Inclusion of slowly digestible starch source is a promising strategy than reducing starch to protein ratio in low protein broiler diets

Abstract

The present study investigated the effects of low protein diets with different starch sources and starch to protein ratio on growth, digestibility, intestinal health, caecal short chain fatty acids (SCFAs), serum cholesterol and triglycerides in broiler chickens. Eight hundred one-day-old male broiler chicks (Ross 308) were randomly allotted to one of 4 dietary treatments with 10 repeats and 20 birds in each repeat. The dietary treatments included 1) a standard protein corn-SBM based diet (SP), 2) a low protein corn-SBM based diet (LPI) without reduced starch: protein ratio, 3) a low protein corn-SBM based diet (LPII) with reduced starch: protein ratio, and 4) a low protein corn-SBM-peas based diet (LPP) and reduced starch: protein ratio. Soy hulls were added in the LPII and LPP diets to reduce starch: protein ratio. During the experiment period from 11 to 24 d, FI was not affected by the dietary treatments (P > 0.05). The BWG was significantly reduced in the LPI diet compared to the SP diet (P < 0.05). Likewise, FCR deteriorated in LPI and LPII but was better in the SP diet followed by the LPP diet (P < 0.05). The apparent total tract digestibility (ATTD) of dry matter (DM) varied significantly among the dietary treatments (P < 0.01). While ATTD of starch was similar for all the diets except the LPP diet wherein the ATTD of starch was significantly lower (P < 0.001). Ether extract digestibility was also significantly different between the SP and LPII dietary treatments (P < 0.01). The AME and AME_n values were significantly lower in the LPP diet compared with other dietary treatments (P < 0.001). Nitrogen retention (%) was increased in all the LP diets compared with the SP diet (P < 0.001), but it was significantly better in both LPII and LPP diets compared to the LPI diet. The data showed that cecal SCFAs production was increased in the LPII and LPP compared to the SP and LPI diets (P < 0.001). Further, the production of acetic, butyric, and propionic acids was substantially higher in the LPP diet (P < 0.001). There was no significant difference in gene expression of Claudin-1 and ZO-1 (P > 0.05). However, MUC-2 and GLUT-1 gene expression were significantly downregulated in the LPI diet (P < 0.05). The concentration of cholesterol and triglycerides was significantly increased in the LPI diet (P < 0.001). In conclusion, the addition of peas as a slowly digestible starch source combined with soy hulls in low protein diet helped to partly recover the growth performance and improved cecal SCFAs production compared to other low protein diets with and without reduced starch: protein ratio in broiler chickens.

INTRODUCTION

The global poultry industry is confronted with a multitude of challenges including high and unpredictable raw material cost and urgency to produce environmentally sustainable feeds. The existing scenario reinforces the need to develop effective nutritional strategies to reduce dietary protein levels in poultry. Ostensibly, low protein diets could reduce the dependency of the poultry industry on soybean meal (SBM) because synthetic amino acids can be considered as an alternative and more sustainable protein source compared to SBM. Furthermore, low protein diets are environmentally beneficial as they reduce nitrogen excretion (Bregendahl et al., 2002; Belloir et al., 2017; Macelline et al., 2020), which is a global concern, especially in Europe. Indeed, some notable progress has been made in this domain, but numerous questions still need to be addressed to reach the objective of the successful development of low protein diets.

A slight reduction in dietary protein is achievable without losing broiler performance by rational inclusion of non-bound amino acids in poultry diets (Belloir et al., 2017; Chrystal, Moss et al., 2020b) Traditionally, dietary protein reduction is accomplished by increasing cereal grains at the expense of SBM and with the addition of (non)-essential synthetic amino acids (Greenhalgh et al., 2020). This strategy invariably widens the starch-to-protein ratio in a typical corn-SBM diet and has certain consequences on the digestive dynamics of both protein and starch in broilers (Liu and Selle, 2017a; Selle and Liu, 2019). Recently, Selle and Liu (2019) observed that the starch: protein ratio increased from 1.55 to 2.57 when dietary protein was considerably decreased from 20 to 15.6%. This sudden transition deteriorated the FCR (1.62 vs. 1.49) in a low protein diet and increased the fat pad weight by around 71%. Greenhalgh et al. (2020) demonstrated better weight gain and feed intake with a numeric improvement in feed conversion ratio (FCR) by capping starch: protein ratio in low protein diets.

Some recent studies showed that highly digestible starch could overburden the proximal gut with glucose in a way that amino acids would compete with glucose for co-absorption with sodium through the Na-dependent transport systems (Liu and Selle, 2017a; Moss et al., 2018). In addition, Moss et al. (2018) also found a negative correlation between starch digestibility and 11 amino acids at the ileum level in low protein diets supporting the proposition that there may be a potential competition between starch and protein digestibility. The transition from high to low protein diets improves starch digestibility at various segments of the small intestine. Thus, the starch digestion rate could be fundamentally important, and slowly digestible starch sources may promote glucose catabolism by the enterocyte and spare amino acids (Moss et al., 2018).

