Internal dose of vanadium in rats following repeated exposure to vanadyl sulfate and sodium orthovanadate via drinking water

Abstract

Vanadium is a ubiquitous environmental contaminant that exists in multiple oxidation states. Humans are exposed to vanadyl (V⁴⁺) and vanadate (V⁵⁺) from dietary supplements, food, and drinking water and hence there is a concern for adverse human health. The current investigation is aimed at identifying vanadium oxidation states in vitro and in vivo and internal concentrations following exposure of rats to vanadyl sulfate (V⁴⁺) or sodium metavanadate (V⁵⁺) via drinking water for 14 d. Investigations in simulated gastric and intestinal fluids showed that V⁴⁺ was stable in gastric fluid while V⁵⁺ was stable in intestinal fluid. Analysis of rodent plasma showed that the only vanadium present was V⁴⁺, regardless of the exposed compound suggesting conversion of V⁵⁺ to V⁴⁺ in vivo and/or instability of V⁵⁺ species in biological matrices. Plasma, blood, and liver concentrations of total vanadium, after normalizing for vanadium dose consumed, were higher in male and female rats following exposure to V⁵⁺ than to V⁴⁺. Following exposure to either V⁴⁺ or V⁵⁺, the total vanadium concentration in plasma was 2- to 3-fold higher than in blood suggesting plasma as a better matrix than blood for measuring vanadium in future work. Liver to blood ratios were 4–7 demonstrating significant tissue retention following exposure to both compounds. In conclusion, these data point to potential differences in absorption and disposition properties of V⁴⁺ and V⁵⁺ salts and may explain the higher sensitivity in rats following drinking water exposure to V⁵⁺ than V⁴⁺ and highlights the importance of internal dose determination in toxicology studies.

Graphical Abstract

Introduction

Vanadium has generated significant interest in the regulatory community due to its ubiquitous presence and potential impact on human health. It is used in unregulated dietary supplements and low levels of vanadium are also present in the majority of foods and in drinking water (Maughan et al., 2004; NTP, 2008; Trevino et al., 2019). Vanadium-containing compounds are being considered as potential antidiabetic, anticancer, and anti-hypersensitive agents (Trevino et al., 2019). Environmental vanadium contamination sources are predominately metallurgical operations and carbon-based fuel combustion (ATSDR, 2012). The element has recently been detected at relatively high concentrations (up to 400 μg/g soil) in fly ash waste, which may over time leach into the environment (Davison et al., 1974; Ruhl et al., 2010; Hesterberg et al., 2015). It was monitored in public water systems by the United States Environmental Protection Agency (EPA) from 2013–2015 under the Third Unregulated Contaminant Monitoring Rule and was found in 73.6% samples, with 3.3% above the reference concentration of 21 μg/L (EPA, 2017). Currently, vanadium is included in the Fourth Contaminant Candidate List by the EPA in 2016 (Richardson and Ternes, 2014; EPA, 2020).

The chemistry of vanadium is complex. It exists in several oxidation states (from +2 to +5) and a variety of species. The oxidation state and the types of species are a function of the concentration, pH, redox potential, and other factors (Crans et al., 2004; ATSDR, 2012; Mutlu et al., 2017; Gustafsson, 2019; Trevino et al., 2019). For example, in low pH and sub oxic and/or mildly reducing conditions, vanadyl (V⁴⁺) and oxocations (VO²⁺) predominate. Under neutral to high pH and oxic conditions, vanadate (V⁺⁵) and corresponding oxoanions (e.g., H₂VO₄^-, HVO₄^2-) predominate with oligomers of H₂VO₄^- and HVO₄^2- (eg., dimers, trimers, tetramers) formed at neutral pH. Hence, V⁴⁺ and V⁵⁺ are the most prevalent oxidation states and water and soil are the main media of transport of vanadium compounds in the environment (ATSDR, 2012; Trevino et al., 2019).

To date, there have been only a few rigorous studies investigating the potential toxicity of vanadium compounds. Vanadium pentoxide can have carcinogenic effects upon inhalation in rats and mice (NTP, 2002; Ress et al., 2003). There are knowledge gaps regarding the relative toxicity of the two most predominant oxidation states (V⁴⁺ and V⁵⁺) found in water. This led the National Toxicology Program to investigate the potential toxicity of V⁴⁺ and V⁵⁺ salts following exposure via drinking water in rodents (NTP, 2020b; NTP, 2020a). As an initial activity in the program, the vanadium species present in aqueous formulations under different pH and at different concentrations were investigated for V⁴⁺ and V⁵⁺ salts (Mutlu et al., 2017). The data demonstrated that oxovanadate species in solutions of sodium orthovanadate and metavanadate salts are similar and that V⁴⁺ and V⁵⁺ oxidation states are stable under acidic (pH 3.5) and neutral (~ pH 7) conditions, respectively. In addition, in the same investigation, the long-term stability of the oxidation states in dose formulations was confirmed under these conditions to ensure the suitability of the drinking water formulations in repeated dose toxicology studies. Based on the findings of this investigation, vanadyl sulfate and sodium metavanadate were selected as test compounds, representing V⁴⁺ and V⁵⁺ oxidation states, respectively (Mutlu et al., 2017; NTP, 2020b; NTP, 2020a). Male and female rats and mice were exposed to 14 d via drinking water up to 2000 mg/L to sodium metavanadate representing V⁵⁺ oxidation state and up to 1340 mg/L vanadyl sulfate representing V⁴⁺ oxidation state (Roberts et al., 2016). Based on overt toxicity and other clinical observations, V⁵⁺ appears to be more toxic in rodents than V⁴⁺ (Roberts et al., 2016). The purpose of the current investigation is to evaluate the oxidation state of vanadium present and internal dose in plasma, blood, and liver of rats collected from the 14-d studies to put the observed toxicological data into greater context.

