Ca Minerals and Oral Bioavailability of Pb, Cd, and As from Indoor Dust in Mice: Mechanisms and Health Implications

Background: Elevating dietary calcium (Ca) intake can reduce metal(loid)oral bioavailability. However, the ability of a range of Ca minerals to reduce oral bioavailability of lead (Pb), cadmium (Cd), and arsenic (As) from indoor dust remains unclear. Objectives: This study evaluated the ability of Ca minerals to reduce Pb, Cd, and As oral bioavailability from indoor dust and associated mechanisms. Methods: A mouse bioassay was conducted to assess Pb, Cd, and As relative bioavailability (RBA) in three indoor dust samples, which were amended into mouse chow without and with addition of CaHPO4,  CaCO3, Ca gluconate, Ca lactate, Ca aspartate, and Ca citrate at 200–5,000μg/g Ca. The mRNA expression of Ca and phosphate (P) transporters involved in transcellular Pb, Cd and As transport in the duodenum of mice was quantified using real-time polymerase chain reaction. Serum 1,25-Dihydroxyvitamin D3 [1,25(OH)2D3], parathyroid hormone (PTH), and renal CYP27B1 activity controlling 1,25(OH)2D3 synthesis were measured using ELISA kits. Metal(loid) speciation in the feces of mice was characterized using X-ray absorption near-edge structure (XANES) spectroscopy. Results: In general, mice exposed to each of the Ca minerals exhibited lower Pb-, Cd-, and As-RBA for three dusts. However, RBAs with the different Ca minerals varied. Among minerals, mice fed dietary CaHPO4 did not exhibit lower duodenal mRNA expression of Ca transporters but did have the lowest Pb and Cd oral bioavailability at the highest Ca concentration (5,000μg/g Ca; 51%−95% and 52%−74% lower in comparison with the control). Lead phosphate precipitates (e.g., chloropyromorphite) were observed in feces of mice fed dietary CaHPO4. In comparison, mice fed organic Ca minerals (Ca gluconate, Ca lactate, Ca aspartate, and Ca citrate) had lower duodenal mRNA expression of Ca transporters, but Pb and Cd oral bioavailability was higher than in mice fed CaHPO4. In terms of As, mice fed Ca aspartate exhibited the lowest As oral bioavailability at the highest Ca concentration (5,000μg/g Ca; 41%−72% lower) and the lowest duodenal expression of P transporter (88% lower). The presence of aspartate was not associated with higher As solubility in the intestine. Discussion: Our study used a mouse model of exposure to household dust with various concentrations and species of Ca to determine whether different Ca minerals can reduce bioavailability of Pb, Cd, and As in mice and elucidate the mechanism(s) involved. This study can contribute to the practical application of optimal Ca minerals to protect humans from Pb, Cd, and As coexposure in the environment. https://doi.org/10.1289/EHP11730


