Cadmium-Induced Proteinuria: Mechanistic Insights from Dose–Effect Analyses

Cadmium (Cd) is a toxic metal that accumulates in kidneys, especially in the proximal tubular epithelial cells, where virtually all proteins in the glomerular ultrafiltrate are reabsorbed. Here, we analyzed archived data on the estimated glomerular filtration rate (eGFR) and excretion rates of Cd (ECd), total protein (EProt), albumin (Ealb), β2-microglobulin (Eβ2M), and α1-microglobulin (Eα1M), which were recorded for residents of a Cd contamination area and a low-exposure control area of Thailand. Excretion of Cd and all proteins were normalized to creatinine clearance (Ccr) as ECd/Ccr and EProt/Ccr to correct for differences among subjects in the number of surviving nephrons. Low eGFR was defined as eGFR ≤ 60 mL/min/1.73 m2, while proteinuria was indicted by EPro/Ccr ≥ 20 mg/L of filtrate. EProt/Ccr varied directly with ECd/Ccr (β = 0.263, p < 0.001) and age (β = 0.252, p < 0.001). In contrast, eGFR values were inversely associated with ECd/Ccr (β = −0.266, p < 0.001) and age (β = −0.558, p < 0.001). At ECd/Ccr > 8.28 ng/L of filtrate, the prevalence odds ratios for proteinuria and low eGFR were increased 4.6- and 5.1-fold, respectively (p < 0.001 for both parameters). Thus, the eGFR and tubular protein retrieval were both simultaneously diminished by Cd exposure. Of interest, ECd/Ccr was more closely correlated with EProt/Ccr (r = 0.507), Eβ2M (r = 0.430), and Eα1M/Ccr (r = 0.364) than with EAlb/Ccr (r = 0.152). These data suggest that Cd may differentially reduce the ability of tubular epithelial cells to reclaim proteins, resulting in preferential reabsorption of albumin.


Introduction
Cadmium (Cd) is an environmental contaminant of continuing public health concern worldwide because it is detectable in most food types; as such, diet forms the main source of exposure in non-occupationally exposed and non-smoking populations [1,2]. Multiple organ systems, including kidneys [1,2], bone [3], liver [4,5] and the central nervous system [6], are susceptible to the toxicity of Cd, even at low body burdens. The cytotoxicity of Cd has been demonstrated in nearly all cell types, such as erythrocytes and the tubular epithelial cells of kidneys, which are known to actively accumulate Cd [1,2,7].
The protein albumin, with a molecular weight of 66 kDa, is synthesized in the liver and ordinarily secreted into the circulation at a rate of 10-15 g per day [12,13]. Catabolism in muscle, the liver, and the kidney proximal tubular epithelial cells balance synthesis, and homeostasis is continued. In good health, the plasma concentration of albumin is between 3.5 g/dL and 5 g/dL, and the average half-life in plasma is 19 days [12,13]. Albumin is not normally filtered by glomeruli, due to its large molecular weight and its negative charge. By means of transcytosis through endothelial cells and podocyte foot processes, albumin enters the urinary space at a rate of 1-10 g per day [8][9][10][11][12][13]17].
Albuminuria is diagnosed when excretion of albumin, measured as albumin-tocreatinine ratio, rises to levels above 20 and 30 mg/g creatinine in men and women, respectively [18][19][20]. The persistence of albuminuria for at least three months is a diagnostic criterion of CKD. A progressive decrease in the eGFR below 60 mL/min/1.73 m 2 , termed low eGFR, is also a diagnostic criterion of CKD [18][19][20].
An elevated excretion of low-molecular-weight proteins, namely, retinol binding protein, β 2 -microglobulin (β 2 M), and α1-microglobulin (α 1 M), has been the most frequently investigated [1,2]. The protein β 2 M, with a molecular weight of 11,800 Da, is synthesized and shed by all nucleated cells in the body [21]. By virtue of its small mass, β 2 M is filtered freely by the glomeruli and is reabsorbed almost completely by the kidney's tubular cells [22]. Cd has been shown to cause a reduction in the tubular maximum reabsorption of β 2 M [23], and increased β 2 M excretion has been used as an indicator of tubulopathy for many decades. However, our previous study showed that β 2 M excretion of 100-299, 300-999, and ≥ 1000 µg/g creatinine was associated with 4.7-, 6.2-, and 10.5-fold increases in the risk of an estimated glomerular filtration rate (eGFR) ≤ 60 mL/min/1.73 m 2 , which is commensurate with CKD [24]. These data suggest that an increased excretion of β 2 M above 300 µg/day, termed β 2 -microglobunuria, could be a consequence of Cd-induced tubulopathy in conjunction with nephron loss, evident from a reduction in the eGFR to 60 mL/min/1.73 m 2 or below [24].
The present study aimed to evaluate the concurrent effects of Cd accumulation in kidneys on GFR reduction and excretion rates of proteins of high and low molecular weights, namely, total protein, albumin, β 2 M, and α 1 M. To enable an accurate assessment of these effects of Cd, we normalized excretion of Cd and all excreted proteins to creatinine clearance (C cr ). This C cr -normalization corrects for differences in the number of surviving nephrons among study subjects, and it depicts an amount of a given chemical excreted per volume of filtrate, which is at least roughly related to the amount of the chemical excreted per nephron [25].

