Impact of low-crude protein and insoluble fiber diets on post-weaning diarrhea, growth performance, intestinal morphology, and gene expression for nursery pigs with natural rotavirus and subject to enterotoxigenic Escherichia coli F18+ experimental infection

TL;DR

This study found that low-protein or fiber-rich diets in young pigs did not reduce diarrhea or improve recovery from gut infections compared to traditional diets with zinc and antibiotics.

Contribution

The study evaluates the effectiveness of low-protein and fiber diets in mitigating post-weaning gut issues in pigs infected with rotavirus and ETEC.

Findings

Low-protein and fiber diets did not reduce scour scores or improve recovery from infections compared to traditional diets.

The PC diet (with zinc oxide and carbadox) significantly reduced scour scores during infections.

LCP and FIB diets lowered TNFα levels but did not translate into biological benefits.

Abstract

This study evaluated the impact of low‑crude protein (LCP) and insoluble fiber (FIB) diets on post‑weaning diarrhea, intestinal damage, and growth in nursery pigs naturally infected with rotavirus and challenged with enterotoxigenic Escherichia coli F18 (ETEC). A total of 240 pigs were randomly assigned to 40 pens (6 pigs/pen) for a 42-d experiment. Pens were randomly assigned to one of five diets: NC (standard nursery diet without zinc oxide, ZnO, or carbadox); PC (NC + 3,750 mg/kg ZnO and 50 mg/kg carbadox); LCP (NC with reduced crude protein and SID Lys); FIB (NC + 8% wheat middlings); and ZINC (NC + 3,750 mg/kg ZnO). Experimental diets were fed in two phases: phase 1 (d 0–7) and phase 2 (d 7–21) and phase 3 (d 21–42) was a common diet for all pigs. Pigs were positive for rotavirus (strains A, B, and C) within the first week post-wean. To model late nursery outbreaks, pigs were orally inoculated with a field strain of ETEC on day 14-, or 0-d post-inoculation (DPI 0). Scour scores were recorded daily for the first 28 d (−14 to 14 DPI). Body weight (BW), average daily gain (ADG), feed disappearance (ADFI), and feed efficiency (G: F) were evaluated for four periods: DPI −14 to −7; −7 to 0; 0 to 7; 7 to 28. Sample pigs were confirmed to be susceptible to ETEC (FUT1GG/GA). Serum TNFα was measured at DPI 0, 3 and 6. Ileal and colonic tissues and colonic digesta were collected from two pigs per pen, euthanized on DPI 3 and 6. Fecal swabs (one/pen) were collected on DPI 0, 2, 4, 6, and 14 to assess shedding of ETEC F18 and STb genomic material. Data were analyzed using mixed model methods (SAS 9.4, Cary, NC) with pen as the experimental unit with the fixed effects of diet, period, and their interaction. Overall growth metrics were unaffected by dietary treatment. Rotavirus and ETEC inoculation increased scour scores, and F18/STb genomic material was detected in feces after inoculation. The PC diet reduced scour scores during both rotavirus and ETEC infections (P < 0.01). Ileal morphology and gene expression improved from DPI 3 to 6 (P < 0.05), independent of diet (P > 0.1). Colonic morphology, inflammation, and putrescine and cadaverine levels were unaffected by diet (P > 0.1). Although LCP and FIB diets lowered circulating TNFα at DPI 0 and 6, these reductions did not translate into measurable biological benefits. Compared with the PC diet, the LCP and FIB diets were ineffective in improving outcomes of the enteric challenge in this study.

Young pigs fed low-protein or fiber-rich diets after weaning didn’t show less diarrhea or faster recovery from gut infections compared to pigs fed a traditional diet with zinc and antibiotics. These findings suggest that such alternative diets may not provide enough protection or support during common post-weaning health challenges.

Figures

Daily fecal consistency score during first 28 d of experiment, corresponding to 14 d pre- and 14 d post- inoculation with ETEC. Pigs were natural infected with rotavirus strains A, B, and C at time of weaning and orally inoculated with ETEC F18 14 d post weaning, indicated by vertical red line. Effects of treatment, day, and treatment × day all significant (P < 0.01). fecal consistency scores were visually scored using a 1–4 scale where 1 = solid, 2 = semi-solid, 3 = semi-liquid, and 4 = liquid.

Tables

Table 1. Dietary and nutritional composition (as-fed basis) of the negative control (NC) diet, used to formulate the positive control (PC) and zinc oxide (ZINC) diets, alongside low-crude protein (LCP) and fiber (FIB) diets for a 42-d nursery study.

	Phase 1 d0–7 (DPI −14 to −7)			Phase 2 d7–21 (DPI −7 to 7)			Phase 3 d21–42 (DPI 7–28)
Ingredient, %	NCa	LCP	FIB	NC 1	LCP	FIB	Common
Corn	31.47	41.55	24.31	45.46	53.03	38.09	59.15
Soybean meal	20.00	11.19	20.00	28.00	21.84	28.00	34.00
Oat Groats	15.00	15.00	15.00	5.00	5.00	5.00	—
DairyLac 80b	12.91	12.91	12.91	5.41	5.41	5.41	—
Wheat Middlings	—	—	8.00	—	—	8.00	—
Fishmeal	5.00	5.00	5.00	2.50	2.50	2.50	—
Animal Plasma	4.88	5.45	4.86	2.27	2.27	2.27	—
Soy oil	3.00	2.37	3.68	3.00	3.00	3.88	3.00
Dried Whey	2.50	2.50	2.50	2.50	2.50	2.50
Monocal Phosphate	1.22	1.23	1.22	0.91	0.94	0.90	0.83
Calcium Carbonate	0.81	0.83	0.80	1.47	1.46	1.46	0.99
L-Lysine (Lys)	0.39	0.41	0.35	0.43	0.41	0.40	0.45
Salt	0.30	0.30	0.30	0.50	0.50	0.48	0.50
VTM premixh	0.40	0.40	0.40	0.40	0.40	0.40	0.40
L-Methionine (Met)	0.25	0.21	0.23	0.26	0.22	0.25	0.25
L-Threonine (Thr)	0.23	0.22	0.21	0.25	0.22	0.23	0.25
Copper sulfate	0.07	0.07	0.07	0.07	0.07	0.07	—
L-Valine (Val)	0.08	0.08	0.06	0.12	0.10	0.09	0.13
L-Tryptophan (Trp)	0.06	0.07	0.05	0.04	0.04	0.03	0.04
L-Isoleucine (Ile)	0.06	0.10	0.04	0.04	0.05	0.02	—
L-Arginine (Arg)	—	0.11	0.02	—	0.05	0.03	—
Corn Starch	1.38	—	—	1.38	—	—	—
Phytasec	—	—	—	—	—	—	0.01
Nutrient composition, calculated
MEd, Mcal/kg	3.39	3.37	3.37	3.35	3.35	3.34	3.34
Crude Protein (CP), %	22.31	19.50	22.85	21.85	19.50	22.39	21.10
Calcium, %	0.93	0.91	0.93	1.00	0.98	1.00	0.78
AvPe, %	0.60	0.60	0.60	0.40	0.40	0.40	0.37
Lactose, %	12.00	12.00	12.00	6.00	6.00	6.00	—
SID Lys, %	1.46	1.30	1.46	1.42	1.26	1.42	1.33
SID Met + Cys: Lys, %	0.58	0.58	0.58	0.58	0.58	0.58	0.58
SID Thr: Lys, %	0.65	0.65	0.65	0.65	0.65	0.65	0.65
SID Trp: Lys, %	0.20	0.20	0.20	0.19	0.19	0.19	0.19
SID Val: Lys, %	0.67	0.67	0.67	0.67	0.67	0.67	0.68
SID Ile: Lys, %	0.55	0.55	0.55	0.55	0.55	0.55	0.55
SID Leu: Lys, %	1.00	1.00	1.01	1.02	1.04	1.03	1.08
SID Arg: Lys	0.70	0.70	0.70	0.82	0.82	0.82	0.96
SID Val: Lys, %	0.67	0.67	0.67	0.67	0.67	0.67	0.68
NDFf, %	4.65	4.94	6.85	5.83	6.07	8.02	7.11
ADFf, %	2.04	1.81	2.33	2.73	2.58	3.02	—
TDFg, %	8.63	7.78	10.47	9.16	8.6	11.0	10.25
Insoluble NSPg,%	6.06	5.59	7.42	6.52	6.22	7.87	7.36
Soluble NSPg, %	1.61	1.33	1.82	1.72	1.53	1.93	1.71

Table 2. Analyzed dietary composition of complete diets on an as-fed basis.

