An oat fiber intervention for reducing PFAS body burden: A pilot study in male C57Bl/6J mice

Abstract

Perfluoroalkyl substances (PFAS) are a major public health concern, in part because several PFAS have elimination half-lives on the order of years and are associated with adverse health outcomes. While PFAS can be transported into bile, their efficient reuptake by intestinal transporter proteins results in minimal fecal elimination. Here, we tested the hypothesis that consumption of oat β-glucan, a dietary supplement known to disrupt the enterohepatic recirculation of bile acids, will reduce PFAS body burdens. Male C57Bl/6J mice were fed diets based on the “What we eat in America” analysis that were supplemented with inulin or oat β-glucan and exposed via drinking water to a seven PFAS mixture (PFHpA, PFOA, PFNA, Nafion Byproduct-2, PFHxS and PFOS) for 6 weeks. One cohort of mice was euthanized at the end of the exposure, and one cohort continued on the experimental diets for 4 more weeks without additional PFAS exposure. The β-glucan fed mice drank significantly more water than the inulin fed mice, resulting in a significantly higher dose of PFAS. Relative to overall exposure, we observed lower serum concentration trends (p<0.1) in β-glucan fed mice for PFHpA, PFOA and PFOS. Additionally, β-glucan fed mice had lower adipose:body weight ratios and liver and jejunum triglyceride concentrations. Hepatic mRNA expression of Cyp4a10, Cyp2b10 and Cyp3a11 were elevated in PFAS exposed mice, with only the expression of Cyp3a11 decreasing following depuration. This pilot study generates support for the hypothesis that oat β-glucan supplementation can reduce PFAS body burdens and stimulate healthful effects on lipid homeostasis.

INTRODUCTION

Perfluoroalkyl substances (PFAS) are ubiquitous and persistent environmental contaminants. Various definitions classify the PFAS chemical class differently (Hammel et al., 2022). By “legacy” PFAS, this refers to the longer-chain perfluoroalkyl acids and their precursors historically used for commercial and industrial uses, including aqueous film forming foam (AFFF) to quell fuel fires at military bases and airports, processing aids in manufacturing, consumer products, and food contact materials. These include perfluorooctanoic acid (PFOA), perfluorononanoic acid (PFNA), perfluorodecanoic acid (PFDA), perfluorooctane sulfonic acid (PFOS), and perfluorohexane sulfonic acid (PFHxS), which have been found to be associated with a number of adverse human health effects, including increased serum low density lipoprotein cholesterol (LDL-C), reduced antibody responses to vaccination, altered liver serum biomarkers, kidney and testicular cancer, pregnancy-induced hypertension, and decreased birth weight (ATSDR, 2021). Historical and ongoing exposure to PFOA and PFOS via drinking water, and to a lesser extent via food and air, has resulted in nearly ubiquitous exposure to people living in the United States (CDC, 2019).

At present, there are no pharmacological interventions and there are only limited non-pharmacological intervention strategies to reduce PFAS body burden. Current strategies to reduce internal levels of PFAS focus primarily on reducing exposures, such as filtration of contaminated drinking water (NASEM, 2021). Few strategies are available to intentionally reduce body burdens of PFAS via enhancing elimination. Blood loss, blood donation, and cholestyramine, an anionic resin used to treat high cholesterol, reduce PFAS body burdens (Ducatman et al., 2021; Gasiorowski et al., 2022; S. Genuis et al., 2010, 2013; S. J. Genuis et al., 2014; Johnson et al., 1984; Lorber et al., 2015; Møller et al., 2024; Taylor et al., 2014; Wong et al., 2014). Recent epidemiological studies investigating dietary patterns and PFAS body burdens show that the consumption of fruits or vegetables and fiber-rich diets are associated with lower serum PFAS concentrations in both adults and children (Dzierlenga et al., 2021; Lin et al., 2020; Seshasayee et al., 2021; Sultan et al., 2023; Y. Wang et al., 2024). Similarly, life-style changes designed to reduce serum cholesterol via increasing dietary fiber were shown to be associated with reductions in serum PFAS (Morgan et al., 2023). We also recently reported that consumption of an oat β-glucan supplement with meals for 4 weeks was associated with a decrease in serum concentrations of legacy PFAS (Schlezinger et al., 2024).

Perfluoroakyl carboxylic acids and sulfonic acids are transported from the gut contents into the body by cells that line the intestines (Kimura et al., 2020; Zhao et al., 2017), leading to efficient initial absorption following ingestion of PFAS (Pizzurro et al., 2019) and reabsorption after excretion of PFAS in bile (Cao et al., 2022). In humans, biliary clearance rates of PFOA and PFOS are significantly higher than urinary clearance rates (e.g., PFOA: urinary clearance = 0.03 ml/kg/day vs. biliary clearance = 1.06 ml/kg/day; PFOS: urinary clearance = 0.02 ml/kg/day vs. biliary clearance = 2.98 ml/kg/day)(Fujii et al., 2015; Harada et al., 2007). However, the reabsorption rates following biliary clearance of PFOA and PFOS also are high (0.89 and 0.97, respectively) (Harada et al., 2007), thus only a small proportion of PFAS excreted in the bile leaves the body in feces (Fujii et al., 2015). While PFAS dissolved in the fluid portion of the blood can enter the glomerular filtrate in the kidney, they can also be reabsorbed from the pre-urine via active transporter proteins in renal proximal tubule cells, also resulting in inefficient PFAS excretion in urine (Niu et al., 2023). Coupled with low levels of urinary excretion, the cycle of excretion in bile and reuptake by gut lining cells via enterohepatic recirculation is an important factor contributing to the long half-life of legacy PFAS in humans (Harada et al., 2007). As such, reducing enterohepatic recirculation is anticipated to reduce body burdens of PFAS and decrease the potential for adverse health effects that result from higher body burdens for longer periods of time.

