Bone as a Possible Target of Chemical Toxicity of Natural Uranium in Drinking Water

1Increased uranium levels in groundwater are associated with uranium-rich ores and high solubility of uranium under oxidizing conditions in soft and bicarbonate-rich waters. Consequently, exceptionally high uranium concentrations have been found in private drilled wells located mostly in the southern part of Finland. We identified earlier a cohort of people who live in that area and use drilled wells for drinking water.

2In long-term exposure, uranium accumulates in the bone and kidneys. The kidney has been considered the main target organ of chemical toxicity of uranium in humans, but effects in other tissues or organs remain poorly known. The intake of natural uranium through drinking water is associated with kidney function, particularly increased fractional excretion of calcium and phosphate in urine. The implications of the accumulation of natural uranium in the bone in humans are not known. The early distribution of uranium in the skeleton is similar to that of calcium. Uranium is assumed to deposit on the bone surface, and the uranyl ion (UO22+) is assumed to be exchanged with calcium ions at the surfaces of bone mineral crystals but not to participate in crystal formation. Gradually, uranium is redistributed in the bone and other tissues. The current biokinetic model of the International Commission on Radiological Protection suggests three compartments for uranium in human bone: bone surface, exchangeable bone volume, and nonexchangeable bone volume. It also suggests that uranium leaves bone surfaces more slowly than does calcium and that the removal from the nonexchangeable bone compartment may occur but with the rate of bone turnover.

3The resemblance of uranium metabolism to that of calcium in bone enables the effects of uranium on bone. Indeed, uranium administration in rats is known to affect the bone. Acute or continuous exposure to uranium may lead to decreased bone formation rate and also increased bone resorption in rats.

4The aim of this study was to assess whether uranium intake through drinking water affects the biochemical markers of bone turnover in humans. The present study extends our previous study, which suggested that uranium exposure is associated with altered proximal renal tubular function. To our knowledge, this is the first report on the possible effects of ingested natural uranium on bone in humans.

5The source population was identified from the drinking water database of STUK–Radiation and Nuclear Safety Authority, with radionuclide analyses of more than 5,000 drilled wells. This study was limited to southern Finland, where uranium concentrations are highest. The study population was a subpopulation of our previous study on effects of natural uranium on kidney function, with a more detailed description published earlier. The first questionnaire was mailed to 798 households. Based on the first questionnaire, 436 persons were selected, with a maximum of two persons per household where a drilled well had been used for drinking water at least for the previous year (median duration of use, 11 years). The second questionnaire was used to collect information on residential history and use of drilled well water and its daily consumption, use of other beverages, smoking history, education, occupation, disease history, and use of medication and herbal products. Seventy-eight percent of the persons who received the second questionnaire agreed to participate in the study (samples were received from these persons). Further information on, for example, bone fracture history and information on menopause and on physical activity was also collected (67% replied to this third questionnaire). We do not have information on type or date of the bone fractures.

6Subjects were excluded if they were < 25 years of age (n = 11); had diabetes mellitus (n = 4); reported long-term use of glucocorticoids (n = 11), thiazide diuretics (n = 7), methotrexate (n = 1), or sodium aurothiomalate (n = 1); were currently pregnant (n = 4); or used effective equipment for removing uranium from well water (n = 4). The final study population consisted of 288 persons from 179 households. Most of the study persons had never smoked, and their average body mass index was 25 kg/m2 (Tables 1 and 2). Twenty-six women used estrogen (oral contraceptives or hormonal replacement therapy) regularly during the previous year, and the women had had two deliveries on average. None of the subjects reported hyperparathyroidism.

7The study protocol was approved by the National Public Health Institute Standing Committee on Ethics (project 8/030399).

8The water, urine, and nonfasting blood samples were collected between 14 September and 1 December 1999. The samples were collected at a time when the study persons had consumed water from the drilled well throughout the previous week. Samples were not taken unless at least 1 week had elapsed since an acute infection. The study persons brought the water and urine samples collected overnight to the laboratory in the morning. At the same visit, blood samples were taken (77% of the samples were taken before 1100 hr). In addition, body weight and height were measured in a standardized fashion. The water and overnight urine samples for uranium analyses were conserved with concentrated HNO3. Water samples were stored at room temperature but serum and urine samples frozen at −20°C until analyzed.

9Uranium in drinking water and urine were analyzed blind with inductively coupled plasma mass spectrometry. The analysis and quality control procedure are described by. The primary measure of uranium exposure was uranium concentration in drinking water (micrograms per liter). In addition, we measured daily intake of uranium from drinking water (volume used × concentration, micrograms), cumulative intake from drinking water (daily intake × duration of the water consumption, grams), and uranium concentration in urine (micrograms per liter or micrograms per millimole creatinine). The exposure variables were highly correlated with each other (Table 3).

