Introduction

During a radiological/nuclear (R/N) emergency, the ability to rapidly triage individuals exposed to potentially life-threatening radiation has been identified as one of the major gaps in radiological medical countermeasures (Blakely et al. 2005). Development of high-throughput, noninvasive and rapid bio-monitoring tools for the screening of mass populations in the event of such a catastrophe has become an obvious need (Lanz et al. 2009). Urine is an ideal noninvasive specimen for bioassay. Recent efforts have been made on the development of high-throughput screening tools for studying urine metabolomics to identify biomarkers for exposure to ionizing radiation (IR) (Lanz et al. 2009; Tyburski et al. 2009a, b; Johnson et al. 2011, 2012). Metabolomics is a recent “omics” technology that focuses on the study of metabolites, which are usually defined as small molecules with a mass less than 1000 Da (Pernot et al. 2012). In these high-throughput metabolomics assays, a number of low molecular weight (<1000 Da) metabolites have been identified as biomarkers of IR exposure in the urine samples from mice, rats and non-human primates (Lanz et al. 2009; Tyburski et al. 2009a, b; Johnson et al. 2011, 2012). In addition, urine contains a large array of proteins that reflect not only the physiology of the kidney and the urogenital tract, but also that of blood (Pieper et al. 2004). Protein in urine originates from glomerular filtration of plasma, excretion from epithelial cells and casts, and formation of urinary exosomes (Barratt and Topham 2007). Therefore, changes in urinary protein concentrations may report directly on the dysfunction of cells within the urinary tract, whereas other problems may be detectable through the transmission of biomarkers from blood into urine. As it is composed of filtered plasma proteins, and proteins from the kidney and other genitourinary organs, urine proteome has been hypothesized to be a useful source of biomarkers for IR exposure (Pernot et al. 2012; Sharma et al. 2010). Both total body irradiation and local kidney irradiation of rats with a single 10-Gy dose of X-rays showed significant changes in urine proteome within 24 h after irradiation (Sharma et al. 2008, 2010). There has been a renewed interest in studying urinary biomarkers to be used as a rapid screening tool for IR exposure in an R/N emergency. However, most of the recent studies have been focused on exposure to external radiation, such as gamma radiation (Lanz et al. 2009; Tyburski et al. 2009a, b; Johnson et al. 2011, 2012) and X-rays (Sharma et al. 2008, 2010). The present study aims to investigate some urinary proteins as a possible clinical biomarker for internal exposure.

For internally deposited radioactive materials, traditional biomarkers that are sensitive to uniform external radiation fields are of little use because of the non-uniform distribution of the radiation dose and the high risk to the organs that are not easily sampled or characterized by the available biological fluids (Brooks 2001; Durante 2003). On the other hand, for internal contamination, it may be possible to develop biomarkers that reflect changes in the organs that are targets for deposition of the radioactive materials (Brooks 2001). In the present study, polonium-210 (210Po) was chosen as the isotope of internal exposure since (a) it has been identified by the International Atomic Energy Agency as one of the radionuclides that can cause harm to human health if used in a malicious activity such as a terrorist attack using a radiological dispersion device (IAEA 2005), (b) it was used to deliberately poison Mr. Alexender Litvinenko in London, UK, in November 2006 (Parkins 2007; Harrison et al. 2007) and (c) our previous experience in studying the metabolism of 210Po in rats (Li et al. 2010, 2011a, b; Sadi et al. 2011).

Materials and methods

Chemicals and standards

210Po standard (PoCl4 in 2 M HCl) was purchased from Eckert and Ziegler Isotope Products (Valencia, California, USA). Conversion of 210Po from chloride to citrate form was carried out as described previously (Li et al. 2010). Sodium chloride, thiomersal and Tween-20 were purchased from Sigma-Aldrich Canada (Oakville, ON, Canada).