	Phase 1 D 0–7 (DPI −14 to −7)					Phase 2 D 7–21 (DPI −7 to 7)					Phase 3 D 21–42 (DPI 7 to 28)
	PC	NC	LCP	FIB	ZINC	PC	NC	LCP	FIB	ZINC	Common
GEa, Mcal/kg	4.01	4.03	3.97	4.08	4.01	3.94	3.99	3.96	4.06	3.99	4.04
CPb, %	21.6	22.4	18.6	22.0	21.2	20.9	20.7	19.1	21.5	20.5	20.0
TDFc	10.3	10.2	8.6	13.0	11.8	14.3	13.3	13.4	14.9	13.2	16.0
Insoluble NSPd	9.2	8.7	7.8	11.1	10.1	12.8	11.7	12.2	12.9	12.0	13.2
Soluble NSPd	1.1	1.5	0.8	2.0	1.7	1.5	1.6	1.2	2.0	1.2	2.7

Table 3. Impact of dietary treatment (trt) on the growth performance of weaned pigs during a 42-d nursery experiment, where pigs were subject to ETEC F18 challenge at 14 d post-wean and had a natural rotavirus infection at the time of weaning.

	Dietary Trt						P- Valuea
	PC	NC	LCP	FIB	ZINC	SEM	Trt	Period	Trt × Period
BW, kg							0.65	<0.01	0.41
DPI −14	5.47	5.47	5.48	5.47	5.48	0.303
DPI −7	6.46	6.21	6.32	6.17	6.35	0.303
DPI 0	8.92	8.51	8.54	8.58	8.52	0.303
DPI 7	12.07	11.53	11.54	11.45	11.61	0.303
DPI 28	25.98	26.35	24.94	25.00	25.14	0.340
ADG, kg							0.45	<0.01	0.78
DPI −14 to −7	0.14	0.11	0.12	0.10	0.13	0.018
DPI −7 to 0	0.34	0.33	0.32	0.34	0.31	0.018
DPI 0 to 7	0.41	0.41	0.40	0.42	0.39	0.018
DPI 7 to 28	0.66	0.64	0.63	0.64	0.61	0.018
Overallb	0.45	0.42	0.42	0.42	0.41	0.013	0.31	–	–
ADFI, kg							0.67	<0.01	0.18
DPI −14 to −7	0.13	0.12	0.13	0.11	0.12	0.021
DPI −7 to 0	0.38	0.35	0.39	0.36	0.39	0.021
DPI 0 to 7	0.55	0.51	0.48	0.55	0.49	0.021
DPI 7 to 28	0.90	0.88	0.88	0.86	0.86	0.021
Overallb	0.57	0.54	0.55	0.53	0.54	0.014	0.28	—	—
G: F							0.21	<0.01	0.01
DPI −14 to −7	1.12a	0.94ab	0.96ab	0.91b	1.00ab	0.038
DPI −7 to 0	0.92	0.94	0.81	0.96	0.79	0.038
DPI 0 to 7	0.77	0.81	0.82	0.81	0.81	0.038
DPI 7 to 28	0.74	0.73	0.72	0.75	0.71	0.038
Overallb	0.78	0.78	0.76	0.80	0.75	0.013	0.15	—	—

Table 4. Impact of dietary treatment (trt) and days post inoculation (DPI) on qPCR cycle threshold value (CT) of genes for F18 fimbriae and STb toxin of Escherichia coli in fecal DNA.

	Day (DPI)		Dietary Trt						P Value
	Daya	SEM	PC	NC	LCP	FIB	ZINC	SEM	Trt	Day	Trt × Day
F18 CT									0.55	<.01	0.02
DPI 0	39.6x	0.58	39.9	39.5	39.1	40.0	39.4	1.43
DPI 2	29.1yz	0.57	28.6	29.6	30.2	28.9	28.1	1.33
DPI 4	28.9z	0.57	26.7	31.8	28.2	27.9	30.0	1.33
DPI 6	30.9y	0.56	34.2a	31.1ab	27.1b	30.3ab	31.7ab	1.33
DPI 14	39.1x	0.63	38.3	40.0	38.9	39.0	39.3	1.33
STb CT									0.70	<.01	0.17
DPI 0	36.4x	0.65	36.8	36.0	37.3	34.9	37.1	1.36
DPI 2	30.2z	0.62	29.7	31.2	31.3	30.5	28.4	1.36
DPI 4	30.0z	0.62	27.2	32.3	29.8	29.8	31.1	1.36
DPI 6	32.6y	0.62	35.0	33.4	29.6	32.4	32.5	1.36
DPI 14	38.2x	0.69	37.8	39.1	38.3	37.0	38.7	1.56

Table 5. Impact of dietary treatment (trt) and day Post-Inoculation (DPI) on serum tumor necrosis factor-alpha (TNFα).

	Day (DPI)		Treatment							P Value
	DPI	SEM	PC	NC	LCP	FIB	ZINC	SEM	Trt		DPI	Trt × DPI
Serum TNFα pg/mL									0.02		0.03	0.85
DPI 0	59.6x	6.67	55.1ab	104.3a	52.6b	46.1b	54.1ab	27.46
DPI 3	40.9y	4.79	41.0	47.5	38.0	39.2	39.3	13.45
DPI 6	56.6x	6.72	56.3ab	111.5a	51.7b	42.2b	42.7b	31.62

Table 6. Impact of dietary treatment (trt) and day post-inoculation (DPI) on ileal mRNA gene expression in pigs challenged with ETEC F18 at 14 d post-weaning.

	Day (DPI)		Treatment						P Value
	DPI	SEM	PC	NC	LCP	FIB	ZINC	SEM	Trt	DPI	Trt × DPI
ZO expressiona									0.63	<0.01	0.40
DPI 3	0.58y	0.056	0.58	0.61	0.50	0.83	0.44	0.14
DPI 6	1.39x	0.131	1.535	1.169	1.355	1.414	1.484	0.327
OCCL expressiona									0.65	<0.01	0.26
DPI 3	0.21y	0.036	0.37	0.19	0.16	0.24	0.15	0.16
DPI 6	0.62x	0.106	0.68	0.83	0.64	0.36	0.74	0.34
CD14 expressiona									0.75	<0.01	0.68
DPI 3	0.40y	0.046	0.54	0.38	0.34	0.43	0.34	0.16
DPI 6	1.08x	0.123	0.98	1.37	0.92	1.00	1.15	0.37

Table 7. Impact of dietary treatment (trt) and day post-inoculation (DPI) on ileal (IL) and colonic morphologya, and colitis probabilityb in pigs challenged with ETEC F18 at 14 d post-weaning.

	Day (DPI)		Treatment						P Value
	DPI	SEM	PC	NC	LCP	FIB	ZINC	SEM	Trt	DPI	Trt × DPI
IL VHa, µm									0.89	<0.01	0.14
DPI 3	263y	8.8	244	273	256	258	284	20.9
DPI 6	315x	8.9	322	295	333	342	284	20.9
IL CDa, µm									0.94	<0.01	0.29
DPI 3	236y	6.4	248	234	244	222	231	15.2
DPI 6	181x	6.4	170	198	166	187	186	15.2
IL VCa									0.85	<0.01	0.13
DPI 3	1.2y	1.20	1.1	1.3	1.1	1.3	1.3	0.18
DPI 6	1.9x	1.91	2.0	1.6	2.2	1.9	1.8	0.18
CO CDa, µm									0.11	0.45	0.70
DPI 3	403	10.2	385	412	417	406	394	23.9
DPI 6	391	10.2	356	406	387	437	367	23.9
Colitisb, %									0.42	0.35	0.55
DPI 3	32	7.8	19	50	28	28	39	17.7
DPI 6	43	8.7	19	50	50	72	28	17.7

Table 8. Impact of dietary treatment (trt) on colonic luminal amino acids and biogenic amines at 3-and 6-d post-inoculation (DPI) with ETEC F18 in pigs inoculated at 14 d post-weaning.