Introduction
Adverse health effects caused by exposure to lead (Pb), 1 cadmium (Cd), 2 and arsenic (As) in the environment are of global concern. 3 Arsenic is classified as a class I human carcinogen [4][5][6] ; Pb was associated with intellectual decrement in children 5-10 y of age 7 ; blood and urine Cd was linked with tubular damage in women 53-64 y of age 8 and lower estimated glomerular filtration rate in adult women. 9 Among pathways, incidental ingestion of cocontaminated indoor dust is an important source for child exposure to Pb, Cd, and As. 10,11 Worrying observations were made that, via atmospheric deposition, high Pb, Cd, and As concentrations of 56,184, 329, and 2,380 mg=kg occurred in indoor dust impacted by mining/smelting in Hunan, China. 11 Children playing indoors may incidentally ingest dust particles via hand-tomouth behavior, thereby being exposed to Pb, Cd, and As. 12 A recent study showed median dust ingestion rates of 0.30-42 and 0:40-43 g=d for toddlers (age 2-3 y) and children (age 4-15 y) living in a cobalt mining area, with 2-fold higher dust ingestion in the dry season than in the rainy season. 13 For children exhibiting elevated blood Pb levels, a strong correlation has been observed with Pb dust concentration. 14 In addition, child blood Pb isotopic composition resembles that of Pb in indoor dust but is significantly different from that of other sources such as food, drinking water, or soil, suggesting that incidental ingestion of indoor dust is the predominant contributor to children's blood Pb. 15 The indoor dust exposure scenario is expected to be more relevant in countries such as China and India, where air pollution leads to higher indoor dust deposition rates. 16 In light of the contribution of indoor dust as an exposure source, it is important to reduce Pb, Cd, and As exposure via indoor dust ingestion.
It is challenging to reduce Pb, Cd, and As exposure simultaneously. Traditional strategies such as remediation of contaminated sites using amendments for in situ immobilization may reduce Pb exposure [17][18][19] but increase As exposure. [20][21][22] In comparison, reducing metal(loid) oral bioavailability via elevating dietary mineral intake may provide a way to simultaneously lower human exposure to coexisting metal(loid)s. Operationally, oral bioavailability refers to the percentage of an oral dose that is absorbed from the gastrointestinal (GI) tract into the systemic circulation, which can vary from 0% to 100%. 23,24 In the gut, metal(loid) oral bioavailability is driven by two processes: solubilization in the GI tract and transepithelial transport. 25 The former can be reflected by metal(loid) solubility in simulated GI fluids via in vitro solubility assays, 26 whereas metal(loid) bioavailability is often measured using in vivo animal bioassays such as mice as a surrogate for humans. 27 Because transepithelial transport drives metal(loid) bioavailability, exploring factors that reduce intestinal metal(loid) transport can help to develop strategies to reduce metal internal exposure and thereby to protect human health. There are no known specific transport proteins for Pb and Cd, which often enter enterocytes through use of the machinery provided for essential metals such as calcium (Ca), iron (Fe), and zinc (Zn). 28,29 Specially, Fe and Zn transporters such as divalent metal transporter 1 (DMT1) and Zrt/Irt-related proteins (ZIPs) and Ca transporters including an apical Ca 2+ influx channel termed transient receptor potential, vanilloid, type 6 (TRPV6), an intracellular Ca 2+ -binding buffering protein calbindin-D 9k , and a highaffinity basolateral Ca 2+ efflux channel termed plasma membrane calcium ATPase 1b (PMCA1b) are involved in transcellular Pb and Cd transport. 29,30 The sharing of Pb/Cd with Ca transporters provides a chance to manipulate dietary Ca intake to reduce Pb/ Cd oral bioavailability. One way is to take Ca minerals, with ∼ 43% of people in the United States reporting supplemental Ca use based on 2003-2006 National Health and Nutrition Examination Survey (NHANES) data. 31 Early studies showed lower Pb or Cd absorption from the GI tract in rats, 32,33 mice, 34 and humans 35 on high-Ca diets or foods and the protective role of Ca in alleviating Pb and Cd toxicity in pigeons. 36 However, the underlying molecular mechanisms have yet to be understood. Besides possible Ca competition with Pb and Cd for shared Ca transporters, adaptation to high levels of dietary Ca may decrease the expression of intestinal Ca transporters to avoid hypercalcemia and maintain Ca homeostasis. 30,37 Calcium homeostatis may be achieved via a series of physiological processes that eventually decrease renal synthesis of 1,25-Dihydroxyvitamin D3 [1,25ðOHÞ 2 D 3 ], 30,38 which acts as a steroid hormone that critically controls the expression of Ca transporters in the intestine. 30,39,40 By this negative feedback loop, Ca minerals may reduce transcellular Pb and Cd transport via Ca transporters and thereby their bioavailability, which remains to be confirmed.
Unknowns also remain regarding the efficacy of different Ca minerals in decreasing the bioavailability of Pb and Cd, which is important for practical use to reduce human metal(loid) exposure. To improve Ca nutrition in humans, both inorganic Ca (CaCO 3 and CaHPO 4 ) and organic Ca (Ca gluconate, Ca lactate, Ca aspartate, and Ca citrate) have been used. In comparison with inorganic Ca, organic Ca is more effective in providing Ca because organic ligands can form a complex with Ca, preventing Ca precipitation with oxalate, phosphate, or phytate under neutral intestinal conditions. 41,42 Although organic Ca may be more effective in stimulating Ca transporter down-regulation, organic ligands may complex with metal(loid)s to increase their solubility in the intestine and lead to higher accumulation in mice. 43 As such, it is unclear whether organic Ca can reduce Pb and Cd oral bioavailability. Moreover, recent studies reported the formation of sparingly soluble Pb-phosphate precipitates, particularly chloropyromorphite [Pb 5 ðPO 4 Þ 3 Cl], in mouse intestine when phosphate (P) and Pb were co-ingested. [44][45][46][47] In light of this, CaHPO 4 may play dual roles in reducing Pb oral bioavailability, with supplied Ca reducing intestinal Ca transporters and P precipitating with Pb. As a consequence, CaHPO 4 may reduce Pb or Cd bioavailability most efficiently among Ca minerals.
Regarding As oral bioavailability, the effects of Ca intake are yet to be explored. When As is ingested, stepwise As transformation occurs in the GI tract, [48][49][50] with inorganic arsenate (As V ) ang arsenite (As III ) being more toxic and readily absorbed than monomethylarsenate (MMA V ) and dimethylarsenate (DMA V ). 51,52 Unlike Pb and Cd, As V is absorbed in the small intestine via P transporters such as the type IIb sodium phosphate (NaPi-IIb) cotransporters in the apical membrane of enterocytes, 53 whereas As III is absorbed via aquaglyceroporins. 54 Hence, Ca may not compete with As directly for absorption transporters. However, in the proximal small intestine, the major site of metal(loid) absorption, 30 a large portion of ingested As may exist as As V and react with Ca cations, forming low-solubility Ca arsenates, 55 thereby decreasing As solubility. Another possible way is via reduced duodenal expression of P transporters related with Ca-inhibited renal 1,25ðOHÞ 2 D 3 synthesis. This can be a side effect of Ca minerals in reducing 1,25ðOHÞ 2 D 3 , because 1,25ðOHÞ 2 D 3 critically controls the intestinal expression of both P and Ca transporters to maintain both phosphorus and Ca homeostasis. 30,56 If this occurs, duodenal transcellular As V transport via P transporters may be reduced, thereby reducing As bioavailability, which remains to be assessed.
The objective of this study was to evaluate the ability of Ca supplementation to decrease the bioavailability of Pb, Cd, and As in orally ingested indoor dust in mice. Of note, particular attention was paid to different types of Ca, the use of real indoor dusts cocontaminated with Pb, Cd, and As, and the molecular mechanisms behind reduced Pb, Cd, and As oral bioavailability. It was hypothesized that Ca minerals can cause lower oral bioavailability of coexisting Pb, Cd, and As in indoor dust with varied efficacy. To achieve this goal, a combination of in vivo mouse bioassays, in vitro solubility assays, duodenal transporter expression quantification, and fecal metal(loid) speciation analyses by X-ray absorption near-edge structure (XANES) spectroscopy and X-ray diffraction (XRD) was used. The specific objectives were to: a) assess effects of dietary supplementation of six Ca compounds on Pb, Cd, and As bioavailability in mice exposed to contaminated indoor dusts and identify the Ca form most efficient in reducing Pb, Cd, and As bioavailability and b) elucidate mechanisms accounting for differences in Pb, Cd, and As bioavailability after Ca administration with a focus on Ca and P transporter expression and metal(loid) species present in the GI tract. Results are critical to the practical application of optimal Ca minerals to protect humans from Pb, Cd, and As exposure in the environment.