10Serum osteocalcin and amino-terminal propeptide of type I procollagen (P1NP) were used as indicators of bone formation, reflecting different stages of osteoblast differentiation. We analyzed osteocalcin using an immunoradiometric assay, which measures human osteocalcin (1–49) and human osteocalcin peptide (1–43) (ELSA-OSTEO; CIS Bio International, Gifsur-Yvette, France). At the 15-μg/L level, the intra- and interassay variations were 2.0 and 3.1%, respectively. P1NP was analyzed with a commercial radio-immunoassay (Procollagen Intact P1NP; Orion Diagnostica, Oulunsalo, Finland); intra- and interassay variations at the 40-μg/L level were 2.0 and 4.9%, respectively.

11Serum type I collagen carboxy-terminal telopeptide (CTx) was used as an indicator of bone resorption. CTx was analyzed with an enzyme immunoassay (Serum CrossLaps One Step ELISA; Osteometer Biotech, Herlev, Denmark); intra- and interassay variations at the 2.4-nmol/L level were 6.4 and 7.2%, respectively.

12In men, the correlation between the log-transformed osteocalcin and P1NP was 0.70; between osteocalcin and CTx, 0.38; and between P1NP and CTx, 0.32. In women, the correlations were 0.63, 0.46, and 0.36, respectively.

13Urine calcium was measured with atomic absorption spectrophotometry (EFOX 5053; Eppendorf, Hamburg, Germany) (detection limit, 0.1 mmol/L). Urine phosphate was measured based on a colored complex with ammonium molybdate (Konelab 60I; Konelab Co., Espoo, Finland) (detection limit, 2.0 mmol/L). Urinary excretions of calcium and phosphate (millimoles per hour) were calculated from the volume of urine divided by the overnight collection time. The mean ± SD excretion of urine was 79 ± 36 mL/hr in men and 78 ± 37 mL/hr in women.

14For all the parameters determined, the observations below the detection limits were recorded as half of the detection limit. An analysis of the residuals indicated that they were not normally distributed and that some of the observations were highly influential. Therefore, the robust regression method using iteratively reweighted least squares (Huber and Tukey bi-square weight function) with rreg routine in Stata/SE 8.1 for Windows (Stata Corp., College Station, TX, USA) was used. Robust regression assigns a weight to each observation, with lower weights given to possible influential observations. Some results from the conventional linear regression are also given in “Results.”

15The analyses were performed separately for men and women. For men, markers of bone metabolism levels were modeled using linear and quadratic terms for age and a variable for current smoking. The model used for women included a categorical age term [< 45 (reference), 45–55, 55–65, and ≥65 years], with additional variables for recent regular estrogen use and current smoking.

16Algebraically these are as follows:

17in which y is the indicator of bone metabolism, α is a constant, b is the regression coefficient, x is continuous uranium exposure, agecat is age category (45–55, 55–65, and ≥65 years), smo is current smoking status, and estro is use of estrogens, the last two being binary indicator (dummy) variables. R2 and p-values of the models are described in Table 4.

18Analyses were also carried out using the above-mentioned models with log-transformed urinary calcium or phosphate excretions as explanatory variables.

19For men, the levels of osteocalcin, P1NP, and CTx tended to decrease with age until about 60 years of age, after which bone turnover appeared to increase gradually but age accounts for a relatively small proportion of the variation in the bone turnover measurements (Figure 1). All p-values were < 0.01 for associations between the linear age variable and all outcome variables. In men, current smoking was associated with decreased levels of osteocalcin (p = 0.01), P1NP (p = 0.11), and CTx (p = 0.03).

20For women, the levels of osteocalcin, P1NP, and CTx were highest in the age group of 55–65 years, but the differences between the age groups were not statistically significant (Figure 1). Smoking in women was not statistically significantly associated with any marker of bone turnover. Estrogen use was associated with significantly decreased levels of osteocalcin, P1NP, and CTx (Figure 1).

21The uranium concentration in water varied from 0.001 to 1,920 μg/L (Table 2), with 27% of the concentrations > 100 μg/L and 59% > 15 μg/L. The median daily intake of uranium from drinking water was 36 μg, and the cumulative intake was 120 mg. The median annual committed equivalent radiation dose of bone surfaces was 0.36 mSv/year (maximum, 41 mSv/year), based on the uranium intake and the average uranium isotope activity ratios measured in Finnish drilled well waters (234U:238U = 2) and dose conversion factors (234U, 7.4 × 10−7 Sv/Bq; 238U, 7.1 × 10−7 Sv/Bq).

22In men, uranium exposure was associated with elevated CTx levels (Figure 2) with the p-values in the robust regression 0.05 for uranium in water, 0.16 for daily intake, and 0.16 for cumulative intake. The corresponding p-values in conventional linear regression analyses were 0.01, 0.02, and 0.03. There was an indication of an association between increased levels of osteocalcin and uranium concentrations in drinking water (p = 0.19; p-value in the conventional linear regression was 0.04). Levels of P1NP were not associated with uranium exposure. Uranium concentrations in urine expressed as micrograms per liter or micrograms per millimole creatinine were not associated with the markers of bone turnover.

23Increased urinary excretion of calcium tended to be associated with increased CTx levels (p-value from the robust regression was 0.10), and some indication was found for increased urinary excretion of phosphate with decreased osteocalcin levels (p = 0.16) in men. The other associations between calcium or phosphate excretion and bone turnover were not close to the statistical significance in the robust regression.