Compound, umol/g		Treatment							P Value
	DPI		PC	NC	LCP	FIB	ZINC	SEM	Trt	DPI	Trt × DPI
Alanine	3		1.81	1.66	2.02	1.50	2.00	0.261	0.29	0.02	0.40
	6		1.31	1.51	1.38	1.44	1.47	0.202
Glycine	3		0.79	0.88	1.21	0.81	0.89	0.174	0.62	0.04	0.83
	6		0.59	0.81	0.66	0.74	0.56	0.108
Valine	3		0.61	0.53	0.71	0.56	0.61	0.117	0.96	0.01	0.72
	6		0.39	0.50	0.42	0.47	0.31	0.082
Leucine	3		0.71	0.56	0.73	0.60	0.68	0.112	0.99	0.01	0.74
	6		0.47	0.55	0.47	0.50	0.38	0.081
Isoleucine	3		0.42	0.40	0.52	0.43	0.44	0.083	0.82	<0.01	0.78
	6		0.27	0.36	0.32	0.34	0.23	0.058
Threonine	3		0.47	0.44	0.60	0.46	0.49	0.076	0.18	<0.01	0.52
	6		0.26	0.39	0.38	0.38	0.23	0.049
Serine	3		0.67	0.72	1.01	0.74	0.74	0.144	0.41	<0.01	0.90
	6		0.41	0.59	0.58	0.59	0.45	0.079
Proline	3		0.50	0.45	0.59	0.45	0.60	0.095	0.73	<0.01	0.66
	6		0.31	0.42	0.38	0.37	0.28	0.066
Aspartic acid	3		0.82	0.95	1.10	0.86	1.02	0.155	0.07	<0.01	0.30
	6		0.44	0.69	0.62	0.79	0.58	0.111
Methionine	3		0.27	0.20	0.29	0.22	0.22	0.049	0.66	0.08	0.63
	6		0.16	0.21	0.21	0.21	0.13	0.036
Hydroxy proline	3		0.02	0.03	0.03	0.03	0.03	0.005	0.19	<0.01	0.98
	6		0.02	0.02	0.02	0.02	0.02	0.004
Glutamic acid	3		4.21	3.52	4.58	2.94	3.84	0.712	0.56	0.02	0.34
	6		2.46	2.82	3.05	3.40	2.67	0.528
Phenylalanine	3		0.35	0.26	0.34	0.28	0.34	0.051	0.68	0.08	0.76
	6		0.28	0.28	0.24	0.24	0.22	0.040
Glutamine	3		0.08	0.05	0.05	0.05	0.10	0.026	0.30	0.94	0.57
	6		0.05	0.06	0.06	0.06	0.08	0.022
Ornithine	3		0.16	0.13	0.20	0.14	0.19	0.033	0.40	<0.01	<0.01
	6		0.10	0.12	0.15	0.15	0.08	0.026
Lysine	3		1.03	0.77	1.09	0.79	1.01	0.148	0.47	0.40	0.38
	6		0.86	0.90	0.95	0.89	0.59	0.127
Histidine	3		0.02b	0.03ab	0.06a	0.06a	0.03ab	0.013	<0.01	0.95	0.03
	6		0.01b	0.07a	0.04a	0.05a	0.04a	0.013
Tyrosine	3		0.30	0.24	0.32	0.26	0.30	0.043	0.65	0.11	0.68
	6		0.26	0.26	0.22	0.21	0.20	0.036
Cadaverine	3		2.14	2.32	3.83	1.19	3.79	1.678	0.27	0.02	0.51
	6		1.62	1.52	0.87	0.75	1.53	0.633
Putrescine	3		0.64	0.68	0.71	0.53	0.54	0.088	0.52	0.38	0.78
	6		0.53	0.58	0.51	0.52	0.88	0.121

Full text

Introduction

Post-weaning diarrhea (PWD) remains a significant health and economic challenge in swine production, contributing to reduced growth performance, elevated morbidity and mortality, and substantial financial losses (Dubreuil et al. 2016; Fairbrother and Nadeau, 2019). The nursery phase is particularly vulnerable due to stressors associated with weaning and exposure to enteric pathogens such as enterotoxigenic Escherichia coli (ETEC), rotavirus, and Clostridium difficile.

These ETEC strains differ from commensal E. coli by expressing virulence factors that enable adhesion to the small intestine and secretion of enterotoxins, resulting in secretory diarrhea and epithelial damage (Dubreuil et al. 2016; Fairbrother and Nadeau, 2019). The most common ETEC strains in weaned pigs carry fimbrial genes F4 (K88) and F18, with 92.7% of ETEC-related PWD cases involving these adhesins (Frydendahl, 2002). Recent diagnostic surveillance at the Iowa State University Veterinary Diagnostic Laboratory (2010–2023) has shown that F18 has increasingly replaced K88 as the predominant fimbrial type detected in clinical cases (Paiva et al. 2025). Following colonization, ETEC produces heat-labile (LT) and heat-stable (ST) enterotoxins that disrupt electrolyte transport, leading to diarrhea, dehydration, impaired growth, and mortality rates up to 30% (Dubreuil et al. 2016).

Rotavirus also contributes to PWD by infecting enterocytes, causing villous atrophy and reduced absorptive capacity, which results in malabsorptive and osmotic diarrhea (Graham et al. 1984; Kapikian et al. 2001). Rotavirus is endemic in sow herds throughout the Midwest, with near-universal seropositivity among adult pigs and frequent shedding during the periparturient period. Co-infections with enterotoxigenic E. coli are common in nursery pigs, compounding the severity of post-weaning diarrhea and complicating management strategies.

Historically, antibiotics and pharmacological levels of zinc oxide (ZnO; 1000–3000 mg/kg) have been used to mitigate PWD and support growth performance (Poulsen, 1995; Carlson et al. 1999; Hill et al. 2000; Cromwell, 2002; Zhu et al. 2017; De Mille et al. 2022). However, concerns regarding antimicrobial resistance and environmental contamination have led to regulatory restrictions on their use (Wei and Yang, 2010; Bednorz et al. 2013), prompting investigation into alternative nutritional strategies.

Low crude protein (CP) diets and dietary fiber inclusion have emerged as promising approaches. Weaning-induced villous atrophy and reduced brush border enzyme activity impair protein digestion, with up to 40% of dietary protein reaching the hindgut undigested (Moeser et al. 2017; Montagne et al. 2007; Högberg and Lindberg, 2004). This undigested protein serves as a substrate for microbial fermentation, producing biogenic amines such as cadaverine and putrescine via decarboxylation of lysine, ornithine, and arginine (Kim et al. 2008; Heo et al. 2009; Linares et al. 2011; Davila et al. 2013; Berthoud et al. 2022). These compounds are associated with mucosal irritation and exacerbation of PWD (Porter and Kenworthy, 1969; Nollet et al. 1999; Chelakkot et al. 2018; Teti et al. 2002). High CP diets may intensify this effect by increasing fermentable substrate availability (Jeaurond et al. 2008), whereas low CP diets have been shown to reduce protein fermentation (Yu et al. 2019), pathogen survival (Opapeju et al. 2009), and scouring incidence (Heo et al. 2009; Heo et al. 2015; Wu et al. 2015; Gao et al. 2020; Limbach et al. 2021; Liu et al. 2022; Kim et al. 2023; Marchetti et al. 2023). Thus, lowering CP is hypothesized to reduce bacterial proliferation and diarrhea by limiting fermentable protein in the hindgut. However, their impact on intestinal atrophy and ETEC F18-specific pathology remains unclear.