Dietary fiber decreases the absorption of bile acids alongside several other food constituents such as cholesterol and fat by adsorbing substances or trapping them in a viscous gel matrix within the small intestine, thereby contributing to overall serum-cholesterol lowering effects (Jesch & Carr, 2017). Numerous studies have shown that daily consumption of gel-forming dietary fibers (e.g., β-glucan found in oats and barley, (Yu et al., 2022)) leads to reductions in LDL-C in the blood because these fibers “trap” bile acids, which are formed from cholesterol in the liver. Trapping of bile acids in the gut lumen thereby decreases the potential for enterohepatic recirculation and enhances their fecal excretion (Silva et al., 2021). Structurally, many PFAS are, by design, amphiphilic (soap-like) molecules that have excellent surfactant properties due to their long hydrophobic tail and the short polar negatively charged hydrophilic head. Since PFAS share biochemical characteristics with bile acids, they are anticipated to behave similarly.

Given that gel-forming fibers enhance bile acid excretion and that legacy PFAS share similar amphiphilic properties to bile acids, we hypothesized that regular consumption of oat β-glucan would reduce PFAS body burdens. To test this hypothesis, male C57Bl/6J mice were exposed to a mixture of seven PFAS in drinking water while being fed a control diet supplemented with inulin or to the experimental diet supplemented with oat β-glucan. To the best of our knowledge, this is the first study to evaluate the influence of oat β-glucan on concentrations of legacy (perfluoroheptanoic acid (PFHpA), PFOA, PFNA, PFOS) and newer replacement PFAS (PFHxS, Nafion Byproduct 2 (NBP2)) in the body.

MATERIALS AND METHODS

Materials.

Sources, CAS numbers and purity information for all PFAS are provided in Table 1. All other reagents were from Thermo Fisher Scientific (Waltham, MA), unless noted.

Table 1.

PFAS source information on final concentration in drinking water

Chemical	CAS Registry Number	Source	Reported Purity	Drinking Water Concentration (mg/L)a
Perfluoroheptanoic acid (PFHpA)	375-85-9	Santa Cruz Biotechnology	≥98%	1.91
Perfluorooctanoic acid (PFOA)	335-67-1	Sigma-Aldrich	95%	0.222
Perfluorononanoic acid (PFNA)	375-95-1	Santa Cruz Biotechnology	≥97%	0.233
Perfluorohexane sulfonate potassium salt (PFHxS)	3871-99-6	Santa Cruz Biotechnology	≥98%	0.317
Perfluorooctane sulfonic acid (PFOS)	1763-21-1	Santa Cruz Biotechnology	≥97%	0.276b
Nafion By-Product 2 (NBP2)	749836-20-2	Synquest Laboratories	95%	0.369

^aConcentrations in replicate stocks and drinking water were confirmed by LC-ESI⁻- MS/MS.

^bPFOS contained an almost equal mass concentration of PFOSA (0.3 mg/L in drinking water; PFOS/PFOSA ~1:1).

In vivo exposure.

All animal studies were approved by the Institutional Animal Care and Use Committee at Boston University and performed in an American Association for the Accreditation of Laboratory Animal Care accredited facility (Animal Welfare Assurance Number: A3316–01). Male, C57BL6/J mice (stock number: 000664, Jackson Laboratories, Bar Harbor, ME) arrived at 3 weeks of age and were acclimated for six days.

Vehicle and treatment water were prepared from NERL High Purity water (23-249-589, Thermo Fisher Scientific), which is prepared using the most efficacious methods to remove PFAS (i.e., reverse osmosis and carbon filtering)(Appleman et al., 2014). Sucrose was added to the drinking water to ensure consumption of PFOA, but the concentration was significantly lower than the sucrose concentration in sugar-sweetened drinks (10–12%) (Sundborn et al., 2019). This sucrose concentration resulted in an approximate daily sucrose intake of 21 mg/day per 20 g mouse, compared to an approximate daily sucrose intake from food of 517 mg/day. Concentrated stock solutions of each PFAS were prepared in NERL water. Aliquots were added to NERL water containing 0.5% sucrose to prepare a mixture of six PFASs at the concentrations shown in Table 1. Significant care was taken to minimize contamination of vehicle drinking water, which was verified by chemical analysis. Glass water bottles with rubber stoppers (WTRBTL, Braintree Scientific, Inc., Braintree, MA) were labeled for Vh or PFAS drinking water and kept segregated through the experiment. Water bottles and stoppers were washed using alkaline soap and scrub brushes that also were kept segregated.

Mice were provided with one of two custom diets (Table 2). Both diets were based on the “What we eat in America (NHANES 2013/2014)” analysis for what adults eat (Research Diets, New Brunswick, NJ)(USDA, 2018). The control diet was supplemented with inulin and the experimental diet was supplemented with oat β-glucan (BranPure^® Oat Bran Fiber, Layer Origin Nutrition, Ithaca, NY).

Table 2.

Diet components (Research Diets, Inc.)