Indoor Dust and Ca Compounds
Three indoor dust (<150 lm) samples (nominally labeled A-C) with varied Pb, Cd, and As bioavailability measured by Zhao et al. 11 were tested. Samples were collected in the field from indoor surfaces (floors, window sills, and furniture) of three homes located in a Pb/Zn mining/smelting area of Hunan, China in 2017. House dust (>30 g) deposited on indoor surfaces of each home was brushed into plastic sealable bags. Following collection, dusts were freeze-dried, debris was removed, and dusts were sieved to <150 lm. Each dust sample was thoroughly mixed using a blender (JYL-C93T; Joyoung) to ensure dust composition would be consistent throughout the experiments. Dusts were measured for concentrations of Pb, Cd, and As using inductively coupled plasmamass spectrometry (ICP-MS; NexION300X, PerkinElmer) following digestion using HNO 3 and H 2 O 2 at 105°C according to U.S. Environmental Protection Agency (U.S. EPA) Method 3050B. 57 The samples varied in Pb (6,469-56,185 lg=g), Cd (83:0-230:0 lg=g), and As concentrations (569-2,644 lg=g) (Table S1), serving as representative dusts to assess impacts of Ca minerals on oral bioavailability of metal(loid)s at a range of concentrations.
Commonly used dietary Ca minerals (CaHPO 4 , CaCO 3 , Ca gluconate, Ca lactate, Ca aspartate, and Ca citrate) of analytical purity or higher grade were obtained from Shanghai Macklin Biochemical Co., Ltd. Calcium minerals were digested using HNO 3 and H 2 O 2 at 105°C according to U.S. EPA Method 3050B 57 and measured for Pb, Cd, and As concentrations in the six Ca compounds using ICP-MS, which ranged from 1.83-55.6, 0.04-0.63, and 0:22-1:22 lg=g, respectively (Table S2).

Mouse Experiments
Following intestinal absorption in animals, metal(loid) accumulates in tissues, especially the liver and kidneys; therefore, they are often used as the bioavailability end points. 27 Based on animal bioassays, metal(loid) bioavailability is typically measured as relative bioavailability (RBA), i.e., the ratio of bioavailability in a test sample to that of a highly soluble toxicity reference value compound (e.g., Pb acetate, Cd chloride, and sodium arsenate). [58][59][60] The RBA calculation accounts for the physiological differences between animals and humans and can be extrapolated to humans. 27 To quantify metal(loid) RBA in the presence of Ca minerals, mouse experiments were performed. The use of mice was approved by the Animal Care and Use Committee of Nanjing University. Throughout the study, powdered low-Ca (200 lg=g) AIN-93G purified rodent diet was used as the basal diet (Jiangsu Medicience Ltd.). The diet was digested using HNO 3 and H 2 O 2 at 105°C according to U.S. EPA Method 3050B 57 and measured for Pb, Cd, and As concentrations using ICP-MS, which were low (350 ± 113 lg=kg Pb, 4:15 ± 2:13 lg=kg Cd, and 32:5 ± 7:07 lg=kg As). Based on the basal diet, the following test diets were prepared: a) diet amended with dust alone (dustdiet); b) diet amended with both dust and Ca (dust+Ca-diet); c) diet amended with Ca alone (Ca-diet); and d) diet amended with Pb acetate, Cd chloride, or sodium arsenate alone (reference-diet). The dust-diet was prepared by adding each dust in the basal diet at 2% (w/w). From the basal and dust-diet, Ca-diet and dust+Ca-diet were prepared by adding each Ca compound, with the elemental Ca concentrations being 200, 1,000, and 5,000 lg=g Ca, respectively. Briefly, mouse basal diet was freeze-dried and ground to powder using a blender (JYL-C93T; Joyoung). Dusts were incorporated into the diet powder at 2% (w/w) and thoroughly mixed. Then, Ca salts were incorporated into dust-containing diet powder to achieve 200, 1,000, and 5,000 lg=g Ca. To diet mixtures, Milli-Q water was added, and the mixtures were then molded into pellets, freeze-dried, and stored at 4°C until use. The Ca concentrations are environmentally relevant when compared to dietary reference intakes for Ca, such as the adequate intake of 500 mg=d Ca for healthy infants age 1-3 y recommended by Institute of Medicine of the National Academies. 30 Similarly, Pb acetate (CAS# 6080-56-4), Cd chloride (CAS# 10108-64-2), or sodium arsenate (CAS# 13464-38-5) was amended in the basal diet to 50, 250, 500 lg=g Pb, 5, 10, 15 lg=g Cd, or 50, 100, 150 lg=g As, respectively.
Female mice (BALB/c, ∼ 18 g body weight, 8 wk old) were purchased from Nanjing Junke Bioengineering Co. Ltd. Mice were housed in polyethylene cages on dry wood-chip bedding and acclimated for 7 d under environmental conditions of 25°C, 50% humidity, and a 12:12 h light:dark cycle, with free access to basal AIN-93G diet and Milli-Q water. After acclimation, the test diet was supplied to mice over 10 d, followed by 1 d of clearance by supplying basal diet according to the method established by Juhasz et al. 61 Female mice are more docile than male mice and therefore commonly used in metal(loid) oral bioavailability studies. [43][44][45][46]61 Mice were housed alone, with one mouse in each cage. For each treatment, three mice were used as biological replicates (n = 3) unless stated otherwise. Before exposure, mouse body weight and diet supplied were weighed, with body weight and diets remaining being measured again at the end of the exposure. During the 10-d period, mice had free access to diets and Milli-Q water. Diet consumption was calculated as the difference in weight between diets supplied and remaining. At the end of the exposure, mice were euthanized with carbon dioxide to collect blood samples, with one half of blood of each mouse collected in anticoagulation tubes and the other half collected in promoting coagulating tubes. Blood in promoting coagulating tubes was allowed to clot at room temperature for 1 h prior to centrifugation (4°C, 10,000 × g, 1 min) to get serum. Mouse liver, kidney, duodenum, feces, and femur samples were sampled and stored at -80 C prior to analyses.