Dietary fiber, particularly insoluble fiber, also plays a role in gut health. Wheat middlings, containing approximately 36.5% total dietary fiber (94.5% insoluble), have been evaluated for their potential to reduce clinical signs during ETEC exposure (Casas et al. 2018). Insoluble fiber increases digesta transit rate (Kim et al. 2012; Molist et al. 2014), potentially limiting pathogen contact time and reducing ETEC colonization. It is hypothesized that increasing the insoluble fiber component of the diet could reduce pathogen colonization and thereby lessen the negative effects of infection.

This study aimed to evaluate the efficacy of low CP and insoluble fiber-rich diets as compared to a positive control (PC) diet containing ZnO and carbadox, and a ZnO-alone diet, in mitigating post-weaning performance losses, PWD, bacterial shedding, and intestinal atrophy following experimental E. coli F18 challenge in pigs with naturally occurring rotavirus infection. We hypothesized that dietary interventions (ie PC, LCP, FIB, ZINC) would reduce scours and intestinal damage while improving overall performance relative to a negative control (NC) diet.

Materials and methods

All experimental procedures adhered to guidelines for the ethical and humane use of animals for research according to the Guide for the Care and Use of Agricultural Animals in Research and Teaching (FASS, 2010) and were approved by the Iowa State University Institutional Animal Care and Use Committee (IACUC #23-022) and Institutional Biosafety Committee (IBC #23-012).

Animals and experimental design

A total of 240 newly weaned mixed-sex pigs from a commercial herd that had not been vaccinated against E. coli (body weight 5.47 ± 0.27 kg; PIC 337 × 1050, Genus, Hendersonville, TN) arrived immediately after weaning at approximately 21 d of age. Pigs were individually weighed and randomly allocated to 40 mixed-sex pens (6 pigs/pen) at a research facility at Iowa State University (Ames, IA). Pens were on raised decks with woven wire flooring and were equipped with a galvanized steel feeder (58 cm of feeder space) and a nipple drinker, allowing for ad libitum access to feed and water throughout the study. Floors under the raised decks were cleaned with water at least three times per week to remove urine and feces. Room temperatures were set to 30 °C and gradually lowered to 22 °C over the 42-d period. Upon arrival all pigs were vaccinated against porcine reproductive and respiratory syndrome virus (PRRSv) and genotyped for fucosyltransferase (FUT1) via cheek swab using Sanger sequencing modified from Frydendahl et al. (2003). Due to the nature of pig source and research facility, resistant pigs (defined as those with the FUT1^AA^ genotype, approximately 28% of the population) were present within the population of pigs used for the study. These resistant pigs were distributed randomly across treatments.

Dietary treatments

Pens were randomly assigned to one of five dietary treatments (8 pens/treatment, Table 1) consisting of a negative control (NC), positive control (PC), low crude protein (LCP), fiber (FIB), and ZnO only (ZINC) diets. The NC was a standard nursery diet without ZnO or carbadox (Mecadox, Phibro Animal Health, Teaneck, NJ). The PC supplemented the NC diet with 3,000 mg/kg added Zn (provided from 3,750 mg/kg ZnO) and carbadox (55 mg/kg, Mecadox, Phibro Animal Health, Teaneck, NJ). The LCP diet had reduced crude protein (CP) while also lowering the standardized ileal digestible (SID) Lys content as compared with the NC diet (1.30 vs 1.46% SID Lys in LCP vs. NC, respectively), all other amino acids were balanced to SID Lys. Arginine was held constant across treatments due to data indicating impacts on early nursery performance (Hagen et al. 2024; Humphrey et al. 2024). The FIB treatment added 8% wheat middlings to the NC diet, while the ZINC treatment included 3,000 mg/kg added Zn (from 3,750 mg/kg ZnO) in the NC diet.

The NC, PC, and ZINC diets were formulated to be nutritionally identical, with the only differences being that corn starch in the NC diet was replaced by carbadox (55 mg/kg from Mecadox 2.5) and/or zinc (3,000 mg/kg Zn provided from 3,750 mg/kg ZnO) in the PC and ZINC diets, respectively. Ingredient adjustments were made in the LCP and FIB diets to meet their respective dietary strategies. Diets were formulated to have a determined level of available phosphorus. Total calcium was allowed to be included at a level similar across the dietary treatments within the phase. All other nutrients were formulated to meet or exceed the National Research Council recommendations for swine (NRC, 2012). Dietary treatments were administered in three phases: phase 1 (d 0–7), phase 2 (d 7–21), and all five dietary treatments were discontinued at the start of phase 3, when a common diet (d 21–42) was provided to all pigs. This common diet contained no pharmacological Zn or carbadox. All diets were provided in mash form. Feed was manufactured at the Iowa State University Swine Nutrition farm feed mill (Ames, IA), and samples were collected at the completion of mixing and stored at −20 °C for subsequent analysis.

Feed samples were ground to 1 mm particle size (Variable Speed Digital ED-5 Wiley Mill; Thomas Scientific, Swedesboro, NJ). In duplicate, diets were analyzed for nitrogen (N; method 990.03; AOAC, 2007; Trumac; LECO Corp., St. Joseph, MI) and gross energy. For N analysis, ethylenediaminetetraacetic acid (9.56% N) was used as the standard for calibration and was determined to contain 9.59 ± 0.02% N. Crude protein was calculated as N × 6.25. Gross energy was determined using an isoperibolic bomb calorimeter (model 6200; Parr Instrument Co., Moline, IL). Benzoic acid (6,318 kcal GE/kg) was used as the standard for calibration and was determined to contain 6,314 ± 8 kcal GE/kg. Total dietary fiber (TDF) was analyzed at the University of Missouri (Columbia, MO) using AOAC 2022.01 (AOAC, 2022) methodology, which distinguishes insoluble (IDF) and soluble (SDF) fiber fractions.

ETEC inoculation

All pigs were orally drenched with a freshly cultured ETEC F18 ^+^ strain (6 mL/pig, approximately 5 × 10^8^ CFU/mL) on day 14 of the experiment, designated as day post inoculation (DPI) 0. This timing was selected to reflect field observations of late-phase nursery outbreaks, which commonly emerge around 2 wks post-weaning as pigs transition to more complex diets and encounter heightened environmental stressors. The ETEC F18 strain (originated from the intestine of a nursery aged pig diagnosed with enteric colibacillosis) used in this study was positive for heat-labile (LT), heat-stable b (STb), and enteroaggregative Escherichia coli heat-stable enterotoxin 1 (EAST1) as determined by PCR at the Iowa State University Veterinary Diagnostic Laboratory (ISU VDL, Ames, IA). The bacterial inoculum was prepared at the ISU VDL (Ames, IA). Briefly, a fresh bacterial culture grown for ^∼^18 h at 37 °C on standard sheep blood agar (Remel, Thermo Fisher Scientific, Lenexa, KS) was used to inoculate two 50 mL bottles of sterile tryptic soy broth (TSB; BD, Sparks, MD, USA), followed by overnight incubation at 37 °C with shaking. The next day, these broth cultures were transferred into two fresh bottles, each containing 450 mL of fresh TSB, and incubated for an additional 5 h at 37 °C with shaking. The optical density at 600 nm (OD_600_) of the culture was adjusted to approximately 0.9 using sterile PBS, corresponding to around 5 × 10^8^ CFU/mL as determined via viable plate count.

Growth and fecal consistency measurements

Pigs and feeders were weighed individually on days 0, 7, 14, 21, and 42 of placement corresponding with day post-inoculation (DPI) −14, −7, 0, 7, and 28. Feed disappearance was recorded to calculate average daily gain (ADG), average daily feed intake (ADFI), and gain to feed ratio (G: F) for each phase. Daily fecal consistency scores were recorded for each pen from DPI −14 to 14, representing the average fecal consistency within the pen where: 1 = solid, 2 = semi-solid, 3 = semi-liquid, and 4 = liquid. Removed pigs were weighed with a date recorded to adjust growth and feed intake data. Mortalities were less than 5% for the whole trial, therefore mortality data was not analyzed or included.