Rapidly digestible starch can increase plasma glucose levels while other nutrients are still being absorbed with potential implications on protein utilization. Therefore, differences in growth performance can be expected in diets with similar contents of digestible nutrients, but with different digestion dynamics (Liu and Selle, 2017b; Weurding et al., 2003a). These differences in the rate and site of starch digestion can affect the small intestine both directly and indirectly, triggering changes in post-prandial metabolism and bacterial fermentation (Bolhuis et al., 2008; Herwig et al., 2019; Zijlstra et al., 2012). Moreover, undigested starch from slowly digestible starch sources can serve as a fermentation substrate to microbiota for the production of volatile fatty acid products in the distal part of the gut (Regmi et al., 2011), thus improving bird health and resilience against zoonotic organisms (Herwig et al., 2019). In the past, field peas (Pisum sativum) corn-based diet showed improved broiler performance compared with a tapioca-corn-based diet, and this performance improvement was attributed to slowly digestible peas starch and improved protein utilization (Weurding et al., 2003a).

Notwithstanding the recent interest in reducing the widening starch: protein ratio in low protein diets, more research is warranted on the potential impact of different types of starch sources and their subsequent impact on growth performance and health index of broilers. Therefore, we hypothesized that slowly digestible starch source and reduced starch: protein ratio may favor growth performance in broilers when fed low protein diets. Hence, the main objective of this study was to explore the effects of starch sources and reduced starch: protein ratio on performance and gut health in low protein broiler diets.

MATERIALS AND METHODS

This study was conducted at the Centre for Expertise and Research in Nutrition (CERN), Adisseo France S.A.S. All experimental procedures and animal care adapted to the conduct of this experiment were under French legislation.

Bird Husbandry and Experimental Diets

Eight hundred, 1-d-old Ross 308 male chicks were obtained from a commercial hatchery. The chicks were weighed and randomly allocated to 1 of 4 dietary treatments. Each dietary treatment had 10 replicates with 20 birds in each replicate pen (1.5 × 1.0 m² each). Each pen contained wood shavings as litter, one hanging feeder, and one bell-shaped water trough. The room temperature was initially set at 32°C for the first week and gradually decreased by 3°C in subsequent weeks. From 1 to 7 d, birds were exposed to 23 h light, and after 7 d, the hours of light were decreased gradually until 4 h of darkness was reached at14 d until the end of the experiment. From 0 to 10 d, a standard common starter diet was provided as crumb, subsequently, birds were shifted to experimental diets which were in pellet form and fed from 11 to 26 d including the last 2 of the digestibility period.

Four isocaloric (ME, 3,000 kcal/kg) experimental diets were formulated with varied starch: protein ratio; SP was a standard corn-SBM diet (21% CP) with starch: protein ratio (1.79), LPI diet was a low CP (18% CP) corn-SBM diet with starch: protein ratio (2.51), LPII was a low CP (18% CP) corn-SBM diet with reduced starch: CP ratio (2.14), LPP diet was also a low CP diet (18% CP) corn-peas-SBM diet with reduced starch: protein ratio (2.14). All diets were manufactured at Euro Nutrition (Saint-Symphorien, France). The composition and the estimated nutrient content of all the experimental diets are presented in Tables 1 and 2. The amino acid levels met or exceeded the nutrient requirement of the breeding company recommendation. Diets and water were provided as ad libitum throughout the experimental period.

Table 1.

Ingredient composition (%) of experimental diets.

Ingredients “as fed” %	SP1	LPI	LPII	LPP
Corn	55.72	68.50	57.45	37.21
Peas	-	-	-	30
Soybean meal 48%	35.23	22.56	22.07	13.71
Soya hulls	-	-	8.18	6.38
Soya oil	4.03	2.02	5.29	5.97
Calcium carbonate	0.83	0.91	0.81	0.86
Monocalcium phosphate	1.92	1.97	2	1.97
Salt	0.23	0.11	0.11	0.14
Adisodium2	0.38	0.53	0.52	0.49
L-Lysine HCl 98%	0.18	0.54	0.54	0.37
L-Threonine 98.5%	0.12	0.28	0.29	0.27
L-Tryptophan 98%	-	0.02	0.03	0.04
L-Valine 96.5%	-	0.21	0.23	0.22
L-Isoleucine 98%	-	0.19	0.20	0.19
L-Leucine 99%	-	-	0.05	0.11
Vitamin-mineral premix3	0.50	0.50	0.50	0.50
Liquid hydroxy methionine4 88%	0.33	0.45	0.47	0.53
L-Arginine 99%	-	0.31	0.33	0.18
Glycine 99%	-	0.37	0.38	0.33
Titanium dioxide	0.50	0.50	0.50	0.50
Total	100	100	100	100

¹SP = standard protein diet; LPI = low protein diet with non-condensed starch: protein ratio; LPII = low protein diet with reduced starch: protein ratio; LPP = low protein peas diet with reduced starch: protein ratio.

