Polylactose Exhibits Prebiotic Activity and Reduces Adiposity and Nonalcoholic Fatty Liver Disease in Rats Fed a High-Fat Diet

ABSTRACT

Background

Prebiotic dietary fibers change the intestinal microbiome favorably and provide a health benefit to the host.

Objectives

Polylactose is a novel fiber, synthesized by extrusion of lactose. We evaluated its prebiotic activity by determining its fermentability, effect on the microbiota, and effects on adiposity and liver lipids in a diet-induced obesity animal model.

Methods

Male Wistar rats (4–5 wk old) were fed normal-fat (NF, 25% fat energy) or high-fat (HF, 51% fat energy) diets containing different fibers (6% fiber of interest and 3% cellulose, by weight), including cellulose (NFC and HFC, negative and positive controls, respectively), polylactose (HFPL), lactose matched to residual lactose in the HFPL diet, and 2 established prebiotic fibers: polydextrose (HFPD) and fructooligosaccharide (HFFOS). After 10 wk of feeding, organs were harvested and cecal contents collected.

Results

HFPL animals had greater cecum weight (3 times greater than HFC) and lower cecal pH (∼1 pH unit lower than HFC) than all other groups, suggesting that polylactose is more fermentable than other prebiotic fibers (HFPD, HFFOS; P < 0.05). HFPL animals also had increased taxonomic abundance of the probiotic species Bifidobacterium in the cecum relative to all other groups (P < 0.05). Epididymal fat pad weight was significantly decreased in the HFPL group (29% decrease compared with HFC) compared with all other HF groups (P < 0.05) and did not differ from the NFC group. Liver lipids and cholesterol were reduced in HFPL animals when compared with HFC animals (P < 0.05).

Conclusions

Polylactose is a fermentable fiber that elicits a beneficial change in the gut microbiota as well as reducing adiposity in rats fed HF diets. These effects of polylactose were greater than those of 2 established prebiotics, fructooligosaccharide and polydextrose, suggesting that polylactose is a potent prebiotic.

Introduction

Obesity continues to be a major public health issue worldwide. The WHO's Health Statistics Report of 2018 estimated that obesity was occurring in 39% of the global population (1). The inflammation and insulin resistance that accompany increased adiposity can lead to many health issues, such as type 2 diabetes (T2D), cardiovascular disease, and nonalcoholic fatty liver disease (NAFLD). NAFLD is of particular interest, because its incidence has been estimated at 75%–90% in adults who are overweight or obese, occurring in 25% of adults worldwide (2, 3).

NAFLD is defined as the accumulation of excess lipids in the liver (2–4). Simple fatty liver (nonalcoholic fatty liver) is the most common form of NAFLD and occurs with little to no inflammation or scarring. Nonalcoholic steatohepatitis (NASH) is the more severe form of NAFLD and occurs with irreversible scarring, or cirrhosis, and fibrosis (2–4). Continued liver lipid accumulation, inflammation, cirrhosis, and fibrosis in NASH lead to chronic liver disease and, eventually, liver failure. Currently, no effective pharmacological treatment exists for NAFLD. However, weight loss, as a result of dietary changes, increased exercise, or bariatric surgery, has been shown to reverse the early stages of NAFLD (2–4). Given the difficulty of achieving weight loss by diet or exercise, and the invasive nature of bariatric surgery, a dietary approach to reverse NAFLD that does not require weight loss would be highly desirable.

Little is known about the molecular mechanisms that lead to NAFLD (5, 6). It is clear that obesity, hypertriglyceridemia, insulin resistance, and T2D are associated with the development of NAFLD (5, 6). A number of factors can result in the accumulation of lipids in the liver, including increased uptake of triacylglycerols from dietary fat and fatty acids from lipolysis of adipose tissue, increased de novo hepatic lipid synthesis, and a reduction in lipid oxidation and mobilization (4, 7, 8). Many of these factors are dysregulated in obesity and hyperinsulinemia. During periods of hyperinsulinemia, many tissues, including the liver, become insulin resistant, leading to hepatic lipid accumulation. This is largely attributable to the upregulation of lipogenic transcription factors, such as sterol regulatory element binding protein-1c (SREBP1c) (4, 7, 8), which results in increased lipid synthesis in the liver. Further, there is a reduction in lipid oxidation and mobilization. With continued lipid accumulation, low-grade inflammation develops that can eventually lead to NASH, fibrosis, and cirrhosis.