Animal experiments

Animal experiments were carried out in the Biological Research Facility, Canadian Nuclear Laboratories, Atomic Energy of Canada Limited (AECL), Chalk River, ON, Canada. Male Wister rats, 49–52 days of age, were purchased from Charles River Canada (Saint Constant, Quebec, Canada). The rats were acclimatized in metabolic cages for 11 days prior to the commencement of the experiment. For the control set, a group of five rats were each intravenously injected (from the tail vein) with 0.3 ml of sterile saline. For the exposed sets, each of three groups (five rats in each group) was intravenously injected with 1601 ± 376 Bq (referred hereafter to as low dose level), 10,846 ± 591 Bq (medium dose level) or 48,467 ± 2812 Bq (high dose level) of 210Po in citrate form (0.3 ml prepared in sterile saline), respectively. The administered activity was calculated based on the weight of the injected polonium citrated solution. After injection, the rats were kept in metabolic cages (each rat in a separate cage) and urine samples were collected every 24 h over 4 days. The absorbed dose to a rat liver over 4 days corresponding to the low, medium and high levels of injected activity was estimated to be 7, 45 and 203 mGy, respectively. This estimation was based on the assumptions that (a) the average body weight of a Wistar (male) rat was taken to be 350 g, (b) the average liver weight was taken to be 3 % of the total body weight, and (c) the cumulated activity of 210Po in the liver over 4 days was considered to be 60 % of the injected activity. The cumulated activity was derived from a previous animal experiment study for binding of 210Po in rat tissues (Lanzola et al. 1973). The weights of each rat, their daily consumption of food and water, and daily excretion of urine and feces were monitored to ensure that they were maintained in good health. All animal experiments were approved by the AECL Animal Care Committee.

Sample collection and pre-treatment

Pre-analytical handling and storage condition of bio-specimens (blood, serum, urine, saliva, cerebrospinal fluid and bronchial lavage fluid) can often influence stability of the biomarkers (Hubel et al. 2011). Urine bio-specimens stored at low temperature (−20 to −80 °C) are known to form a precipitate (mainly composed of calcium oxalate dihydrate) after thawing to room temperature (Saetun et al. 2009). In addition to calcium ions, these precipitates have been shown to deplete urinary proteins during storage at −20 °C in as short as 12 h (Vittinghus 1990). The addition of preservatives has been demonstrated to adequately preserve urinary proteins for an extended period subsequent to the freeze/thaw process (Vittinghus 1990; Tencer et al. 1994, 1997). Thiomersal (a bacteriostatic factor) and Tween-20 (a non-ionic surfactant factor) used in the present study have been shown to adequately preserve β2-microglobulin, orosomucoid, albumin, transferrin and immunoglobulin G (IgG) in urine samples for their subsequent analysis with an enzyme-linked immunosorbent assay (ELISA) (Vittinghus 1990).

The urine sample collected daily from each rat was transferred from the urine collection tube of the metabolic cage to a 50-ml polypropylene centrifuge tube. The volume and weight of the urine sample were recorded. The urine samples were centrifuged at 1500g for 10 min to remove any suspended/colloidal material. A 5 ml aliquot of a pre-centrifuged urine sample was then transferred to a 15-ml polypropylene tube, and a preservative solution (a mixture of 1 ppm thiomersal and 1 % Tween-20 in deionized water) was added at 10 % (v/v). It was stored at 4 °C in a cold room until analyzed for creatinine concentration and measured for 210Po. A second aliquot (5 ml) of the pre-centrifuged urine sample was transferred to an Amicon® Ultra-15 centrifugal filter device (3000 NMWL, Millipore Canada Ltd., Etobicoke, ON, Canada). Ten milliliters of deionized water was added to the urine in the filter device, capped, gently shaken to mix well and mounted on a centrifuge. Ultrafiltration of the urine sample was carried out according to the desalting procedure described by the manufacturer in the user guide. The ultra-filtered concentrate (~1 ml) collected from the filter device was quantitatively transferred to a 15-ml polypropylene tube. The filter device was rinsed twice each with a 0.5 ml aliquot of 11.2 mM sodium chloride solution, and rinse solutions were added to the ultra-filtrated concentrate. A preservative solution (a mixture of 1 ppm thiomersal and 1 % Tween-20 in deionized water) was added to the ultra-filtered concentrate at 10 % (v/v). It was stored at 4 °C in a cold room until analyzed for urinary proteins and measured for 210Po. During the shipment of the samples (from the Canadian Nuclear Laboratories, Chalk River, ON, Canada, to the Radiation Protection Bureau, Ottawa, ON, Canada), the sample containers were packed in dry ice.