²Adisodium = sodium sulfate.

³The vitamin-mineral premix provided per kilogram of diet: vitamin A, 10,000 IU; vitamin D3, 4500 IU; vitamin E, 100 IU; vitamin B12, 20.0 µg; vitamin B6, 4.0 mg; vitamin K3, 3.0 mg; vitamin B1, 2.5 mg; vitamin B2, 8.0 mg; nicotinic acid, 40.0 mg; pantothenic acid, 20.0 mg; folic acid, 1.5 mg; biotin, 0.3 mg; choline chloride, 550 mg; Fe, 60.0 mg; Mn, 80.0 mg; Se, 0.2 mg; Zn, 80.0 mg; Cu, 15.0 mg; I, 1.0 mg.

⁴Liquid hydroxy methionine 88% = Rhodimet AT88, France S.A.S.

Table 2.

Nutrient composition (%) of experimental diets.

Nutrients %	SP	LPI	LPII	LPP
Crude protein	21.01	18.00	18.00	18.00
Crude fat	6.58	4.84	7.85	8.01
Crude fiber	3.37	2.91	5.48	5.58
Starch	37.63	45.14	38.49	38.49
Starch: crude protein	1.79	2.51	2.14	2.14
Digestible Lys	1.15	1.15	1.15	1.15
Digestible TSAA	0.87	0.87	0.87	0.87
Digestible Thr	0.77	0.77	0.77	0.77
Digestible Trp	0.23	0.18	0.18	0.18
Digestible Arg	1.26	1.23	1.23	1.23
Digestible Val	0.87	0.87	0.87	0.87
Digestible Ileu	0.80	0.78	0.78	0.78
Digestible Leu	1.57	1.30	1.27	1.27
Digestible Phe	0.91	0.70	0.67	0.68
Digestible His	0.49	0.38	0.37	0.37
Digestible Gly	0.72	0.91	0.93	0.92
Digestible Gly+Ser	1.57	1.57	1.57	1.57
AMEn (kcal/kg)	3000	3000	3000	3000
Ca	0.87	0.87	0.87	0.87
Na	0.22	0.22	0.22	0.22
K	0.92	0.69	0.74	0.78
Cl	0.22	0.22	0.22	0.22
dEB (mEq/kg)	285	227	240	250
Available P	0.45	0.45	0.45	0.45
Analyzed values (%)
CP	21.1	17.6	17.8	17.8
Starch	39	47	39	41
Crude fat	6.7	5	7.5	8.1
Crude fiber	2.9	2.2	4.7	4.7
Amylose	18.90	16.90	20.01	22.31
Amylopectin	81.10	83.10	79.99	77.69

Data Collection and Digestibility Assay

Body weight and feed intake were measured at the beginning and end of the trial period to calculate body weight gain and FCR. A digestibility assay was conducted from 25 to 26 d of the experimental period using titanium dioxide (TiO₂) as a digestibility marker. We separated 28 kg of complete mash feed before pelleting for each treatment. Before the start of the digestibility study, 100 g of complete mash feed collected at the start of the trial was mixed thoroughly with specified quantity of TiO₂ followed by mixing this premix with 1 kg of mash feed again. Finally, this premix was added in remaining feed, thoroughly mixed and pelleted. Same steps were followed for each dietary treatment. On 25 d, 3 birds closer to the mean body weight of each replicate pen were selected and shifted to digestibility cages. Birds were fed experiment diets mixed with TiO₂ from 25 to 26 d. The first 24 h served as an adaptation period. The following day, birds continued to receive TiO₂ mixed diets from 08:00 to 14:00 h in the afternoon. Then the collection of excreta continued until 50g of fresh excreta sample per bird was obtained. Fresh excreta samples from each replicate pen were pooled, freeze dried, and stored for subsequent analysis. At 26 d, 10 birds per treatment were euthanized by electronarcosis and exsanguination. Blood samples were collected, stored in ice, and subsequently centrifuged at 3,000 x g for 15 min at 4°C to harvest serum. The serum samples were stored at −20°C until further analysis. Afterward, the body cavity was excised to collect ileal tissue samples for gene expression analyses. Briefly, 1 cm of tissue was plunged in a 1.5 mL Eppendorf containing 500 µL of RNALater (Sigma-Aldrich, St. Louis, MO) and store at −20°C for subsequent analysis. Later, 20-30 mg of tissue was extracted using PureLink RNA/DNA Mini kit (Invitrogen, Renfrewshire, UK). The purity of total RNA was measured on Take3 microplate (BioTek Instruments GmbH, Bad Friedrichshall, Germany) and the quality on a non-denaturing 2% agarose gel (E-Gel EX Gels, Invitrogen, Renfrewshire, UK). Reverse transcription was done with 1 µg of total RNA with High Capacity RNA-to-cDNA Kit (Applied Biosystems, Waltham, MA). The cDNA was diluted 10 times and 5 µL was used (20 µL final volume) with 200 nM of primers on a QuantiStudio 5 (Applied Biosystems, Waltham, MA) following optimal conditions for each primer set.