Recent studies have shown that there is a relation between the gut microbiota, NAFLD, and obesity (9, 10). Increased large intestinal abundance of the phylum Firmicutes and a reduction in Bacteroidetes [or increased ratio of Firmicutes to Bacteroidetes (F:B)] has been correlated with increased adiposity in humans (11, 12) and animals (13), although its usefulness as a biomarker of obesity has been challenged (14). Similarly, human studies have shown correlations between lean subjects and increased abundances of Bifidobacterium (15) and/or Lactobacillus (16, 17). Thus, changes in the abundance of specific taxa of large intestinal bacteria appear to have an important influence on host health.

Increased production of SCFAs may mediate these host health benefits. Large intestinal bacteria ferment dietary fibers and other nondigestible carbohydrates to produce SCFAs, which can be taken up into the circulation and utilized by host tissues (18, 19). SCFA production differs based on the type of carbohydrate presented to the large intestinal microbiome (20) as well as an individual's particular microbiome (21). SCFAs, primarily composed of butyrate, propionate, and acetate, are utilized differently after their absorption in the large intestine. Butyrate is primarily utilized for energy in colonocytes. The remaining SCFAs, including unutilized butyrate, are transported through the portal circulation, with propionate and butyrate taken up by hepatocytes (22), where propionate is used as a gluconeogenic precursor (23, 24). Acetate is primarily taken up by the adipose and skeletal muscle, where it is utilized as a substrate for cholesterol, fatty acid, glutamate, and glutamine synthesis (24, 22). Understanding how a nondigestible carbohydrate changes the microbiome, subsequently affecting SCFA production and altering hepatic lipid metabolism, will likely be important for developing effective dietary approaches to preventing or reversing NAFLD.

Prebiotics are a potential dietary intervention that can alter the gut microbiome by providing a substrate for SCFA production. Prebiotics are defined as indigestible substrates that are fermented in the large intestine, elicit a beneficial change in the gut microbiome, and provide a health benefit to the host (25). The objective of the present study was to investigate the prebiotic potential of a novel dietary fiber, polylactose. Polylactose is produced by polymerizing lactose by reactive extrusion, using heat and citric acid as catalysts (26). Purified polylactose analyzes as a dietary fiber (26). It contains oligomers with a degree of polymerization (DP) of 3–11 monomeric sugars, with the bulk of the oligomers falling in the range of 3–5 DP. Because several established prebiotic dietary fibers are oligomers, we hypothesized that this polylactose preparation might also perform as a prebiotic dietary fiber, as defined above. Therefore, we assessed its fermentability, its effect on the gut microbiota and on cecal SCFA, and whether it produced host health benefits, focusing on its ability to reduce liver lipids and adiposity.

Methods

Preparation of polylactose

Polylactose was produced on a twin-screw extruder as previously described (26, 27). It was subsequently solidified, ground, rehydrated in distilled water, and purified on a mixed-bed carbon and ion exchange column. The filtration column consisted of layers of diatomaceous earth, a 1:1 mixture of Ambersep 200 (H+) and Amberlite FPA 53 (OH−), and activated carbon (Norit GAC 1240 PLUS, Cabot Norit Americas, Inc.). The eluate and 1 L of rinse were collected and pooled before spray drying. This produced a white, almost tasteless, powder. Polylactose fiber, lactose, glucose, and citric acid were measured in the final product using commercial kits (Integrated Total Dietary Fiber Assay Kit, Lactose/Sucrose/D-Glucose Assay Kit, and Citric Acid Assay Kit, respectively; Megazyme). The final product contained ∼50% soluble fiber, 22% free lactose, residual glucose and citric acid, and other unidentified materials.

Animals and treatments

Seventy-two male Wistar rats (initial body weight 75–99 g, 4–5 wk of age) were purchased from Envigo Laboratories. Animals were singly housed in hanging wire cages and kept on a 12-h light/dark cycle. Food and water were provided to the animals ad libitum for the duration of the study. All animal use procedures were approved by the University of Minnesota Animal Care and Use Committee.