Sample collection

Two pigs per pen, closest to the average of the pen and susceptible to ETEC (FUT1^GG^ or FUT1^GA^), were selected as sample pigs. The first (random of the two selected pigs) was euthanized via captive bolt and exsanguination on 3 DPI of the experiment. A 40 cm section of terminal ileum (IL) was collected. The proximal 20 cm was immediately formalin fixed for histological evaluation, while the distal portion was washed with phosphate-buffered saline (pH 7.4) and frozen on dry ice for gene expression analysis. Additionally, a 10 cm section of spiral colon tissue was formalin fixed, and colonic contents were collected and stored at −20 °C for biogenic amine analysis. The formalin-fixed tissues were stored in 10% neutral buffered formalin solution at room temperature for 48 h, and then transferred to 70% ethanol before being submitted to the ISU VDL (Ames, IA) for tissue processing and slide preparation for histopathology.

The second pig euthanized had blood collected on DPI 0, 3, and 6, and fecal swabs were collected on DPI 0, 2, 4, 6. This pig was euthanized on day 20 (DPI 6) following the same procedure as above. Blood was collected via sterile jugular venipuncture using 10 mL serum vacutainer tubes (Becton Dickinson, Franklin Lakes, NJ) and allowed to clot for at least 6 h at room temperature. An additional fecal swab was taken from one susceptible pig (FUT1^GG or GA^) per pen on D 28 of the experiment (DPI 14) for additional ETEC shedding measurement.

Fecal ETEC F18 and STb qPCR

Extraction of total fecal DNA from fecal swabs collected on DPI 0, 2, 4, 6, and 14 from ETEC F18 susceptible pigs, was conducted using a commercially available DNA extraction kit (Powersoil DNEAsy kit, Catalog: 74014, Qiagen, Germantown, MD). Concentration of DNA was quantified using a spectrophotometer (ND-100; NanoDrop Technologies, Inc., Rockland, DE) and samples were diluted to 25 ng/µL with nuclease-free water. Twenty-five ng of fecal DNA was amplified using a duplex real-time quantitative polymerase chain reaction (qPCR) assay for two specific virulence genes of the ETEC F18 inoculum: F18 fimbriae (GenBank: M61713.1) and STb (STII) toxin (GenBank: M35586.1). Primers were designed by Applied Biosystems (Thermo Fisher Scientific, Waltham, MA) under the following assay IDs F18: APAAKXE tagged with VIC (Catalog 4448508) and STb: AP9HVJN tagged with FAM (Catalog 4331348). Reactions were carried out in duplicate at 10 µl using 1x TaqPath^TM^ BactoPure^TM^ Microbial Master Mix with ROX as a passive reference dye (Catalog A52700, Thermo Fisher Scientific, Waltham, MA), 1x of each gene specific assay, and 25 ng of fecal DNA on the QuantStudio^TM^ 3-Real-Time PCR machine (Thermo Fisher Scientific, Waltham, MA). The PCR reactions consisted of a hold stage at 60 °C for 30 s and 95 °C for 2 min, followed by 40 cycles of 95 °C denaturing for 10 s and 60 °C annealing for 30 sec, the 40 cycles were followed by another hold stage of 60 °C for 30 sec. Undetectable samples were set as a CT of 40 (maximum cycles) for statistical analysis.

Serum TNFα analysis

Serum was separated by centrifugation (2000 × g for 10 min at 4 °C), collected, and stored at −80 °C for subsequent TNFα analysis. Serum collected from FUT1-positive pigs on DPI 0, 3, and 6 was analyzed for porcine tumor necrosis factor alpha (TNFα) using a commercially available kit (Thermo-Fisher, Waltham, MA, CAT #ES24RBX10) according to manufacturer’s protocol. The intra-assay coefficient of variation (CV) was set at 10%, and the inter-assay CV at 15%.

Histological analysis

Formalin-fixed ileum and colon tissues were processed and embedded in paraffin wax at the ISU VDL (Ames, IA). Sections (5 μm) were cut and stained with hematoxylin and eosin. Three transverse sections from the ileum and one from the colon were examined using images taken at 20X magnification with the DP80 Olympus Camera mounted on an OLYMPUS BX 53/43 microscope (Olympus Scientific, Waltham, MA). In the ileum, ten well-oriented villi and their corresponding crypts were measured from each of the three sections using OLYMPUS cellSens Dimension 1.16 software (Olympus Scientific, Waltham, MA) for a total of 30 measurements. These measurements for ileal villus height (VH), crypt depth (CD), and the villus-to-crypt (VC) ratio were averaged for each animal. In the colon, crypt depth (CD) was determined by averaging 10 measurements from well-defined crypts.

Colonic inflammation

Fixed colon tissues, previously stained with hematoxylin and eosin for structural measurements, were examined by an ISU VDL pathologist who was blinded to treatment. Colon samples were scored on a scale from 0 to 3: 0 for no inflammation, 1 for very mild suppurative inflammation limited to lymphoglandular structures of the mucosa, 2 for mild suppurative inflammation including rare crypt abscesses and rare neutrophils in the lamina propria, and 3 for moderate suppurative inflammation including numerous crypt abscesses, neutrophils in the lamina propria, and attenuation of enterocytes within the mucosa. Due to low instances of colitis scores 1 or 3, scores ≥ 1 were considered as observed colitis for statistical analysis.

GC-MS biogenic amine and amino acid analysis

Colonic digesta samples (300 mg), which were previously frozen, were submitted to the W.M. Keck Metabolomics Research Laboratory (RRID: SCR_017911; Ames, IA) for targeted analysis of free amines and amino acids via gas chromatography–mass spectrometry (GC-MS). Sample preparation and analysis followed Pearce et al. (2024) and the recommended EZ: faast Kit protocol for physiological amino acids (Phenomenex, Torrance, CA), with minor modifications.

Briefly, samples were weighed into 2 mL microcentrifuge tubes and spiked with 400 µL of internal standard reagent (0.2 mM norvaline, 10% 1-propanol, 20 mM hydrochloric acid), followed by 1,000 µL of cold 25% 1-propanol (HPLC grade 1-propanol: LC-MS grade water, 1:3; Fisher Scientific, Waltham, MA). Two 2.8 mm ceramic beads were added prior to homogenization using a Bead Mill 24 Homogenizer (Thermo Fisher Scientific, Waltham, MA). Samples were sonicated in an ice-cold water bath (Model 2510, Branson Ultrasonics, Brookfield, CT) for 10 min, vortexed for 5 min, and centrifuged at 15,000 × g for 7 min. Supernatants were recovered and 200 µL of each extract was subjected to solid-phase extraction, derivatization, and GC-MS preparation per kit instructions. Standards for putrescine, cadaverine, and amino acids (Sigma-Aldrich, St. Louis, MO) were prepared identically.

The GC-MS analysis was performed using an Agilent 7890 GC coupled to a 5975 Mass Selective Detector (Agilent Technologies, Santa Clara, CA) with a Zebron ZB-AAA column (10 m × 250 µm; Phenomenex). One microliter of sample was injected in splitless mode (10:1 split) at 280 °C. The oven temperature was programmed from 70 °C (held for 0.5 min) to 320 °C at 30 °C/min, held for 3.167 min. Helium served as the carrier gas at 1.2 mL/min. The MS transfer line was maintained at 280 °C, with electron ionization at 70 eV. Source and quadrupole temperatures were set at 230 °C and 150 °C, respectively. Mass spectra were collected from m/z 45 to 450.

Compound identification and quantification were performed using AMDIS (National Institute of Standards and Technology, Gaithersburg, MD) and a manually curated retention-indexed GC-MS library. Additional identification was supported by the NIST20, Wiley 11, and EZ: faast spectral libraries. Final concentrations were calculated by integrating peak areas relative to the internal standard and standard curves and normalized to digesta sample mass. Missing values were imputed as half the minimum detected concentration for each compound to account for values below the detection limit while minimizing bias (Giskeødegård and Lydersen, 2022).

Ileal gene expression qPCR

Approximately 30 mg of ileal tissue was homogenized using the Qiagen Tissuelyser II (Germantown, MD), followed by total RNA extraction using the Qiagen RNeasy Mini Kit (Catalog: 74104, Qiagen, Germantown, MD). The RNA concentration was assessed using a spectrophotometer (ND-100; NanoDrop Technologies, Inc., Rockland, DE), and all samples had 260:280 nm ratios above 1.8. Complementary DNA (cDNA) was synthesized from 0.8 μg of isolated RNA using the QuantiTect Reverse Transcription Kit (Catalog: 205311, Qiagen, Germantown, MD). Post-synthesis, all cDNA samples were diluted tenfold with nuclease-free water.