The reaction mixture for the qPCR containing 5 μL of the diluted cDNA was mixed with 10 μL PowerUp SYBR Green (Applied Biosystems, Waltham, MA), forward and reverse primers (final concentration of 0.2 µM for each primer), and sterile water according to the manufacturer's instructions. qPCR was performed following cycling parameters: initial denaturalization at 95°C for 10 min for 1 cycle followed by 40 cycles of amplification at 95°C for 15s, 60°C for 60s, and 72°C for 20s and a final step of 72°C, 10 min. GAPDH was used as a housekeeping gene. Data were analyzed using the efficiency-corrected 2−ΔΔCt method (Pfaffl, 2001). The primers used for quantifying target genes have been given in Table 6.

Table 6.

Primer sequences of reference and target genes used for PCR analysis.

RNA Target	Gene Name	Direction	Oligonucleotide sequence (5′-3′)	Accession no./Reference
GAPDH	Glyceraldehyde-3-phosphate dehydrogenase	F	TGTGACTTCAATGGTGACAGC	NM_204305
		R	GCTATATCCAAACTCATTGTCATACC
MUC-2	Mucin 2	F	ATGCGATGTTAACACAGGACTC	BX_930545
		R	GTGGAGCACAGCAGACTTTG
Claudin-1	Claudin1	F	CATACTCCTGGGTCTGGTTGGT	NM_001013611.2
		R	GACAGCCATCCGCATCTTCT
ZO-1	Zonula Occuldin1	F	CTTCAGGTGTTTCTCTTCCTCCTC	XM_413773
		R	CTGTGGTTTCATGGCTGGATC
GLUT-1	Glucose transporter 1	F	AGATGACAGCTCGCCTGATG	(Slawinska et al., 2019)
		R	GTCTTCAATCACCTTCTGCGG

The cecal contents from the same birds were extracted, put into Eppendorf tubes, kept on ice, until freezing, and stored at -20°C. The cecal contents were kept frozen at −20°C until further analyses.

Chemical Analyses

The diet and excreta samples were analyzed for dry matter (DM), nitrogen, ether extract (EE), titanium dioxide, gross energy (GE), and starch to calculate AME and AME_n and apparent total tract digestibility (ATTD) of starch and ether extract in lyophilized excreta samples. Starch in diet and excreta samples was analyzed using enzymatic colorimetry method (NF V18-121). Dry matter and crude protein were measured using the method AOAC procedure (method 934.01; AOAC, 2006). GE content was determined using a standardized methodology (method 940.15 and 990.03, respectively) (AOAC, 1990). The short chain fatty acids (SCFAs) were analyzed as free acids, using pivalic acid (Sigma Aldrich, St. Louis, MO) as an internal standard. For this, 400 µL of fermentation fluid and 2.4 mL of 1.0 mM pivalic acid solution were mixed, vigorously shaken for 5 min, and then centrifuged at 3,000 × g for 10 min. Then 800 µL of the supernatant and 400 µL of saturated oxalic acid solution were mixed, incubated at 4°C for 60 min, and centrifuged at 18,000 × g for 10min. The supernatant was analyzed by gas chromatography (Agilent Technologies, Santa Clara, CA) using a glass column packed with 80/120 Carbopack B-DA/4% Carbowax stationary phase, helium as a carrier gas, and a flame ionization detector. The acids quantified were acetic, propionic, butyric, valeric, and lactic acid (Kettunen et al., 2015).

The serum samples were analyzed by Labeo Laboratory (1, route de Rosel – Saint-Contest 14053 Caen Cedex 4, France) for triglycerides and cholesterol.

