Study population and collection of samples

Ten 24-h urine samples from cancer patients of the Antofagasta region were collected before Pt-based drugs chemotherapy was initiated. Ninety 24-h urine samples from cancer patients were collected immediately after the infusion of the Pt-base drugs; and ten 24-h urine samples were collected from patients subjected to chemotherapy drugs involving drugs that did not contain a metal. All samples were provided by the Oncological Centre of Antofagasta Regional Hospital (Region II, Chile). The 90 cancer patients they are distributed in the following cases: 32 lung cancer, 16 ovarian cancer, 12 cervical–uterine cancer, eight testicular cancer, five thymus cancer, three oesophageal cancer, three gastric cancer, three Hodgkin lymphomas, two vagina cancer, two bladder cancer, and one each of gallbladder, tonsil, salivary gland, and osteosarcoma cancers. The dose calculation for chemotherapy had been made by the BSA method [24]. Three of the four patients with ovarian cancer initially treated with cisplatin were again treated with cisplatin after 28 days, whereas the fourth patient was treated with carboplatin. Cancer others than lung was categorised as “others cancers”; only patients in this group were treated with carboplatin. Samples were collected between May and August 2005.

Chemicals, solutions, materials and instruments

The chemicals employed were Instra (J.T. Baker, Phillipsburg, NJ, USA) and Suprapur (Merck, Whitehouse Station, NJ, USA), both of which were of trace-element grade. Solutions were prepared in deionised water that was subsequently distilled in quartz (conductivity, <0.5 μMHO/cm at 25°C). The deionisation was undertaken in a deioniser with two ion-exchange cartridges—a Metex 01506-45 and a Research 01506–35 (Cole-Parmer, Waltham, MA, USA)—connected in tandem.

The glass materials used were sequentially treated for 24 h with 2-M solutions of hydrochloric acid and nitric acid prepared from Merck pro analysis quality grade chemicals. After being washed with water, the glass materials were treated for another 24 h with a 0.025 M disodium salt of ethylenediamine tetraacetic acid (EDTA) solution prepared from Titriplex III (Merck), after which they were rinsed with water.

Pt(II) standard dilutions were prepared from a primary solution of Certipur (Merck) and those of Se(IV) were prepared from a primary solution of Titrisol (Merck); the concentration of both primary solutions was 1000 mgL^–1. An intermediate 50 mgL^–1 standard solution was made, which was used to prepare the working dilutions. Standard As solutions were prepared from a 1000 mgL^–1 primary solution of disodium arsenate Titrisol (Merck), from which a 500 mgL^–1 intermediate standard dissolution was prepared for the working dilutions. To generate arsenic hydride, 10 M chlorhydric acid solution was used, which was prepared from Instra (J.T. Baker). A 3% p/v sodium borohydride solution was prepared in a 1% p/v sodium hydroxide solution, both of which were Merck pro analysis grade quality chemicals.

Urine samples were transported in a cooler at 4–5 °C to an acclimatised and pressurised Pre-treatment Sample Room equipped with a laminar-flow hood (Labconco Purifier, class II; Labconco, Kansas City, MO, USA). Samples were microfiltered in a Nalgene system through a Millipore cellulose ester membrane (pore size, 0.45 μm; diameter, 47 mm; Millipore, Bedford, MA, USA). Samples were then stored at –20 °C. Immediately before analyses of metals were carried out, the samples were thawed and again microfiltered, this time through a Millipore cellulose ester membrane of pore size 0.22 μm protected by a borosilicate glass pre-filter (Advantec-MFS, Japan) (pore size, 0.7 μm).

Matrix-matched ionic strength surrogate urine was prepared by diluting subtidal seawater thrice with deionised glass-distilled water. As a reference, the measured average salinity of urine samples was 13 parts per thousand (ppt). This was measured in a previously calibrated Orion Ion Meter 1260 (Thermo Scientific, Waltham, MA, USA).