Materials and Methods

Materials:

Deionized water (18 MΩ cm⁻¹) was from Pure Water Solutions (Hillsborough, NC, USA). Semiconductor grade methanol, concentrated phosphoric acid, ethylenediaminetetraacetic acid (EDTA) (1M), concentrated ammonium hydroxide, and filter tubes (10 kDa cutoff) were procured from Sigma-Aldrich (St. Louis, MO, USA). Tetrabutylammonium hydroxide (1M, TBA(OH)) was obtained from Fluka, Buchs, Switzerland). Vanadyl sulfate trihydrate (99.5%;) and sodium metavanadate (≥99%) were procured from NOAH Technologies Corporation (San Antonio, TX, USA) and MP Biomedicals, LLC, (Santa Ana, CA, USA), respectively. Optima-grade concentrated high-purity nitric acid and Ultrex-grade and 30% high-purity, non-stabilized hydrogen peroxide were from Fisher (Waltham, MA) and J.T. Baker (Center Valley, PA), respectively. National Institute of Standards and Technology (NIST)-traceable stock solutions of yttrium (Y), scandium (Sc), praseodymium (Pr) and vanadium were from High Purity Standards (Charleston, SC). Glycine (UltraPure Grade) and concentrated hydrochloric acid (Ultrex purity) were from Fisher Scientific (Hampton, NH). FaSSIF/FeSSIF/FsSSGF powder was from Biorelevant.com (London, UK). Control Hsd:Sprague Dawley® SD® rat whole blood, plasma, and liver (BioIVT, Hicksville, NY) and control Sprague Dawley rat plasma samples collected with lithium heparin anticoagulant (BioReclamation IVT, Baltimore, MD, USA) were used for method development. All other materials were obtained from commercial sources.

Instrumentation.

For the analysis, of V⁴⁺ and V⁵⁺ species, a Waters ACQUITY UPLC (Milford, MA, USA) was used for chromatographic separation equipped with an ACQUITY UPLC BEH C₁₈ column (1.7 μm particles; 2.1 X 50 mm, Waters) and Empower 3 V 7.10.00.00 software (Waters). Isocratic elution with a flow rate of 0.6 mL/min and ~ 10,000 psi pressure were used. Preliminary speciation experiments were performed on a Thermo (Bremen, Germany) X-Series II ICP-MS equipped with a concentric glass nebulizer and Peltier-cooled glass spray chamber, measuring ⁵¹V in reaction cell mode with ammonia gas to minimize the interference from the ³⁵Cl¹⁶O mass. All remaining speciation analyses were performed with an Element 2 sector field inductively coupled plasma mass spectrometer (SF-ICP-MS) (Thermo, Bremen, Germany) equipped with an ESI PC3 Peltier-cooled cyclonic spray chamber (Electron Scientific, Appleton, WI, USA) to reduce oxide formation. The SF-ICP-MS was operated using Element ICP-MS software version 3.1.2.242 and tuned by monitoring the ⁵¹V signal to optimize instrument response. Analysis was performed in medium resolution mode to resolve the spectral interference from ³⁵Cl¹⁶O mass. Chromatographic data were collected using the Xcalibur Export Plugin V 1.0 and processed using Xcalibur V 2.0.6 (Thermo, Franklin, MA, USA).

For analysis of total V in samples collected as part of the 14-day toxicity study, a DigiPREP graphite heating block from SCP Science (Champlain, NY) was used for sample digestion prior to analysis. Total V was measured in study samples on the X-Series II ICP-MS system described above.

14-Day Toxicity Study Design

All drinking water formulations of vanadyl sulfate and sodium metavanadate were solutions and were stable during study period as described in Mutlu et al. (Mutlu et al., 2017). The 14-d study design details and toxicity data are published previously and hence will not be given in details here (Roberts et al., 2016; Roberts et al., 2018). All study data can be found in CEBS (CEBS data). Briefly, groups of 5 male and 5 female Hsd:Sprague Dawley® SD® rats (5 animals/cage) were exposed to vanadyl sulfate (0, 83.8, 167.5, 335, 670, and 1340 mg vanadyl sulfate/L, ~ pH 3.5) and sodium metavanadate (0, 125, 250, 500, 1,000, and 2,000 mg sodium metavanadate/L, ~ pH 7) via drinking water for 14 d. Blood was collected in K₃EDTA from up to 5 animals per group from 0 (control), 83.8, 335, and 1340 mg/L vanadyl sulfate groups and 0 (control), 125, and 500 mg/L sodium metavanadate groups on the morning of the study termination on day 15 for the determination of vanadium species and internal dose of vanadium. Samples for the 2000 mg/L sodium metavanadate exposure group were not available due to early deaths of animals in this group. Approximately 2 mL of blood was frozen at −80 °C immediately following collection. The remainder of the blood was centrifuged, and the resulting plasma was stored at −80°C. While still under anesthesia, rats were humanely terminated by exsanguination and livers were collected and snap frozen in liquid nitrogen and transferred to storage at −80°C until analysis.

Analysis of V⁴⁺ and V⁵⁺ in rat plasma

The method previously used to quantitate V⁴⁺ and V⁵⁺ in aqueous samples was adapted (Kilibarda et al., 2013). Briefly, mobile phase for speciation was prepared at 18 mM EDTA, 0.50 M TBA⁺, 20 mM phosphoric acid, and 4% methanol and adjusted to pH ~ 7 with concentrated ammonium hydroxide. Solvent calibration standards were prepared daily by dissolving vanadyl sulfate trihydrate and sodium metavanadate in mobile phase over the concentration range of 0.25–15 ng V/mL and were used for analyte quantitation. Solvent quality control (QC) standards were prepared at a target concentration of 6 ng/mL V⁴⁺, V⁵⁺, and combined V⁴⁺ and V⁵⁺.