Descriptive Characteristics of Study Subjects According to eGFR
Among 405 cohort participants, 190 (46.9%) were residents of a low-exposure area (Bangkok), and 215 (53.1%) persons lived in a high-exposure area of Mae Sot District, where Cd pollution was endemic ( Table 1).
The overall mean age was 44.6 years. Of the total participants, 45.9% were smokers, including those who had quit less than 10 years ago. The percentages of females, hypertension, and diabetes were 51.4%, 13.8%, and 2.7%, respectively. The overall % of subjects with proteinuria, defined as urine protein ≥ 20 mg/L was 28.6%. The % of proteinuria was 30.6% when protein excretion (E Prot ) was normalized to C cr and E Prot /C cr × 100 ≥ 20 mg/L was a cutoff value. By eGFR stratification, 207 (51.1%), 147 (36.3%), and 51 (12.6%) had eGFR values > 90, 61-90, and ≤60, mL/min/1.73 m 2 , respectively. For simplicity, the eGFR values > 90, 61-90, and ≤60 mL/min/1.73 m 2 were referred to as high, moderate, and low, respectively. The distributions of men and women across these three eGFR groups were similar. The low-eGFR group was the oldest, with the highest % of smoking, diabetes, and hypertension.
Mean plasma creatinine, mean urine creatinine, mean urine protein, and mean urine Cd concentrations all were highest, middle, and lowest in the low-, moderate-, and high-eGFR groups, respectively (p < 0.001). Mean plasma total protein and mean plasma albumin in the high-and moderate-eGFR groups did not differ. Mean values of E Cd /E cr and E Prot /E cr were highest, middle, and lowest in the low-, moderate-, and high-eGFR groups, as were mean values of E Cd /C cr and E Prot /C cr (p < 0.001). Table 2 provides results of the multiple regression modeling of protein excretion as log[(E Prot /C cr × 10 5 ) where Cd excretion was incorporated as log[(E Cd /Ccr × 10 5 ) along with other five independent variables (age, diabetes, sex, hypertension, and smoking). Age, E Cd /C cr , diabetes, sex, hypertension, and smoking together accounted for 30.6%, 37%, and 25.9% of total variation in E Pro t/C cr among all subjects (p < 0.001), men (p < 0.001), and women (p < 0.001), respectively. Age and E Cd /C cr were independently associated with E Prot /C cr in men and women. A positive association of E Prot /C cr and E Cd /C cr was stronger in men (β = 0.376, p < 0.001) than women (β = 0.179, p = 0.050).

Predictors of Protein Excretion
An additional regression analysis of E Prot /C cr was undertaken to assess a potential influence of plasma protein/albumin levels using data from 190 subjects of the low-exposure group. Plasma protein, plasma albumin, age, log [(E Cd /C cr ) × 10 5 ], sex, and smoking altogether did not explain the E Prot /C cr variability (adjusted R 2 = 0.008, p = 0.281). Associations of E Prot /C cr with plasma protein, plasma albumin, and other independent variables were all not significant (p > 0.05). Table 3 provides results of multiple regression models of eGFR in which Cd excretion as log[(E Cd /Ccr × 10 5 ) was incorporated together with five over independent variables (age, diabetes, sex, hypertension, and smoking). Age, E Cd /C cr , diabetes, sex, hypertension, and smoking together accounted for 65%, 68%, and 63.5% of total variation in eGFR among all subjects (p < 0.001), men (p < 0.001), and women (p < 0.001), respectively. Lower eGFR values were associated with older age (β = −0.558, p < 0.001) and higher E Cd /C cr (β = −0.266, p < 0.001). An inverse relationship between eGFR and E Cd /C cr Cd excretion rates in women (β = −0.334, p < 0.001) was stronger than men (β = −0.178, p = 0.012). Table 4 provides results of logistic regression modeling where the prevalence odds ratios (POR) for proteinuria and low eGFR were determined for subjects of three Cd exposure groups. The independent variables included age, diabetes, sex, hypertension, and smoking. The POR for proteinuria was lower at younger ages (POR = 0.923, p < 0.001), as was the POR for low eGFR (POR = 0.888, p < 0.001). The POR for proteinuria was increased 4.6-fold in those with (E Cd /C cr ) × 100 > 8.28 µg/L filtrate (p < 0.001). Normotension was associated with a 64% decrease in POR for low eGFR (p = 0.016). The effect of hypertension on the POR for proteinuria was insignificant (p = 0.055). A dose-effect relationship was seen between POR for low eGFR and E Cd /C cr ; the POR for low eGFR was increased 4.6-fold in those with (E Cd /C cr ) × 100 ranging between 2.72 and 8.28 µg/L filtrate (p = 0.035), and the POR for low eGFR rose to 5.1 at E Cd /C cr ) × 100 > 8.28 µg/L filtrate (p < 0.001).

