.1°C and 87.5 ± 4.2%, respectively. The lighting programmes were 23L: 1D during the first week and 12L: 12D from the second to the last week of trial.

All dietary treatments consisted of corn-soy-bean meal (SBM) without an antibiotic and a coccidiostat as T1 (basal diet: BD), 200 ppm of amoxicillin in BD (T2), probiotics at the levels of 2.5 × 10⁷ and 5.0 × 10⁷ cfu/kg BD (T3 and T4, respectively). The probiotic product in this study was manufactured by a local company (K.M.P. Biotech Co., Ltd., Thailand). The probiotic product contained drum-dried spore-forming bacteria (Bacillus subtilis KMP-BCP-1 and Bacillus licheniformis KMP-9), which was declared to be 1.0 × 10¹⁰ viable spores/kg of each Bacillus spp. The probiotic groups consisted of equal numbers of each strain of Bacillus. The certain amount of probiotics was thoroughly mixed with the BD in the mixer machine. Experimental diets were in a mash form and divided into 3 phases, starter (d 1–21), grower (d 22–35) and finisher (d 36–42). All nutrients met the requirement for Ross broiler: Nutrition specifications (Aviagen Inc. 2019) as shown in Table 1.

Table 1. Ingredient and chemical composition of basal diet (g/kg DM unless classified otherwise).

Ingredient	Starter	Grower	Finisher
Corn	436.4	459.0	489.7
Cassava	60.0	80.0	100.0
Full-fat soybean	85.0	80.0	75.0
Soybean meal	320.0	281.3	240.0
Saltstone 50%	5.4	5.6	5.5
Limestone	13.1	12	11.1
Monodicalcium phosphate	14.8	14.4	13.3
Rice bran oil	48.5	51.8	53.6
Antimould agent^1	1.0	1.0	1.0
DL-Methionine	2.6	2.4	2.4
L-lysine HCl	0.3	0	0
Choline chloride 75%	1.9	1.5	1.4
Absorbent toxins^2	2.0	2.0	1.0
Premix^3	8.0	8.0	5.0
Sodium bicarbonate	1.0	1.0	1.0
Chemical composition
ME (kcal/kg)	3,050	3,100	3,150
CP	215.0	200.0	180.0
EE	65.0	70.0	75.0
Fibre	35.0	40.0	40.0
Ash	57.6	50.6	51.5
Ca	9.0	8.5	8.0
P	4.0	3.8	3.5
Methionine %	0.63	0.55	0.54
Lysine %	1.44	1.29	1.16
Threonine %	0.88	0.80	0.71
Arginine %	1.52	1.37	1.22
Sodium %	0.18	0.18	0.16
Potassium %	0.90	0.90	0.85

¹Propionic acid.

²Hydrate Sodium Calcium Aluminosilicate, HSCAS.

³Premix provided the following per kg of diet: Vitamin A 12,500 IU, Vitamin D₃ 16,000,000 IU, Vitamin E 100 mg, Vitamin K₃ 4.0 mg, Vitamin B₁ 3.0 mg, Vitamin B₂ 9.0 mg, Vitamin B₆ 6.0 mg, Vitamin B₁₂ 0.03 mg, Niacin 80.0 mg, Pantothenic acid 18.0 mg, Folic acid 2.5 mg, Biotin 0.40 mg, Vitamin C 200.0 mg, Zinc 96.0 g, Iron 64.0 g, Copper 12.8 g, Iodine 0.80 g, Cobalt 0.16 g, Selenium 0.16 g.

CP, crude protein; EE, ether extract; Ca, calcium; P, phosphorus.

Growth performance parameters i.e. feed intake (FI), BW, BW gain, ADG, FCR and mortality rate were determined on d 21 and 42. Broiler chickens in each pen were weighed at d 1, 21 and 42. Total FI of the specific phase was divided by total BW gain (the weight of both live and dead chickens) during that phase defined as FCR.