Plasma and QC samples were prepared using a method adapted from Levine, et al. (Levine et al., 2015) in control Sprague Dawley rat plasma. Briefly, 0.25 mL of plasma was aliquoted into an acid-washed, glass centrifuge tube and mixed with 2.25 mL of eluent (or eluent and spiking solution mixture in the case of matrix QC samples). Samples were then gently mixed by vortex and centrifuged at ~3,170 × g for ~7 minutes. Supernatants were transferred to 10 kDa cutoff filter tubes and centrifuged again at ~3,170 × g for ~20 minutes. The resulting filtrate was analyzed by SF-ICP-MS.

As a pilot investigation evaluating vanadium species in samples collected from animals exposed to V⁴⁺ and V⁵⁺, 12 plasma samples (n=1/exposure concentration/sex) were prepared (except no vanadium was added) and analyzed similarly to the procedure described above. Solvent calibration standards were prepared as similar to described above over the concentration range of 0.25–15 ng V/mL which is equivalent to 2.5–150 ng V/mL plasma. Plasma QC samples were prepared at target concentrations of 5 and 100 ng V/mL plasma for both V⁴⁺ and V⁵⁺.

Investigation of V⁴⁺ and V⁵⁺ oxidation state stability in vitro

The stability of V⁴⁺ and V⁵⁺ oxidation states under the conditions of the gastrointestinal (GI) tract was investigated using synthetic gastric and intestinal fluid spiked with respective V⁴⁺ and V⁵⁺ species. Synthetic gastric fluid was prepared by dissolving glycine and concentrated hydrochloric acid in deionized water to produce a final concentration of 0.42 M hydrochloric acid and 0.40 M glycine at a pH of 1.5 (Bradham et al., 2011). FaSSIF/FeSSIF/FsSSGF powder was used to prepare synthetic intestinal fluid at pH ~6.50 (Galia et al., 1998). Triplicate samples were prepared by diluting individual and combined V⁴⁺, V⁵⁺ stock solutions in synthetic gastric and intestinal fluids at nominal 15 ng/mL and 75 ng/mL concentrations, respectively. At selected times up to 48 h, samples were diluted 5-fold with the mobile phase (18 mM EDTA, 0.50 M TBA+, 20 mM phosphoric acid, and 4% methanol adjusted to pH ~ 7) to quench interconversion of oxidation states and analyzed by SF-ICP-MS.

Samples were also analyzed for total vanadium by ICP-MS after being stored for 456 h in gastric fluid and 216 h in intestinal fluid samples. These time periods were not intended to mimic biological transit time in animals, but rather to provide a measure of the total metal concentration after prolonged storage at the simulated gastric and intestinal conditions as a measure of total vanadium stability. Stock solutions were prepared using a NIST-traceable 10 μg V/mL stock solution. A NIST-traceable stock solution of Pr was also used as the internal standard source. Calibration standards were prepared over a range of 0.05 – 25 ng V/mL in 5% (v/v) nitric acid-matrix match digested samples. Samples were prepared in a clean plastic hood to prevent contamination and were digested in a graphite digestion block. Gastric fluid samples were digested by addition of 0.5 mL of concentrated nitric acid to 2 mL of each sample, then heating at 90 °C for 1 h. Samples were then spiked with internal standard and diluted to volume with deionized water. Intestinal fluid samples were digested by addition of 0.5 mL concentrated nitric acid to 0.2 mL of each sample, then heating at 90 °C for 1 h. Samples were then spiked with internal standard, Pr, and diluted to volume with deionized water. Samples were further diluted by a factor of 11 to alleviate matrix effects. Matrix blanks were prepared the same as the gastric and intestinal fluid samples and analyzed alongside them to monitor background vanadium concentration.

Analysis of total vanadium concentration following exposure of male and female rats to vanadyl sulfate and sodium metavanadate

A method was qualified to measure total vanadium in blood, plasma and liver. Stock standards at 100 ng V/mL were prepared using commercial vanadium standards by diluting in 2.5 mL nitric acid and deionized water to 50 mL. Internal standards (Y, SC, and Pr) were prepared similarly at 100 ng/mL. Solvent calibration standards were prepared from this as appropriate using the same diluents to give standards in the calibration range 0.05–25 ng/mL. Matrix QC samples were prepared in blood (40 and 750 ng V/mL blood) and plasma (40 and 400 ng V/mL plasma) by aliquoting 0.25 mL of the blank matrix into 15-mL tubes and fortifying with vanadium standard solutions to target concentrations as given. Liver QC samples were prepared by transferring liver (0.1962 to 0.4989 g) to 15-mL plastic tubes and fortifying with vanadium to contain target concentrations of 0.04 and 0.40 ng V/mL extract (corresponding tissue concentrations ranged from 20 to 442 ng/g depending on the mass of liver used).

To blood and plasma QC samples, 0.5 mL concentrated nitric acid and 0.25 mL 30% hydrogen peroxide were added and digestion tubes were lightly capped and allowed 30 min of pre-digestion time at ambient temperature in a clean plastic hood. Samples were digested in a heating block at 75°C for 60 min and allowed to cool to room temperature. Internal standard solution of 0.1 mL containing Y and Sc or Pr was added and all tubes were diluted to a final volume of 10 mL with deionized water, capped, and mixed by vortex for analysis. For liver samples, 0.5 mL concentrated nitric acid was added and digestion tubes were lightly capped and allowed 30 min of pre-digestion time at ambient temperature in a clean plastic hood. Samples were digested in a heating block at 95°C for 30 min, digests cooled to room temperature, and 0.25 mL 30% hydrogen peroxide was added. Tubes were returned to the graphite heating block and further digested at 95°C for 60 min. Internal standard was added as described for plasma, and all tubes were final volume of 10 mL with deionized water, capped, and mixed by vortex. Liver digests were further diluted 1:25 prior to analysis. Study samples and method blanks were prepared similar to QC standards described above except no vanadium standard was added.