Excretion Rates of Various Proteins and Cadmum in the High-Exposure Group
We analyzed the urine composition of 215 subjects who resided in a Cd-contaminated area in an attempt to quantify the influence of the number of surviving nephrons and Cd exposure. By eGFR stratification, there were 33 (15.3%), 131 (61%), and 51 (23.7%) who had eGFR > 90, 61-90, and ≤60 mL/min/1.73 m 2 , respectively (Table 5).
Relative to the high-eGFR group, the excretion of creatinine tended to rise in the moderate-and the low-eGFR groups (p = 0.054), while the urinary Cd concentrations (µg/L) in three eGFR groups showed no variation (p = 0.646) ( Table 5). Consequently, the Cd body burdens of subjects in the three eGFR groups were not distinguishable (p = 0.079).
Distinctively, differences in body burden of Cd were apparent when E Cd was normalized to C cr ; the mean E Cd /C cr was highest, middle, and lowest in the low-, moderate-, and high-eGFR groups (p < 0.001). As expected, those in the low-eGFR group excreted β 2 M, α 1 M, albumin, total protein, and Cd at the highest rates. The % of E β2M /E cr ≥ 1000 µg/g creatinine were 25.6%, and more than half (66.7%) of those with such a high E β2M excretion were in the low eGFR group. The % of E β2M /E cr ≥ 1000 µg/g creatinine was lowest, middle, and highest in the high-, moderate-, and low-eGFR groups, respectively (p < 0.001).
Only one subject (0.5%) had E Alb /E cr > 300 mg/g creatinine. E Alb /E cr values in the majority of subjects (84.2%) were < 30 mg/g creatinine. Only 15.3% had microalbuminuria (E Alb /E cr values of 30-300 mg/g creatinine). The % of those with microalbuminuria was lowest, middle, and highest in the high-, moderate-, and low-eGFR groups, respectively (p < 0.001).

Correlation Analysis of Age, BMI, and Chemical Constiuents of Urine
We next undertook a correlation analysis of nine parameters that included age, BMI, excretion of β 2 M, α 1 M, albumin, total protein, and Cd (Table 6).
With the exception of BMI, age correlated positively with all other variables. BMI was negatively correlated with age, E β2M /C cr , and E α1M / cr . The excretion of creatinine correlated positively with age, E Cd /C cr , and E Prot /C cr . This parameter did not correlate with E Alb /C cr , E β2M , or E α1M /C cr . E Cd /C cr was more closely correlated with E Prot /C cr (r = 0.507), E β2M (r = 0.430), and E α1M /C cr (r = 0.364) than with E Alb /C cr .
Among three proteins quantified, β 2 M and α 1 M represented low-molecular-weight proteins. Albumin represented a high-molecular-weight protein, while urinary total pro-tein was a measure of all proteins. As expected, E Pro /C cr was highly correlated with all three proteins, β 2 M, α 1 M, and albumin (r = 0.508 − 0.546). Table 6. The Pearson correlation analysis of chemical compositions of urine samples from 215 residents of a high-exposure area.