Calculation and Statistical Analysis

The ATTD of starch, AME, and AMEn were calculated by the following equation:

$$\text{ATTD}\text{starch}(\%)=1-\left\{\frac{Ti{O}_{2}indiet}{Ti{O}_{2}inexcreta}\times \frac{Nutrientinexcreta}{Nutrientindiet}\right\}\times 100$$

$$\text{AMEn}\left(\text{kcal}/\text{kg}\text{diet}\right)=GEdiet-\left[GEexcreta\times \left(\frac{Ti{O}_{2}indiet}{Ti{O}_{2}inexcreta}\right)-8.22\times \left\{N\%indiet-\left[N\%inexcreta\times \frac{Ti{O}_{2}indiet}{Ti{O}_{2}inexcreta}\right]\right\}\right]$$

The data were subjected to one-way ANOVA. The body weight at 10 d was used as a covariate for subsequent statistical analysis of growth parameters. The significant means were compared by Tukey's test. All the data were analyzed using JMP 17.0 software. The significance level was set at P < 0.05. Gene expression data were checked for normality using Shapiro-Wilk test and significance was checked using Wilcoxon test.

RESULTS

Growth Performance

The growth performance data is shown in Table 3. During the experiment period from 11 to 24 d, we did not see any effect of dietary treatments on feed intake (P > 0.05). Birds fed the LPI had lower BWG compared to the SP diet (P < 0.05). Dietary treatments differed significantly in terms of FCR (P < 0.05). The FCR was better in the SP diet followed by the LPP diet (P < 0.05). In contrast, birds fed both LPI and LPII diets had poor FCR compared with the SP and LPP diets (P < 0.05).

Table 3.

Effect of different dietary treatments on growth performance of broiler chickens (11–24 d).

Items	SP	LPI	LPII	LPP	SEM	P-value
BW1 d10, g	374	377	376	378	4.46	0.94
Feed intake, g	1356	1387	1383	1386	9.10	0.351
BW d 24, g	1409a	1378b	1381ab	1405ab	8.75	0.01
BWG2 d 24, g	1032a	1001b	1004ab	1028ab	7.70	0.01
FCR	1.31c	1.38a	1.37a	1.35b	0.004	<0.001

SP = standard protein diet; LPI = low protein diet without reduced starch: protein ratio; LPII = low protein diet with reduced starch: protein ratio; LPP = low protein peas diet with reduced starch: protein ratio.

^a,b,cMean values with different superscripts in a row differ significantly.

¹BW: body weight.

²BWG: body weight gain.

Nutrient Digestibility and Apparent Metabolizable Energy

The data on ATTD of nutrients, AME, and AME_n is shown in Table 4. The ATTD of DM was significantly different among the dietary treatments (P < 0.001). DM digestibility was significantly lower in the LPP diet while the LPI diet had highest DM digestibility. The data on ATTD of starch showed statistically significant differences among the dietary treatments. We observed similar ATTD of starch for all the diets except the LPP diet wherein the ATTD of starch was significantly lower compared with all other dietary treatments (P < 0.001). Likewise, the data on AME and AME_n showed significant differences among dietary treatments (P < 0.001). Nevertheless, the highest AME and AME_n values were observed in the LPI diet followed by the SP diet and the LPII diet. In line with the ATTD of starch, the AME and AME_n values were significantly lower in the LPP diet (P < 0.001). There were also significant differences in nitrogen retention percentage (Nret %) among the dietary treatments (P < 0.001). Nret % was increased in the LPII and LPP diets and was significantly higher compared with the LPI and SP diets. For ATTD of ether extract, significant difference was observed between the SP and LPII dietary treatments (P < 0.01). Ether extract digestibility was higher in the SP diet compared to the LPII diet (96.3 vs. 93.1%).

Table 4.

Effect of dietary treatments on AME, AME_n and apparent total tract digestibility of nutrients (26 d).

Items	SP	LPI	LPII	LPP	SEM	P-value
DM1 %	73.3b	78.1a	70.3c	68.3d	0.39	<0.001
AME2, kcal/kg	3090b	3179a	3071b	2881c	18.58	<0.001
AMEn3, kcal/kg	2912b	3019a	2906b	2718c	17.30	<0.001
Nitrogen retention %	64.2c	67.3b	71.0a	70.3a	0.68	<0.001
Ether extract %	96.3a	94.7ab	93.1b	95.0ab	0.005	0.007
Starch, %	97.9a	98.7a	98a	95.5b	0.22	<0.001

Abbreviations: SP, standard protein diet; LPI, low protein diet without reduced starch: protein ratio; LPII, low protein diet with reduced starch: protein ratio; LPP, low protein peas diet with reduced starch: protein ratio.

^a,b,c,dMean values with different superscripts in a row differ significantly.

¹DM: dry matter.

²AME: apparent metabolizable energy.

³AME_n: Nitrogen-corrected AME.

Cecal Short Chain Fatty Acids

The data on cecal SCFAs is shown in Table 5. The dietary treatments significantly affected the cecal SCFAs production in birds. Birds fed the LPP diet had significantly higher amounts of cecal total SCFAs compared with other dietary treatments (P < 0.05). The pairwise comparison showed that cecal SCFAs production was increased in the LPII and LPP compared to the SP and LPI (P < 0.001). Further, the production of acetic, butyric, and propionic acids was substantially higher in the LPP diet (P < 0.001). However, valeric acid was significantly reduced in the LPI diet compared to other dietary treatments (P < 0.05).