All samples were analyzed by ICP-MS. Calibration curves were generated by plotting the vanadium signal corrected for Y internal standard signal as a function of concentration. Internal standard correction was done by monitoring the signal of a constant concentration of internal standard in all standards and samples and calculating the ratio of internal standard signal in each sample to the signal from the initial blank. This ratio was used to adjust the intensity of the analyte signal, which accounts for instrument drift over time. A linear regression was used to relate signal intensity to analyte concentration. The concentration of vanadium in samples were calculated using the linear regression, dilution, and sample volume/weight. Blood and plasma data were reported as ng/mL blood or plasma and liver data were reported as ng/g liver.

Results

Pilot study sample analysis for V⁴⁺+ and V⁵⁺

Solvent calibration standards showed baseline separation of V⁴⁺ and V⁵⁺ in ~1 min. A representative chromatogram for a solvent calibration standard is shown in Figure 1 showing rapid elution of V⁴⁺ and V⁵⁺ with baseline separation. Calibration curves for both V⁴⁺ and V⁵⁺ oxidation states had correlation coefficient values ≥ 0.99 and relative errors (%RE) were within ±13% for both species in all standards in the range of 2.5 – 150 ng/mL. V⁴⁺ and V⁵⁺ were not detected in control plasma, suggesting background levels were negligible.

Figure 1.

Representative chromatogram of a calibration standard at 5 ng/mL V⁴⁺ and V⁵⁺ from Sector Feld-Inductively-Coupled Plasma Mass Spectrometry (SF-ICP-MS) analysis showing separation of V⁴⁺ and V⁵⁺ species.

To determine which V⁴⁺+ and V⁵⁺ species are present in plasma following exposure of animals to vanadyl sulfate and sodium metavanadate via drinking water, a pilot analysis was conducted. Representative study samples (n=1/sex/group) were selected to cover control and exposed animals. In chromatograms, only a peak corresponding to V⁴⁺ was observed, regardless of whether the animals were exposed to vanadyl (V⁴⁺) or vanadate (V⁵⁺), suggesting either conversion of V⁵⁺ to V⁴⁺ in vivo and/or potential instability of V⁵⁺ during plasma storage and sample preparation. The V⁴⁺ concentrations quantified in this pilot investigation are given in Table 1. In vanadyl sulfate (V⁴⁺) exposed animals, V⁴⁺ concentration increased with the exposure concentration of the compound with females having higher levels at the higher exposure concentration than males. Concentrations of V⁴⁺ in sodium metavanadate (V⁵⁺) exposed animals also increased with the exposure concentration although the increase was less than exposure concentration proportional which correlates with the decreased water consumption and subsequently the chemical consumption (Roberts et al., 2016). There was no apparent sex-difference in internal concentration of vanadium following exposure to sodium metavanadate (Table 1).

Table 1.

Concentration of V⁴⁺ and V⁵⁺ species in plasma following exposure of male and female rats via drinking water to vanadyl sulfate (V⁴⁺) and sodium metavanadate (V⁵⁺) for 14 d^a.

Test Compound	Sex	Exposure Concentration (mg/L)	V4+ Concentration (ng/mL)
Vanadul sulfate (V4+)	M	0	NDb
	M	83.8	48.7
	M	1340	625
	F	0	ND
	F	83.8	18.9
	F	1340	1440
Sodium metavanadate (V5+)	M	0	ND
	M	125	147
	M	500	284
	F	0	ND
	F	125	124
	F	500	281

^aV⁵⁺ species were not detected in any sample. Values given are n=1/sex/exposure concentration.

^bND, not detected.

Stability of V⁴⁺ and V⁵⁺ oxidation states in vitro

To investigate whether the absence of V⁵⁺ species in sodium metavanadate exposed animals were due to disposition in vivo and/or instability of species during and following sample collection and analysis, a series of in vitro studies were conducted using conditions mimicking the GI conditions by spiking V⁴⁺, V⁵⁺ or combined V⁴⁺ and V⁵⁺. Data are presented as mean determined concentration and percent relative error (RE) estimated using nominal and determined concentrations.

Data for V⁴⁺ and V⁵⁺-spiked simulated gastric fluid samples are given in Table 2. Figure 2A illustrates the change in composition overtime in V⁵⁺ spiked samples. The presence of a peak representing V⁴⁺ in sample of V⁵⁺ at 0 h demonstrates that the some V⁵⁺ conversion happens during the sample preparation stage and/or analysis; the calculated concentration was 36% below the nominal value (Table 2 and Figure 2A). By 48 hours, ~ 95% of V⁵⁺ present in simulated gastric fluid was converted (RE −95%) (Table 2). A comparison of total vanadium concentration measured in the samples (13.9 ng/mL) to a sum of the V⁴⁺ and V⁵⁺ species measured (average 0–48 h value ~ 11.6 ng/mL) indicates that a small fraction of vanadium is not accounted for by the sum of the V⁴⁺ and V⁵⁺ species (Table 2). To further illustrate the behavior of V⁵, representative chromatograms from synthetic gastric fluid spiked with both V⁴⁺ and V⁵⁺ are shown in Figure 2B clearly demonstrating the decrease in V⁵⁺ and corresponding increase in V⁴⁺ over time. By 48 h, the peak corresponding to V⁵⁺ was close to the detection limit. In contrast, in V⁴⁺ spiked gastric fluid samples, no other peaks were observed in the chromatograms; the RE estimated for V⁴⁺ were ≤ ± 14% at each time point demonstrating that V⁴⁺ is stable in gastric fluid (Table 2). The concentration determined as V⁴⁺ and total vanadium are similar (Table 2).

Table 2.

Measured concentration of V⁴⁺ and V⁵⁺ in gastric fluid over time^a.