All animals were adapted to a purified rodent diet (modified AIN-93G diet) for 2 wk before the start of experimental diets (Supplemental Table 1). At the start of week 3, the animals were randomly assigned into 6 groups (12 rats/group), which were fed 1 of 6 experimental diets for 10 wk. The 6 experimental diets were as follows: normal-fat cellulose (NFC; modified AIN-93G diet), high-fat cellulose (HFC; 9% cellulose), high-fat polylactose (HFPL; 6% polylactose fiber, 3% cellulose), high-fat matched free lactose (HFMFL; 2.6% lactose, matched to the amount of free lactose in the HFPL diet, 6.4% cellulose), high-fat polydextrose (HFPD; 6% PD, 3% cellulose), and high-fat fructooligosaccharide (HFFOS; 6% FOS, 3% cellulose). The prebiotic dose was chosen as one that is potentially attainable in the human diet. Supplemental Table 1 shows the composition of the diets. Food intake and body weight were measured weekly. A 3-d fecal collection was conducted at week 8 of the experimental diets. Feces were lyophilized and stored at −20°C until further analysis.

After 10 wk feeding the experimental diets, food-deprived animals were anesthetized with isoflurane and blood collected by cardiac puncture. Blood was centrifuged at 3000 × g for 20 min at 4°C and plasma aliquots were stored at −80°C until use. Ceca were harvested, cecal pH was measured with a spear-tip electrode, and contents were expressed into sterile 15-mL conical tubes, frozen on dry ice, and stored at −80°C. Empty ceca were rinsed, weighed, and discarded. One epididymal fat pad from each animal was harvested, rinsed, weighed, frozen on dry ice, and stored at −80°C. Livers were excised, rinsed, weighed, frozen on dry ice, and stored at −80°C until further analysis.

Blood glucose control and gluconeogenesis potential

Tolerance testing was performed during the final 2 wk of feeding the experimental diets after overnight (8-h) deprivation of food. Insulin-tolerance tests (ITTs; 1 IU insulin/kg body weight) and pyruvate-tolerance tests (60 mg sodium pyruvate/kg body weight) were conducted during week 9, with each being administered intraperitoneally. Oral-glucose-tolerance tests (OGTTs) (0.5 g glucose/kg body weight) were conducted during week 10. For all tolerance testing, blood glucose was measured using blood glucose test strips and a glucometer (AlphaTrak, Zoetis Inc.) at 0, 15, 30, 60, 90, and 120 min. AUCs were calculated using the trapezoidal method.

Liver lipids and cholesterol

Approximately 1 g of liver from each animal was homogenized and lipids were extracted using chloroform:methanol (2:1) (28). Extracted lipids were dried under N₂ gas and quantified gravimetrically. Lipid samples were reconstituted with 10 mL chloroform:methanol (2:1) and cholesterol was measured enzymatically using a previously described method (29).

Circulating adipokines and lipids.

Circulating leptin and adiponectin were quantified by ELISA (Millipore Sigma EZRL-83K and abcam ab10874, respectively) using plasma samples from the time of harvest. Plasma triacylglycerols were measured spectrophotometrically using a commercial kit (T7532120, Pointe Scientific Inc.).

Microbiome sequencing

Cecal contents were thawed on ice and mixed before use. DNA was extracted from ∼180–220 mg of contents using the Qiagen Stool Extraction kit and DNA eluates were quantified spectrophotometrically. DNA eluates were then submitted to the University of Minnesota Genomics Center for sequencing of the 16S ribosomal subunit with the V5/V6 region amplicon on an Illumina MiSeq using 2 × 300-bp paired end reads. Data were analyzed using DADA2 and phyloSeq in R (The R Foundation).

SCFAs and bile acids

SCFAs were extracted from cecal contents and measured by GC (30–32). Bile acids were extracted from fecal samples, partially purified (33), and total bile acids quantified enzymatically (34).

Statistical analysis

All data except microbiota data were analyzed by 1-factor ANOVA followed by Duncan's multiple range test using SAS version 9.4 (SAS Institute). Homogeneity of variances was assessed using Levene's test. Any variable with unequal variances between the groups was log transformed before statistical analysis. Duncan's multiple range test was used to conduct pairwise comparisons among all groups, reducing the chance of making type II errors. To examine for trends in differences, Student's t tests were used to compare groups of interest. Pearson correlation analyses were conducted in SAS for SCFAs and adiposity measures. AUC was calculated in R and analyzed by ANOVA and Student's t test in SAS. Microbiota data were analyzed in R (35), using the DADA2 and phyloseq packages for β-diversity, taxonomic abundance, and statistical analyses.