2.2. Chemical analysis of diet

The diet samples were placed in a forced-air oven at 60°C for 24 h. After drying, the samples were ground through a 0.42 mm screen in a mill and then analysed for dry matter (DM), crude protein (CP), ether extract (EE), fibre, ash, calcium (Ca) and phosphorus (P) by the standard method of the association of official analytical chemists (AOAC 2012). The gross energy content was determined by total combustion of the samples with an adiabatic bomb calorimeter (model PARR1281, PARR Instrument Corp., U.S.A.).

2.3. Apparent ileal digestibility

Acid insoluble ash (AIA), Celite™ (Lompoc, CA, U.S.A.), was added to the diets at the level of 20 g/kg as an indigestible marker on d 18–21 and d 39–42 for the determination of coefficient of apparent ileal digestibility (CAID). CAID was calculated according to Angkanaporn et al. (1996) as follow:$\mathrm{CAID}=1-[(\mathrm{nutrient}/\mathrm{AIA})\mathrm{ileal}\mathrm{content}/(\mathrm{nutrient}/\mathrm{AIA})\mathrm{diet}]$

Two broiler chickens per pen (12 birds per treatment) were randomly selected to euthanatize using an intracardial injection of pentobarbital sodium on d 21 and 42. The ileum (from the Meckel’s diverticulum to the ileocecal junction) was ligated and removed from the remaining parts of the gastrointestinal tract. The ileal digesta in each pen of the same replicate was manually forced, pooled and stored in sealed bags at −20°C. Analyses of the experimental diets and digesta were performed according to the methods of AOAC (2012). Moreover, ileal samples were analysed for DM, CP and EE according to Angkanaporn et al. (1996). All analyses were performed in 6 replicates.

2.4. Caecal bacterial population

After collection of ileal content, the composite caecal content was collected and kept in the sterile plastic bottle, then immediately placed on icebox until analysis. The caecal samples were transported to a microbial laboratory within a day. Bacterial counts were determined by specific methods: Bacillus spp. using Health Protection Agency, National Standard Method F 15: 2005, Lactic acid bacteria (LAB) using ISO 15214:1998 as well as E. coli and Salmonella spp. using guideline ISO 7251:2005 and ISO 6579-1:2017 respectively.

Bacillus spp was determined on MYP agar and blood agar. The plates were incubated under anaerobic conditions at 30°C for 24–48 h. E. coli count was operated on double – lauryl sulfate tryptose broth and single – lauryl sulfate tryptose broth and incubated for 24 h at 37°C. Examination of LAB number was performed on MRS agar, after 72 h incubation at 30°C under anaerobic conditions. Detection of Salmonella spp. was conducted following selective culture and nutritive enrichment (selenite enrichment, MSRV, incubation at 41. 5°C, 24 h) on XLD and BGA agar plates. Incubation was operated for 24 h at 37°C.

2.5. Jejunal histomorphology

The procedures of jejunal histomorphology were performed according to Awad et al. (2009) and Tsirtsikos et al. (2012). Jejunal segments were removed approximately 2 cm from the middle part of jejunum (between bile duct entry and Meckel’s diverticulum). All jejunal samples were fixed in 10% neutral buffered formalin solution and then embedded in paraffin wax. The histological samples were cut into 6 µm sections, stained by haematoxylin and eosin, and examined using light microscopy (BX50, Olympus, Tokyo, Japan) supplied with a digital camera Micropublisher 5.0 (QImaging, Surrey, Canada). The tissue micrographs were taken by selected program of Image Pro™ Plus version 6 (Media Cybernatics Inc., MD., U.S.A.).

The determinations of villus height (VH), crypt depth (CD) and VH/CD were observed for 8 villi and their adjoining crypts per cut. Each treatment had 6 samples with 4 cross-sections of each sample. The VH was determined from the villi tip to the villus crypt junction, while the CD was measured the depth of the invagination between two villi. Images were evaluated by a technician blinded to the treatment (Awad et al. 2009).

2.6. pH and ammonia-nitrogen measurement

The pH of the jejunal digesta was determined by insertion of a pH probe into the lumen using pH meter (Laboratory pH/ORP metre: SP-2100, Suntex Instruments Co. Ltd., Taiwan) on d 21 and 42 (Veerle et al. 2004).