Measurement of 210Po in the urine samples

A 0.25 ml aliquot from each of the preserved urine samples (both whole urine and ultra-filtered concentrate) was transferred to a liquid scintillation vial (polypropylene, 20 ml) and mixed with 20 ml of a liquid scintillation cocktail (Optiphase HiSafe 3, PerkinElmer, Inc., Woodbridge, ON, Canada). Measurement of 210Po was carried out on a Hidex 300 SL liquid scintillation counter (LSC) with α/β discrimination capabilities (Hidex Oy, Turku, Finland) with a counting time of 1 h. Calibration of the LSC for the counting efficiency (disintegration per second/counts per second) was carried out by a second measurement of each sample after addition of 0.05 ml aliquot of a known activity of 210Po standard. Poisson counting statistics was used for the measurement of counting uncertainty. The measurement uncertainty was calculated from the combined standard deviation of counting uncertainty and uncertainty in counting efficiency as the other components (volume of sample, counting time) of uncertainty were considered negligible. Measurement uncertainty (expressed as 2 sigma) was <5 % in most of the 210Po measurements.

Measurement of creatinine in the urine samples

Creatinine concentration in the preserved whole urine samples was determined through a QuantiChrom™ creatinine assay kit (DICT-500, BioAssay Systems, Hayward, California, USA) following manufacturer’s instructions. The creatinine assay is based on kinetic Jaffe reaction that utilizes a red-colored complex formed by the interaction of alkaline picrate and creatinine. Urine samples and creatinine standards (500 µg/ml) were diluted tenfold with deionized water in 1-ml Eppendorf tubes. The measurement of creatinine was carried out on a clear bottom 96-well plate by a UV–VIS Plate Reader (SpectraMax Plus384, Molecular Devices Corporation, Sunnyvale, California, USA). Optical density (OD) was measured immediately (0 min) and after 5 min at 510 nm wavelength. The creatinine concentration in a urine sample was calculated as

$${\text{Concentration}}_{\text{sample}} = \frac{{{\text{OD}}_{{{\text{sample}}\,5\,\hbox{min} }} - {\text{OD}}_{{{\text{sample }}\,0\,\hbox{min} }} }}{{{\text{OD}}_{{{\text{standard}}\,5\,\hbox{min} }} - {\text{OD}}_{{{\text{standard}}\,0\,\hbox{min} }} }} \times {\text{Concentration}}_{\text{standard}}$$

Urine sample from each rat was measured in duplicate, and the result was expressed as average and standard deviation from duplicate measurement.

ELISA of the ultra-filtered urine samples

The double-antibody, sandwich ELISA was carried out on the ultra-filtered concentrate of the urine samples for four proteins, namely albumin, alpha-1-acid glycoprotein, IgA and IgG. The ELISA kits (together with the analysis protocols) for these four rat proteins were purchased from GenWay Biotech Inc. (San Diego, California, USA). The measurement of the proteins in the polystyrene microtiter wells was performed using a UV–VIS Plate Reader according to the analysis protocol supplied by the ELISA kit manufacturer. A calibration curve (absorbance at 450 nm vs concentration of protein) was built from the measurement of seven different concentrations (each in duplicate) for each of the protein standards. A four-parameter logistic curve was used for the calibration curve fitting. Concentration of a protein in an ultra-filtered concentrate of a urine sample was calculated from the calibration curve. Sample dilution factors (based on the whole urine samples) for albumin, alpha-1-acid glycoprotein, IgG and IgA were 400, 50, 10 and 10, respectively. Urine sample from each rat (at a specific collection period) was measured in duplicate, and the result was expressed as average and standard deviation from duplicate measurement.

Statistical analysis

Statistical differences (p < 0.05) between treatment and control groups were determined by a repeated-measures design, one-way analysis of variance (ANOVA) with Tukey–Kramer post hoc multiple comparison test using GraphPad InStat version 3.00 for Windows 95 (San Diego, CA, USA, www.graphpad.com).