	Inulin Diet (D22022202)		β-Glucan Diet (D22022203)
Component	%gm	% kCal	%gm	% kCal
Protein	15.7	14	15.3	14
Carbohydrate	51.8	47	51.1	48
Fat	18.5	38	18.0	38
Insoluble Fiber	4.4		4.3
Soluble Fiber	2	1	2
kcal/gm		4.4		4.3
Ingredient	gm		gm
Casein	158.5		144.58
L-Cystine	4.5		4.5
Corn Starch	158.3		162.8
Maltodextrin 10	100		100
Sucrose	201.7		201.7
Cellulose	40		24.2
Inulin	18.5		0
BranPure Oat Bran			66
Soybean Oil	25		25
Lard	92		92
Butter	50.4		49.83
Mineral Mix S10026	10		10
DiCalcium Phosphate	13		13
Calcium Carbonate	5.5		5.5
Potassium Citrate, 1 H2O	16.5		16.5
Vitamin Mix V10001	10		10
Choline Bitartrate	2		2
Cholesterol	1		1
FD&C Yellow Dye #5	0.025
FD&C Red Dye #40			0.05
FD&C Blue Dye #1	0.025

Mice were provided diets and administered vehicle (0.5% sucrose) drinking water or PFAS mixture drinking water ad libitum for 6 weeks, an exposure period we have used previously to examine the toxicological effects of PFOA (J. Schlezinger et al., 2020). After 6 weeks of exposure, one cohort of Vh and PFAS exposed mice was euthanized (exposure cohort), and one cohort of Vh and PFAS exposed mice (depuration cohort) continued in the experiment for 4 additional weeks with all mice receiving Vh water and their assigned diet. Food and water consumption were determined on a per cage basis each week. Body weight was measured weekly. Prior to euthanasia, the drinking water was changed to unamended water, and mice were fasted for 6–8 hours. At euthanasia (terminal isoflurane and exsanguination), blood was collected via cardiac puncture, and serum was separated. Liver was weighed and aliquots were collected for gene expression and triglyceride analyses. Sections of jejunum were squeezed to remove gut contents and isolated for gene expression (5–10 cm from the stomach) and triglyceride (10–15 cm from the stomach) analyses. The jejunum was chosen for these analyses as it is the area of the small intestine responsible for dietary fat absorption (Booth et al., 1961). All samples were flash frozen in liquid nitrogen and stored at −80°C.

PFAS chemical analyses.

The analytical method targeted the six PFAS in the mixture - PFHxS, PFHpA, PFOA, PFOS, PFNA, and Nafion Byproduct 2 (NBP-2), which is a small subset of the 22 PFAS species routinely quantified by the analytical method in this study as described below (and Table S1). Perfluorooctane sulfonamide (PFOSA) was found to be a major contaminant in the PFOS standard (~1:1 ratio) and was also quantified. Furthermore, we quantified and report on PFHpS concentrations, which was found as a contaminant in one of the two batches of drinking water. For the purpose of this study, we report quantitative results on the aforementioned eight PFAS species. Information on standards/materials and samples preparation protocols are provided in the Supplemental Material. The PFAS species were quantified by liquid chromatography – negative electrospray ionization - tandem-mass spectrometry (LC-ESI⁻- MS/MS) in an Applied Biosystems API4000 triple quadrupole mass spectrometer using the isotope dilution method. Chromatographic separation was accomplished on a Shimadzu LC20 series stack using a Luna C18, 3μm, 100×4.6mm analytical column (Phenomenex). Mobile phases were 10mM ammonium acetate in D.I. water (A) and 10mM ammonium acetate in methanol (B). Chromatographic gradient was 10% B the first min, to 65% B at 2 min, to 99% B at 15 min, hold 99% B to 20 min, followed by 5 min post-column equilibration. Sample injection volume was 10μL. Background PFAS contamination was eliminated using an online delay column (Phenomenex Luna C18, 50×4.6 mm, 3μm). A diverter valve was used (VICI, Valco Instrument Co. Inc.) to divert the front and back-end of the chromatographic run to waste.

The scheduled MRM mode (Table S1) was used for data acquisition of the target set of PFAS. Twenty percent of samples were analyzed as true blind duplicates. Analyte recovery was 95–102% at serum concentrations of 0.5–5 ng/mL (<7% relative standard deviation), with calibration curve coefficients of R >0.999 for all analytes. Limits of detection ranged from 1 to 50 pg/mL, equivalent to 10–500 fg injection on the column (Table S2). Tested and certified PFAS-free labware was used throughout the chain of sample acquisition, processing, and analysis. Quality control included laboratory blanks, blind sample duplicates, random blanks to check for carryover or cross-contamination, and a well-characterized aqueous film forming foam sample. Additional details about the analytical method parameters are provided in the Supplemental Material, Table S1 and S2.

Triglyceride analysis.

Liver and jejunum total triglyceride concentrations were assayed as described (Norris et al., 2003). We prepared saponified neutralized liver and jejunum extracts and generated a standard curve with glycerol standard solution (Sigma G7793). Triolein content was assessed using free glycerol reagent (Sigma F6428). Triolein was converted to triglyceride and divided by wet organ weight to derive organ lipid content (mg triglyceride / g organ).

Gene expression analysis.

Total RNA was isolated from liver and jejunum samples via column-based extraction and genomic DNA removal using the Direct-zol RNA Miniprep Kit (Zymo Research, Orange, CA). cDNA was synthesized from total RNA using the iScript^™ Reverse Transcription System (BioRad, Hercules, CA). All qPCR reactions were performed using the PowerUp^™ SYBR Green Master Mix (Thermo Fisher Scientific, Waltham, MA). The qPCR reactions were performed using a QuantStudio 3 Real-Time PCR System (Applied Biosystems, Carlsbad, CA): UDG activation (50°C for 2 min), polymerase activation (95°C for 2 min), 40 cycles of denaturation (95°C for 15 sec) and annealing (various temperatures for 15 sec), extension (72°C for 60 sec). The primer sequences and annealing temperatures are provided in Table S3. Relative gene expression was determined using the Pfaffl method to account for differential primer efficiencies (Pfaffl, 2001), using the geometric mean of the Cq values for beta-2-microglobulin (B2m) and 18sRNA (R18s) for liver and R18s for intestine. The average Cq value from two livers or jejunums from female C57/BL6J mice was used as the reference point. Data are reported as “Relative Expression.”