Colonic digesta samples from 2 broiler chickens in each replicate were collected in the plastic bag and immersed in ice during the transportation to laboratory. Samples on d 21 and 42 were determined for the concentrations of ammonia-nitrogen (NH₃-N) using the phenol-hypochlorite method (Strickland and Parsons 1972).

2.7. Blood parameters

Two broiler chickens per replicate were bled via jugular or wing vein. The whole blood samples were collected into tubes containing EDTA as an anticoagulant, complete blood count (CBC) was then analysed including heterophil to lymphocyte (H/L) ratio as an indicator of non-specific stress (Gross and Siegel 1983). The differential leukocyte counts were determined manually in blood smear in terms of the differential percentage of white blood cells.

2.8. Statistical analysis

Experimental data were analysed by one-way ANOVA according to the general linear model (GLM) procedure of the SAS version 9.4 (SAS Institute Inc. Cary, NC. 2015). The broiler pen was defined as the experimental unit. The logarithmic conversion of the microbial counts was conducted prior to doing statistical analyses. Difference among means of treatment was compared using the Duncun’s new multiple range test. Statistical significance was based on P < 0.05.

3. Results

3.1. Growth performance

There were no significant differences in ADG, FI and FCR among groups during d 1–21. Broiler chickens supplemented with probiotics at both levels or antibiotic had a better ADG and FCR (P < 0.05) than those in the control group (Table 2) during the overall period (d 1–42). However, there were no significant differences between the probiotic groups and amoxicillin group during d 22–42 and d 1–42. No significant difference was found in FI among treatments throughout the experimental periods. Mortality rates in the probiotic groups were lower than amoxicillin and control groups (P = 0.053).

Table 2. Growth performance in broiler chickens for 3 periods (starter, finisher and overall).

Parameters	Treatment				SEM	P-value
	T1	T2	T3	T4
Starter 1–21 d
Initial BW (g/b)	41.75	42.09	41.67	42.03	0.48	0.833
ADG (g/b/d)	39.7	41.5	40.2	40.6	0.30	0.309
Feed intake (g/b/d)	63.3	64.0	63.6	65.9	0.50	0.296
FCR	1.59	1.54	1.58	1.62	0.01	0.101
Finisher 22–42 d
ADG (g/b/d)	66.6^b	74.9^a	74.5^a	74.6^a	1.10	0.008
Feed intake (g/b/d)	150.9	157.2	149.1	148.0	1.70	0.246
FCR	2.29^b	2.00^a	2.10^ab	1.98^a	0.04	0.034
Overall 1–42 d
Final BW (g/b)	2275.0^b	2485.4^a	2450.3^a	2466.7^a	1.20	0.005
ADG (g/b/d)	53.2^b	58.2^a	57.4^a	57.7^a	0.60	0.007
Feed intake (g/b/d)	107.6	110.6	106.3	106.9	1.00	0.446
FCR	2.03^b	1.90^a	1.86^a	1.85^a	0.02	0.029
Mortality (%)	5.00	6.7	1.7	0.00	1.00	0.053

^a,bMeans values with different superscripts in the same row differ significantly (P < 0.05).

ADG, average daily gain; BW, body weight; FCR, feed conversion ratio.

3.2. Apparent ileal digestibility

No significant difference was found in nutrient CAID among groups on d 21. Broiler chickens in the high probiotic inclusion group (T4) showed the improved CAID of DM and CP (P < 0.05) compared with those in the control group (T1) but did not differ from those in the amoxicillin group (T2) on d 42. There were no significant differences in CAID of EE among groups at d 21 and 42 (Table 3).

Table 3. Coefficient apparent ileal digestibility (CAID) of nutrients in broiler chickens on d 21 and 42.

Items	Treatment				SEM	P-value
	T1	T2	T3	T4
d 21
DM	0.65	0.69	0.68	0.68	0.01	0.087
CP	0.78	0.81	0.80	0.80	0.77	0.428
EE	0.75	0.78	0.77	0.77	1.63	0.908
d 42
DM	0.70^b	0.75^a	0.73^a	0.74^a	0.01	0.002
CP	0.75^b	0.81^a	0.79^ab	0.80^a	0.76	0.034
EE	0.76	0.80	0.79	0.80	0.83	0.387

^a,bMeans values with different superscripts in the same row differ significantly (P < 0.05).