24In women, uranium exposure was not associated with any indicators of bone turnover (Figure 2). Urinary excretion of neither calcium nor phosphate was associated with bone markers.

25Those 32 study persons who reported a history of any bone fractures in adulthood were not statistically significantly more exposed to uranium than those without such history (median cumulative doses of uranium of 124 mg in those with fractures vs. 117 mg in those without). There were no differences in the levels of the markers of bone metabolism among those with or without past fractures (data not shown).

26The uranium exposure covered a wide range of concentrations in this study. More than half of the study persons used drinking water with uranium concentration exceeding 15 μg/L, which is the new provisional World Health Organization (WHO) guideline value for uranium in drinking water.

27In men, chronic uranium exposure indicated by uranium level in drinking water as well as daily and cumulative uranium intakes tended to be associated with the increased levels of the bone resorption marker CTx and to a lesser degree of the bone formation marker osteocalcin. The association of uranium exposure and CTx reached a marginal significance at the 5% level in the robust analysis that down-weights the influence of possible outliers and was significant in the conventional regression analysis. This finding may indicate that bone is a possible target of chemical toxicity of natural uranium.

28In contrast to men, no statistically significant associations with uranium exposure and the measured bone turnover markers were observed in women. In women, subtle effects may be masked by other strong determinants of bone turnover, such as menopause and hormone use. Potential confounding factors including menopausal status, recent body weight changes, physical activity, and calcium and vitamin D supplementation could not be effectively controlled in the present study, although they all have an influence on bone metabolism.

29The subtle effects of uranium on bone markers in men could be explained by different mechanisms. Accumulation of uranium in bone may have a local effect on bone metabolism or structure. Direct effect of uranium on bone has been shown in animal studies, with accumulation of uranium into bone. In laboratory animals exposure has been shown to modify bone formation and resorption. Direct bone effect is also supported by the present observation that the change in bone markers was more strongly associated with uranium concentration in drinking water and daily or cumulative intake than with uranium concentration in urine. Concentration in water and intake likely describe long-term exposure to uranium and consequently accumulation to bone better than concentration in urine, which reflects recent exposure.

30Uranium might have an effect on bone also via its influence on kidneys. Chronic renal insufficiency has been reported to affect bone metabolism. On the other hand, elevated levels of bone markers have been observed in patients with kidney damage due to chronic exposure to cadmium, another metal toxic to the kidney. Cadmium-induced osteoporosis is associated specifically with tubular damage, including increased excretion of calcium in urine. However, uranium in drinking water does not cause severe kidney damage or cytotoxicity even at high exposure levels, nor had the subjects in the present study significant kidney insufficiency. We have earlier shown that intake of natural uranium in drinking water is associated with increased fractional excretion of calcium and phosphate. By increasing leakage of calcium into urine by disturbing its tubular reabsorption, uranium exposure could secondarily lead to bone resorption. However, the increased excretion of calcium and phosphate in urine has been suggested to be associated most strongly with the recent uranium exposure (uranium in urine), and in the present study CTx was associated with long-term exposure to uranium (concentration in water, and daily and cumulative intake). Accordingly, on the basis of earlier animal studies and the present results, we propose that the relationship between uranium intake and bone markers reflects the direct effect of uranium on bone.

31Only uranium was analyzed from the water samples. It is possible that other elements or constituents in drinking water confound the results. To be a confounding factor, it should be associated with both uranium concentration and the outcome measures. However, other heavy metals, including cadmium and lead, occur extremely rarely in substantial concentrations in Finnish drilled wells and are not correlated with uranium concentrations (Kurttio et al., unpublished data). Therefore, other elements in drinking water are very unlikely to confound the results.

32In this study population, drinking water is expected to be the predominant source of uranium, especially among those with elevated uranium concentrations in well water. The study persons had used drinking water from the drilled wells with measured uranium concentrations for at least 1 year. Therefore, a steady state uranium exposure can be anticipated.

33Uranium concentration in the private wells drilled in bedrock may vary considerably over time, and therefore a spot sampling may not accurately represent the long-term uranium exposure. Additionally, the daily and cumulative intakes of uranium are based on study persons’ own estimates on their drinking water consumption, which also adds uncertainty. Although urinary uranium concentration is unaffected by these sources of uncertainty, it is limited mainly to current uranium exposure. Yet there is a high correlation between uranium exposure indicators.

34Substantial variation in age complicates the interpretation of the results because several age-dependent factors influencing the bone turnover may mask possible effects of uranium. As was seen in this study, the levels of bone turnover markers remain approximately stable from 25 years of age to menopausal age (~ 55 years) in women and to 65 years of age in men. Obviously focusing on limited ages would facilitate the interpretation of the results.

35We found some evidence for an association between increased bone turnover and exposure to natural uranium through drinking water among men. The fact that similar effects were not observed in women may be due to other stronger factors in bone metabolism of women that may mask the effects of uranium. This study suggests that in addition to kidneys, bone may be another target for uranium toxicity.