Results

Polylactose reduced adiposity with no significant changes in energy intake or body weight

Daily mean energy intake was significantly greater in HFC than in HFPL (P < 0.05). However, the energy intakes for NFC, HFMFL, HFPD, and HFFOS did not differ from either HFC or HFPL (Table 1). Final body weights did not differ between the groups (P = 0.31). The epididymal fat pads of HFPL and NFC weighed less than those of HFC (P = 0.034) (Table 1). Further, epididymal fat pad weight of HFPL was significantly less than those of HFMFL, HFPD, and HFFOS (P < 0.05). Plasma leptin concentration in HFPL and NFC was less than in HFMFL. Plasma leptin concentration, which has been shown to correlate with total body fat (36, 37), was highly correlated with epididymal fat pad weight (r = 0.69, P < 0.0001). Plasma adiponectin concentration did not differ between the groups (P = 0.88).

TABLE 1.

Energy intake, final body weight, cecal pH, and tissue weights in rats fed normal-fat diet or high-fat diet alone or with matched free lactose, polylactose, polydextrose, or fructooligosaccharide for 10 wk¹

Variable	NFC	HFC	HFPL	HFMFL	HFPD	HFFOS
Energy intake, kJ/d	377.4 ± 14.3ab	410.1 ± 12.9a	363.7 ± 15.9b	371.6 ± 12.8ab	382.2 ± 13.7ab	385.6 ± 14.0ab
Initial body weight, g	197.4 ± 3.9	196.2 ± 3.9	192.4 ± 3.6	189.7 ± 3.9	188.8 ± 5.0	195.5 ± 5.6
Final body weight, g	490.0 ± 11.8	534.8 ± 14.8	496.5 ± 12.5	517.8 ± 17.5	514.9 ± 19.6	523.0 ± 13.4
Empty cecum weight, g/100 g FBW	0.119 ± 0.003c	0.109 ± 0.007c	0.324 ± 0.022a	0.104 ± 0.004c	0.156 ± 0.008b	0.152 ± 0.009b
Cecal pH	6.91 ± 0.66a	6.98 ± 0.06a	6.05 ± 0.08c	6.86 ± 0.06a	6.54 ± 0.08b	6.98 ± 0.06a
Epididymal fat pad weight, g/100 g FBW	1.076 ± 0.068bc	1.322 ± 0.064a	0.943 ± 0.055c	1.335 ± 0.098a	1.301 ± 0.081a	1.220 ± 0.072ab
Liver weight, g/100 g FBW	3.060 ± 0.119c	3.447 ± 0.162ab	3.195 ± 0.104bc	3.350 ± 0.108abc	3.126 ± 0.082bc	3.541 ± 0.079a

¹ n = 12. Values are means ± SEMs. Means within a row without a common letter are statistically different, P < 0.05. FBW, final body weight; HFC, high-fat cellulose; HFFOS, high-fat fructooligosaccharide; HFMFL, high-fat matched free lactose; HFPD, high-fat polydextrose; HFPL, high-fat polylactose; NFC, normal-fat cellulose.

Polylactose was vigorously fermented in the large intestine

Most fermentation by the gut microbiota in the rat occurs in the first part of the large intestine, the cecum. Fermentation lowers the pH of the cecal contents and increases the mass of the cecal tissue (38). Cecal pH was lower in HFPL and HFPD than in all other groups, but pH was lower in HFPL than in HFPD (P < 0.0001) (Table 1). Empty cecum weight was dramatically greater in HFPL than in all other groups (P < 0.0001). HFPD and HFFOS also had a significantly greater empty cecum weight than HFC and NFC. Total cecal SCFA content was increased in HFPL compared with all other groups (P < 0.0001). This was due to greater amounts of acetate and propionate in the cecal contents of HFPL than in all other groups (P < 0.0001) (Figure 1). Total butyrate content in the cecal contents did not differ between the groups. Total cecal contents SCFA was strongly correlated with empty cecal weight (r² = 0.788, P < 0.0001, for all animals). Total cecal acetate and propionate were also individually highly correlated with empty cecum weight (r² = 0.792 and 0.811, respectively, P < 0.0001), whereas total cecal contents butyrate showed no correlation to empty cecum weight (r² = −0.105, P = 0.40). Epididymal fat pad weight did not correlate with total cecal contents SCFA (r = −0.17, P = 0.18) nor with any individual SCFA (r = −0.18, P = 0.13 for acetate; r = −0.21, P = 0.09 for propionate; r = 0.13, P = 0.29 for butyrate). Likewise, plasma leptin concentration did not correlate with total cecal contents SCFA (r = −0.008, P = 0.95) nor with any individual SCFA (r = −0.02, P = 0.87 for acetate; r = −0.004, P = 0.97 for propionate; r = −0.02, P = 0.90 for butyrate).

FIGURE 1.