Statistical analyses.

Data are presented as data points from individual mice or as mean ± standard deviation (SD). It should be noted that one β-glucan-PFAS mouse consumed significantly more water than other mice (explained in legend of Figure 1). Statistical significance was determined using Student’s t tests, Two-Factor ANOVAs, or Three-Factor ANOVAs (Prism 8, GraphPad Software Inc., La Jolla, CA). Statistical significance was evaluated at an α= 0.05 for all analyses.

Figure 1. Effect of diet on food and water consumption.

Three-week-old male C57BL6/J mice were treated with either vehicle (Vh, NERL water with 5% sucrose) or PFAS drinking water for 6 weeks. After 6 weeks, one cohort of mice was euthanized and one cohort of mice continued in the experiment for 4 additional weeks with all mice receiving Vehicle water and continuing on their assigned diet. During treatment and depuration, the mice were fed an American Diet supplemented with either inulin or β-glucan (see Table 1). Food and water consumption were determined on a per cage basis weekly and then divided by the total weight of the mice in the cage and by 7 days to determine the consumption rate per gram of mouse. One mouse in the β-glucan-PFAS group drank significantly more water than all other β-glucan mice (24.5 mg/g mouse vs. 5.5 ml/g mouse, respectively), was moved to single housing and was not included in the calculation for this analysis. The vertical dashed line indicates when exposure ended. Data are presented as mean ± SD. N = 3–4 cages. Diet significantly increased water consumption (p = 0.0185, Two-Factor Repeated Measures ANOVA).

RESULTS

PFAS concentrations in drinking water

We began by designing a drinking water-based PFAS exposure regimen. We chose to include both PFCAs (PFHpA, PFOA, and PFNA) and PFSAs (PFHxS, PFOS and NBP-2). PFOA, PFNA, PFHxS and PFOS are PFAS commonly measured in people (Graber et al., 2018). Short-chain PFAS (i.e., PFHpA) are now abundant in indoor environments (Zheng et al., 2023). Perfluoroalkyl ether sulfonates (i.e., NBP-2) are a considerable concern for residents living in communities whose drinking water has been contaminated by the fluorochemical industry (Wallis et al., 2023). Perfluoroalkyl ether sulfonates are used to manufacture polymers with ionic properties (e.g., Nafion) and can be released during the manufacturing process or use (Saleeby et al., 2021). The mixture was designed based on equimolar (1 μM) concentrations of PFOA, PFNA, PFHxS and PFOS; PFHpA and NBP-2 were included at higher concentrations, based on their shorter or unknown half-lives in rodents, respectively (Chang et al., 2012; Lou et al., 2009; Russell et al., 2015; Sundström et al., 2012; Tatum-Gibbs et al., 2011). The total PFAS drinking water concentration (3.4 mg/L) was designed to mimic the exposure level of PFOA we have previously shown to generate significant toxicological effects following 6 weeks of exposure, while generating a human-relevant, albeit an occupational level, body burden (J. Schlezinger et al., 2020; J J Schlezinger et al., 2021). Measured concentrations of PFAS in drinking water are reported in Table 1. We found that PFOSA was a significant contaminant of the PFOS stock solution and therefore the PFAS drinking water (0.34 mg PFOSA/L drinking water). In addition, PFHpS was quantified as a contaminant in one of the two PFAS drinking water batches (0.006 mg PFHpS/L drinking water). All other 14 PFAS monitored by the method (and listed in Table S1 footnote) were, with minor exceptions, at concentrations below their respective limits of detection in drinking water (<1–10 ng/L).

Food and water consumption by diet type

We also designed a control and an experimental diet that differed in fiber type. For 19–30 year old people, the recommended daily intake of fiber is 25 and 38 g/day, for females and males respectively, or approximately 14 g total fiber/1000 kcal (IOM, 2005). Using this guideline, we designed the mouse diets to have a total fiber content of 14.9 g total fiber/1000 kcal. There are no dietary reference intakes for soluble and insoluble fibers. For context, broccoli and green peas have ratios of soluble fiber to insoluble fiber of 1:9, an unpeeled apple and spinach have ratios of 1:7, and oats, barley and avocado have a ratio of 1:2 (Marlett & Cheung, 1997). Both diets were designed to have soluble to insoluble fiber ratios of 1:2. Both diets contained 4% insoluble fiber in the form of cellulose and hemicellulose, polysaccharides that are the main components of plant cell walls (IOM, 2005). The control diet contained 2% inulin, a non-gel-forming soluble fiber that is a polydisperse β-(2,1)-linked fructan (IOM, 2005). The experimental diet contained 2% β-glucan (from oats), a gel-forming soluble fiber that is a homopolysaccharide (IOM, 2005).

Food and water consumption were measured per cage weekly over the course of the experiment. Food consumption was not influenced by diet (Figure 1A). Water consumption was influenced by diet, with mice consuming the β-glucan diet drinking significantly more water than mice consuming the inulin diet (p = 0.02, Figure 1B). As a result, we calculated an estimate of total PFAS exposures in the inulin and β-glucan diet groups. Mice consuming the β-glucan diet had a significantly greater exposure to PFAS than mice consuming the inulin diet (Figure 2).

Figure 2. Estimate of total PFAS dose in mice consuming inulin and β-glucan diets.