DM, dry matter; CP, crude protein; EE, ether extract.

3.3. Caecal bacterial population

Broiler chickens fed with probiotics (T3 and T4) had higher level of LAB and the ratio of LAB to E. coli (P = 0.001) compared with those fed with amoxicillin (T3) but did not differ from those in the control group (T1) on d 21. T3 and T4 had higher (P < 0.001) numbers of Bacillus spp. compared with T1 and T2 at the age of 42 d. Moreover, the numbers of LAB in T4 significantly increased (P = 0.01) compared with those in T1 and T2. The numbers of E. coli in T2 and T4 were lower than those in T1 (P = 0.037) on d 42. The ratio of LAB to E. coli in T4 on d 42 was the highest (P = 0.001) but was not different from that in T3. However, Salmonella spp. was not detected in all groups at both periods (Table 4).

Table 4. Caecal bacterial populations in broiler chickens (log₁₀ cfu/g) on d 21 and 42.

Parameters	Treatment				SEM	P-value
	T1	T2	T3	T4
d 21
Bacillus spp.	4.67	5.27	4.54	4.84	0.11	0.069
LAB	7.24^ab	6.80^b	7.65^a	7.82^a	0.10	0.001
E. coli	4.97	4.52	4.69	4.67	0.07	0.150
Salmonella spp.	ND	ND	ND	ND	-	-
LAB/E. coli ratio	1.29^a	1.04^b	1.37^a	1.40^a	0.04	0.001
d 42
Bacillus spp.	3.58^c	4.18^b	4.58^a	4.63^a	0.09	<0.001
LAB	6.90^bc	6.69^c	7.30^ab	7.57^a	0.10	0.010
E. coli	5.07^b	4.53^a	4.75^ab	4.55^a	0.08	0.037
Salmonella spp.	ND	ND	ND	ND	-	-
LAB/E. coli ratio	1.19^b	1.15^b	1.33^ab	1.49^a	0.03	0.001

^a,b,cMeans values with different superscripts in the same row differ significantly (P < 0.05).

ND, not detected; LAB, lactic acid bacteria.

3.4. Jejunal histomorphology

At the age of 21 and 42 d, broiler chickens supplemented with both levels of probiotics had increased (P < 0.001) villus height compared with those in the other groups. Broiler chickens fed with amoxicillin had the lowest crypt depth (P ≤ 0.001) while there were no statistically significant differences for those in the other groups on d 21 and 42. The ratios of villus height to crypt depth in jejunum increased in groups fed with T4 and T2 (P = 0.019) compared with the control group on d 21 and 42. The group fed with T3 had higher ratio of the villus height to crypt depth on d 21 compared with the control group (P < 0.001) while it was not different from the remaining groups on d 42 (Table 5).

Table 5. Jejunal morphology in broiler chickens on d 21 and 42.

Parameters	Treatment				SEM	P-value
	T1	T2	T3	T4
d 21
Villus height (µm)	546.36^b	511.57^b	719.06^a	697.50^a	1.15	<0.001
Crypt depth (µm)	80.01^a	53.67^b	82.61^a	80.48^a	3.03	<0.001
VH/CD ratio	6.83^b	9.53^a	8.70^a	8.67^a	0.26	<0.001
d 42
Villus height (µm)	616.54^b	554.19^c	675.28^a	686.85^a	3.56	0.001
Crypt depth (µm)	102.72^a	68.59^b	89.49^a	82.59^ab	3.45	0.001
VH/CD ratio	6.00^b	8.08^a	7.55^ab	8.32^a	0.27	0.019

^a,b,cMeans values with different superscripts in the same row differ significantly (P < 0.05).

VH, villus height; CD, crypt depth.

3.5. pH and ammonia-nitrogen

There were no significant differences among groups in jejunal pH and colonic NH₃-N concentrations on d 21 and 42 (Table 6). The pH values were in the ranges of 6.09–6.27 on d 21 and 5.80–6.31 on d 42. Broiler chickens fed with amoxicillin group (T2) tended to have higher NH₃-N (P = 0.05) than those with the high level of probiotic group (T4) on d 21.