For real-time quantitative polymerase chain reaction (qPCR) gene -specific primers (IDT, Coralville, IA) for Ribosomal protein-L19 (RPL19, endogenous reference gene), zona occludens-1 (ZO), occludin (OCCL), and CD14 were diluted to 10 µM with nuclease-free water. The sequence (5′–3′) of primers and probes for target genes RPL19, ZO, OCCL, and CD14 genes are reported in Becker et al. (2020). Each qPCR reaction consisted of 10 µL iQ SYBR Green Supermix (Catalog: 1708880, Bio-Rad Laboratories, Inc., Hercules, CA), 1 µL each of forward and reverse primers, 5 µL nuclease-free water, and 3 µL cDNA, totaling a 20 µL reaction volume. Samples were analyzed in duplicate, alongside no-reverse transcriptase negative controls and a pooled cDNA reference sample in every 96-well plate.

Amplification was carried out using a Real-time PCR Detection System (iQ5; Bio-Rad Laboratories, Inc., Hercules, CA) with cycling conditions comprising a 5-min initial denaturation at 95 °C, followed by 40 cycles (95 °C for 30 s, 55 °C for 30 s, and 72 °C for 30 s). A dissociation curve analysis validated the amplification specificity of each reaction, with optical detection conducted at 55 °C. Analysis of amplification plots utilized Optical System Software version 2.0 (iQ5; Bio-Rad Laboratories Inc., Hercules, CA) to derive cycle threshold (CT) values for each reaction. Abundance of mRNA abundance in individual samples was normalized relative to RPL19 and the pooled sample, and fold change was determined using the 2^-ΔΔCT^ method (Livak and Schmittgen, 2001).

Statistical analysis

The data were analyzed using mixed model methods (Proc Mixed, SAS 9.4, SAS Inst., Cary, NC) with pen as the experimental unit and the fixed effects of treatment, time, and their interaction. Starting body weight served as a covariate for growth performance. Pen was treated as a repeated measure, employing the spatial power covariance structure for growth performance due to unequal time points, and the autoregressive covariance structure for all other outcomes.

Testing for normality and homoscedasticity of the studentized residuals was conducted using the Univariate procedure. Serum TNFα, colon digesta metabolites, and PCR gene expression data analysis utilized log-transformation to ensure normality, with least squares (LS) means and standard error of the mean (SEM) back transformed for reporting. Statistical outliers were identified as data points deviating greater than three studentized residuals from the model and were subsequently excluded from the analysis (three pens excluded from overall GF analysis).

The probability of colitis (score ≥ 1) on DPI 3 and 6 was assessed using a binomial logistic regression model (PROC LOGISTIC, SAS 9.4, Cary, NC), incorporating fixed effects of dietary treatment, DPI, and their interaction. Firth’s Penalized Likelihood correction was applied to address quasi-complete separation, which can bias maximum likelihood estimates in sparse or imbalanced data. Predicted probabilities of colitis, back transformed from the logit scale, are reported.

For significant fixed effects (P < 0.05), LS means were separated using pairwise t-tests (PDIFF option, SAS 9.4, SAS Inst., Cary, NC), with a Tukey post-hoc adjustment for multiple comparisons. Results were considered significant at P ≤ 0.05 and suggestive of a trend at P > 0.05 and ≤ 0.10.

Results

Diet analysis

Differences in ingredient composition led to variability in gross energy (GE), even though calculated metabolizable energy (ME) was similar across diets and equal between the LCP and FIB diets in phase 1 (Table 2). Reduced soybean meal inclusion lowered GE in the LCP diets, whereas the addition of wheat middlings increased GE in the FIB diets. As intended, crude protein levels were lower in the LCP diets. Total dietary fiber analysis showed that the FIB diets contained the highest levels of TDF, IDF, and SDF among treatments and were greater than the NC; however, in phase 2 the PC diet analyzed higher than expected for TDF and IDF.

Pig performance and genetic status

Pigs arrived from the sow farm at 5.47 ± 0.27 kg and exited the nursery at 25.26 ± 7.37 kg after 42 d. The overall ADG was 0.38 ± 0.17 kg, ADFI was 0.49 ± 0.25 kg, and G: F was 0.83 ± 0.14. Dietary treatments alone did not significantly affect BW, ADG, or ADFI during any period or across the entire experiment (Table 3, P > 0.1). A significant treatment by period interaction (Table 3, P = 0.01) was observed for G: F, with pigs fed the FIB diet having a 23% reduction compared to PC in the first week post-wean pigs with other treatments remaining intermediate. This reduction in G: F did not persist throughout the experiment with no effect on overall G: F (Table 3, P > 0.1).

Pigs were blindly assigned to pens prior to obtaining genotyping results, with 28% of the pig population testing negative (FUT1^AA^) for the FUT1 receptor, and resistant to F18 attachment. Due to randomization, resistant pigs were randomly distributed across treatments, and no pen had more than 3 resistant pigs.

Clinical signs

Following signs of scouring immediately post-weaning, pigs tested positive for rotavirus strains A, B, and C on day 4 post-weaning via PCR analysis of fecal and intestinal tissue samples (ISU VDL, Ames, IA). Elevated fecal consistency scores during the pre-ETEC F18 period reflected the clinical impact of rotavirus infection. Daily fecal consistency scores revealed significant effects of treatment and day (Fig. 1; P < 0.01), as well as a significant treatment × day interaction (Fig. 1; P < 0.01). Following ETEC F18 inoculation, fecal consistency scores increased across all groups, peaking between 2- and 3-DPI. Pigs fed the PC diet consistently exhibited lower fecal scores throughout both the natural rotavirus outbreak and the post-ETEC F18 challenge period compared to all other treatments.

Daily fecal consistency score during first 28 d of experiment, corresponding to 14 d pre- and 14 d post- inoculation with ETEC. Pigs were natural infected with rotavirus strains A, B, and C at time of weaning and orally inoculated with ETEC F18 14 d post weaning, indicated by vertical red line. Effects of treatment, day, and treatment × day all significant (P < 0.01). fecal consistency scores were visually scored using a 1–4 scale where 1 = solid, 2 = semi-solid, 3 = semi-liquid, and 4 = liquid.

Fecal DNA ETEC F18 and STb shedding

Baseline qPCR analysis of fecal DNA indicated no detectable shedding of ETEC F18 or STb genomic material at DPI 0 (Table 4). Gene shedding peaked between DPI 2 and 4, with CT values of 29.1 and 28.9 for the F18 gene, and 30.2 and 30.0 for the STb gene, respectively. A significant treatment × day interaction was observed for F18 expression (P = 0.02; Table 4). While CT values were comparable across treatments at DPI 0, 2, 4, and 14, pigs fed the LCP diet exhibited higher F18 gene shedding than the PC group at DPI 6. By DPI 14, shedding of both genes had returned to baseline levels.

Tissue morphology, inflammatory markers, and metabolites

Analysis of serum TNFα revealed lower concentrations at DPI 3 compared to DPI 0 and 6 (Table 5, P = 0.03), or lower circulating TNFα corresponding with peak scouring. Dietary treatment also influenced circulating levels (Table 5, P = 0.02). Pigs fed LCP and FIB diets had significantly lower concentrations than the NC group at DPI 0, while PC and ZINC groups remained intermediate. By DPI 6, TNFα concentrations in the LCP, FIB, and ZINC groups were again significantly lower than NC, with the PC group remaining intermediate. Thus, dietary effects on circulating TNFα were evident before and after peak infection (DPI 0 and 6), but not at peak infection (DPI 3), when values converged across treatments.

Ileal CD decreased between DPI 3 and 6, while VH and VC increased, demonstrating improved villus structure from peak to post‑infection (Table 7, P < 0.01). No significant interactions were observed between dietary treatment and DPI on ileal morphology. In terms of ileal gene expression, significant effects were attributed to DPI (Table 6, P < 0.01), but no significant impact of dietary treatment was detected (P > 0.1). Expression of ZO, OCCL, and CD14 was higher at DPI 6 than at DPI 3 across all treatments, reflecting improvements following peak infection (scouring).