Mice were exposed to Vh and PFAS drinking water and fed diets as described in Figure 1. Total PFAS exposure was estimated by calculating the total water consumed per g of mouse over the six-week exposure period (specific to the cage each mouse was in). Water consumption was then multiplied by the concentration of each PFAS in drinking water and exposure summed across PFAS (PFHpA, PFOA, PFNA, PFHxS, PFOS, NBP-2, PFOSA, PFHpA). Data are presented as mean ± SD. Data are based on 4 mice in the inulin group and 3 mice in the β-glucan group. One mouse in the β-glucan group drank significantly more water than all other mice, was moved to single housing and was not included in the calculation. The total exposure in that mouse was estimated to be 45,594 μg/kg. *** - Significantly different from inulin (p < 0.0002, two-tailed, unpaired, Student’s t test).

Serum PFAS by diet and treatment group

Serum PFAS concentrations were measured in mice euthanized at the end of the six-week exposure (exposure cohort; Expo) and following a four-week depuration (depuration cohort; Dep) (Table 3). Serum PFAS concentrations were significantly higher in serum of mice receiving PFAS drinking water at the end of the six-week exposure, regardless of diet (p<0.0001 for treatment for all PFAS in Table 3, Three-Factor (treatment, diet, time) ANOVA). The rank order of serum:drinking water concentration ratios following the exposure period was PFHxS ≈ PFOA > PFOS ≈ PFNA > NBP2 >>> PFHpA. There were significant decreases in serum PFAS concentrations at the end of the four-week depuration relative to the Expo cohort, regardless of diet for PFHpA, PFOA, PFHxS, PFOS and NBP2 but not for PFNA (Table 3, p<0.0001 for time, Three-Factor (treatment, diet, time) ANOVA). PFHpA was below the limit of detection in serum (<0.008 ng/mL) after the 4-week depuration. A significant decrease was not observed for PFHpS or PFOSA following cessation of treatment.

Table 3.

Serum PFAS concentrations (ng/ml serum) in mice after exposure (expo) and depuration (dep).

			Total PFAS		PFHpA****		PFOA****		PFNA		PFHxS**		PFOS***		NBP2*		PFHpS		PFOSA
			Ave.	SD	Ave.	SD	Ave.	SD	Ave.	SD	Ave.	SD	Ave.	SD	Ave.	SD	Ave.	SD	Ave.	SD
Vh	Inulin	6 wk expo	172	171	1	1	10	4	8	14	1	1	23	3	1	1	<0.006		106	150
		+ 4 wk dep	629	267	2	2	15	10	10	11	33	37	25	12	4	3	<0.006		569	231
	β-Glucan	6 wk expo	604	85	1	1	9	1	10	8	9	6	31	6	4	2	<0.006		540	77
		+ 4 wk dep	1052	316	2	1	23	34	136	260	2	1	24	5	<0.015		<0.006		865	124
PFAS	Inulin	6 wk expo	41834	18151	1022	197	5807	2025	2892	2057	9561	5822	4659	1202	3859	2624	67	19	13967	9035
		+ 4 wk dep	23402	11874	0	0	2011	1205	2188	1541	4575	2735	2915	1086	1841	1473	71	25	9801	8029
	β-Glucan	6 wk expo	41121	8627	1094	184	5706	934	4110	879	13750	2773	4924	769	3371	866	99	11	8070	4528
		+ 4 wk dep	25473	10520	1	1	1866	658	3147	2130	4542	2103	2933	481	1615	1065	64	11	11305	6498

N = 3–4. Data from two mice were excluded for all PFAS because the overall PFAS concentrations were more than two standard deviations from the mean (one β-glucan-PFAS-exposure mouse and one inulin-PFAS-depuration mouse). Effects of diet, treatment and time (exposure vs. depuration) were tested using a Three-Factor ANOVA. For treatment x time:

^*p < 0.05,

^**p < 0.01,

^***p < 0.001,

^****p < 0.0001.

a – LOD = limit of detection (reported in Table S2).

Because of the difference in water consumption, and therefore PFAS uptake between the diet groups, we compared the effect of diet on PFAS serum concentrations by calculating the ratio of the serum concentration to the estimated PFAS total dose. After six weeks of exposure and diet, serum PFAS concentrations, relative to total dose, tended to be lower in the β-glucan fed mice for PFHpA, PFOA and PFOS (Figure 3). There was no effect of diet on depuration of PFAS following cessation of treatment. Although, for serum concentrations of PFHpS not adjusted for total dose, a significant decrease in serum concentration occurred in β-glucan-fed mice during the depuration period (Table 3, p=0.03 for treatment x diet x time, Three-Factor (treatment, diet, time) ANOVA).

Figure 3. Effect of diet on serum concentrations of PFAS.

Mice were exposed to Vh and PFAS drinking water and fed diets as described in Figure 1. A. Total exposure to each PFAS was calculated as described in Figure 2. The serum PFAS concentration, determined in each mouse, was then divided by the total exposure estimate for that mouse. B. Effects of die and time (exposure vs depuration) on PFAS serum concentrations were tested using a Two-Factor ANOVA. * p<0.05 (Šídák’s test). Data are from individual mice, with the mean indicated by a line. N = 3–4. Abbreviations: Exposure cohort (Expo), depuration cohort (Dep).

General and organ-specific toxicity

We measured several endpoints for signs of general toxicity. While PFAS are known to be immunotoxic (Zhang et al., 2023), exposure was not associated with splenic or thymic atrophy in this exposure scenario (data not shown). For body weight gain (Figure 4A), neither diet nor PFAS exposure was a significant factor. Longer time in the experiment (as assessed by comparing the exposure and depuration cohorts) was associated with greater body weight gain. Adipose content of the mice appeared upon visual inspection to differ by diet; therefore, we collected and weighed the perigonadal fat pads in the Dep cohort (Figure 4B). The adipose:body weight ratio was significantly lower in the β-glucan fed mice than the inulin fed mice, regardless of PFAS exposure.