Table 5.

Effect of dietary treatments on cecal short chain fatty acids (SCFAs), gene expression, cholesterol and triglyceride in broiler chickens (26 d).

Items	SP	LPI	LPII	LPP	SEM	P-value
SCFAs mmol/kg	85.22c	75.69c	144.52b	183.69a	30.73	<0.001
Acetic acid mmol/kg	52.61c	54.07c	103.17b	128.37a	17.34	<0.001
Propionic acid mmol/kg	11.77ab	6.68b	16.14a	15.57a	5.96	0.004
Butyric acid mmol/kg	12.59b	9.07b	18.13ab	26.67a	8.10	<0.001
Valeric acid mmol/kg	1.68a	0.80b	1.75a	2.19a	0.72	0.001
Claudin-1	0.98a	0.96a	0.60a	0.65a	0.31	0.020
GLUT-1	1.47a	1.02b	1.33ab	1.25ab	0.26	0.007
MUC-2	2.03a	1.07b	1.32ab	1.49ab	0.72	0.04
ZO-1	1.16a	1.02a	0.89a	0.97a	0.20a	0.105
Cholesterol g / l	1.41b	1.79a	1.43b	1.41b	0.12	<0.001
Triglycerides g / l	0.74c	1.14a	0.81bc	0.96ab	0.15	<0.001

Abbreviations: SP, standard protein diet; LPI, low protein diet without reduced starch: protein ratio; LPII, low protein diet with reduced starch: protein ratio; LPP, low protein peas diet with reduced starch: protein ratio.

^a,b,cMean values with different superscripts in a row differ significantly.

Gene Expressions

The data on Claudin-1, GLUT-1, MUC-2, and ZO-1 is shown in Table 5. There was no significant difference in gene expression of Claudin-1 and ZO-1 among all the dietary treatments (P > 0.05). However, MUC-2 gene expression was significantly downregulated in the LPI diet compared with the SP diet (P < 0.05). Likewise, GLUT-1 expression was also significantly downregulated in the LPI diet compared with the SP diet (P < 0.05). However, the MUC-2 and GLUT-1 data in the LPII and LPP were comparable to the SP and the LPI diets.

Serum Cholesterol and Triglycerides

The data on serum cholesterol and triglyceride is shown in Table 5. The serum cholesterol concentration was significantly increased in the LPI diet compared to the other dietary treatments (P < 0.001). Similarly, the concentration of triglycerides in serum was also higher in the LPI diet compared to the SP and LPII diets while the LPP diet had comparable serum triglyceride concentration with other dietary treatments except SP diet (P < 0.05).

DISCUSSION

In the recent past, the development of low protein diets for poultry has become an area of intense research. Feeding low protein diets has produced conflicting reports on growth performance in broiler chickens (Chrystal, Moss et al., 2020a; Maynard et al., 2021; Lambert et al., 2023). In our study, we did not observe any difference in feed intake of broilers despite a -3% pts difference in crude protein level. Contrarily, weight gain was reduced in LPI compared with the SP diet but a numeric increase in weight gain was observed in the LPP diet compared to other LP diets. The FCR was significantly poor in all LP diets notwithstanding the starch source and starch: protein ratio. However, it was still significantly better in the LPP diet with reduced starch: protein ratio compared with the LPI and LPII diets with and without reduced starch: protein ratio, respectively. This is an interesting outcome, and it validates the positive effects of slowly digestible starch sources on the growth performance of broilers (Weurding et al., 2003a) fed low protein diets. Previously, Deschepper and Groote (1995) reported that decreasing the dietary protein contents by approx. 4% pts significantly reduced the weight gain and deteriorated the FCR despite similar levels of essential amino acids in both diets. The experimental diets in our study were balanced in essential amino acids, but still, a numeric difference in feed intake in all low protein diets compared to the SP diet deteriorated the FCR. Previously, Bezerra et al. (2016) posited that the low protein diets may limit the bird's ability to synthesize non-essential amino acids (NEAAs) that could impair the protein synthesis process owing to low nitrogen pool. Therefore, the birds fed low protein diets would use essential amino acids (EAAs) for the synthesis of NEAA creating a deficiency of these EAAs which compromises not only growth performance (Dean et al., 2006) but also the carcass composition in the growing phase (Bregendahl et al., 2002).