Content of individual SCFAs in cecal contents in rats fed normal-fat diet or high-fat diet alone or with matched free lactose, polylactose, polydextrose, or fructooligosaccharide for 10 wk. Values represent mean ± SEM, n = 12. Labeled means without a common letter are significantly different, P < 0.05. HFC, high-fat cellulose; HFFOS, high-fat fructooligosaccharide; HFMFL, high-fat matched free lactose; HFPD, high-fat polydextrose; HFPL, high-fat polylactose; NFC, normal-fat cellulose.

Polylactose reduced fatty liver

Liver weights were greater in HFC and HFFOS than in NFC (P < 0.05) (Table 1). HFPL had less liver lipids than HFC (P = 0.0035) and trended less than HFMFL (P = 0.075), HFPD (P = 0.075), and HFFOS (P = 0.087) (Figure 2A). Liver cholesterol was also decreased in HFPL compared with HFC (P = 0.0028) and was similar to NFC (Figure 2B). Plasma triacylglycerol concentration was lower in all high-fat (HF) diet groups than in NFC (P < 0.0001) (Table 2). Among the HF diet groups, plasma triacylglycerol concentration was lower in HFFOS than in HFC (P < 0.0001). Daily fecal total bile acid excretion was not different between the HF-fed groups, but excretion was less in HFMFL and HFPD than in NFC (P = 0.016 and P = 0.004, respectively) (Table 2).

FIGURE 2.

Liver lipids (A) and liver cholesterol (B) in rats fed normal-fat diet or high-fat diet alone or with matched free lactose, polylactose, polydextrose, or fructooligosaccharide for 10 wk. Values represent mean ± SEM, n = 12. Labeled means without a common letter are significantly different, P < 0.05. HFC, high-fat cellulose; HFFOS, high-fat fructooligosaccharide; HFMFL, high-fat matched free lactose; HFPD, high-fat polydextrose; HFPL, high-fat polylactose; NFC, normal-fat cellulose.

TABLE 2.

Plasma glucose, hormone, and triacylglycerol concentrations, cecal total SCFA content, and fecal bile acid excretion in rats fed normal-fat diet or high-fat diet alone or with matched free lactose, polylactose, polydextrose, or fructooligosaccharide for 10 wk¹

Variable	NFC	HFC	HFPL	HFMFL	HFPD	HFFOS
Fasting plasma glucose, mg/dL	115 ± 4.2	116 ± 5.0	115 ± 7.3	118 ± 4.0	125 ± 4.2	128 ± 3.8
Fasting plasma insulin, ng/mL	4.99 ± 0.57b	6.44 ± 0.64ab	6.58 ± 1.12ab	7.03 ± 0.67ab	8.45 ± 1.17a	5.62 ± 0.50b
Plasma leptin, ng/mL	9.18 ± 1.26b	16.8 ± 1.94a	12.7 ± 1.92ab	20.5 ± 3.13a	15.1 ± 2.34ab	14.2 ± 1.97ab
Plasma adiponectin, µg/mL	19.4 ± 3.56	23.4 ± 3.52	23.3 ± 3.39	25.2 ± 3.85	20.0 ± 3.47	22.4 ± 3.68
Plasma triacylglycerols, mmol/L	1.77 ± 0.14a	1.24 ± 0.12b	0.91 ± 0.14bc	1.02 ± 0.09bc	1.01 ± 0.09bc	0.87 ± 0.09c
Total SCFAs, µmol/cecum	95.3 ± 5.7c	89.4 ± 8.6c	286 ± 30.7a	88.5 ± 9.3c	149 ± 12.4b	131 ± 16.6bc
Fecal bile acid excretion, µmol/d	39.1 ± 2.5a	32.6 ± 3.0ab	26.6 ± 4.3ab	27.1 ± 1.9b	24.6 ± 2.3b	33.4 ± 2.4ab

¹ n = 12, except fecal bile acid excretion, where n = 7–12. Values are means ± SEMs. Means within a row without a common letter are statistically different, P < 0.05. HFC, high-fat cellulose; HFFOS, high-fat fructooligosaccharide; HFMFL, high-fat matched free lactose; HFPD, high-fat polydextrose; HFPL, high-fat polylactose; NFC, normal-fat cellulose.