Figure 4. Effect of diet and PFAS exposure on body weight gain (A) and adiposity (B).

Mice were exposed to Vh and PFAS drinking water and fed diets as described in Figure 1. At euthanasia, (total body weight (6-week exposure and 4-week depuration groups) and perigonadal adipose tissue weight (4-week depuration group) were determined. Data are from individual mice. N = 3–4. Effects of treatment and time (exposure vs depuration) on weight gain for each diet were tested using a Two-Factor ANOVA (time contributed 33–41% of the variation, p<0.05). Effects of diet and treatment on adiposity were tested using a Two-Factor ANOVA (treatment contributed 70% of the variation, p<0.0001). * p<0.05, *** p<0.001 (Šídák’s test). Abbreviations: Exposure cohort (Expo), depuration cohort (Dep)

The liver is a major target organ of PFAS, with PFAS inducing hepatomegaly and lipid accumulation (e.g., (Das et al., 2017)). As expected, PFAS exposure caused an increase in liver:body weight ratio (Figure 5A). The effect of PFAS on liver:body weight ratio decreased over the depuration period (Figure 5A). Liver triglyceride concentrations were not significantly affected by PFAS exposure (Figure 5B). However, while liver triglyceride concentrations increased over time in inulin fed mice through the depuration period, they stayed constant in β-glucan fed mice (Figure 5B). Not surprisingly, we observed PFAS-induced changes in many lipid handling genes we previously demonstrated were PPARα-dependent (Figure S1)(J J Schlezinger et al., 2021).

Figure 5. Effect of diet and PFAS exposure on liver weights (A) and triglyceride content (B).

Mice were exposed to Vh and PFAS drinking water and fed diets as described in Figure 1. A. At euthanasia, total body weight and liver weights were determined. B. Total triglyceride concentrations were determined from measuring triolein content in saponified liver extracts. C. Effects of diet, treatment and time (exposure vs depuration) on liver:body weights and liver triglycerides were tested using a Three-Factor ANOVA. ** p<0.01, **** p<0.0001 (Šídák’s test). Data are from individual mice. N = 3–4. Abbreviations: Exposure cohort (Expo), depuration cohort (Dep) and treatment (Tx)

Soluble dietary fiber, especially β-glucans specifically are known to have positive effects on gut health, via modifying glucose and lipid uptake and the gut microbiome (Gill et al., 2021). Therefore, we assessed lipid accumulation in the jejunum. Both PFAS treatment and the β-glucan diet were associated with lower triglyceride concentrations in the jejunum (Figure 6).

Figure 6. Effect of diet and PFAS exposure on triglyceride accumulation in intestine.

Mice were exposed to Vh and PFAS drinking water and fed diets as described in Figure 1. A. Total triglyceride concentrations were determined from measuring triolein content in saponified neutralized jejunum extracts. B. Effects of diet, treatment and time (exposure vs depuration) on organ weights were tested using a Three-Factor ANOVA. Data are from individual mice. N = 3–4. Abbreviations: Exposure cohort (Expo), depuration cohort (Dep) and treatment (Tx).

Nuclear receptor-regulated gene expression

We also measured the mRNA expression of nuclear receptors known to be activated by PFAS and their target genes in liver and jejunum, including PPARα (Cyp4a10), CAR (Cyp2b10) and PXR (Cyp3a11) (Robarts et al., 2023). At the end of the exposure period, PFAS had no significant effect on hepatic mRNA expression of Ppara, Nr1i3 (CAR) or Nr1i2 (PXR) (Figure 6), although, all were significantly induced by PFAS exposure following the depuration period (Figure 6). Cyp4a10 and Cyp2b10 were induced in liver to similar levels at the end of both the exposure and depuration periods (Figure 6). Cyp3a11 also was induced in liver by PFAS exposure, but the level of induction decreased between the end of the exposure period and the end of the depuration period (Figure 6). In jejunum, mRNA expression of Ppara, Nr1i3 (CAR) and Nr1i2 (PXR), and their target genes were variable, with little discernable pattern of effect (Figure S2).

DISCUSSION

PFAS pose a significant health threat, in part, because of their long half-lives of elimination. New approaches are needed to reduce human PFAS body burdens. Here, we tested the hypothesis that a gel-forming, soluble fiber, oat β-glucan in the diet can reduce PFAS body burdens by reducing PFAS absorption and reabsorption in the intestine. While our analysis was complicated by the fact that mice fed the β-glucan diet drank more water and thus received a higher PFAS dose than mice fed the inulin diet, results provide support for the hypothesis. Further, given the overall 10-week PFAS exposure, the results provide new insights into PFAS accumulation and elimination, as well as toxicodynamic effects. Finally, consumption of the oat β-glucan diet appeared to have additional positive effects on lipid homeostasis.

PFAS toxicokinetics: Mixtures, fiber in diets, excretion pathways

Differences in biliary and/or urinary elimination are hypothesized to lead to exceptionally long human PFAS half-lives. As discussed above, biliary clearance rates of PFOA and PFOS are significantly higher than urinary clearance rates in humans (Harada et al., 2007). However, the reabsorption rates following biliary clearance of PFOA and PFOS also are high (Harada et al., 2007), thus only a small proportion of PFAS excreted in the bile may leave the body in feces. Modeling studies suggest that binding to ASBT, a transporter protein that moves bile acids and PFAS from the gut lumen into enterocytes, is a strong predictor of PFAS half-lives in humans (Cao et al., 2022). Thus, we hypothesize that a gel-forming, dietary fiber will reduce human half-lives of longer chain PFCAs and PFSAs. This hypothesis is supported by the recent report that cholestyramine administration (4 grams, 3 times daily) for 12 weeks significantly reduced serum concentrations of several PFAS in people in the following order: PFOS>PFDA> PFNA>PFOA=PFHxS (Møller et al., 2024). Cholestyramine is an anionic resin used to treat high LDL-C by trapping bile acids; it binds bile acids in the gut and enhances their elimination from the body in feces.