Vanadium Species Spiked (Concentration)	Time (hours)	Mean Determined V4+ (ng V/mL)	V4+ % REb	Mean Determined V5+ (ng V/mL)	V5+ % RE	Mean Total [V] (ng V/mL)c	Total V % RE
V4+ (15 ng/mL)	0	15.2 ± 0.1	1.1	0.44 ± 0.03	NAe	15.6	4.0
	4	14.4 ± 0.1	4.3	NDd	NA	14.4	−4.3
	23	14.8 ± 0.3	1.6	ND	NA	14. 8	−1.3
	48	12.9 ± 0.1	−14	ND	NA	12.9	−14
	456g					15.0 ± 0.3	0.33
V5+ (15 ng/mL)	0	3.4 ± 0.2	NA	9.6 ± 0.1	−36	13. 0	−13
	4	6.1 ± 0.2	NA	4.5 ± 0.1	−70	10. 6	−29
	23	10.1 ± 0.2	NA	2.0 ± 0.1	−87	12. 1	−19
	48f	9.9 ± 0.3	NA	0.72 ± 0.02	−95	10. 6	−29
	456g					13.9 ± 0.3	−7.4
V4+ and V5+ (15 ng/mL each)	0	17.0 ± 0.2	13	10.6 ± 0.2	−30	27.6	−8.0
	4	19.9 ± 0.2	33	5.3 ± 0.1	−65	25.2	−16
	23	24.7 ± 0.1	65	2.4 ± 0.1	−84	27. 2	−9.5
	48	22.0 ± 0.4	47	1.2 ± 0.1	−92	23. 1	−23
	456g					29.5 ± 0.3	−1.8

^aValues given are average ± standard deviation (n=3).

^bRE, percent relative error estimated using the nominal and determined concentration.

^cMean Total [V] (ng V/mL) is calculated as: Mean Determined V(IV) + Mean Determined V(V).

^dND = No peak detected.

^eNA = Not applicable.

^fData are for n=2.

^gMeasurement of total vanadium concentration at t = 456 hours was performed by ICP-MS.

NOTE: Some values may not be able to be replicated due to rounding for significant figures.

Figure 2.

Conversion of V⁵⁺ to V⁴⁺ in gastric fluid. A) Measured concentration of V⁴⁺ and V⁵⁺ over time following spiking with 15 ng/mL V⁵⁺. Error bars represent the standard deviation for n=3. The dashed bar at the top shows total V measured by ICP-MS at t=456 hours. B) Overlaid chromatograms of species formed following spiking with combined V⁴⁺ and V⁵⁺.

Data for V⁴⁺ and V⁵⁺-spiked synthetic intestinal fluid are given in Table 3. Graphical presentations of time-dependent species distribution in V⁴⁺ and V⁵⁺-spiked samples are given in Figures 3A and 3B, respectively. Samples spiked with V⁴⁺ only, showed a RE of −57% at t=0 h indicating loss of V⁴⁺ with a concomitant increase of V⁵⁺, indicating that the loss was due to conversion of V⁴⁺ to V⁵⁺ during sample preparation and/or analysis. By 4 h, the determined concentration of V⁴⁺ was ~ 2% of nominal (RE −98%) demonstrating rapid conversion of V⁴⁺ to V⁵⁺ (Table 3, Figure 3A). In contrast, V⁵⁺ in intestinal fluid was stable over time with RE 0.5–14% (Table 3, Figure 3B); a peak for V⁴⁺ was not present in chromatograms. Collectively, the data demonstrate that V⁵⁺ was stable under the conditions of intestinal fluid. Concentration estimated as a summation of V⁴⁺ to V⁵⁺ were within ≤± 14% (Table 3). Total vanadium concentrations measured by ICP-MS at the end of the experiment were 74.4–90.4 ng/mL (75 ng/mL nominal) for individual spikes and 170 ng/mL combined (150 ng/mL nominal) for combined spikes and was close to the nominal values.

Table 3.

Measured concentration of V⁴⁺ and V⁵⁺ species in intestinal fluid over time^a.

Sample ID	Time (hrs)	Mean Determined V4+ (ng V/mL)	V4+ % REb	Mean Determined V5+ (ng V/mL)	V5+ % RE	Mean Total [V] (ng V/mL)c	Total V % RE
V4+ (75 ng/mL)	0	32.1 ± 1	−57	43.5 ± 1	NA	76.0	1.3
	4	1.33 ± 0.06	−98	72.5 ± 2	NA	73.7	−1.7
	23	NDd	NAe	69.2 ± 0.9	NA	69.2	−7.7
	48	ND	NA	64.1 ± 0.9	NA	64.1	−14
	216a					74.4 ± 2	0.6 7
V5+ (75 ng/mL)	0	ND	NA	85.5 ± 4	14	85.5	14
	4	ND	NA	85.0 ± 2	13	85.0	13
	23	ND	NA	81.5 ± 1	8.4	81.5	8.4
	48	ND	NA	75.0 ± 1	0.5	75.0	−0.1
	216a					90.4 ± 3	21
V4+ and V5+ (75 ng/mL each)	0	29.1 ± 0.5	−61	134 ± 3	79	164	9.1
	4	1.77 ± 0.1	−98	158 ± 1	110	159	6.2
	23	0.990 ± 0	−99	153 ± 2	104	154	2.8
	48	ND	NA	139 ± 3	85	139	−7.6
	216a					170 ± 3	13

^aValues given are average ± standard deviation (n=3).

^bRE, percent relative error.

^cMean Total [V] (ng V/mL) is calculated as: Mean Determined V(IV) + Mean Determined V(V).

^dND = No peak detected.

^eNA = Not applicable.

^fMeasurement of total vanadium concentration at 216 h was performed by ICP-MS.

NOTE: Some values may not be able to be replicated due to rounding for significant figures.

Figure 3.

Conversion of V⁴⁺ to V⁵⁺ in intestinal fluid. Measured concentration of V⁴⁺ and V⁵⁺ in synthetic intestinal fluid samples over time A) 75 ng/mL V⁴⁺-spiked samples B) 75 ng/mL V⁵⁺-spiked samples. Error bars represent standard deviation for n=3. The dashed bar at the top of both plots shows total V measured by ICP-MS at t=216 hours.