Polylactose-fed animals trended toward improved glycemic control

There were no significant differences between the groups in fasting plasma glucose (Table 2). Fasting plasma insulin trended toward being greater in HFC than in NFC (P = 0.0755). Plasma insulin was significantly greater in HFPD than in HFFOS and NFC, but HFC, HFPL, and HFMFL did not differ from HFPD or HFFOS (P < 0.05). The ITT showed no differences in insulin sensitivity, as measured by the glucose response (P = 0.9431, data not shown). However, HFPL showed a trend toward a smaller AUC for the OGTT than HFC (P = 0.056) when compared by Student's t test (Figure 3).

FIGURE 3.

OGTT in rats fed normal-fat diet or high-fat diet alone or with matched free lactose, polylactose, polydextrose, or fructooligosaccharide for 10 wk. (A) Mean OGTT curve by diet group. (B) AUC for the OGTT. Values represent mean ± SEM, n = 12. Labeled means without a common letter are significantly different, P < 0.05. *HFC compared with HFPL, P = 0.056 by Student's t test. HFC, high-fat cellulose; HFFOS, high-fat fructooligosaccharide; HFMFL, high-fat matched free lactose; HFPD, high-fat polydextrose; HFPL, high-fat polylactose; NFC, normal-fat cellulose; OGTT, oral-glucose-tolerance test.

Polylactose elicited a beneficial change in the gut microbiota

β-Diversity analysis of the gut microbiota revealed 3 distinct clusters: control groups (NFC, HFC, and HFMFL), commercial prebiotics (HFPD, HFFOS), and polylactose (HFPL). The HFPL cluster was most dissimilar from the control groups and in line with, yet separated from, the prebiotic cluster. This indicates greater change in the gut microbiota when compared with the prebiotic groups, as was shown by the Bray–Curtis dissimilarity measure of β-diversity as visualized in Figure 4. Statistical analysis of the Bray–Curtis dissimilarity matrix revealed that 33% of the microbial variation between animals was due to diet (R² = 0.33, P < 0.001). Taxonomic analysis (Figure 5) revealed that HFPL contained greater Bifidobacterium species and Gammaproteobacteria than all other HF groups (P < 0.0001). Lactobacillus species were also increased in HFPL compared with HFC (P < 0.04). Akkermansia muciniphila was increased in HFPL when compared with all other HF groups (P < 0.0001), and increased in HFFOS when compared with the control groups (NFC, HFC, and HFMFL; P < 0.0001). Lastly, F:B was reduced in HFPL compared with all other groups (P < 0.0001) (Figure 5). F:B correlated significantly but inversely with cecal total SCFA content (r = −0.377, P = 0.0021) and with cecal acetate and propionate content (r = −0.399, P = 0.0012 and r = −0.383, P = 0.0018, respectively) but not with cecal butyrate content (r = 0.176, P = 0.164).

FIGURE 4.

Bray–Curtis PCoA for β-diversity of bacteria in cecal contents of rats fed normal-fat diet or high-fat diet alone or with matched free lactose, polylactose, polydextrose, or fructooligosaccharide for 10 wk. Each point represents an individual animal, coded by diet. There is a statistically significant difference between the diet groups, by analysis of dissimilarities (adonis) (R² = 0.33, P < 0.001). HFC, high-fat cellulose; HFFOS, high-fat fructooligosaccharide; HFMFL, high-fat matched free lactose; HFPD, high-fat polydextrose; HFPL, high-fat polylactose; NFC, normal-fat cellulose; PCoA, principal coordinate analysis.

Figure 5.

Select taxonomy in cecal contents of rats fed NF diet or HF diet alone or with MFL, PL, PD, or FOS for 10 wk. (A) Bifidobacterium (genus), (B) Lactobacillus (genus), (C) Gammaproteobacteria (class), (D) Akkermansia muciniphila, (E) Firmicutes:Bacteroidetes ratio (phyla). Abundance represented as percent (proportion of taxa sequence reads/total reads per sample). Values represent mean ± SEM, n = 11–12. Labeled means without a common letter are significantly different, P < 0.05. NFC, normal fat cellulose control; HFC, high fat cellulose control; HFPL, high fat polylactose; HFMFL, high fat matched free lactose; HFPD, high fat polydextrose; HFFOS, high fat fructooligosaccharide.

Discussion

Here we present the first study, to our knowledge, of the prebiotic activity of polylactose. Prebiotic dietary fibers are fibers that produce beneficial changes in the gut microbiota and that provide a health benefit to the host (39). Polylactose, a novel, synthetic dietary fiber, fulfills the criteria for a prebiotic, in that it elicited a beneficial change in the gut microbiota in a diet-induced obesity animal model and provided several health benefits, specifically reducing adiposity and fatty liver, and trending toward improvement of glycemic control. In contrast, PD- and FOS-fed animals, who presented only a modest change in the microbiota, showed no change in adiposity, liver lipids, or improvement in glycemic control at the dietary concentration of 6% used in the present study (P > 0.05).