Indeed, the gastrointestinal tract is metabolically very active and utilizes around 20% of all dietary energy to support vital functions of digestion and absorption (Cant et al., 1996; Weurding et al., 2003a). Therefore, the metabolic activity of the small intestine is a determinant of the nutrient supply to the body (Cant et al., 1996). Further, intestinal transport of absorbed nutrients coincides with their partial catabolism in the gut (Vaugelade et al., 1994). Glutamine and glucose are preferentially used to provide energy for the small intestine (Fleming et al., 1991; Fleming et al., 1997; Aderibigbe et al., 2020), however, other amino acids can also serve as energy sources for gut tissues. There is a reasonable possibility that starch will be digested in the proximal section of the small intestine and less glucose will be available to the distal part of the small intestine in diets rich in rapidly digestible starch. Such a scenario will promote the oxidation of more amino acids to fulfill the energy demands of the gut (Aderibigbe et al., 2020). This explanation is supported by the result on GLUT 1 expression which was downregulated in birds fed the LPI diet high in starch contents compared with the SP diet. Many intestinal transporters are mainly driven by substrate concentration (Cheeseman and Harley, 1991) and low expression of GLUT 1 potentially indicates that starch was already digested in the proximal section of the small intestine leading to a very little substrate, thereby low GLUT 1 expression in ileum of the birds fed the LPI diet. Alternatively, a slowly digestible starch source will ensure glucose supply to the distal part of the small intestine and would potentially spare the amino acids from being oxidized (Weurding et al., 2003a).

In our study, we observed significant differences in the ATTD of DM and starch. The LPP diet had lower ATTD of DM and starch compared to other dietary treatments. Pea starch is usually high in amylose which is less digestible (Weurding et al., 2001). According to Moran (1982), starch digestibility is closely linked with granule surface area, starch structure, and degree of crystallinity. The raw materials differ in granule size wherein small granules are more rapidly digested than larger granules (Franco et al., 1992). Not surprisingly, low starch digestibility in the LPP diet was accompanied by low AME and AME_n contents but more than 200 kcal difference in AME value between the LPP and mean AME value of other experimental diets was surprising. Nonetheless, there are some possible explanations for this extraordinary finding. Firstly, the DM and starch digestibility values were significantly lower in the LPP diet compared with other dietary treatments. Despite a positive correlation between extent of starch digestibility in the small intestine and AME in broiler chicken (Rogel et al., 1987; Wiseman et al., 2000), site and rate of starch digestion is also extremely important (Weurding et al., 2003b). Hence, both slow and rapidly digestible starch sources can have similar extent, but their site and rate of starch digestion will differ with potential implications on metabolism and growth (Weurding et al., 2003b). Secondly, insoluble fiber coming from soy hulls might have increased the passage rate that combined with the slow digestion rate of starch from peas increased the starch fraction escaping enzymatic digestion as validated by high production of SCFAs in the LPP compared with other experimental diets. Thus, more starch fraction would become available for microbial fermentation providing less energy compared to enzymatic starch digestion (Weurding et al., 2003b). Interestingly, AME and AME_n were both significantly higher in the LPI diet compared with the LPII and SP diets. This suggested a potential surge in energy availability in birds fed low protein corn-SBM diet. Although we observed low AME and AME_n values in the LPP diet compared with other dietary treatments still an improvement in FCR compared with other low protein diets could be construed as an indication of the lower energy requirement of bird fed low protein diets. Despite this significant improvement in the FCR of the LPP diet compared to other low protein diets with and without reduced starch: protein ratio, we noticed around 3% deterioration in the FCR of birds fed the LPP diet compared with the SP diet mainly due to higher feed intake in the LPP diet. Low protein diets formulated without reducing starch: protein ratio are characterized by very high levels of rapidly digestible cereal grains and exceedingly low-fat levels. Such low protein diets induce relatively high feed intake resulting in overconsumption of energy (Belloir et al., 2017). Many studies on low protein diets have reaffirmed that broilers receiving such diets tended to have a higher abdominal fat deposition (Bregendahl et al., 2002; Belloir et al., 2017; Cappalaere et al., 2021; Chrystal, Moss et al., 2020a) indicating changes in energy metabolism (Chrystal, Moss et al., 2020a; Maynard et al., 2021) with a probable surge in energy availability to the animal. This explanation is further supported by our data on serum cholesterol and triglycerides wherein the LPI diet had increased values for these parameters. In line with our results, Attia et al. (2017) and Lambert et al. (2023) also reported that blood cholesterol was significantly increased in low protein diets. Nevertheless, we observed that reduced starch: protein ratio in low protein diets significantly reduced the cholesterol and triglycerides levels comparable to the SP diet. Likely, the prevalent energy system does not consider the altered characteristics of low protein diets with high starch, low protein, and low fat levels, and a potential switch in energy source for the birds. The selection of energy systems should be explored in depth to produce a more precise system for broilers (Noblet et al., 2022) that can help to reduce abdominal fat or energy storage under the form of cholesterol and triglycerides (Lambert et al., 2023).