Pb-, Cd-, and As-RBA Calculation
After collection, liver, kidney, and femur samples of mice fed dust-diet, dust+Ca-diet, Ca-diet, and reference-diet were freezedried and ground to powder. Dried liver (0:3 g), kidney (0.04), femur (0:02 g), and fresh whole blood (0:15 g) samples were digested using HNO 3 and H 2 O 2 at 105°C according to U.S. EPA Method 3050B. 57 Digested samples were then diluted with Milli-Q water and analyzed for Pb, Cd, and As concentrations by ICP-MS. Calcium concentrations in whole blood, liver, and kidney digest samples were measured using inductively coupled plasmaoptical emission spectrometry (ICP-OES; Optima 5300DV, PerkinElmer). For mice fed reference-diet, strong linear correlations between Pb, Cd, and As dosage level and metal(loid) concentration in mouse tissue were generated to establish dose-response curves ( Figure S1) serving as references to calculate metal(loid)-RBA in dusts. Based on metal(loid) accumulation in liver and kidneys of mice fed dust-and dust+Ca-diet, the relative bioavailability (RBA) of Pb, Cd and As in indoor dust samples was calculated as follows. 62 where, MCðliver + kidneyÞ dust and MCðliver + kidneyÞ reference are Pb, Cd, or As concentration (in micrograms per gram) in liver and kidneys of mice exposed to dust and reference; DL dust and DL reference are Pb, Cd, or As dosage level (in micrograms per gram of body weight per day) from dust and reference exposures. For each mouse, metal(loid) dosage levels were calculated by multiplying daily diet consumption rate (grams per day) with diet metal(loid) concentrations (micrograms per gram) and normalized with body weight.

RNA Isolation and Quantitative Real-Time Polymerase Chain Reaction (RT-PCR)
The duodenum is the major site of metal(loid) absorption. 30 mRNA expression of Ca and P transporters in the duodenum of mice fed Ca-diet was quantified. Briefly, total RNA was isolated and purified from 1-cm portions of duodenum (approximately 10 mg) using the FastPure Cell/Tissue Total RNA Isolation Kit V2 (Nanjing Vazyme Biotech Co., Ltd.) per manufacturer's instructions. RNA concentration and purity were measured using a Nanodrop spectrophotometer (NanoDrop One; Thermo Scientific). Reverse transcription of RNA (1 lg) to cDNA was performed using the HiScript III RT SuperMix for quantitative real-time polymerase chain reaction (qPCR) (+gDNA wiper) Kit (Nanjing Vazyme Biotech Co., Ltd.

Fecal Metal(Loid) Speciation Analyses
Metal(loid) speciation in the feces of mice fed dust A with Ca were characterized. Mice fed dust A was selected because its Pb RBA was affected by CaHPO 4 to the greatest extent among the three test dusts. First, a sequential extraction protocol based on Tessier et al. 64 was employed to fractionate Pb and Cd in dust and mouse feces to the exchangeable, carbonate-bound, Fe/Mn oxides-bound, organic-bound, and residual fractions.
Second, for dust A and mouse feces, Pb XANES spectra were collected at the Pb L III -edge (13,035 eV) on beamline BL11B at the Shanghai Synchrotron Radiation Facility (SSRF), and As spectra were collected at the As K-edge (11,867 eV) on beamline 1W1B at the Beijing Synchrotron Radiation Facility (BSRF). The electron beam energy was 3.5 and 2.5 GeV, with a ring current of 200 and 250 mA in SSRF and BSRF, respectively. Energy calibration was carried out prior to each batch of measurements, respectively, with reference Pb foil and Au foil for Pb and As XANES analysis. Various Pb standards [Pb 5 ðPO 4 Þ 3 Cl, Pb 3 ðPO 4 Þ 2 , PbSO 4 , PbS, ðPbCO 3 Þ 2 ÁPbðOHÞ 2 , plumbojarosite, and Pb adsorbed to goethite, birnessite, fulvic, and humic acid] 18 and As standards [scorodite (FeAsO 4 Á2H 2 O), arsenopyrite (FeAsS), orpiment (As 2 S 3 ), realgar (As 4 S 4 ), Ca arsenate, arsenate and arsenite adsorbed to ferrihydrite] 65 were used as reference spectra. Spectra were collected in fluorescence (samples) and transmission modes (standards). Spectra were collected at 12,840-13,685 eV for Pb and 11,670-12,465 eV for As with step of 5 and 3 eV. Lead and As XANES data were calibrated against the reference foil on the same energy grid and normalized, and the background was removed by spline fitting using Athena software (version 0.9.20; Demeter). Pb and As speciation was quantified by least-squares linear combination fitting (LCF) of the K-edge XANES spectra to reference standards over 13,000-13,120 eV (Pb) and 11,850-11,910 eV (As). 66,67 For dust A and mouse feces, X-ray diffraction (XRD) spectra were also collected using a PERSEE XD-3 diffractometer with Cu-Ka (k = 1:54056 nm) radiation at 35 kV and 30 mA in a step scan mode between 5 and 85°2h with 0.02°step at a scanning rate of 2°per min. XRD spectra of Pb standards were also collected for comparison with samples.

Intestine Tissue As Speciation Analyses
To test whether ingested dust-As could be absorbed via P transporters as As V in the intestine, As speciation in the duodenum tissue and the remaining sections of the small intestine (mixed jejunum, ileum, and cecum) was measured. For mice fed dust A, the intestine tissue samples were freeze-dried, ground, and ultrasonically extracted with methanol/Milli-Q water (1:1 v/v) for 2 h. The extraction was repeated 2 times. The extracts were centrifuged at 3,577 × g for 10 min. Arsenic species in the resulted solutions were separated and identified using high-performance liquid chromatography (HPLC; Waters e2695) coupled with ICP-MS. The mobile phase was 8 mMðNH 4 Þ 2 HPO 4 and 8 mM NH 4 NO 3 at pH 6.2. An anion exchange chromatography column (Hamilton PRP-X100, 250 mm × 4:1 mm, 10 lm particle size) was used to separate As species. 50 Duplicate samples (n = 3), duplicate analysis (n = 3), spiked samples (n = 3), and check values (n = 3) were included for quality assurance and quality control purposes.