Feeding studies of PD and FOS have shown a change in the gut microbiota (40–42) or greater fermentation in the cecum (40, 43). This is consistent with our findings that both PD and FOS were fermented, as indicated by a reduction in cecal pH, an increase in empty cecum weight, and a change in the gut microbiota. However, these differences occurred to a much lesser degree than with polylactose. This suggests that polylactose is a better substrate for certain taxa in the gut, such as Bifidobacterium and Lactobacillus species, leading to greater fermentation. This is further corroborated by the greater content of total and individual SCFAs in the cecal contents of polylactose-fed animals, specifically an increase in acetate and propionate, than in the PD and FOS groups. Other studies have shown a similar increase in total SCFAs and propionate with PD and FOS feeding compared with an HFC control group (43, 44). An HF diet has been reported to suppress SCFA production, relative to a normal-fat (NF) diet, but FOS or other highly fermentable fibers fed as part of the HF diet restored SCFAs to concentrations found in the NF diet group (43, 44). However, in the present study, cecal contents SCFA did not differ between the NF and HF groups.

Rodent studies feeding PD or FOS at dietary concentrations similar to that used in the present study, between 2.5% and 10%, have found variable effects on physiological endpoints. Although adiposity was sometimes reduced (40, 41, 44, 45), final body weight typically was reduced as well, suggesting that the prebiotic fibers reduced body size but did not necessarily change body composition. In the present study, we found that although all groups had statistically similar final body weights, rats fed polylactose had greatly reduced epididymal fat pad weights compared with all other HF groups, indicating a change in body composition. Leptin is an adipokine whose plasma concentration correlates directly with adiposity (36, 37), which was confirmed in the present study. Although plasma leptin was significantly lower in the HFPL group than in 1 of the HF control groups (HFMFL), it did not differ significantly from the HFC group. Oral administration of propionate to mice has been found to increase plasma leptin concentration (46) and therefore could have attenuated the reduction in plasma leptin expected from the lower degree of adiposity in the HFPL group. However, in the present study, no correlation was found between plasma leptin and either cecal contents propionate or total SCFA.

A common feature of the diet-induced obesity animal model is impaired glycemic control. The AUC of the OGTT of the polylactose-fed animals was similar to that of the NF control group (NFC) and showed a trend for a reduction compared with the HFC group (P = 0.056 by Student's t test). Neither PD nor FOS at 6% of the diet reduced fasting plasma glucose or the AUC of the OGTT. This suggests that polylactose prevented the impairment of glycemic control experienced by rats fed an HF diet, whereas the commercial prebiotics PD and FOS did not. Consistent with our finding, FOS fed at 5% on an HF background did not reduce OGTT AUC, fasting glucose, or fasting insulin (P > 0.05) (43). Similarly, no reduction in fasting insulin and fasting glucose in rats fed 10% FOS on an HF background was found by Cluny et al. (40). In contrast, 21% FOS was reported to reduce the OGTT AUC and ITT AUC when compared with an HF control group (42). This suggests that a high dietary concentration of FOS is needed to improve glycemic control in animals fed HF diets, a dietary concentration that is likely unattainable in a human diet.

The HF diet–induced obesity model also presents with increased hepatic lipid accumulation (47, 48). Fatty liver is the first step in development of NAFLD, a condition with few effective treatments (2–4, 49). Simple fatty liver can be reversed by dietary- or exercise-based interventions (2–4, 47). Presently, there is little evidence that commercial prebiotics reverse fatty liver, particularly at attainable dietary amounts (42, 44, 45). Although these studies demonstrated reduced liver weights, only Nakamura et al. (44) found reduced liver lipids. In contrast, when FOS was fed at a high dose (21%), liver weight and total liver cholesterol were reduced but liver lipids were not (45). In the present study, FOS-fed animals did not exhibit a reduction in liver weight, liver lipids, or liver cholesterol concentration. However, polylactose feeding led to lower liver lipid accumulation. Liver cholesterol concentration was lower in both the polylactose and PD groups than in the HFC group.