Results

Urinary excretion of 210Po in rats

Excretion of 210Po in urine from rats was daily measured over 4 days after administration (intravenous injection) of 210Po-citrate, as presented in Fig. 1 (open circles with a solid line). Urinary excretion of 210Po is presented as percent of injected activity (left Y-axis scale). After 210Po administration, the urinary excretion of 210Po was rapidly decreased from 24 h to 48 h (from 0.22 ± 0.03 to 0.07 ± 0.02 %, p < 0.0005, n = 5); thereafter, it remained steady (no statistical difference for 48 vs 72 h and 72 vs 96 h) over the measured period. Results for the measured 210Po in the ultra-filtered concentrate urine (expressed as percent of 210Po in the whole urine) are also presented in Fig. 1 (open squares with a dashed line; right Y-axis scale). The excretion of the 210Po in the ultra-filtered urine (from 54.0 ± 9.2 to 36.7 ± 11.9 % of the whole urine) was observed from 24 to 48 h (p < 0.025, n = 5); thereafter, it remained steady (no statistical difference for 48 vs 72 h and 72 vs 96 h) over the measured collection period.

Fig. 1
figure 1

Urinary excretion of 210Po in rats over the first 4 days after injection; solid line (left Y-axis) represents % 210Po in urine/injected activity, and dashed line (right Y-axis) represents % 210Po in the ultra-filtered concentrate/whole urine. The 210Po activity in each collection period is derived from the mean of five rats each measured in duplicate with an error bar of 1 standard deviation (SD). For % 210Po in urine/injected activity: p < 0.0005 (24 vs 48 h), p > 0.8 (48 vs 72 h) and p > 0.5 (72 vs 96 h). For % 210Po in the ultra-filtered concentrate/whole urine activity: p < 0.025 (24 vs 48 h), p > 0.2 (48 vs 72 h) and p > 0.2 (72 vs 96 h)

Full size image

Effect of ultrafiltration and presence (in vitro) of 210Po on ELISA of urinary proteins

Experiments were carried out to test whether the ultrafiltration and the presence (in vitro) of 210Po will affect the ELISA results for the protein measurements. Preserved urine samples from five control rats (those administered with saline, but not with 210Po) were pooled together. A 20 ml aliquot of the pooled urine sample was split into four 5 ml fractions. Two of these fractions (5 ml each) were spiked with 210Po (1 Bq/ml). Two fractions (one not spiked and the other spiked with 210Po) were ultra-filtered, and ELISA for albumin, alpha-1-acid glycoprotein, IgA and IgG was carried out on each of these fractions (whole urine and ultra-filtered concentrate), as described above. There was no significant difference for measured albumin concentrations between the whole urine and ultra-filtered concentrate for both the un-spiked and 210Po spiked samples (Fig. 2), and similar results were obtained for other measured proteins as well (data not shown).

Fig. 2
figure 2

ELISA results for albumin concentration (mg/l) in whole urine, ultra-filtered concentrate, 210Po spiked whole urine and 210Po spiked ultra-filtered concentrate samples. Each measurement was carried out in triplicate; error bar represents 1 SD. p > 0.5 (whole urine vs ultra-filtered concentrate), p > 0.8 (ultra-filtered concentrate vs Po-210 spiked while urine) and p > 0.5 (Po-210 spiked whole urine vs Po-210 spiked ultra-filtered urine) by t test

Full size image

Urinary excretion of albumin, alpha-1-acid glycoprotein, IgG and IgA

Results for the measured protein concentrations in the daily urine samples from rats are presented in Table 1 for albumin and alpha-1-acid glycoprotein and in Table 2 for IgA and IgG. Concentration of each protein in a 24-h urine sample was normalized with creatinine concentration and expressed as milligram of protein per gram of creatinine. Creatinine normalization served to control for the differences in glomerular filtration, urine concentration and urine volume among different rats. Concentration of a protein in a urine sample at a specific collection period was expressed as average and combined standard deviation from five rats. p values for comparison of the means between the control and an exposed set for a specific collection period are also presented in Tables 1 and 2. As indicated by the p values in Table 1, a statistically significant decrease in the excretion of albumin concentration (creatinine normalized) was observed when the urine samples collected at 24 and 96 h from the internally exposed sets were compared to the control set. A decrease in urinary excretion of albumin concentration (creatinine normalized) for different injected activity levels is shown in Fig. 3. The dose-dependent decrease in creatinine-normalized albumin concentration (expressed in terms of injected activity), as shown in Fig. 4, appears to provide a reasonable fit for a power trend-line relationship (y = ax −b). Although the dose-dependent decrease in albumin concentration was statistically significant (based on the two-tailed p value from an unpaired t test) between the low (1601 ± 376 Bq of 210Po) and medium (10,846 ± 591 Bq of 210Po) dose levels for both 24- and 96-h urine samples, a statistically significant decrease was not observed between the medium (10,846 ± 591 Bq of 210Po) and the high (48,467 ± 2812 Bq of 210Po) levels of injected activity for the 24-h urine samples. As a result, the decrease in urinary excretion of albumin at 96 h seems to provide a better dose-dependent correlation (r 2 = 0.965) than the one observed at 24 h (r 2 = 0.917). Also, no statistically significant, dose-dependent excretion of alpha-1-acid glycoprotein, IgA and IgG was observed when the exposed sets were compared to the control set.