Metal(Loid) Solubility Assessment
To further evaluate metal(loid)-RBA differences across Ca minerals, dust A was selected and assessed for metal(loid) solubility in a simulated gastrointestinal fluid in the presence of Ca minerals. The Solubility Bioaccessibility Research Consortium (SBRC) assay, i.e., the standard U.S. EPA Method 1340, 68 was followed. The assay was performed in triplicate for each test. In the assay, 0:1 g of dust A was added to 10 mL of gastric fluid (0:43 M glycine at pH 1.5) containing 0, 40, 200, or 1,000 mg=L Ca. These Ca concentrations corresponded to 0, 200, 1,000, and 5,000 lg=g Ca amended in diet at Pb:Ca molar ratios of ∼ 1:0, 1:2, 1:10, and 1:50. After 1-h extraction at 37°C, the assay was transitioned to the intestinal phase by adjusting solution pH to 7.0 and adding bile and pancreatin (1.75 and 0:50 g=L). After 4-h extraction, solutions were centrifuged (956 × g, 15 min, 25°C) and filtered (0:45 lm). The extracts (0:1 mL) were then diluted using 0:1 M HNO 3 (9:9 mL) and analyzed for solubilized Pb, Cd, and As using ICP-MS; Fe concentration in undiluted extracts was analyzed using ICP-OES as described elsewhere in the "Materials and Methods" section. By dividing the amount of solubilized Pb, Cd, and As by total metal(loid) concentration in dust added to the gastric fluid, metal(loid) bioaccessibility in the intestinal fluid was calculated. Postextraction dust residues were collected, freeze-dried, and characterized by XRD as described in the "Fecal Metal(loid) Speciation Analyses" section.

Quality Assurance and Control
Indoor dust and soil standard reference materials (SRMs 2584 and 2710a) from the National Institute of Standards and Technology Analysis were included in analyses. Pb, Cd, and As recoveries from the SRMs were 86%-97%, based on U.S. EPA Method 3050B. 57 Arsenic RBA in SRM 2710a (41:2 ± 13:1%) and Pb RBA in SRM 2,584 (74:1 ± 8:83%) were consistent with previous reports ( ∼ 40% and 70%). 69,70 During ICP-MS or ICP-OES analysis, each sample was analyzed in triplicate, with check and spiked solutions being randomly included (95%-105% recoveries). Data are reported as mean and standard deviation of three replicates. Two-sided Student's t-test or ANOVA with Tukey's test was performed to test significant (p < 0:05) differences between Ca and control treatments (version 9.1.3 for Windows; SAS Institute Inc.). Graphs were created using SigmaPlot (version 9.1.3; Systat Software, Inc.) and GraphPad Prism (version 8.0; GraphPad Software, Inc.).
We evaluated the body weight of mice exposed to dust A, B, or C with or without various concentrations of Ca minerals. In comparison with control (No Ca), there were generally no differences in body weight with increasing exposure to Ca minerals. Mice exposed to dust B and 5,000 lg=g Ca as CaHPO 4 did have significantly higher body weight in comparison with control. (Figure S2).
In general, As RBA was lower in mice fed Dust A, B, or C amended with Ca minerals than in those fed dust alone. Mice fed Ca aspartate or Ca gluconate appeared to have the lowest As RBA among mice fed the various other Ca minerals ( Figure  1C,F). In mice fed Dust A, B, or C and exposed to Ca aspartate and Ca gluconate (5,000 lg=g Ca), As RBA (3.00%-6.98%, 11.6%-14.4%, and 2.58%-8.78%) was 41%-72% lower than those exposed to dust alone (10:9 ± 0:87%, 19:6 ± 4:54%, and 9:21 ± 1:72%). The As RBA in mice fed dust with Ca aspartate and Ca gluconate were significantly (p < 0:05) lower than that in mice exposed to dust and any one of the other four Ca minerals (14%-63% lower in comparison with the control).
Similar results were obtained when assessing blood and bone metal(loid) concentrations ( Figures S3 and S4). In mice exposed to each of the indoor dust samples, coexposure to CaHPO 4 (5,000 lg=g Ca) resulted in the lowest blood Pb and Cd concentrations (41%-58% and 38%-60% lower than those exposed to dust alone), whereas coexposure to Ca aspartate or Ca gluconate (5,000 lg=g Ca) resulted in the lowest blood As concentration (39%-73% lower than control) ( Figure S3). Similarly, mice exposed to dust and CaHPO 4 (5,000 lg=g Ca) had 74%-91% lower Pb concentrations in femur than those exposed to dust alone, whereas mice exposed to dust and Ca aspartate (5,000 lg=g Ca) had 27%-59% lower As concentration in femur than those exposed to dust alone ( Figure S4).
Similar to results seen with dust-Pb, mice fed diets amended with 1,000 lg=g Pb as Pb acetate and CaHPO 4 (5,000 lg=g Ca) had 75, 76, and 95% lower Pb concentrations in blood, liver, and kidney of mice than in those exposed to the Pb-amended diet alone ( Figure S5).