In the current study, serum concentrations of PFHpA, PFOA and PFOS all trended toward being lower in β-glucan diet fed versus inulin diet fed mice. However, not all PFAS serum concentrations were influenced by diet. This likely resulted from differences in the proportions that were excreted in urine versus bile. In mice, while PFHxA and PFHpA are rapidly cleared from the blood, PFCAs with 8 or more carbons have very slow elimination from serum (Fujii et al., 2015; Hundley et al., 2006). Correspondingly, PFHxA and PFHpA are largely excreted in urine (≈100%), significantly less PFOA, PFNA and PFDA are excreted in urine (7%, 1%, and 0.3% respectively) and almost no PFASs with ≥ 11 C are excreted in urine (≤ 0.1%) (Fujii et al., 2015). Fecal excretion for all PFAS tested in the Fujji et al. study was very low (≤ 5%); however, PFAS concentrations in bile were not measured (Fujii et al., 2015). Considerably less data are available on the urinary, biliary and fecal excretion of PFSAs in mice. In toxicokinetic studies, less than 3% of the administered PFHxS dose was recovered in urine or feces over any 24-hr collection period with more excreted in urine than in feces (Sundström et al., 2012). For PFOS, less than 0.7% of the administered dose was recovered in urine or feces over any 24-hr collection period with slightly more excreted in urine than in feces (Chang et al., 2012). Given that PFOA and PFOS are largely excreted in bile/feces in mice, this may explain why the depurating effect of β-glucan was stronger for these PFAS. These observations are important to consider in the context of PFAS toxicokinetics in humans who are exposed to a varying mixture of PFAS and fiber uptake in their diets.

Furthermore, with regard to PFAS toxicokinetics, we demonstrate that the body burden of PFAS was significantly influenced by the depuration period. Regardless of diet, the overall PFAS concentrations decreased in serum (≈ 58%) over the 4-week depuration period. PFHpA, PFOA, NBP2, PFHxS and PFOS serum concentrations decreased significantly over the depuration period, while serum concentrations of PFNA, PFOSA, and PFHpS did not. PFHpA was nearly undetectable after the depuration period, which is consistent with the short half-life of elimination (0.7 days) measured in male rats (Russell et al., 2015). The significant reductions in serum concentrations of PFOA, PFHxS and PFOS also are consistent with their known half-lives of elimination in mice, 19, 28 and 37 days, respectively (Chang et al., 2012; Lou et al., 2009; Sundström et al., 2012). It was expected that PFNA serum concentrations would have also decreased significantly, as the elimination half-life in the mouse is 30 days (Tatum-Gibbs et al., 2011). There was a decrease in the PFNA serum concentrations over the depuration period, but the variability in the data was high. Little is known about the toxicokinetics of NBP2, PFOSA or PFHpS in rodent models.

PFAS toxicodynamics: Immediate and sustained effects

Few studies have followed the toxicodynamics of PFAS after both an exposure and depuration period. Physiological characteristics of the liver changed from the end of the PFAS exposure period to the end of the depuration period. Considering the average drinking water consumption rate in this study, the average daily total PFAS dose was ≈ 0.39 mg/kg day for inulin fed mice and ≈ 0.55 mg/kg day for the β-glucan fed mice over the 6-week exposure period. This exposure resulted in a significant increase in liver:body weight ratio, a wellestablished outcome of PFAS exposure in rodents (ATSDR, 2021). The effect was significantly diminished after the depuration period, likely reflecting the decrease in body burden described above. However, liver:body weight ratio was not associated with liver triglyceride concentrations. A number of studies have investigated the effect of PFOA and PFNA on hepatic triglyceride accumulation across different mouse models and study designs (Table S4). Most studies report that PFOA and PFNA (at doses ≥ 0.1 mg/kg/day) increase liver hepatic triglyceride concentrations (Attema et al., 2022; Das et al., 2017; Nakagawa et al., 2012; Nakamura et al., 2009; J J Schlezinger et al., 2021; Tan et al., 2013; J. Wang et al., 2015). It is likely that the dose/timing of exposure and whether or not the mice express human or mouse PPARα play a role in determining the ultimate effect of PFAS on hepatic triglyceride accumulation, (Nakagawa et al., 2012; Nakamura et al., 2009). Studies have also investigated the ability of PFOS and PFHxS to increase hepatic triglyceride concentrations (Table S4). Low level exposures to these PFSAs (≤ 1 mg/kg/day) did not modulate hepatic triglyceride concentrations (Pfohl et al., 2020; Wan et al., 2012). Higher exposures to PFSAs (≥ 3 mg/kg/day) increased hepatic triglyceride concentrations (Bijland et al., 2011; Das et al., 2017; Huang et al., 2020). Thus, we hypothesize that lack of effect of PFAS exposure on hepatic triglyceride concentrations in this experiment can be explained by the relatively low overall dose and the presence of both PFCAs and PFSAs, which have different potencies to induce triglyceride accumulation. A likely explanation for the PFAS-induced increase in liver:body weight ratio is the induction of peroxisome proliferation (Su et al., 2022; Wolf et al., 2008).