Total vanadium concentration following exposure of male and female rats to vanadyl sulfate and sodium metavanadate

Study data can be found at: https://manticore.niehs.nih.gov/cebssearch/paper/15022/private/392054BXYZ (Please note that the reviewer link will be replaces with the public link when the manuscript is accepted for publication by the journal.)

Since V⁵⁺ was not present in plasma samples either due to conversion in vivo and/or instability during sample collection and/or storage, quantitating V⁴⁺ and V⁵⁺ species in samples following exposure of rodents to vanadyl sulfate and sodium metavanadate could not be achieved. Therefore, total vanadium concentration was determined using ICP-MS in plasma, whole blood, and liver. The analytical method used for the analysis of gastric and intestinal fluid samples were qualified to quantitate total vanadium in study samples. The solvent calibration curve range of 0.05–25 ng V/mL corresponds to 10 – 5,000 ng V/mL plasma, 20–10,000 ng V/mL blood, and 50–25,000 ng V/g liver based on the sample preparation and dilution protocol used. Analytical method qualification data are given in Table 4. The method was linear with a correlation coefficient > 0.99. Based on matrix QC samples, the accuracy determined as %RE of the nominal concentration for blood was 7.3–15.1, and precision determined as %RSD was 3.1–3.3. Similarly, the method performance was acceptable for plasma and liver, although for liver the estimated %RE ranged from −4.8 to 27.0 (Table 4).

Table 4.

Analytical method qualification data for total vanadium.

Parameter	Data
Solvent calibration Range (ng/mL)	20–2000
LOD (ng/mL) a	3.6
LOQ (ng/mL) b	20
Correlation Coefficient (r)	>0.99
Blood Intraday Precision (%RSD) c,d	3.1–3.3
Blood Intraday Accuracy (Mean %RE) c,e	7.3 – 15.1
Batch QC sample precision (Mean %RSD) c,d
Blood	2.2 – 3.1
Plasma	2.4 – 2.8
Liver	1.6–7.6
Batch QC sample accuracy (Mean % RE)c,d
Blood	4.7 – 11.7
Plasma	17.4 – 18.0
Liver	−4.8 – 27.0

^aLOD = limit of detection (determined as the 3X standard deviation of LOQ for n=7).

^bLOQ = limit of quantitation.

^cPrecision and accuracy determined for triplicate QCs: blood, 40, and 750 ng V/mL blood; plasma, 40 and 400 ng/mL plasma; liver, 20 – 51.0 ng/g liver low concentration QC samples and 295 – 442 ng/g liver high concentration QC samples.

^d%RSD = percent relative deviation.

^e%RE = percent relative error.

^fND, not determined.

Total vanadium concentration in plasma, blood, and liver following exposure of rats to vanadyl sulfate and sodium metavanadate via drinking water for 14 d are given in Table 5. Dose of compound consumed estimated based on average cage water consumption and exposure concentration and also total vanadium consumed based on % vanadium in compound were previously reported (Roberts et al., 2016; Roberts et al., 2018) and are also presented in Table 5. Plasma, blood, and liver vanadium concentrations normalized to total vanadium consumed (concentration of vanadium in sample per unit dose of vanadium consumed) are also presented in the table for comparison across compounds, exposure concentrations, and sexes. Following exposure of male and female rats to vanadyl sulfate, the concentration of vanadium in blood, plasma, and liver increased more than proportional to the dose of vanadium, as evident from the values normalized to vanadium consumed. In blood, normalized values were 14.2, 12.1, and 29.6 (ng V/mL)/mg V consumed for males and 6.7, 14.6, and 23.7 (ng V/mL)/mg V consumed for females, for 83.8, 335, and 1340 mg vanadyl sulfate/L exposure concentrations. A similar pattern was observed in plasma and liver. The concentration in plasma was 2- to 3-fold higher than in whole blood. The values in liver were the highest with normalized values of 72.2, 80.9, 141 and 45.3, 67.3, and 141 (ng V/g)/mg V consumed for males and females, respectively. In general, there was no apparent sex-difference in vanadium concentrations in male and female rats.

Table 5.

Concentration of total V in whole blood, plasma, and liver of rats following exposure to vanadyl sulfate (V⁴⁺) and sodium metavanadate (V⁵⁺) via drinking water for 14 d^a.

Sex	Exposure Concentration (mg/L)	Dose (mg/kg)b	Vanadium Consumed (mg/day)c	Blood (ng V/mL)	Plasma (ng V/mL)	Liver (ng V/g)	Blood (ng V/mL)/(mg V consumed)d	Plasma (ng V/mL)/(mg V consumed)d	Liver (ng V/g)/(mg V consumed)d
Vanadyl sulfate (V4+)
Male	0	NAe	NA	BDf	BD	BD	NA	NA	NA
	83.8	8.7	2.7	38.4 ± 5.31	74.6 ± 9.69	195 ± 23.9	14.2	27.6	72.2
	335	30.9	9.7	117 ± 10.6	223 ± 14.0	785 ± 58.9	12.1	23.0	80.9
	1340	51.9	16.3	483 ± 95.7	914 ± 172	2290 ± 263	29.6	56.1	141
Female	0	NA	NA	BD	BD	BD	NA	NA	NA
	83.8	10.2	3.2	21.3 ± 3.89	37.0 ± 6.39	145 ± 12.1	6.70	11.6	45.3
	335	34.6	10.9	159 ± 30.0	282 ± 51.5	734 ± 85.6	14.6	25.9	67.3
	1340	53.7	16.8	398 ± 206	1200 ± 534	2360 ± 459	23.7	71.4	141
Sodium metavanadate (V5+)
Male	0	NA	NA	BD	BD	BD	Na	NA	NA
	125	12.9	5.4	173 ± 12.7	297 ± 17.9	1040 ± 85.2	32.0	55.0	193
	500	39.1	16.3	445 ± 46.6	759 ± 79.9	2700 ± 294	27.3	46.6	166
Female	0	NA	NA	BD	BD	BD	NA	NA	NA
	125	14.5	6.0	189 ± 12.4	341 ± 31.4	750 ± 44.0	31.5	56.8	125
	500	42.5	17.7	586 ± 105	1050 ± 202	2400 ± 284	33.1	59.3	136

^aFull data sets can be found in Chemical Effects in Biological Systems : https://manticore.niehs.nih.gov/cebssearch/paper/15022/private/392054BXYZ. Data given are average and ± standard error for up to n=5 animals.