Effects of Ca Minerals on Duodenal Ca and P Transporter Expression
To test our hypothesis that the effects of Ca minerals on the bioavailability of Pb, As, and Cd were associated with the expression of transporters in the gastrointestinal tract, we fed mice one of the six Ca minerals (200-5,000 lg=g Ca food) for 10 d, isolated the duodenum, and measured mRNA expression of apical Ca 2+ influx channel termed transient receptor potential, vanilloid, type 6 (Trpv6), intracellular Ca 2+ -binding buffering protein calbindin-D 9k , high-affinity basolateral Ca 2+ efflux channel termed plasma membrane calcium ATPase 1b (Pmca1b), and the type IIb sodium phosphate (NaPi-IIb) cotransporters. Overall, in comparison with the control, lower expression of Ca transporters was observed in mice fed organic Ca minerals than in those fed CaHPO 4 and CaCO 3 (Figure 2A-C; summary data can be found in Excel Table S2). Regardless of the concentration of Ca mineral in the diet, mice exposed to CaHPO 4 did not exhibit significantly lower mRNA levels of Trpv6, calbindin-D 9k , or Pmca1b (and in some cases exhibited significantly higher levels). In contrast, mice fed Ca gluconate, Ca lactate, Ca aspartate, or Ca citrate generally had lower expression of these three transporters, depending on the concentration of Ca in feed. Particularly, mice fed with the four organic Ca compounds at 5,000 lg=g Ca showed 72%-92%, 65%-67%, and 16%-49% lower expression of these three transporters ( Figure 2G). In terms of NaPi-IIb, mice fed CaCO 3 , Ca aspartate, and Ca citrate had 54%-88% lower duodenal mRNA expression than control mice, whereas mice fed CaHPO 4 , Ca gluconate, and Ca lactate did not exhibit significant differences in expression, although those fed Ca gluconate and Ca lactate trended toward lower expression ( Figure 2D,G).

Effects of Ca Minerals on Ca Concentrations in Tissues
In general, mouse blood, kidney, and liver Ca concentrations were not significantly higher in mice co-fed Ca minerals than controls (fed dust alone), although mice fed some Ca compounds tended to have higher Ca concentration in tissues ( Figure S7-S9). Regardless of Ca supplementation, Ca concentrations in blood of mice fed Dusts A, B, and C were maintained at average of 34, 58, and 68 lg=g, whereas Ca concentrations were maintained at 231, 211, and 184 lg=g in kidney and at 39, 104, and 110 lg=g in liver.

Effects of Ca Minerals on Pb, Cd, and As Solubility in the Intestine
Dust A was selected to investigate effects of Ca minerals on Pb, Cd, and As solubility in the intestine using in vitro assays. The fold differences were calculated as ratios of mean metal(loid) RBA in mice fed Ca to that in the control mice. For Ca-treated mice, values in the heat maps <1:00 suggested transporter expression down-regulation and lower 1,25ðOHÞ 2 D 3 and CYP27B1 than control mice, whereas values >1:00 suggested transporter expression upregulation and higher 1,25ðOHÞ 2 D 3 and CYP27B1 than control mice. Summary data can be found in Excel Table S2. Note: Ca, calcium; P, phosphate; RBA, relative bioavailability; SD, standard deviation.

Lead and As Speciation in Mouse Feces and Intestinal Tissue
In vivo formation of Pb 5 ðPO 4 Þ 3 Cl was demonstrated by charactering Pb speciation in mouse feces following exposure to dust A with CaHPO 4 at 5,000 lg=g Ca. Dust A was selected because its Pb RBA was affected by CaHPO 4 to the greatest extent among the three test dusts. Based on sequential extractions, 96:6 ± 0:12% Pb and 29:2 ± 3:17% Cd were in the residual fraction in feces of CaHPO 4 mice, significantly greater than 78:4 ± 3:57%-86:0 ± 0:34% Pb and 9:71 ± 0:05%-18:3 ± 4:00% Cd in control mice and mice supplied with other Ca minerals ( Figure 4B; summary data can be found in Excel Table S4). XANES analyses showed that, in comparison with the control mice (fed dust A alone), Pb 5 ðPO 4 Þ 3 Cl occurred in the feces of mice fed with dust A and 5,000 lg=g CaHPO 4 -Ca, contributing to 16% Pb in the feces ( Figure 4C). Based on XRD, Pb 5 ðPO 4 Þ 3 Cl and Pb 3 ðPO 4 Þ 2 were also detected in the feces of dust A CaHPO 4 -exposed mice ( Figure 4E).
When assessing As speciation using XANES, Ca-As precipitates were not observed in mouse feces following exposure to dust A with Ca aspartate. The feces of Ca aspartate-treated mice showed As speciation contribution from arsenate and arsenite adsorbed to ferrihydrite (48% and 10%), FeAsS (28%), and As 4 S 4 (13%) but not from Ca arsenate ( Figure 4D).
In the duodenum tissue, 21%-51% of total As was As V , and speciation was not significantly different in mice fed Ca minerals than in control mice (32%) ( Figure 5A). Similarly, in the jejunum, ileum, and cecum tissue of control and Ca-treated mice, 23%-66% of As was As V (Figure 5B; summary data can be found in Excel Table S5). . The differences were calculated as differences in metal(loid) solubility between Ca and control treatments and divided by metal(loid) solubility of control assessment. The asterisk indicates significant (p < 0:05, two-sided Student's t-test) difference in solubility in comparison with control. Values of 0% suggested no differences between Ca and control treatments; negative values suggested lower metal(loid) solubility with Ca minerals; positive values suggested higher metal(loid) solubility with Ca minerals. Data are reported as mean ± SD for each group. Summary data can be found in Excel Table S3. Note: As, arsenic; Ca, calcium; Cd, cadmium; Fe, iron; Pb, lead; SBRC, Solubility Bioaccessibility Research Consortium; SD, standard deviation.

Discussion
Limited studies have measured the bioavailability of Pb, Cd, and As contained in orally ingested indoor dust. 27,70 In mice fed diets amended with one of three indoor dust samples with varied Pb-, Cd-, and As RBA, the effect of cofeeding one of six Ca minerals on Pb, Cd, and As RBA was assessed (Figure 1). An interesting finding was that the RBAs of all the three metal(loid)s contained in the orally ingested dusts were lower in mice fed Ca minerals. However, cofeeding mice with the different Ca minerals resulted in different effects on Pb or Cd bioavailability and on As bioavailability, suggesting that the Ca minerals may differentially interact with Pb or Cd and with As. Among Ca minerals, mice fed a diet amended with dust and CaHPO 4 exhibited the lowest Pb RBA and Cd RBA, whereas those fed a diet amended with dust and Ca aspartate exhibited the lowest As RBA. These results suggest different mechanisms for Pb or Cd reduction and for As RBA reduction.