Digestion of urine samples

Urine samples and controls were mineralised using the wet method in Teflon bombs with 10 ml of nitric acid (Instra; J.T. Baker, USA) for 2 h at 150 °C in a homemade ceramic oven with an internal temperature sensor and external temperature control. After the samples were cooled, the oven was opened and an additional 0.25 mL of perchloric acid and 0.50 mL of sulphuric acid were added to the Teflon bombs with the digested samples; both acids were of Suprapur grade (Merck). Samples were then reheated under the same conditions of time and temperature as described above. Finally, samples were transferred to Erlenmeyer flasks and heated at 300 °C under a gas extraction hood to eliminate excess acids. Sample volumes were made up to 50 mL with deionised water, and then micro-filtered through a 0.22-μm Millipore membrane with a 0.7-μm fibreglass pre-filter (Advantec-MFS, Japan).

To determine total As concentrations, 5-mL aliquots of samples and controls were digested to semi-reflux under a temperature programme between 85 °C and 300 °C [45]. After cooling, samples and controls were diluted volumetrically with 0.5 M HCl prepared from Instra quality grade chemical (J.T. Baker, USA).

Determination of total Pt and Se in urine through inductively coupled plasma optical emission spectrometry (ICP–OES)

With the aim of meeting the matrix-matching conditions, aliquots of three samples of digested urine were titrated with NaOH secondary standard solution, yielding an average acidity equivalent to a pH of 2. To make compatible the matrix-matching conditions of the analytical blanks, seawater was diluted 8-fold, 10 mL of sample was diluted to a final volume of 50 mL and, when necessary, the solution was adjusted to pH 2 with Instra nitric acid (J.T. Baker, USA). The bimetallic Pt–Se standard was also subsequently prepared in this medium.

Determination of total As in urine by hydride-generation atomic absorption spectrometry (HGAAS)

Determination of As was undertaken using an atomic absorption spectrophotometer (GBC 909 AA; GBC Victoria, Australia) equipped with a continuous-flow hydride generator (GBC HG 3000) and a electrothermal atomisation system (GBC EHG 3000). GBC quartz cells and an As-boosted current lamp from Photron (Narre Warren, Victoria, Australia) were also used. The optimized conditions for measuring As through HGAAS are presented in Table 1. Multiple-standard-addition methodology was applied, for which an intermediate 500 mg/L standard solution was prepared in 0.5 M HCl from a primary standard solution (Titrisol; Merck). Additions from this solution were made to surrogate-sample urine, which had been made up to 2.0 μg/L in As. The addition volumes were 50, 100, 150 and 200 μL; the concentrations added were therefore 2, 4, 6 and 8 ng/mL, respectively.

The baseline was adjusted with the analytical blank solution i.e., diluted surrogate urine. QC tests were conducted to evaluate the recovery of As from As-spiked real samples and measurements of secondary standards. Urine samples of patients underwent the procedures and analytical methods described above.

Urine and serum creatinine levels and calculation of creatinine clearance

Levels of urine and serum creatinine were measured using the Jaffé method [46]. Creatinine clearance rate (CrCl) was calculated using the Cockcroft–Gault equation. Serum creatinine is the blood clearance or endogenous blood creatinine depuration [47], their values were provided by the Oncological Center of Antofagasta Regional Hospital. Urine and serum CrCl values are used to evaluate renal dysfunction. The estimated CrCl for normal renal function should be >80 mL/min [48].

Statistical analyses

Statistical analyses were conducted using the Statistica 9.1 computer programme (StatSoft, Tulsa, OK, USA). p < 0.05 was considered significant.

Determination of Pt, Se and As

The calibration curves for ICP–OES determinations of total Pt and Se concentrations in urine were linear. In the case of Pt, the parameters of the calibration equation were: intercept = 93.45 counts · s^–1; slope of 1.079 counts × s^–1/concentration; and r = 0.9998. For Se, the parameters were: intercept = 55.47 counts · s^–1; slope = 1.720 counts × s^–1/concentration; and r = 0.9880.