Variables
Age BMI E Cd /C cr E Prot /C cr E Alb /C cr E β2M /C cr E α1M /C cr   Stratification of subjects by eGFR uncovered differences in renal handling of albumin vs. β2M and α1M in response to Cd. Higher Eβ2M and Eα1M values were associated with higher ECd in both eGFR groups (Figure 1a,b). The associations of Eβ2M, Eα1M, and ECd all were stronger in the low-eGFR group than the high-eGFR group. In comparison, the relationship between EAlb and ECd was insignificant in both eGFR groups (Figure 2a). An association between EAlb and EProt was particularly strong in the low-eGFR group (Figure 2b), as were the associations of EAlb and Eα1M (Figure 2d) and of EProt and Eα1M (Figure 1d). In comparison, the relationships between EAlb and Eβ2M were similar in the two eGFR groups (Figure 2c), as were the relationships between EProt and Eβ2M (Figure 1c). Stratification of subjects by eGFR uncovered differences in renal handling of albumin vs. β 2 M and α 1 M in response to Cd. Higher E β2M and E α1M values were associated with higher E Cd in both eGFR groups (Figure 1a,b). The associations of E β2M , E α1M , and E Cd all were stronger in the low-eGFR group than the high-eGFR group. In comparison, the relationship between E Alb and E Cd was insignificant in both eGFR groups (Figure 2a). An association between E Alb and E Prot was particularly strong in the low-eGFR group (Figure 2b), as were the associations of E Alb and E α1M (Figure 2d) and of E Prot and E α1M (Figure 1d). In comparison, the relationships between E Alb and E β2M were similar in the two eGFR groups (Figure 2c), as were the relationships between E Prot and E β2M (Figure 1c).

Discussion
A defective tubular reabsorption of the low-molecular-weight filterable protein β 2 M, referred to as tubular proteinuria, has been the most frequently reported adverse effect of environmental exposure to Cd. Analytical and biochemical epidemiologic research dissecting Cd-induced proteinuria is lacking. As demonstrated in the present study, mechanistic insights into proteinuria following long-term exposure to Cd and its accumulation in tubular epithelial cells of kidneys emerged when excretion rates of various proteins and Cd were normalized to creatinine (C cr ). C cr -normalization corrects for differences among subjects in number of surviving nephrons. A conventional method of normalization of the rate of excretion of a given substance to creatinine excretion (E cr ) corrects for urine dilution, but it introduces high variance into the dataset, because E cr is not related to Cd exposure nor the function of kidneys. This has led to erroneous data interpretations in the past.
As shown in Table 5, the exposure levels of Cd, measured as E Cd /E cr among groups of residents of a high-exposure area, did not differ statistically (p = 0.641); subjects in all three eGFR groups excreted a similarly high E Cd /E cr of 10 µg/g creatinine. As a result of the indistinguishable body burden, a dose-effect relationship between excretion of various proteins and a Cd exposure measure could not be established.
In comparison, Cd exposure, measured as E Cd /C cr , increased from 8.1 to 9.7 and then 17 µg/L of filtrate in subjects with high, moderate, and low eGFR, respectively (p < 0.001). Excretion rates of total protein, Alb, β 2 M, and α 1 M all increased in parallel to E Cd /C cr increment. In effect, a clear dose-response between E Cd /C cr and excretion rates of total protein, Alb, β 2 M, and α 1 M was evident.
A further analysis of the relationship of the exposure measure (E Cd /C cr ) and excretion rates of total protein, Alb, β 2 M, and α 1 M, shown in Figures 1 and 2, provided further evidence for preferential reabsorption of albumin. An association between E Alb and E Cd was absent in both low-and high-eGFR groups (Figure 2a). This finding may be interpreted to suggest that most Alb is reabsorbed, consistent with a current view that albumin is reabsorbed almost completely and that a fraction of it is returned to the systemic circulation [10][11][12][13][14][15].
In an experimental study, Cd was found to disable the cubilin/megalin receptor system of albumin endocytosis, leading to albuminuria [26]. In addition, Cd diminished expression of megalin and ClC5 channels [27]. Cd may also increase glomerular permeability to albumin, as shown in another study, where a non-cytotoxic concentration of Cd (1 µM) increased the permeability of human renal glomerular endothelial cells in monolayers and caused the redistribution of the adherens junction proteins vascular endothelial-cadherin and β-catenin [28,29].
Of concern, CKD is a progressive syndrome with high morbidity and mortality and affects 8% to 16% of the world's population, with increasing incidence worldwide [30][31][32][33]. Here, we have shown that Cd simultaneously increased risks of low eGFR and proteinuria 4-to 5-fold (Table 4). Indeed, associations between Cd exposure and low eGFR and albuminuria were repeatedly observed among participants in the U.S. National Health and Nutrition Examination Survey (NHANES) undertaken between 1999 and 2016 [34][35][36][37]. Increased E alb was associated with a urinary Cd level as low as 0.22 µg/L in a study of participants in NHANES 2009-2012, aged ≥20 years [38], while an increase in risk of CKD among NHANES 1999-2006 was associated with E Cd /E cr values ≥ 1 µg/g creatinine [34]. In a Spanish population study, a urinary Cd excretion (E Cd ) of 0.27 µg/g creatinine was associated with an increase in the risk of albuminuria by 58% [39]. In another study, E Cd > 1.72 µg/g creatinine was associated with elevated E alb in Shanghai residents [40]. In a systematic and meta-analysis of pooled data from 28 studies, Cd exposure was found to increase the risk of proteinuria by 48% [41]. The pathogenesis of Cd-induced proteinuria, especially in low environmental Cd exposure conditions, requires further study, especially in experimentation, given that the glomerular and tubular causes of albuminuria may not be distinguishable in epidemiologic studies.
The results of the present study have strongly implicated Cd exposure as a risk factor for CKD. Minimization/avoidance of Cd exposure from smoking and habitual consumption of foods containing high levels of the metal are warranted. Kidney fibrosis after chronic exposure to Cd has been demonstratable in experimental studies [42,43]. Evidence from the synchrotron imaging of metals in human kidney tissue samples is in line with Cd-induced fibrosis [44]. The degree of tubular atrophy was positively associated with the level of Cd accumulation in a histopathological examination of kidney biopsies from healthy kidney transplant donors [45]. A prospective cohort study of Japanese residents in an area similarly polluted by Cd reported a 49% increase in mortality from kidney failure among women after adjustment for potential confounding factors [46].