Table 1 Creatinine-normalized albumin and alpha-1-acid glycoprotein concentrations in the daily urine samples from rats for different administered activity of 210Po
Full size table
Table 2 Creatinine-normalized IgA and IgG concentrations in the daily urine samples from rats for different administered activity of 210Po
Full size table
Fig. 3
figure 3

Decrease in urinary excretion of albumin concentration (creatinine-normalized) for different injected activity levels at 24 and 96 h after administration of 210Po. The data were derived from the mean of five rats each measured in duplicate, and an error bar represents 1 SD. For details see text

Full size image
Fig. 4
figure 4

Dose–response (injected activity–urinary excretion of albumin) curve at 24 and 96 h after administration of 210Po. The data were derived from the mean of five rats each measured in duplicate, and an error bar represents 1 SD. An asterisk indicates significant difference (p < 0.05). For details see text

Full size image

Discussion

Urinary excretion of 210Po observed here in the whole urine from rats over the first 4 days (Fig. 1, solid line) resembled the excretion pattern from our previous work on studying metabolism of 210Po in rats (Li et al. 2011a). However, the presence of 210Po in the ultra-filtered concentrate was rather unexpected as the relatively small-sized polonium ions/species were thought to pass through the 3000 Da membrane filter during the ultrafiltration. As shown in Fig. 1 (dashed line), 54–22 % of the 210Po (present in the whole urine) was found to be present in the ultra-filtered concentrate over the first 4 days of excretion. This indicates that there is a significant portion of 210Po that remains bound to the high molecular weight components (>3.000 Da) in the urine. Since the membrane filters (Amicon® Ultra-15) used in this study (for ultrafiltration) are typically used for protein purification, it is possible that the 210Po found in the ultra-filtered concentrate could be bound to the urinary proteins. Binding of 210Po to protein has been reported in liver cytosol from rats that were injected with the 210Po (Lanzola et al. 1973). 210Po was also found to have high binding affinity to metallothionein in rat’s liver and to ferritin, hemocyanin and metal-binding enzymes in marine invertebrates (Aposhian and Bruce 1991; Finger and Smith 1987; Durand et al. 2002).

Radiation-induced changes in the serum proteins characterized by a decrease in albumin, pre-albumin and γ-globulins and by an increase in α- and β-globulins have been observed in previous studies following external irradiation (Walden and Farzahen 1990). Serum levels of IgG and IgA were found to be sensitive to IR exposure; a 70–80 % decrease in IgG and IgA concentration at 5–7 days post-irradiation was observed in mice receiving 8–10 Gy of X-ray irradiation (Walden and Farzahen 1990). As a glomerular ultrafiltrate of blood within kidney, urine may contain protein components that are similar to those found in blood plasma/serum. As a result, pathological changes in human organs that may be reflected in the blood plasma/serum may also be reflected and detected in the urinary proteome (Gonzalez-Buitrago et al. 2007). In fact, the work presented here was inspired by a recent study by Sharma et al., where a total body X-ray irradiation (10 Gy) of rats showed a decrease in urinary excretion of albumin and immunoglobulins, but with a concomitant increase in alpha-1-acid glycoprotein (Sharma et al. 2008). It was therefore interesting to investigate whether internal exposure from an alpha emitter produced similar changes to the expression of urinary protein biomarkers that could potentially be utilized as a screening tool for triaging an internal contamination in an R/N emergency. Unlike uniform total body irradiation (TBI) from the gamma and X-ray radiation, internal contamination may result in non-uniform distribution of the dose due to tissue-/organ-specific distribution of the radionuclide. As a result, an internal contamination may result in biomarkers that are different from external exposure. Moreover, being present at a very close proximity to the cells of a tissue/organ, a relatively low level of an internal exposure (as compared to the external exposure from gamma or X-ray emitters) from an alpha emitter may result in significantly larger biological effect (for example, double-strand DNA break) due to its high linear energy transfer (LET) alpha particles.