^bData were originally reported in Roberts, et al. 2016 and 2018 (8,9). Values given for dose (mg vanadium salt/kg) are group averages estimated from water consumption (and subsequently the chemical consumption) and exposure concentration), and animal body weight.

^cData were originally reported in Roberts, et al. 2016 and 2018 (8,9). Total vanadium consumed was estimated from chemical consumption and assuming vanadium comprises 31% (w/w) of vanadyl sulfate and 41.7% (w/w) sodium metavanadate.

^dConcentrations normalized to vanadium consumed are presented.

^eNA, not applicable.

^fBD = Below the limit of detection (Whole blood = 3.60 ng/mL; Plasma = 1.80 ng/mL; liver = 1.64 ng/g).

Following drinking water exposure to vanadium via sodium metavanadate at 125 and 500 mg/L in male and female rats, the concentration of vanadium in blood, plasma, and liver increased proportional to the dose of vanadium consumed (Table 5). Normalized concentrations in blood were 32.0 and 27.3 (ng V/mL)/mg V consumed in males and 31.5 and 33.1 (ng V/mL)/mg V consumed in females demonstrating lack of sex difference. A similar pattern was observed in plasma. The concentration in plasma was 2-fold higher than in whole blood. The values in liver were the highest with normalized values of 193 and 166 (ng V/g)/mg V consumed for males and 125 and 136 (ng V/g)/mg V consumed for females demonstrating slightly higher values in males compared to females (Table 5).

Discussion

The complex chemistry of vanadium makes the studies investigating the disposition of V⁴⁺ and V⁵⁺ compounds in biological systems very challenging. A recent report detailed an UHPLC method for the rapid speciation of V⁴⁺ and V⁵⁺ with high peak resolution, designed for fresh water samples (Kilibarda et al., 2013). We adapted this method for plasma samples to investigate V⁴⁺ and V⁵⁺ species and internal dose in animals exposed via drinking water to vanadium salts. A pilot investigation indicated that only V⁴⁺ was present in plasma regardless of the vanadium compound used. These results are consistent with the literature reports that V⁴⁺ is the most stable form in biological matrices (Chasteen et al., 1986; Yasui et al., 2000; Frausto da Silva and Williams, 2001; Trevino et al., 2019).

In order to assess whether this observation was due to conversion of V⁵⁺ to V⁴⁺ in vivo or during sample storage and/or analysis, we designed a series of in vitro studies in simulated gastric and intestinal fluids. Analysis of synthetic gastric fluid samples spiked with vanadium salts suggests that V⁵⁺ is converted to V⁴⁺ under these conditions. Starting at 4 h, a third peak was observed as a shoulder on the V⁵⁺ peak, which may correspond to another vanadium species or complex that has not yet been identified. Its concentration was integrated along with the measured V⁵⁺ concentration in all samples as it was not sufficiently resolved to quantify separately. Conversion of V⁵⁺ to V⁴⁺ at low pH was not observed in tap water formulations at pH 3.5 (Mutlu et al., 2017) and hence the phenomenon observed in the gastric solution may be due to the lower pH (1.5) of the medium. In the simulated intestinal fluid, the opposite behavior was observed where V⁵⁺ was stable and V⁴⁺ was converted to V⁵⁺. This speciation behavior is consistent with previous studies where oxidation state and species present were shown to depend on the pH, concentration, and oxic conditions with V⁴⁺ predominating at lower pH and sub-oxic and/or mildly reducing conditions while V⁵⁺ species predominate at higher pH and oxic conditions (ATSDR, 2012; Mutlu et al., 2017; Gustafsson, 2019; Trevino et al., 2019).

The results of the pilot plasma and in vitro experiments point to challenges in vanadium speciation in biological samples. Hence, to determine internal exposure following exposure to vanadium compounds, we qualified a method using ICP-MS to quantitate total vanadium. Using this method, blood, plasma, and liver concentrations of vanadium following repeated exposure of male and female rats to vanadyl sulfate and sodium metavanadate were determined as a part of studies investigating the toxicity of vanadium salts (Roberts et al., 2016; Roberts et al., 2018). Early deaths were observed in 2000 mg/L exposure group of sodium metavanadate studies and hence biological samples were not available.

Lower water consumption was observed in exposed groups relative to control group for both compounds and at nearly all concentrations likely due to palatability (Roberts et al., 2016; Roberts et al., 2018). For example, at the highest sodium metavanadate concentration reported here (500 mg/L), the water consumption was 31.2–32.3% lower than the corresponding controls. At the highest vanadyl sulfate concentration of 1340 mg/L, the water consumption was 77.2–78.5% lower than the control group (Roberts et al., 2016; Roberts et al., 2018). The daily vanadium consumed (mg V/day) per unit compound exposure in drinking water (mg compound/L) estimated using data in Table 5 for sodium metavanadate (0.0326–0.0432 (mg V/day)/(mg compound/L)) were only marginally higher compared to vanadyl sulfate (0.0122–0.0322(mg V/day)/(mg compound/L)).