Neither dietary treatment nor DPI significantly affected the probability of colitis, nor the CD in colon tissues collected at DPI 3 and 6 (Table 7, P > 0.1), indicating colonic morphology remained stable across treatments and time.

Biogenic amine concentrations in colonic contents were measured at both DPI 3 and 6 (Table 8). Dietary treatment did not influence cadaverine or putrescine (Trt P > 0.1). However, DPI significantly impacted cadaverine, with concentrations lower at DPI 6 compared to DPI 3 (P < 0.05). Histidine was affected by treatment (P < 0.01) and by a treatment × DPI interaction (P = 0.03). PC diets showed the lowest overall values, while NC and ZINC diets were low at DPI 3 and higher at DPI 6. In contrast, LCP and FIB diets remained high at both time points. Ornithine, along with several other amino acids, decreased in concentration between DPI 3 and 6 (P < 0.05), including glutamic acid, aspartic acid, proline, serine, threonine, leucine, valine, glycine, and alanine.

Discussion

Clinical signs

Pigs arrived at the nursery with a natural rotavirus exposure, common in commercially sourced pigs. This resulted in scouring from approximately day 2 through 9 post-weaning. Rotavirus damages enterocytes and disrupts absorptive capacity, resulting in both secretory and malabsorptive diarrhea (Graham et al. 1984; Kapikian et al. 2001). During this period, diagnostic testing indicated that rotavirus was the primary driver of the early scouring. Recovery began around day 10 post-weaning, prior to ETEC F18 inoculation at day 14, suggesting that most pigs had cleared the rotavirus infection by the time of challenge. This natural exposure added biological relevance by mimicking commercial nursery conditions, where rotavirus is a common co-factor in PWD (Fairbrother and Nadeau, 2019).

The PC diet consistently maintained lower fecal consistency scores during the rotavirus phase, indicating that the combination of carbadox and ZnO may have aided in reducing scouring even in the context of viral-induced diarrhea. Carbadox’s broad-spectrum antimicrobial properties likely contributed to controlling bacterial infections, modulating immune responses, reducing inflammation, and supporting intestinal recovery (He et al. 2020). Pharmacological ZnO can enhance intestinal integrity, morphology, tight junction expression, mucosal repair, and immune function (Hahn and Baker, 1993; Pearce et al. 2015; Zhu et al. 2017; Zhang and Guo, 2009; Hu et al. 2012; Shen et al. 2014), which may also have contributed to these effects during the rotavirus outbreak. While pharmacological ZnO alone has been shown to reduce scouring and improve gut function (Shelton et al. 2011; Mores et al. 1998; Yi et al. 2023), in this study, the ZINC diet did not reduce fecal consistency during the rotavirus phase.

One of the primary clinical signs of ETEC infection is secretory diarrhea, caused by enterotoxins disrupting fluid and electrolyte secretion in intestinal epithelial cells (Nagy and Fekete, 2005; Dubreuil et al. 2016). The ETEC F18 challenge successfully induced scouring and shedding of the inoculated ETEC, confirming challenge success. Fecal consistency scores peaked between 2- and 3-DPI aligning with previous studies (Sugiharto et al. 2014; Rossi et al. 2014; Yokoyama et al. 1997; Duarte et al. 2020; Kim et al. 2019; He et al. 2022). Across treatments, the severity of scouring remained mild compared to field outbreaks. Unlike field outbreaks, which often involve multiple pathogens and environmental stressors, this study introduced only ETEC F18 under controlled conditions. Clinical signs of ETEC infection are multifactorial, requiring not only pathogen presence but also stressors such as feed restriction, temperature fluctuations, and pig mixing (Dewey et al. 1995; Amezcua et al. 2002; Laine et al. 2008; Moredo et al. 2015). These factors were minimized in the current study, likely contributing to the subclinical response. Genetic resistance to ETEC F18 (FUT1^AA^) was evenly distributed across treatments, minimizing bias. However, resistant pigs within pens may have reduced overall fecal consistency by limiting colonization and diarrhea. Although FUT1 affects susceptibility (Meijerink et al. 1997), few historically compared clinical outcomes between genotypes after F18 challenge, making its impact difficult to quantify. Recent work has begun to address this gap, with Welch et al. (2025) showing that FUT1‑resistant pigs-maintained growth and had reduced mortality following F18 challenge, and Due (2024) reporting greater intestinal attachment, diarrhea severity, and poorer growth in susceptible genotypes.

Fecal shedding of ETEC F18 and STb peaked between DPI 2 and 4 and resolved by DPI 14, coinciding with peak scouring. This also aligns with previous studies where peak fecal excretion typically occur between DPI 3 and 5 and resolve by DPI 9 to 11 (Zúñiga et al. 1997; Verdonck et al. 2002; Tiels et al. 2008; Nadeau et al. 2017; Li et al. 2019). In contrast to more severe challenge models such as Becker et al. (2020), which reported 23% mortality and prolonged shedding peaking at 7 DPI, the current study observed only 3% mortality and no known effect on growth performance. Shedding was quantified using a qPCR targeting two genes specific to the inoculated strain, allowing differentiation from resident ETEC populations.

Dietary impact

Across treatments, the PC diet, supplemented with carbadox and ZnO, consistently reduced fecal consistency scores during both rotavirus and ETEC phases, outperforming all other diets. This is consistent with work by He et al. (2020), where carbadox supplementation during ETEC F18 reduced diarrhea incidence from 14.89 to 8.66%. Carbadox can accelerate pathogen clearance, reduce circulating TNFα, and restore mucosal integrity, demonstrating protection against ETEC (He et al. 2020). In the current study, PC‑fed pigs showed a steeper numerical increase in F18 and STb shedding around DPI 2–4, followed by a more rapid decline by DPI 6, suggesting a faster clearance trend compared with other treatments. Despite these improvements in scouring and indications of quick shedding, growth performance was not enhanced. Previous studies have shown that carbadox can improve growth and intestinal morphology during ETEC challenge, particularly through increases in villus height (He et al. 2020; Kim et al. 2021; Kim et al. 2022). These benefits are thought to arise from reductions in intestinal inflammation, which would otherwise divert energy toward immune activation and away from growth (Klasing and Johnstone, 1991; Johnson, 1997). In the current study, the subclinical nature of the ETEC challenge likely limited intestinal inflammation, as evidenced by circulating TNFα remaining below baseline (<100 pg/mL; Barba-Vidal et al. 2017; López-Colom et al. 2019) for most treatments. The lack of a measurable systemic immune response suggests that inflammation was largely localized within the intestine and did not reach a magnitude that would normally compromise or reallocate energy from growth (Becker et al. 2020).

The ZINC diet, containing pharmacological levels of Zn (as ZnO), demonstrated growth patterns and ETEC shedding similar to the PC diet but did not reduce fecal consistency scores during the rotavirus phase or improve tight junction expression. While ZnO has been shown to enhance intestinal morphology, barrier integrity, mucosal repair, and feed intake (Hahn and Baker, 1993; Pearce et al. 2015; Zhu et al. 2017; Zhang and Guo, 2009; Hu et al. 2012; Shen et al. 2014; Shelton et al. 2011; De Mille et al. 2022), the lack of fecal consistency improvements in this study suggests that ZnO alone may not fully replicate the benefits observed when combined with carbadox. Historical reductions in fecal consistency associated with ZnO appear not to be directly linked to antimicrobial activity against ETEC itself (Mores et al. 1998; Yi et al. 2023), indicating that ZnO may primarily support intestinal repair and barrier maintenance rather than direct pathogen suppression. The additional antimicrobial properties provided by carbadox may have enhanced pathogen clearance and reduced inflammation, contributing to the superior performance of the PC diet in controlling scouring. Although ZnO supported growth and shedding outcomes similar to PC, its inability to reduce diarrhea during rotavirus or ETEC challenge highlights the potential need for combination strategies under multifactorial stress conditions.