The analytical validation and QC results for Pt and Se by ICP–OES and of As by HGAAS analyses in 24-h urine samples of cancer patients indicated that the techniques and methodologies applied in the work were analytically suitable. Surrogate urine prepared from contaminant-free seawater was important for optimisation of the analytical parameters and appropriate representativeness of the results.

Inter-element relationship of Pt, As and Se in urine of cancer patients from the Antofagasta region of Chile treated with Pt-based drugs

Recent developments in multi-elemental analytical techniques and methodologies have enabled exploration of inter-elemental relationships and toxicities of essential elements and heavy metals, that which coexist in environmental and biological matrices [49–54]. Knowledge in this area is mainly based upon single elements; a better-known case is that of As, which has been associated with several diseases, including cancer [55–58]. However, it has become increasingly clear that simple, single-element models are often inadequate to explain the relationship between disease and nutrition, element toxicity, or deficiency of elements in food. Improved understanding of the synergistic and/or antagonistic interactions between trace elements is needed [59]. A multi-elemental hypothesis may be a better approach to understand these relationships.

Renal function before and after cancer treatment with Pt-based drugs is an important parameter. Before starting anti-cancer treatment, it is advisable that the CrCl is >60 mL/min [47, 60]. Patients with lower values have an increased risk of severe suppression of the bone marrow. The CrCl for normal renal function is > 80 mL/min [47, 61]. The CrCl values before Pt-based drug chemotherapy for the control group 2 (baseline patients; see below) were 93.3–114.6 mL / min; serum and urine creatinine values were 0.6–0.9 and 0.8–1.1 mg / dL, respectively. Evidently, Pt-based drug chemotherapy impacts the CrCl and the metal/creatinine quotient values, both parameters are considered controversial [62, 63]. According this approach, after chemotherapy, 82 of 90 cancer patients had kidney damage.

Statistical parametric treatment

Although the number of patient with cancer of the control groups was low, one can observe that in both cases the Se levels tend to be higher than those of As. However, in control group 2, the range of As and Se levels tends to be comparable or slightly smaller to than those of control group 1. In Region II (Antofagasta) Se concentrations have been shown to be higher than those As in the umbilical cords and placentas of normal and malformed newborn babies, and in tissues of patients’ subjected to cardiovascular surgeries [64, 65]. This characteristics was also observed in the present study, inasmuch that the Se and As concentrations in group 3 were comparable with those of control 1 and control 2 groups.

Figure 1 shows the one-way ANOVA profiles (p ≤ 0.05) for calculated drug supplied Pt and Pt in 24-h urine samples from lung-cancer patients treated with cisplatin (group 1); the Shapiro–Wilk normality test parameter are also shown. The one-way ANOVA urine profiles for Se of patients other than lung cancer treated with cisplatin (group 2) did not exhibit significant variations (p ≤ 0.05), but urine profiles of As did (p ≤ 0.05). In both cases, the data did not follow a normal distribution.

Regarding the calculated drug-supplied platinum and the Pt, Se and As in urine, the results presented in Table 3 indicate that the decreasing tendency of supplied Pt was: carboplatin (cancer other than lung, 695 mg) > cisplatin (lung cancer, 163 mg) > cisplatin (cancer other than lung, 120 mg); the tendency of Pt amount in urine was the same as that of supplied Pt. The tendency of excreted Se amount in urine was carboplatin (cancer other than lung, 1.59 μg) > cisplatin (cancer other than lung, 1.22 μg) ≈ cisplatin (lung cancer, 1.28 μg). With respect to excreted As amounts in urine, the tendency was cisplatin (cancer other than lung, 71 ng) > carboplatin (cancer other than lung, 40.8 ng) > cisplatin (lung cancer, 23.5 ng). Hence, with Pt –base drugs and chemotherapy not involving metal-base drugs, Se amounts were always higher than As amounts in urine. This inter-element relationship suggests that, in general, anti-cancer drugs induce higher secretion of Se than of As, and that this relationship was independent of the nature of the drug used in the treatment.