Table 6. pH in jejunum and ammonia-nitrogen in colon on d 21 and 42.

Age of broilers	Treatment				SEM	P-value
	T1	T2	T3	T4
pH
d 21	6.16	6.27	6.09	6.17	0.03	0.277
d 42	5.98	6.31	5.80	5.92	0.10	0.304
NH3-N (mg/l)
d 21	29.88	43.17	40.15	28.13	2.51	0.050
d 42	18.97	25.00	15.24	18.21	2.08	0.432

NH₃-N, ammonia-nitrogen.

3.6. Blood parameters

The feed additives had no effect on haematological parameters except for total white blood cells (WBC) and basophil (Table 7). The low level of probiotic group (T3) had lower total WBC (P < 0.05) than the control group (T1). Basophil in the amoxicillin group (T2) was also significantly elevated (P < 0.05) compared with the high level of probiotic group (T4).

Table 7. Blood parameters of broiler chickens in each group at d 42.

Blood parameters	T1	T2	T3	T4	SEM	P value
RBC (×10^6 cells/µl)	2.60	2.51	2.39	2.43	0.04	0.262
Haemoglobin (g%)	9.39	9.75	9.63	10.21	0.12	0.080
Haematocrit (%)	28.00^b	29.08^ab	28.58^ab	30.50^a	0.34	0.050
WBC (cells/ µl)	14,043^a	9,112^ab	8,383^b	9,928^ab	0.53	0.020
Heterophil (%)	57.25	53.17	58.08	55.42	1.09	0.400
Basophil (%)	6.50^ab	8.25^a	6.00^ab	4.58^b	0.46	0.030
Eosinophil (%)	1.00	0.75	0.33	1.17	0.14	0.140
Lymphocyte (%)	31.67	35.25	33.50	36.83	1.07	0.360
Monocyte (%)	3.58	2.58	2.08	2.00	0.24	0.060
Heterophil/Lymphocyte (H/L) ratio	1.81	1.51	1.73	1.50	0.08	0.400

^a,bMeans values with different superscripts in the same row differ significantly (P < 0.05).

RBC, red blood cells; WBC, white blood cells; H, Heterophil; L, Lymphocyte.

4. Discussion

The supplementation of mixed strains of probiotics (Bacillus subtilis KMP-BCP-1 and Bacillus licheniformis KMP-9) at the levels of 2.5 × 10⁷ and 5.0 × 10⁷cfu/kg BD could improve ADG and FCR in the overall period similar to amoxicillin addition. Specifically, the high level of probiotic or antibiotic inclusion had a better influence on FCR than the low level of probiotic inclusion on d 22–42. For this trial, positive growth performances of broiler chickens are in agreement with those of previous studies that fed with probiotics i.e. 1.0 × 10⁹ cfu/kg feed of B. subtilis PB6 (Teo and Tan 2007), 1.5 × 10⁵ cfu/g feed of three B. subtilis strains (Amerah et al. 2013), 0.2 g/kg feed of B. subtilis spore (4.0 × 10⁹ cfu/g DSM 17299) (Zaghari et al. 2015), 1.0 × 10⁶ cfu/g of B. licheniformis (Zhou et al. 2016), 500 g/MT with 2.0 × 10⁹ cfu/g of B. subtilis B21 or Bacillus licheniformis B26 (Musa et al. 2019). Moreover, quails fed with a multi-strain probiotic (Protexin) at the levels of 0.5 and 1.0 kg/t feed improved BW at sexual maturity compared with those in the control group (Ayasan et al. 2006).

The better results of broiler performances in the probiotic groups were in accordance with the tendency of low mortality rate and improvement of the CAID (DM and CP) as well as the VH/CD ratio. Teo and Tan (2007) also found the same tendency of lower mortality rate in the inclusion of 10⁸ and 10⁹ cfu/t feed of B. subtilis PB6 (CloSTAT) product compared with negative control and Zn bacitracin groups. The higher mortality rates were found in negative control (T1) and amoxicillin (T2) groups in the last week of this trial due to abrupt heat wave. An average temperature in that period (d 36–42) was approximately 35°C. The two strains of Bacillus in our experiment were likely to alleviate the adverse effect of heat stress (e.g. laboured breathing, panting, pale combs/wattles, lifting wings away from body, lethargy etc.) in some extent. This might be because the supplementation of B. subtilis in the diet of heat-stressed broiler chickens assisted to revitalize the impaired villus-crypt structure and improve the colonization of helpful intestinal bacteria (Al-Fataftah and Abdelqader 2014).