Low crude protein diets aim to reduce undigested protein reaching the hindgut, thereby minimizing harmful fermentation products such as biogenic amines like putrescine and cadaverine, which can irritate the mucosa and promote diarrhea (Porter and Kenworthy, 1969; Nollet et al. 1999; Chelakkot et al. 2018; Linares et al. 2011; Davila et al. 2013; Berthoud et al. 2022). Previous studies have shown that high crude protein diets increase biogenic amine concentrations, exacerbating mucosal irritation and diarrhea risk (Pieper et al. 2012; Yu et al. 2019). Elevated crude protein levels can also contribute to osmotic diarrhea by increasing solute load in the colon, triggering water secretion to balance luminal osmolality (Wu et al. 2015; Marchetti et al. 2023). In the current study, crude protein was reduced from 20.5% to 17.65% in phase 1 and from 19.5% to 17.15% in phase 2, representing a reduction of less than 3% compared to the NC diet. This modest reduction may have been insufficient to meaningfully reduce metabolite production. Concentrations of cadaverine (1.82 ± 0.44 μM/g) and putrescine (0.61 ± 0.03 μM/g at DPI 6) were lower than previously reported in studies with greater crude protein reductions. For example, Pearce et al. (2024) reduced crude protein from 24% to 17.4%, reporting cadaverine at 7.7 ± 7.7 μM/g and putrescine at 1.74 ± 0.30 μM/g. These comparisons suggest that the NC diet may not have provided a high enough crude protein baseline to induce measurable colonic fermentation. Consequently, the LCP diet did not reduce fecal consistency scores during rotavirus or ETEC infection, nor did it improve colonic inflammation. The incidence of colitis at DPI 3 and 6 also did not differ between LCP and NC, further supporting the lack of effect on colonic health.

Pigs fed the LCP diet exhibited prolonged fecal shedding of ETEC F18 through DPI 6, whereas pigs in the other treatments had returned to baseline by this time. Despite this extended detection, pigs fed the LCP diet did not experience more severe clinical signs, indicating that higher qPCR-based shedding does not necessarily reflect worsened disease expression in this model. The qPCR detects strain-specific DNA and is a sensitive tool for tracking exposure dynamics, although it does not distinguish between DNA from live and dead bacteria (Byun et al. 2013). Histologically, no differences in ileal morphology or barrier gene expression (ZO, occludin, CD14) were observed at DPI 3 or 6, and numerical increases in villus-to-crypt ratio did not translate into functional improvements in fecal consistency, gene expression, or growth. These findings align with previous studies showing that LCP diets can improve villus height and crypt depth following ETEC infection (Opapeju et al. 2009; Yue and Qiao, 2008; Fan et al. 2017), but suggest that in this study, small crude protein reductions were insufficient to enhance intestinal recovery. Some studies have reported improved barrier integrity with LCP diets, including Opapeju et al. (2009) who compared 22.5% versus 17.6% crude protein, and (Wang, 2018) who used reductions exceeding 4%. In contrast, the current study reduced crude protein by 2.85% in phase 1 and 2.35% in phase 2, which may have been insufficient to elicit similar effects. Although LCP diets have reduced scouring in ETEC F4 models (Heo et al. 2009), the current results did not replicate those outcomes for ETEC F18. The lack of non-challenged controls limits differentiation between natural post-weaning intestinal maturation and diet-induced recovery.

The FIB diet, which incorporated wheat middlings as a source of insoluble fiber, was expected to accelerate digesta passage, reduce bacterial colonization, and offset rotavirus-induced reductions in gut motility (Kim et al. 2012; Molist et al. 2014; Ramig, 2004; Bertschinger et al. 1979; Casas et al. 2018). Insoluble fiber has also been associated with increased villus height and feed intake (Hedemann et al. 2006; Molist et al. 2014). However, the FIB diet did not reduce scouring during rotavirus or ETEC challenge, nor did it improve pathogen clearance or intestinal morphology. Reduced feed efficiency in the first week post-weaning likely reflected energy dilution combined with limited early voluntary intake (De Jong et al. 2014; Li and Patience, 2017). Lower circulating TNFα in FIB pigs compared to NC at DPI 0 and 6 may suggest reduced inflammation; however, this did not correspond with improvements in scouring, growth, or intestinal recovery, highlighting the limited biological relevance of this change. These results emphasize that although insoluble fiber can theoretically enhance passage rate and reduce pathogen contact time, the inclusion level and timing used in this study were insufficient to achieve these outcomes.

Morphological and inflammatory impact

Morphological improvements were observed across all dietary treatments between DPI 3 and 6, including increased villus height, decreased crypt depth, and a higher villus‑to‑crypt ratio. These changes are consistent with recovery following ETEC infection. ETEC‑induced damage is driven by enterotoxins that stimulate fluid loss from enterocytes (Berberov et al. 2004) and trigger apoptosis of infected cells (Schmitz et al. 1999). Previous studies have documented villus blunting (Liu et al. 2013; Becker et al. 2020; Sun et al. 2021), crypt deepening (Pluske et al. 2018), and reduced villus‑to‑crypt ratios (Liu et al. 2013; Becker et al. 2020) during ETEC challenge. In this study, improvements in villus‑to‑crypt ratio were driven by decreased crypt depth and increased villus height across all diets. Although the absence of a non‑challenged control complicates interpretation, the timing of peak scouring and shedding could suggest that these improvements reflect recovery from ETEC rather than natural post‑weaning development.

Similar to the morphological changes, expression of barrier-related genes (ZO, OCCL, CD14) increased between DPI 3 and 6, likely reflecting both mucosal recovery and post-weaning maturation (Bauer et al. 2011). Dietary treatments did not significantly affect these gene expression changes or morphological outcomes. Numerical differences among LCP- or FIB-fed pigs did not correspond with improvements in fecal consistency, pathogen clearance, or growth, suggesting limited functional impact.

Circulating TNFα concentrations remained below typical baseline levels (<100 pg/mL; Barba-Vidal et al. 2017; López-Colom et al. 2019), with only transient elevations in NC at DPI 0 and 6. In pigs, TNFα plays a significant role in bacterial infections and digestive disorders (Pauli, 1995). This pattern may suggest that inflammatory responses were localized to the intestinal mucosa rather than systemic. However, the lower mRNA expression of CD14 at DPI 3 compared to DPI 6 partially refutes this speculation. CD14 is a co-receptor that plays a critical role in the innate immune response by recognizing lipopolysaccharide (LPS) and facilitating immune cell activation (Pålsson-McDermott and O’Neill, 2004), and has previously been shown to increase in ileal tissue during peak infection (Becker et al. 2020). As gene expression reflects transcriptional potential rather than protein activity or immune cell recruitment, further work is needed to clarify the functional relevance of these findings.

Despite differences in fecal consistency and pathogen shedding, none of the dietary treatments improved growth performance compared to the NC diet. This indicates that the nutritional strategies tested did not enhance productive performance under the conditions of this experiment. Although the PC diet reduced scouring and supported faster pathogen clearance, these benefits did not translate into measurable growth advantages relative to NC. It is possible that the ETEC challenge did not impair growth performance due to its subclinical nature, as growth trajectories remained stable following inoculation. Similar outcomes have been reported in other ETEC F18 challenge studies where diarrhea was induced without significant reductions in growth performance (McLamb et al. 2013; Sun et al. 2021). However, without a non-challenged control group, it is difficult to determine whether this reflects true resilience, effective recovery, or a lack of pathogenic impact.

Conclusions

In conclusion, the ETEC F18 challenge increased scouring post‑inoculation without affecting mortality or growth performance. Prior to inoculation, pigs were naturally exposed to rotavirus, which also contributed to scouring. The PC diet effectively reduced fecal consistency scores during both rotavirus and ETEC F18 challenge, demonstrating its ability to manage enteric disease under these conditions. In contrast, the LCP and FIB diets, as formulated in this study, did not provide benefits in controlling scouring or improving morphological and inflammatory parameters. These outcomes are specific to the context and conditions of this study. Because these strategies did not work independently, future research should explore multifactorial tools and combinations of approaches to improve resilience against enteric disease.