The amounts of Pt, As and Se in 24-h urine samples from cancer patients expressed as metal/g creatinine showed the same distribution tendency and statistical significance as non-normalised values (Table 4). The comparison with metal/g creatinine values for healthy subjects (i.e., As, 33.3 μg/g creatinine [66, 67]; Se, (women), 9.8 μg/g creatinine; Se (men), 13.5 μg/g creatinine [68]; and Pt, 3.3 ng/g creatinine [69]), along with the results obtained in this work, revealed that the values for cancer patients were lower than those for healthy persons. However, the values obtained for Se and Pt were higher than the values for healthy persons. The comparison of metal/g creatinine values between cancer patients from this work with those in other studies i.e., As, 116 μg/g creatinine (patients with arsenicism without cutaneous lesions); As, 121.2 μg/g creatinine (patients with arsenicism with skin cancer) [66]; Se, 206.7 μg/g creatinine [70]; and Pt 0.30–4.18 μg/g creatinine [42], indicated that, in this work, metal/g creatinine-reported values were lower than those of persons with skin cancer reported in the literature. For Se, the values reported in the literature were higher than those reported here and, for Pt from cisplatin, the values were within the range of those reported previously. However, in the case of Pt from carboplatin, the metal/g creatinine values obtained in this work were approximately threefold higher than the values in the literature [42]. Owing to a multiplicity of factors involved in creatinine excretion, the analytical reproducibility of the heavy metal/creatinine quotient value is not satisfactory; this parameter is also poorly representative in the clinical scenarios [63].

For 24-h urine samples, there were no significant differences between the excreted Pt amounts for lung-cancer patients and for patients with cancer other than lung cancer treated with cisplatin despite the significant differences in the amounts of supplied Pt. Analytically, the excretion of As was less than of Se, which could suggest that Pt favours Se excretion.

Multivariate statistical treatment

The cluster technique was applied to the involved variables (i.e., calculated amounts of Pt supplied through Pt-based drugs and the urinary amounts of Pt, As and Se excreted in 24-h urine after the cancer patients chemotherapy with cisplatin and carboplatin). Cluster analysis is a multivariate statistical methodology applied to datasets if there is a lack of information about the classes comprising the data. The basic objective is to group the variables by similarities [71]. The technique is based on two aspects: the way the distance between the variables is measured (metric) and the groupings or clusters (linkage or amalgamation rule) [72]. The technique was applied to the most representative sets of values in accordance the Shapiro–Wilk normality test. Ward’s method was applied with Pearson’s 1-r approximation to measure the distance between variables and the groupings or clusters.

Figure 2a shows the dendrogram of the variables related to 24-h urine values after the first cycle of treatment with cisplatin to lung-cancer cases as well as patients with cancer other than lung (groups 1, 2). In addition, 4 patients with ovarian cancer were considered after a second cycle of treatment. Carboplatin was used to treat only patients with cancer other than lung cancer, and only one patient had a second treatment cycle. Figure 2a shows two main groupings [Pt supplied - Pt excreted] and the amounts of As and Se excreted [Se, As excreted]. From application of the bidimensional grouping technique (Figure 2b), the quantitative extension of supplied Pt and the corresponding excretions of Pt (μg), Se (μg) and As (ng) in urine were compared. We concluded that urinary excretion of Pt after the infusion of cisplatin and carboplatin was low in the short term (even if more than one cycle drug was applied). Hence, Pt would initially be predominantly distribute and bio-accumulate in the tissues.