The probiotic treatments especially at the level of 5.0 × 10⁷ cfu/kg BD promoted the CAID of CP and DM on d 42. It is in accordance with previous studies. For instance, the probiotic inclusion enhanced the CAID of CP and ether extract (Mountzouris et al. 2010). Greater CAID of energy, CP and DM was also observed in probiotic supplementation (Li et al. 2008). Moreover, CAID of most essential amino acids was improved in the group of multi-strain probiotics (Zhang and Kim 2014). It was likely that there were more nutrients remaining available for absorption and growth for chickens receiving probiotic treatment. Moreover, B. subtilis and B. licheniformis could produce extracellular enzymes e.g. α-amylase and proteases, resulting in enhancing the nutrient digestion processes (Carlisle and Falkinham 1989). The amoxicillin addition as a positive control also improved CAID and performances. It could be due to the increased nutrient availability by the suppression of metabolic activities and the growth of gut microorganisms i.e. E. coli with the concurrent change in the VH/CD ratio.

The probiotics at both levels increased the jejunal VH and VH/CD ratio on d 21 and 42. The present results are consistent with several researchers who performed probiotic trials (Samanya and Yamauchi 2002; Kim et al. 2012). Intestinal morphology including VH and VH/CD ratio is an indicator of gut health in poultry. A high ratio is related to longer villi in which epithelium is sufficiently matured with active functions. Fan et al. (1997) reported that VH and VH/CD ratio were directly correlated with the epithelial turnover. The longer villi were related to the activation of cell mitosis (Samanya and Yamauchi 2002). The probiotics, consisted of a combination of several bacteria, significantly affected the villi growth of jejunum and ileum, resulting in creating extended surface areas to absorb nutrients (Ledezma-Torres 2015). Our results revealed that the potential of mixed local strains of Bacillus probiotics was able to improve the jejunal histomorphology and the apparent ileal digestibility in broiler chickens.

The higher level of probiotics (5.0 × 10⁷ cfu/kg diet) in this study significantly increased the numbers of LAB and Bacillus spp. but significantly decreased the number of E. coli in caecum on d 42. However, there were not significantly different in those numbers between two probiotic treatments with different levels. After Bacillus subtilis spores were ingested for 20 h, vegetative cells outnumbered spores all over the chicken GI tract. Spore-based probiotics possibly play a role in the chickens via metabolically active mechanisms (Cartman et al. 2008). The results of bacterial numbers in intestine were variable in previous studies. For example, the number of Lactobacillus was increased while that of E. coli was decreased in broiler chickens fed with the 3-bacterial species probiotics (Zhang and Kim 2014). However, the addition of B. subtilis PB6 (10⁸–10⁹ cfu/t feed) did not affect the populations of Lactobacillus and Bifidobacterium but tended to reduce (P < 0.10) Clostridium and E. coli populations (Teo and Tan 2007). Moreover, there was a significant reduction in the intestinal E. coli population but no increase in the number of Lactobacillus was observed (Molnár et al. 2011). The supplementation of 10⁷–10⁹ cfu multi-microbe probiotic product/kg diet significantly reduced the caecel numbers of Coliforms and Clostridium but did not show any influence on those of total anaerobic bacteria and Bifidobacterium in broiler chickens (Kim et al. 2012). The multi-microbe probiotic products with two different prepared methods (i.e. submerged liquid and solid substrate fermentations) had significantly lower the caecal coliform numbers than the negative control group (Shim et al. 2010). In this trial, the LAB/E. coli ratio on d 42 was increased substantially in the caecum of broiler chickens consuming probiotic diet, especially the higher level of probiotics. It was consistent with the finding of Molnár et al. (2011) in the ileum. The amoxicillin group on d 42 had a significant reduction in most of the caecal bacterial numbers particularly the E. coli population and the LAB/E. coli ratio. It could be possible that both strains of Bacillus had crucial mechanistic role against pathogens in the GI tract compared to AGP fed group. The numbers of two Bacillus strains in this probiotic product could be sufficient to suppress the colonization and persistence of E. coli in the distal part of gastrointestinal tract. In general, the undesirable microbial populations competing with the hosts for nutrients are reduced in the alimentary tract. The present study demonstrated that this probiotic product had the potential to promote growth performance because of the increment of the VH, VH/CD ratio and CAID possibly via promoting secretion of digestive enzymes and development of digestive system (Lee et al. 2010) including improvement of beneficial bacteria and suppression of pathogenic bacteria (Song et al. 2014). The proposed modes of action involved the balance of gut microflora, immunomodulation, contention for adhesion sites and production of antimicrobial agents (Lee et al. 2010; Khan and Naz 2013).