Down-Regulated Duodenal Ca and P Transporters
Studies show that Pb and Cd are absorbed from the intestine via Ca transport proteins, whereas As V is absorbed by P transporters. 29,53 We showed that in mice fed certain Ca minerals, the mRNA of specific Ca and P transporters in the duodenum was lower than in control mice, suggesting that Ca minerals were contributors to the lower Pb, Cd, and As RBA in mice administered Ca minerals. Specifically, we observed lower mRNA expression of Ca (Trpv6, calbindin-D 9k , and Pmca1b) and P transporters (NaPi-IIb) in the duodenum of mice fed Ca-diet, especially in mice fed organic Ca ( Figure 2). When organic Ca (Ca gluconate, Ca lactate, Ca aspartate, and Ca citrate) is ingested, ligands such as gluconate, amino acids, or citrate can complex Ca to increase its Ca solubility in the GI tract. 42 Higher Ca solubility would stimulate the body to decrease the expression of Ca transporters in the intestine to a greater extent to maintain Ca homeostasis. 30 The Ca homeostasis hypothesis was supported by the results that mouse blood, kidney, and liver Ca concentrations were not higher in mice co-fed Ca minerals than controls ( Figure S7-S9). Lower levels of duodenal Ca transporter mRNA in mice fed Ca minerals were associated with lower serum PTH expression and lower renal CYP27B1 activity, which coincided with lower 1,25ðOHÞ 2 D 3 (Figure 2; Figure S6). Under increased Ca intake and absorption, transient increase in serum Ca levels acts via a negative feedback loop to lower PTH secretion, which can suppress renal CYP27B1 activity and renal 1,25ðOHÞ 2 D 3 production. 30 The reduced 1,25ðOHÞ 2 D 3 is the key that regulates the down-regulated intestinal expression of Ca transporters. We also measured lower NaPi-IIb mRNA expression in mice fed organic Ca (Figure 2). Studies show that 1,25ðOHÞ 2 D 3 is a common key factor that drives the expression of both Ca and P . For Pb and As speciation analyses using XRD and XANES, feces of three replicate mice were pooled and mixed thoroughly before analyses. Summary Pb and As fractionation data can be found in Excel Table  S4. Note: As, arsenic; Ca, calcium; Cd, cadmium; Pb, lead; SD, standard deviation; SRBC, Solubility Bioaccessibility Research Consortium; XANES, X-ray absorption near-edge structure; XRD, X-ray diffraction.
transporters in the duodenal epithelium to maintain Ca and phosphorous homeostasis. 30 In addition to impacts on Ca transporters, the Ca-induced 1,25ðOHÞ 2 D 3 reduction can also decrease the expression of P transporters in the intestine.

Mechanisms Underlying Pb and Cd RBA Reduction with CaHPO 4
In this study, mice fed CaHPO 4 did not exhibit significantly lower mRNA expression of Ca transporters ( Figure 2); however, these mice had lower Pb and Cd RBA than those who were not fed Ca ( Figure 1A,B). The data suggested a unique mechanism for Pb and Cd RBA reduction with CaHPO 4 . Recent studies based on mouse models show that Pb RBA may be reduced in the presence of phosphate due to the reaction between soluble Pb and phosphate in the GI tract to form insoluble Pb phosphates {e.g., chloropyromorphite ½Pb 5 ðPO 4 Þ 3 Cl}. [44][45][46][47] The effectiveness of CaHPO 4 in reducing Pb and Cd RBA was likely attributed to potential precipitation of Pb or Cd phosphates in the intestine. This idea is supported by our current findings, which show 99% and 82% lower Pb and Cd solubility with the presence of CaHPO 4 in intestinal fluid (1,000 mg=L Ca) ( Figure  3A,B), and the formation of Pb 5 ðPO 4 Þ 3 Cl in collected dust residues ( Figure 4A). More important, in feces of mice co-fed dust A and CaHPO 4 , XANES and XRD analyses showed occurrence of Pb 5 ðPO 4 Þ 3 Cl ( Figure 4C,E), demonstrating in vivo formation of Pb phosphate minerals. Lead phosphate minerals are the least soluble Pb species [e.g., K sp of Pb 5 ðPO 4 Þ 3 CI = 1 × 10 -84:4 ], 71 so its formation can sharply decrease Pb solubility in the intestine, thereby reducing Pb bioavailability. [44][45][46][47] These results suggested that Pb phosphate precipitation rather than decreased Ca transporter expression was the key contributor to the lower Pb RBA in the presence of CaHPO 4 . It is possible that a similar mechanism was responsible for reduction in Cd RBA.

Pb and Cd RBA Reduction with Organic Ca
In comparison with mice fed CaHPO 4 , those fed organic Ca exhibited lower duodenal mRNA expression of Ca transporters ( Figure 2); however, these mice exhibited higher Pb and Cd RBA in comparison with those fed CaHPO 4 ( Figure 1A,B). In some cases, such as in mice fed dust A or C with Ca citrate (5,000 lg=g Ca), higher Pb RBA was observed relative to the control ( Figure 1D). This difference in effect of organic Ca was possibly due to the ability of organic ligands to enhance Pb solubility in the intestine ( Figure 3A), counteracting the role of down-regulated transporters. The higher Pb solubility corresponded to a concurrent higher soluble Fe ( Figure 3D). By complexing with Fe, lactate, Figure 5. Impact of Ca minerals on As speciation in mouse intestine tissue. Arsenic speciation in the duodenum (A) and the remaining section of the small intestine (B, mixed jejunum, ileum, and cecum) of mice (mean ± SD; n = 3) receiving diets amended with dust sample A alone (Control) or in combination with Ca hydrogen phosphate (CaHPO 4 ), Ca carbonate (CaCO 3 ), Ca gluconate, Ca lactate, Ca aspartate, and Ca citrate (5,000 lg Ca=g). There was no significant (p > 0:05; two-tailed Student's t-test) difference in As speciation among treatments. Summary data can be found in Excel Table S5. Note: Ca, calcium; SD, standard deviation. citrate, and gluconate inhibited Fe precipitation in intestinal fluids, 72 thereby reducing Pb coprecipitation with Fe in the neutral intestinal environment and leading to increased Pb solubility. This effect may have been offset by a decrease in Pb transport across the GI barrier via lower Ca transporter expression (as shown in Figure  2), resulting in the net effect of lower Pb RBA in comparison with the control.