For an internal exposure with 210Po, a dose-dependent decrease in urinary excretion of albumin was observed in rat urine collected at 24 h (for low and medium dose levels) and 96 h (for low, medium and high dose levels) after exposure. The following discussion attempts to provide a plausible explanation for the observation from this study based on the existing knowledge of metabolism and biokinetics of 210Po. Lanzola et al. (1973) have studied the subcellular distribution of 210Po in rat tissues over periods of up to 47 days after intravenous, oral or intramuscular introduction of the radionuclide in citrate or colloidal form. The majority (50–80 %) of the total tissue 210Po was found in the cytosol of liver, kidney and testes irrespective of the time after injection (over periods of up to 47 days), the route or form in which the radionuclide was administered. Relative distribution (percent of injected dose) of 210Po among these three tissues follows the order, liver > kidney > testes. Similar tissue distribution results were also observed in a previous study on the metabolism of 210Po in rats, where 9–18, 2.9–9.7, 0.9–3.3 and 0.5–1.5 % of the injected activity was localized in the liver, kidney, spleen and lung, respectively, 4 days following intravenous administration (Li et al. 2011a). Early distribution of polonium in target organs and tissues as determined in baboons, dogs and a human subject indicates that, in general, the liver (~35 %), kidney (~5 %), red bone marrow (~5 %) and spleen (~2 %) concentrate polonium more than other tissue except for temporary deposition in the lung after inhalation of an insoluble form (Leggett and Eckerman 2001; Ansoborlo et al. 2012). These values have since been revised after the International Commission on Radiological Protection (ICRP) Publication 67 model, with a distribution corresponding to liver (30 %), kidney (10 %), bone marrow (10 %) and spleen (5 %) (ICRP 1993). These studies (Li et al. 2011a; Leggett and Eckerman 2001; Ansoborlo et al. 2012; ICRP 1993) suggest that after internal exposure, a relatively high proportion of 210Po localize in the liver compared to the other organs or tissues. One of the major metabolic functions of liver is the formation of plasma proteins. Essentially all of the albumin and fibrinogen of the plasma proteins, as well as 50–80 % of the globulins, are formed in the liver (Hall and Guyton 2011). Plasma proteins are a mixture that contains albumin (~78 %), globulin (~21 %) and fibrinogen (~1 %) (Hall and Guyton 2011). Thus, albumin is the major protein among other plasma proteins produced by the liver. A number of recent studies have demonstrated possible depletion of albumin in plasma due to oxidative damage in cases of liver dysfunction/damage (Oettl et al. 2008, 2013; Stauber et al. 2014). Since urine is the glomerular filtration product of plasma in the kidney, depletion of albumin in urinary excretion, as observed in this study, could be linked to dysfunction/damage in albumin formation in the liver where 210Po preferentially localizes.

It is noteworthy that, unlike pathological proteinuria (indicated by an increase in urinary excretion of albumin) resulting from the malfunction of the glomerular function of kidney (Gonzalez-Buitrago et al. 2007) and normal physiological proteinuria resulting from strenuous exercise (Poortmans and Jeanloz 1968), the observed decrease in albumin excretion is consistent with previously observed radiation-induced changes in the concentration of albumin in serum/plasma (Walden and Farzahen 1990) and in urine of rats (Sharma et al. 2008). However, it is necessary to verify the consistency of the observed dose-dependent decrease in albumin beyond 96 h after injection. A follow-up animal experiment is currently underway to study the effect of 210Po exposure on the excretion of urinary proteins for an extended period. Further study focusing on the metabolomics functioning of secreted liver proteins may lead to identification of novel biomarkers for internal polonium exposure.