Small differences in energy intake over time can lead to differences in body weight and adiposity. In the present study, the HFC group had a significantly greater energy intake (P < 0.05) than the HFPL group as well as a heavier fat pad weight. However, the HFMFL and the HFPL groups had essentially identical energy intakes but fat pad weight was significantly less in the HFPL group than in the HFMFL group. Thus, changes in the large intestinal microbiome are a more likely explanation for the decrease in adiposity and liver lipids and modest improvement in glycemic control found in the group fed polylactose. This group had greatly increased taxonomic abundance of Bifidobacterium and Lactobacillus species and Gammaproteobacteria in the cecal contents, all of which are fermenters of galactose (50, 51), 1 of the 2 monosaccharides composing lactose, which is used to synthesize polylactose. Bifidobacterium and Lactobacillus are established probiotic organisms and their increased abundance has been correlated with reduced fatty liver (9, 10, 52–55). Interestingly, fecal Gammaproteobacteria abundance increases after bariatric surgery (56), which is known to reduce fatty liver (57), yet Gammaproteobacteria abundance was slightly but significantly greater in obese children with NAFLD than in healthy normal-weight children (58). Thus, the influence of Gammaproteobacteria on fatty liver warrants further investigation.

Galactose fermenters produce large amounts of acetate and propionate, typically the 2 most abundant SCFAs (59–61), which were in much greater abundance in the cecal contents of the polylactose group relative to all other groups. After uptake into the portal circulation, acetate and propionate are used by the liver for energy metabolism and also regulate lipid synthesis, largely via activation (phosphorylation) of AMP kinase by acetate, increasing lipid oxidation in the liver (24). In addition, polylactose-, PD-, and FOS-fed animals all exhibited an increase in A. muciniphila when compared with the nonprebiotic control animals. A. muciniphila is suggested as a beneficial gut bacterium (62) and has been associated with reduced adiposity and improved glycemic control in animal and human studies (9, 63–65). Although A. muciniphila was increased in abundance in all 3 prebiotic groups, polylactose-fed animals had a statistically significantly greater abundance than the other prebiotic groups.

The mechanism by which polylactose may reduce fatty liver and adiposity is uncertain. Although viscous dietary fibers are effective at reducing fatty liver and adiposity in animal models, independently of their fermentability (66, 67), polylactose consumption does not increase small intestinal viscosity (data not shown). F:B shows a modest correlation with BMI in humans (68) and is positively associated with increased adiposity in human and animal studies (68, 69). Only the polylactose group reduced F:B relative to the NFC and HFC groups, consistent with this association. How changes in F:B might mediate changes in fatty liver and adiposity is unknown. The correlation in the present study of F:B with the cecal content of acetate and propionate suggests that SCFAs may mediate the effect of F:B on adiposity and fatty liver. Feeding acetate, propionate, or butyrate to mice as part of an HF diet has been found to attenuate the accumulation of adipose tissue relative to the HF diet alone (70). However, our results suggest there is a threshold of SCFA production needed to reduce adiposity and fatty liver, because only polylactose, which produced much greater amounts of cecal SCFA than PD or FOS, reduced adiposity and fatty liver.

In conclusion, our findings demonstrate that polylactose, a novel dietary fiber, behaves as a prebiotic in rats fed an HF diet. That is, polylactose is a dietary fiber that changes the large intestinal microbiota in a beneficial manner, provides health benefits to the host, specifically reducing adiposity and liver lipids, and thus fulfills the definition of a prebiotic dietary fiber (39). Further, polylactose produced these effects at a dietary concentration at which other prebiotic dietary fibers did not provide health benefits. Additional studies of the potential benefits of polylactose consumption seem warranted.

Notes

Supported by the Midwest Dairy Association (to TCS). The funder had no role in the design, implementation, analysis, and interpretation of the data.

Author disclosures: The authors report no conflicts of interest.

Supplemental Table 1 is available from the “Supplementary data” link in the online posting of the article and from the same link in the online table of contents at https://academic.oup.com/jn/.

Abbreviations used: DP, degree of polymerization; F:B, Firmicutes to Bacteroidetes ratio; FOS, fructooligosaccharide; HF, high fat; HFC, high-fat cellulose; HFFOS, high-fat fructooligosaccharide; HFMFL, high-fat matched free lactose; HFPD, high-fat polydextrose; HFPL, high-fat polylactose; ITT, insulin-tolerance test; NAFLD, nonalcoholic fatty liver disease; NASH, nonalcoholic steatohepatitis; NF, normal fat; NFC, normal-fat cellulose; OGTT, oral-glucose-tolerance test; PD, polydextrose; T2D, type 2 diabetes.