Figures 3a (lung-cancer) and 3b (cancer other than lung cancer) show the statistical dispersion plots according to the linear correlation model for the excretion of Se and As compared with Pt excretion after cisplatin treatment (groups 1, 2). For lung-cancer patients (Figure 3a), the tendency of As excretion was contrary to the increased excretion of Pt. With regard to Se, only a small group of patients showed increased excretion of Se accompanied by increased excretion of Pt. For most cases, Se was distributed above and below the 95% confidence interval. According to our hypothesis, As excretion (cancer agent) was low compared with the Se excretion, and was negatively correlated with homeostatic excretion of Pt. Se excretion (protective agent against cancer) was comparatively higher than of As, but only a few cases were positively correlated with Pt excretion. Therefore, lung tumours would bind more strongly to As than Se.

For Pt from cisplatin infused in patients with cancer other than lung cancer (Figure 3b), excretion of As increased linearly with increased excretion of Pt, and the excretion of Se was similar to that described for lung-cancer patients treated with cisplatin (Figure 3a). According to our hypothesis, excretion of the cancer agent (As) compared with the excretion of Se was also low, but was positively correlated with the homeostatic excretion of Pt. In turn, the assumed protective agent against cancer (Se) did not correlate with Pt excretion, but its level of excretion was higher than seen in lung-cancer patients. In the case of patients with cancer other than lung cancer treated with carboplatin, the dispersion plot for excretion of As respect to excreted Pt exhibited inverse tendency (r = –0.2713), but Se did not correlate with excreted Pt.

Regard to supplied Pt through cisplatin to patients with cancer other than lung cancer, excretions of Se and As did not exhibit any tendencies. However, in the case of Pt supplied through cisplatin to lung-cancer patients, the excreted As tended to exhibit an inverse linear tendency (r = –0.2918) with respect to supplied Pt. In turn, with regard to supplied Pt by treatment with carboplatin in patients with cancer other than lung cancer, As tended to exhibit a direct linear tendency (r = 0.3987), and Se tended to exhibit an inverse linear tendency (r = –0.2388). These functional inter-element associations can be important for the hypothesis outlined in the presently work. Therefore, differences appeared to exist between the relationships and inter-element associations among As, Se and Pt for patients treated for lung cancer and those treated for cancer other than lung with Pt-based drugs. However, either with Pt-based drugs or those not containing Pt, the excretion of Se was greater than that of As. This phenomenon could enhance the effect of carcinogens such as arsenic, reducing the protective effect of Se, and thereby lead to metastasis [28, 32, 73].

Pt and As are not essential elements for humans, but Se is considered essential and an anti-cancer agent [32]. Consequently, the information obtained in this work about urinary levels of As, Se and Pt immediately after cancer patients were subjected to chemotherapy with Pt-based drugs, could contribute to a better understanding of the adverse effects of Pt, and help in the more accurate estimation of drug doses [24, 74]. Consideration of urinary levels of carcinogens such as As and cancer-fighting nutrients such as Se could allow more individualised or dedicated strategies for predictive calculations of drug dosages to be applied to cancer patients.

A well-established correlation has been found between As exposure in the Antofagasta region of Chile and the risk of cancer of the bladder, lung and skin [75]. Consequently, urinary levels of As and Se after each treatment cycle with Pt-based chemotherapeutic drugs could be important for a better understanding of the carcinogenic mechanism of As exposure [76] and about the hypothetic cancer-protective effects of Se. Genetic and epigenetic processes appear to be involved in cancer mechanisms [77, 78]. Epigenetic studies focus on the stress that As provokes in cells without affecting DNA, but directly influences the genome expression. In this complementary mechanism, As participates as an ‘opportunistic carcinogen’ that enhances the genotoxicity and mutagenicity of other environmental stressors [75]. During Pt-based drug treatments, patients are exposed to an acute impact of Pt. This subjects cellular homeostasis to deliberate stress by Pt, inducing toxic and, perhaps hormesic responses. Among cancer patients exposed chronically to arsenic and treats with Pt-drugs based, Se, As and Pt can participate synergistically in the regulated and unregulated activation of apoptosis, phagocytosis, and cellular homeostasis involved in toxicological and hormesis mechanisms [18, 20, 77, 79–81].