PFAS exposure also modulated mRNA expression of PPARα, CAR and PXR, as well as their target genes. Studies in PPARα null mice show that CAR and PXR pathways are activated by PFAS (Rosen et al., 2017; Su et al., 2022; Wen et al., 2019). Studies with PXR, CAR and FXR null mice show that after PPARα, CAR is likely the second most significant contributor to PFDA-induced changes in gene expression(Abe et al., 2017; Cheng & Klaassen, 2008). RNA expression of Ppara, Nr1i3 (CAR) or Nr1i2 (PXR) were all significantly induced by PFAS exposure following the depuration period. CAR and PXR expression have previously been shown to be inducible by PFAS after extended (8+ week) exposures (Cheng & Klaassen, 2008; Li et al., 2019). Given that PFAS body burdens remain significant at the end of the depuration period, sustained continued induction of gene expression is likely being driving by the continued presence of the PFAS in the liver. As expected, mRNA expression of PPARα (Cyp4a10), CAR (Cyp2b10) and PXR (Cyp3a11) target genes was upregulated following the exposure period and remained significantly elevated after the depuration period. Interestingly, only the expression of CYP3a11 decreased over the depuration period, suggesting that PXR may be less potently activated by PFAS than PPARα or CAR. This is in contrast to what was reported for recovery from PFOS exposure in rats in which PROD activity (pentoxyresorufin-O-depentylation, a marker for CYP2B) was elevated only for 28 days following cessation of exposure while testosterone 6β-hydroxylation, a marker for CYP3A activity, remained elevated for the full 84 day recovery period (Elcombe et al., 2012). The sustained induction of PPARα, CAR and PXR target genes appears to be in contrast to PFAS-induced effects on genes involved in fatty acid metabolism (e.g., Cd36, Acox1, Cpt1a, Scd), as these are only transiently induced by PFOA or PFOS (Li et al., 2019; Seacat et al., 2003; Uy-Yu et al., 1990).

In addition to potentially reducing PFAS body burdens, oat β-glucan appeared to have beneficial effects on lipid homeostasis. In the mice fed the β-glucan diet for 10 weeks, there was a significantly lower amount of body fat similar to what was observed following an 18 week feeding trial with 45 kcal % fat and 10 weight % β-glucan in the diet of C57BL6/J mice (Howard et al., 2024). Evidence supports an anti-obesogenic effect of oat β-glucan in humans, as well (Mathews et al., 2023). Furthermore, triglyceride content of the liver and jejunum were lower in mice fed the β-glucan diet. One potential explanation for these observations is that oat β-glucans may reduce the uptake of fatty acids in the intestine (Drozdowski et al., 2010)

Study limitations

This study has several limitations. First, it was unexpected that mice consuming the β-glucan diet consumed more water; thus, preventing us from achieving comparable PFAS exposures in the control and treatment groups. Second, there was a significant amount of variability in the serum PFAS concentrations. This likely resulted from slight differences in the preparation of drinking water weekly, rather than for the entire study-duration. Third, we hypothesize that the most effective administration of the oat fiber would be as a supplement consumed just prior to a meal so that the fiber concentration in the gut lumen would be at its highest when bile is released. This is not feasible in a sub-chronic mouse study. Last, we chose to use a mixture of PFAS. However, we do not know how different PFAS may influence interactions with transporter proteins (e.g., differential affinity leading to differences in competitive binding) or how they may each influence the expression of transporter proteins.

Practical and effective interventions are needed to reduce PFAS body burdens. Soluble, gelforming dietary fibers are a safe and effective intervention used to reduce serum cholesterol by interrupting the enterohepatic recirculation of bile acids, which are synthesized from cholesterol in the liver. Here we tested the hypothesis that consumption of the soluble, gel-forming fiber in oats (β-glucan) could also reduce PFAS body burdens. The results of this pilot study are supportive of the hypothesis, which will need to be tested further by addressing the limitations outlined above.

Figure 7. Effect of diet and PFAS exposure on liver nuclear receptor-related gene expression.

Mice were exposed to Vh and PFAS drinking water and fed diets as described in Figure 1. Following isolation of RNA from liver, gene expression was determined by RT-qPCR. A. Ppara and its target gene. B. Nr1i3 and its target gene. C. Nr1i2 and its target gene. D. Effects of diet, treatment and time (exposure vs depuration) on organ weights were tested using a Three-Factor ANOVA. * p<0.05, ** p<0.01, *** p<0.001, **** p<0.0001 (Šídák’s test). Data are from individual mice. N = 3–4. Abbreviations: Exposure cohort (Expo), depuration cohort (Dep) and treatment (Tx)

HIGHLIGHTS.

• In mice exposed to PFAS in a mixture, the rank order of serum:drinking water concentrations was PFHxS ≈ PFOA > PFOS ≈ PFNA > NBP2 >>> PFHpA.

• We observed lower serum concentration trends in β-glucan fed mice for PFHpA, PFOA and PFOS.

• Activation of PXR, but not CAR or PPARα, decreased significantly over a 4-week depuration period.

• β-glucan fed mice had lower adipose:body weight ratios and liver and jejunum triglyceride concentrations.

ABBREVIATONS

CAR

Constitutive androstane receptor

LC-ESI⁻ -MS/MS

liquid chromatography – negative electrospray ionization - tandem-mass spectrometry

LDL-C

Low density lipoprotein cholesterol

NBP-2

Nafion Byproduct-2

PFAS

Per and polyfluoroalkyl substances

PFCA

Perfluorocarboxylic acid

PFHpA

Perfluoroheptanoic acid

PFHxA

Perfluorohexanoic acid

PFOA

Perfluorooctanoic acid

PFNA

Perfluorononanoic acid

PFHxS

Perfluorohexane sulfonic acid

PFOS

Perfluorooctane sulfonic acid

PFOSA

Perfluorooctanesulfonamide

PFSA

Perfluorosulfonic acid

PPARα

Peroxisome proliferator-activated receptor alpha α

PXR

Pregnane x receptor