Lf was obtained from Tatua Co-operative Dairy Company, Limited (Tatua, New Zealand). To deplete iron, citric acid was added to the Lf solution and adjusted to pH 2.1 to obtain a final iron saturation of 7%. (NH₄)₂Fe(SO₄)₂ and HCO₃⁻ were added to the Lf solution to obtain a 100% iron-saturated Lf solution, as previously described (8).

Dextran sulphate sodium (DSS) was provided as a 36–50 kDa reagent by MP Biomedicals (Santa Ana, CA, USA). All reagents required for quantitative (q)PCR analysis were obtained from Promega Corporation (Madison, WI, USA).

Equal numbers of male ten-week-old BALB/c mice, with a mean weight of 22.0±2.0 g, were obtained from Vital River Laboratory Animal Technology Co., Ltd. (Beijing, China). Mice were housed in clean filter-top cages under standard conditions of 50±10% humidity and an equal 12-h dark/light cycle and fed with standard mouse chow for 7 days. Mice were randomly divided into the following 4 groups: i) Normal, normal diet in the absence of any special treatment; ii) model, permitted to drink 2.5% DSS for modeling on days 15–21, plus normal saline (100 μl/10 g) by gavage on days 1–28; iii) apo-Lf, 2.5% DSS for modeling on days 15–21, plus apo-Lf (100 mg/kg/d) by gavage on days 1–28; and iv) holo-Lf, 2.5% DSS for modeling on days 15–21, plus holo-Lf (100 mg/kg/d) by gavage on days 1–28. On day 28, all mice were sacrificed.

To reflect the general conditions of the mice, disease activity index (DAI) scores were determined by an investigator blinded to the protocol. The extent of loss in body weight, fecal character, fecal occult blood or hematochezia was assessed as previously described (9). On day 28 following treatment, the colon was removed in the region that spanned the colo-cecal junction to the anus. The length of the resected tissue was measured, rinsed with 5 ml 0.01 mol/l PBS (pH 7.4) to remove fecal remnants and dissected longitudinally at the mesenteric attachment. Next, 5 mm of the distal colon was removed and fixed for 48 h in PBS buffered with 10% formalin. The tissue was then processed for paraffin embedding, cut into 5-μm sections and stained with hematoxylin and eosin. Other sections of the colon were preserved in liquid nitrogen. The study was approved by the Ethics Committee of the College of Veterinary Medicine, Inner Mongolia Agricultural University.

The following parameters were evaluated by statistical methods: Weight loss, fecal character, fecal occult blood and hematochezia (Table I) (9).

Histological scores were given based on previously described criteria (10), with slight modifications. An extra score was added, which was denoted as 3.5 and represented observation of total crypt loss, but with evidence of retention to the surface epithelium. In addition, the lamina propria and sub-mucosa showed a serious inflammatory cellular infiltration in this group (Table II). Whole tissue specimens on slides were evaluated at magnification ×200 using light microscopy. Each microscopic field was evaluated and given a score. If more than one score was present in any given field of view, the scores were multiplied by the estimated percentage in the field and the sum of those values was calculated. At the conclusion of this evaluation process, the average score of the microscopic fields was recorded and analyzed statistically. The grading score was based on previously described criteria (10).

MPO activity was measured as an indicator of neutrophil accumulation in colonic mucosa as previously described (11).

All primers (Table III) were designed using Primer Premier 5.0 software (Premier Biosoft, Palo Alto, CA, USA) and the primers were manufactured by Shinegene Molecular Biotechnology Co., Ltd. (Shanghai, China). The annealing temperatures ranged between 58 and 53°C for 45 min and amplification was performed using 30 cycles for each gene of interest.

Statistical analysis was performed using SPSS version 13.0 (SPSS, Inc., Chicago, IL, USA) software program for Windows. Data are presented as the mean ± SD. Data sets were analyzed by one-way analysis of variance (ANOVA) and Fisher’s protected LSD post-hoc test. Those values with a significant difference were further analyzed by the Student-Newman-Keuls test. P<0.05 was considered to indicate a statistically significant difference.

Among the 4 groups of mice on day 1, no difference in body weight was observed (as determined by ANOVA). However, body weight increased gradually in the normal group. Mice in the model group showed weight loss over 1–14 days due to the gavage treatment received and a significant weight loss was observed from day 15 due to double stimulation by treatment with DSS and gavage. Treatment with holo-Lf and apo-Lf was observed to inhibit weight loss (Fig. 1).

Length of the colon in the model group was significantly decreased compared with the normal group (P<0.05; Fig. 2). These observations indicate that Lf was capable of preventing shortening of the colon.

Weight loss, fecal character, fecal occult blood and hematochezia were evaluated individually as previously described (12). The highest DAI score was observed in the model group (Fig. 3). Lf also promoted a beneficial effect in experimental colitis. Low DAI scores were observed in the Lf groups (Fig. 3).

Histological changes in mice of the 4 groups were evaluated individually as previously described (10). The highest score was observed in mice of the model group and a lower score was observed in the Lf treatment groups when compared with the model group; this was particularly true in the context of mice in the apo-Lf treatment group (Fig. 4A and B).

Maximal MPO activity was observed to be 2.78±0.39 U/g in the model group, which was significantly higher than that noted in the control group (0.93±0.13 U/g) (P<0.05). MPO activities also decreased following Lf treatment (Fig. 5).

IL-1β mRNA levels were observed to be lowest in the normal group and at the highest levels in the model group (Fig. 6). In addition, levels of IL-1β mRNA were observed to be lower in the Lf treatment groups than in the model group (2.27±0.79 vs. 6.63±1.57; P<0.05). TNF-α mRNA levels were lowest in the normal group and highest in the model group (Fig. 6). In addition, TNF-α mRNA levels were observed to be lower in the Lf treatment group compared with the model group (1.73±0.46 vs. 5.23± 1.05; P<0.05; Fig. 6).