The SCFAs have numerous beneficial effects on the host including higher mucin production and suppression of gastric apoptosis by acetate (Liu et al., 2017) while butyrate can promote intestinal development and maintain intestinal integrity (Sun and O’Riordan, 2013). In addition, SCFAs have been reported to decrease plasma cholesterol concentrations in rodents and humans (Demigné et al. 1995; Fushimi et al., 2006; Kondo et al., 2009). Generally, the SCFAs are synthesized by the gut microbiota by fermenting the undigested carbohydrates (Besten et al., 2013; Li et al., 2018). In this context, cecum is an active fermentation site where carbohydrates including cellulose, starch, and polysaccharides escaping enzymatic digestion of the small intestine could be fermented. Our results also showed a significant effect on SCFA production wherein SCFAs were high in the LPP diet compared with other dietary treatments. The combined effect of peas that are relatively high in amylose and soy hulls rich in non-starch polysaccharides (75%) out of which 60% are insoluble (Lo, 1989) appears to be the most likely explanation for high production of SCFAs in the LPP diet. As reported previously, insoluble fibers increase intestinal transit rate thus promoting substrate for microbial fermentation and SCFAs production (Scapini et al., 2018). The inclusion of peas facilitated an elevated synthesis of metabolites derived from the saccharolytic pathway (Czerwiński et al. 2010). It has been reported that acetyl-CoA serves as a main precursor unit of cholesterol through an intricate metabolic process in which 3-hydroxy-3-methylglutaryl-CoA reductase is the rate-limiting enzyme (Rodwell et al., 1976). Further, in vitro studies demonstrated that propionate decreased cholesterol synthesis by reducing the enzyme activity of hepatic 3-hydroxy-3-methylglutaryl-CoA synthase and 3-hydroxy-3-methylglutaryl-CoA reductase (Bush and Milligan, 1971; Rodwell et al., 1976). The in vivo experiments employing ³H₂O as a marker demonstrated that the rate of total cholesterol synthesis was reduced in rat livers by dietary propionate supplementation (Hara et al., 1999; Wright et al., 1990). These expositions align well with what we observed in our study. Clearly, the SP diet along with LPII and LPP diets with reduced starch: protein ratio and higher fiber levels due to soyhulls had higher propionate and lower serum cholesterol levels. The influence of acetic acid in cholesterol homeostasis is not completely clear but Fushimi et al. (2006) reported that acetic acid influenced serum cholesterol levels. Rats fed a diet with 1% (w/w) cholesterol had significantly lower serum cholesterol levels when at 0.3% (w/w) supplemented acetate in their diets. In our study, propionate seems to be better linked with cholesterol level compared with acetate concentration.

Tight junctions (TJs) act as a barrier against the passage of microorganisms, toxins, and other unwanted compounds while ensuring the movement of important ions and solutes (Macelline et al., 2020; Barekatain et al., 2023). We did not see any significant changes in gene expression of Claudin-1, and ZO-1. However, the MUC-2 expression was significantly downregulated in the LPI diet compared with the SP diet. In a recent study, Barekatain et al. (2023) reported increased intestinal permeability and poor intestinal barrier function in birds receiving severely reduced protein diets (20.4% vs. 15.6%) but there was no significant difference in Claudin-1, ZO-1, and MUC-2 expression between control and moderately low protein diets (20.4% vs. 17.6%) consistent with what we observed in our study except for MUC-2 which was reduced in our study in the LPI diets. Similarly, Macelline et al. (2020) also showed that broilers fed low protein diet supplemented with essential amino acids were able to maintain TJs in their GITs with no effect on ZO-1 and Claudin-1 expression. TJ proteins are very complex and dynamic and can be influenced by numerous factors, such as residual feed, commensal bacterial, and external factors (Barekatain et al., 2023). Therefore, it is fair to assume that the effect of low protein diets on TJs is still inconclusive and further work at the protein level compared with the mRNA level may elucidate the underlying mechanism.

CONCLUSIONS

Our study demonstrated that the combination of slowly digestible starch sources, such as peas, with soy hulls as insoluble fiber, helped to regain a portion of lost performance compared to the SP diet. Moreover, it significantly improved growth performance compared to other LP diets, both with and without reduced starch: protein ratio. It also enhanced global cecal SCFAs production, making it a promising strategy for the development of low protein diets in broiler chickens. However, due to the limited availability of peas as a feed raw material globally, future research should focus on determining the appropriate ratio of rapidly and slowly digestible starch contents from more commonly used cereals and its optimum combination with insoluble fiber fraction to develop more practical recommendations for universal application of low protein broiler diets.

DISCLOSURES

The authors are employees of Adisseo France S.A.S.