Mechanisms Underlying As RBA Reduction with Ca
Arsenic and Ca are absorbed via different transporters 29,53 ; therefore, their competition for absorption transporters in the intestine is unlikely. Similar to reactions between Pb and phosphate, Ca cations and arsenate anions may react to form low-soluble Ca arsenates under neutral intestinal conditions, thereby decreasing As solubility and bioavailability. However, the mechanism was ruled out because XANES showed the absence of Ca-As precipitates in mouse feces following exposure to dust A with Ca aspartate ( Figure 4D).
Instead, Ca-induced P transporter down-regulation may be the key contributor to the reduced As RBA (Figures 1C and  2D). Following ingestion, a large portion of As can be observed as As V , as suggested by high percentage of As V in mouse intestinal tissue ( Figure 5). Through reduced duodenal NaPi-IIb, transepithelial As V transport may be inhibited, thereby leading to decreased As RBA. Among six Ca minerals, Ca aspartatetreated mice showed the most pronounced lower renal CYP27B1 and serum 1,25ðOHÞ 2 D 3 ( Figure 2E,F), which was associated with 70%-88% lower NaPi-IIb expression in the duodenum ( Figure 2D). In addition, Ca aspartate was not associated with higher As solubility when measured using the in vitro SBRC method in contrast with CaHPO 4 , Ca gluconate, or Ca citrate ( Figure 3C). Therefore, we propose that Ca aspartate was most effective in reducing As RBA presumably via Ca-decreased P transporters.

Conceptual Models of Ca Minerals in Reducing Metal(Loid) RBA
Based on data in this study, we hypothesize a mechanistic understanding of the ability of different Ca minerals to mitigate Pb, Cd, and As exposure as proposed in Figure 6. We propose that without Ca supplement, Pb or Cd and As were readily absorbed via Ca and P transporters in the duodenum ( Figure  6A), and that with CaHPO 4 supplement, expression of Ca transporters in the intestine was not reduced. However, we hypothesize that reduction in Pb or Cd bioavailability occurred due to the precipitation of Pb or Cd phosphate phases under the neutral intestinal conditions ( Figure 6B). With organic Ca minerals, high Ca solubility and bioavailability may act via a negative feedback loop to suppress renal 1,25ðOHÞ 2 D 3 production to avoid Ca overabsorption, which subsequentially resulted in down-regulated expression of Ca and P transporters in the intestine, consequently contributing to reduced Pb, Cd, and As bioavailability. The mechanism may be antagonized by the role of organic ligands in increasing metal(loid) solubility in the intestinal tract ( Figure 6C). Organic anions, including gluconate, lactate, citrate, and aspartate, may complex Pb and Cd, increasing their solubility in the intestine, thereby reducing the impact of decreased Ca transporters on bioavailability reductions. Organic Ca, particularly Ca aspartate, provided a highly soluble form of Ca to stimulate the greatest down-regulation of P transporters, whereas aspartate did not increase As solubility, thereby leading to the lowest bioavailable As.

Health Implications
Different from efforts that rely on eliminating anthropogenic sources to lower the health risk of metal(loid)-contaminated indoor dust, 73,74 this study assessed the ability of Ca minerals to lower metal(loid) oral bioavailability. An interesting finding was that, in mice fed dietary Ca minerals, oral bioavailability of coexisting Pb, Cd, and As in indoor dust was lower (Figure 1). Data indicate that dietary Ca supplement can be a valuable strategy to protect humans from toxic metal(loid) coexposure. More important, different Ca minerals had different effects on the bioavailability of the metals. To reduce children's exposure to Pb and Cd, CaHPO 4 is likely the best choice because mice fed this Ca species had lower Pb and Cd oral bioavailability than those fed other Ca minerals (Figure 1). To reduce children's exposure to As, Ca aspartate is likely the best choice ( Figure 1). When aimed to reduce children's exposure to coexisting Pb, Cd, and As, a dual supplement approach (CaHPO 4 and Ca aspartate taken alternately or in a mixture) may be a promising approach.
In comparison with traditional remediation strategies such as in situ metal(loid) immobilization by amendments, [75][76][77] dietary supplementation with Ca minerals is a cost-effective strategy that is easy to implement and more acceptable to the public. In addition, the strategy can improve Ca nutrition and may address Ca malnutrition, which threatens a large portion of the world's population, 78,79 and thereby may achieve a win-win result: reducing toxic metal exposure and improving Ca nutrition. Although this study focuses on indoor dust, the Ca strategy may also be effective in reducing Pb, Cd, and As exposure from other sources, such as soil ingestion, 80,81 diet (especially rice consumption), [82][83][84] and water drinking, [85][86][87] which warrants future investigations. In addition, the findings may also provide a guide to reducing Pb, Cd, and As bioavailability in livestock as a means to reduce human exposure to these toxicants via consumption of livestock products.
One limitation of this study is that the role of Ca supplements in reducing oral bioavailability of varying Pb, Cd, and As compounds (e.g., Pb carbonate and Pb oxide) with high bioavailability was not elucidated. Also, further studies are needed to assess impacts of Ca supplements on gut microbiota to see the potential contribution of gut microbiota to reduced Pb, Cd, and As bioavailability. Additionally, impacts of other important variables, such as sex, age, and diet, on the performance of Ca supplements deserve investigation. Given potential differences between humans and mice, conducting cohort studies is a foremost priority to assess the effectiveness of Ca supplements for intervention on human Pb, Cd, and As exposure.