Following exposure to vanadyl sulfate (V⁴⁺), total vanadium concentrations in blood, plasma, and liver increased more-than proportional to the vanadium dose at the highest exposure concentration in males and in mid and high exposure concentrations in females as evident by the values normalized to vanadium consumption (Table 5). There was no apparent sex difference in concentration in any of the matrices investigated. Plasma concentrations were 2- to 3-fold higher than blood suggesting that the amount of vanadium crossing the red blood cell membrane was negligible. This observation combined with the complexity of use of blood in analytical methods point to plasma as a better matrix to assess systemic exposure to vanadium. Liver to blood ratios were approximately 5–7 demonstrating retention of vanadium in the liver. Unlike following exposure to V⁴⁺, following exposure of male and female rats to vanadium via sodium metavanadate (V⁵⁺), the internal dose of total vanadium increased linearly with the vanadium dose consumed. As with V⁴⁺ exposure, liver had the highest concentration followed by plasma and blood. Liver to blood ratios were ~ 6 in males and ~ 4 in females. Distribution of vanadium to tissues including bone in animals following oral exposure has been reported previously (Parker and Sharma, 1978; Nakai et al., 1995; Edel et al., 2006). Following exposure of rats to sodium metavanadate or vanadyl sulfate to 50 ppm in drinking water, kidney had the highest concentration followed by bone, liver, and muscle. As observed in the current study, the vanadium concentrations measured using flame atomic absorption spectrometry were higher following administration of sodium metavanadate than following exposure to similar levels of vanadyl sulfate (Parker and Sharma, 1978).

Based on overt toxicity and other clinical observations following exposure via drinking water for 14 d, sodium metavanadate (V⁵⁺) appears to be more toxic in rodents than vanadyl sulfate (V⁴⁺) (Roberts et al., 2016). Total vanadium intake estimated based on the water consumption didn’t explain the differential toxicity observed. In our investigation, in general, the concentration of total vanadium in blood, plasma, and liver of rats for a unit dose of vanadium consumed was higher following exposure to V⁵⁺ than V⁴⁺. This suggests potential differences in the absorption, distribution, metabolism, and excretion ADME) and toxicokinetic (TK) properties of V⁴⁺ and V⁵⁺ compounds leading to the observed differences in the internal dose of total vanadium in rats. Differential ADME properties of V⁴⁺ and V⁵⁺ compounds have been suggested based on data from limited studies and is summarized in Trevino et al. (Trevino et al., 2019). It has been suggested that V⁵⁺ compounds (i.e., corresponding oxoanions are H₂VO₄^-, HVO₄^2-) are partially reduced to V⁴⁺ (i.e. oxocation VO₂⁺) in the stomach following oral exposure, and in the intestine under slightly basic conditions forms a sparingly soluble vanadium oxohydroxide, VO(OH)₂. However, oxoanioans H₂VO₄^- and HVO₄^2- are absorbed 3 to 5 times higher than VO₂⁺ leading to higher systemic vanadium levels following exposure to V⁵⁺ than V⁴⁺. The extent to which these species are formed is also determined by other factors such as changes in redox conditions in the local environment and presence of food, enzymes, or other biomolecules (Trevino et al., 2019). In blood, vanadium species bind to serum proteins during transport, particularly albumin and transferrin (Sanna et al., 2009a; Sanna et al., 2009b). Vanadate/vanadyl can undergo redox reactions in extracellular fluid as mentioned above. Vanadates are present in the blood as either H₂VO₄^- or HVO₄^2- due to its microenvironment and are not expected to form oligovanadates (Pessoa et al., 2015). In intracellular space, the speciation again can vary depending on the specific physiological conditions and hence interconversion between species and oxidation states are likely constantly happening depending on the microenvironment. Taken collectively, these make investigating ADME behavior of vanadium species in vivo complex and difficult to predict. Regardless, the difference in internal total vanadium concentration observed in the current investigation, in part, may explain the observed differences in toxicity in rodents following exposure to V⁴⁺ and V⁵⁺ salts and highlights the importance of determining the internal dose in toxicology studies. However, it should be noted that the concentrations presented in the current investigation are for a single timepoint following exposure and may not reflect the totality of exposure for direct and fair comparisons between V⁴⁺ and V⁵⁺ salts and highlights the importance of generating data from well-designed TK and ADME studies.

Conclusions

Vanadium speciation may play a significant role in its toxicity, but little is known about the speciation in vivo. Measurement of vanadium species with regards to the oxidation state following exposure of rats to vanadyl sulfate (V⁴⁺) or sodium metavanadate (V⁵⁺) showed only the presence of V⁴⁺ in plasma suggesting conversion of V⁵⁺ to V⁴⁺ in vivo and/or instability of V⁵⁺ species. Based on total vanadium concentration, the internal exposure to vanadium was higher in male and female rats following exposure via a V⁵⁺ salt compared to V⁴⁺ and is consistent with the higher sensitivity observed in rats following exposure to sodium metavanadate (V⁵⁺) than to vanadyl sulfate (V⁴⁺). These data point to potential differences in absorption and disposition properties of V⁴⁺ and V⁵⁺ salts and highlights the importance of internal dose determination in toxicology studies.

Figure 4.

Conversion of V⁴⁺ to V⁵⁺ in intestinal fluid. Overlaid chromatograms of species formed following spiking with combined V⁴⁺ and V⁵⁺ at 150 ng/mL. (A) t = 0 h, and (B) t = 4 h. Key: black solid – V(⁴⁺); gray solid – V(⁵⁺); black dashed – both V⁴⁺ and V^5+.

Highlights:

• Vanadium is an environmental contaminant and exists in V⁺⁴ and V⁺⁵ oxidation states.

• V⁺⁵ was converted to V⁺⁴ in rodents following exposure via drinking water.

• Total vanadium in blood and tissues per unit exposure was higher for V⁺⁵ than V⁺⁴.

• Data explains higher toxicity of V⁵⁺ compared to V⁴⁺ following exposure in rodents.

• Data highlights the importance of internal dose determination in toxicology studies.