The supplementation of two strains of probiotics (Bacillus subtilis KMP-BCP-1 and Bacillus licheniformis KMP-9) did not affect digesta pH in jejunum though a previous report found that Bacillus could produce a number of volatile compound and organic acids (Lei 2011). They may have a minor role in altering jejunal luminal pH since major site of fermentation in poultry is in the large intestine.

Ammonia is a main odour pollutant and can induce a harmful effect on poultry health and production. The group with high level of two mixed strains of Bacillus tended to reduce NH₃-N concentration compared with the antibiotic group on d 21. In accordance with this result, multiple probiotics decreased the faecal ammonia concentration in broiler chickens (Yoon et al. 2004). The ammonia concentrations in excreta were decreased when broiler chickens were supplemented with probiotics based on Lactobacillus plantarum (5.0 × 10⁷ cfu/g) and Bacillus subtilis (2.0 × 10⁷ cfu/g) at the level of 0.1% in diet (Hassen and Ryu 2012) including with 10⁵ cfu/kg diet of B. subtilis UBT-MO₂ (Zhang et al. 2013). The feasible mechanism came from the increased digestibility (Li et al. 2008) via the release of digestive enzymes (Hassen and Ryu 2012), resulting in less substrate for the microbial fermentation in the large intestine (Yan et al. 2012).

Stress indicators (H/L ratio) were not different in broiler chickens fed with or without probiotic supplemented diets. This current result is in accordance with previous study. H/L ratio was not affected by dietary multi-bacterial probiotic supplementation or stocking density (Cengiz et al. 2015). In contrary, the probiotic supplementation could reduce this ratio in overcrowding broiler chickens (Karthiayini and Philomina 2014). These broiler chickens in the open house for this experiment suffered from negative effects of heat stress with H/L ratio rose above 0.8 (Gross and Siegel 1983). Most values retrieved in this experiment were within the normal range reported by Jain (1993) except WBC and basophil. The number of WBC of the control group was significantly higher than that of the low level of probiotic group (P = 0.02). It was possible due to the stressful condition related to the caecal number of E. coli. Basophil of the high level of probiotic group was lower than that of the amoxicillin group. Low percentage of basophil presumably resulted from parasitic disturbances or early inflammatory and immediate hypersensitivity reactions. A higher percentage of basophil was characteristic of a pro-inflammatory response and possibly be the impact of sensitivity to antibiotics or the presence of bacteria eliciting an immune response (Neveling et al. 2017).

Diet composed of mixed local strain probiotics (Bacillus subtilis KMP-BCP-1 and Bacillus licheniformis KMP-9) at the level of 5.0 x10⁷ cfu/kg feed improved the growth performance of broiler chickens in a tropical region by enhancing balance of caecal bacteria, jejunal histomorphology and ileal digestibility.

Acknowledgements

The authors thank The Panuspokphand Co., Ltd. as well as The K.M.P. Biotech Co., Ltd. for their material supports and Dr. Paisan Tienthai, Department of Veterinary Anatomy, Faculty of Veterinary Science, Chulalongkorn University, for his technical support in laboratory.

Disclosure statement

No potential conflict of interest was reported by the author(s).