Participants
We assembled data from 405 persons (197 men and 208 women) who participated in the larger population-based studies undertaken in a low-exposure area (Bangkok), and an endemic area of Cd contamination in the Mae Sot District, Tak Province, Thailand [47]. The Institutional Ethical Committee, Chiang Mai University and the Mae Sot Hospital Ethical Committee approved the study protocol. At the time of recruitment, all participants had lived at their current addresses for at least 30 years, and all gave informed consent prior to their participation. Exclusion criteria were pregnancy, breastfeeding, a history of metal work, and a hospital record or physician's diagnosis of an advanced chronic disease. Smoking, diabetes, hypertension, regular use of medications, educational level, occupation, and family health history were ascertained by questionnaire. Diabetes was defined as fasting plasma glucose levels ≥ 126 mg/dL or a physician's prescription of antidiabetic medications. Hypertension was defined as systolic blood pressure ≥ 140 mmHg, diastolic blood pressure ≥ 90 mmHg, a physician's diagnosis, or prescription of antihypertensive medications.

Collection and Analysis of Blood and Urine Samples
Simultaneous blood and urine sampling is required to normalize E Cd to C cr . Accordingly, second morning urine samples were collected after an overnight fast, and whole blood samples were obtained within 3 h after the urine sampling. Aliquots of urine, whole blood, and plasma were stored at −20 • C or −80 • C for later analysis. The assay for urine and plasma concentrations of creatinine ([cr] u and [cr] p ) was based on the Jaffe reaction. The assay of β 2 M in urine ([β 2 M] u ) was based on the latex immunoagglutination method (LX test, Eiken 2MGII; Eiken and Shionogi Co., Tokyo, Japan). The assay of urinary albumin ([alb] u ) was based on a turbidimetric method (UniCel ® DxC800 Synchron system, Beckman Coulter, Fullerton, CA, USA).
For the Bangkok group, the urinary concentration of Cd ([Cd] u ) was determined by inductively coupled plasma mass spectrometry (ICP/MS, Agilent 7500, Agilent Technologies, Santa Clara, CA, USA) because it had the high sensitivity required to measure very low Cd concentrations. Multi-element standards (EM Science, EM Industries, Inc., Newark, NJ, USA) were used to calibrate the Cd analyses. The accuracy and precision of those analyses were ascertained with reference urine (Lyphochek ® , Bio-Rad, Sydney, Australia). The low limit of detection (LOD) of urine Cd, calculated as 3 times the standard deviation of blank measurements, was 0.05 µg/L. The Cd concentration assigned to samples with Cd below the detection limit was the detection limit divided by the square root of 2 [48].
For the Mae Sot group, [Cd] u was determined with atomic absorption spectrophotometry (Shimadzu Model AA-6300, Kyoto, Japan). Urine standard reference material No. 2670 (National Institute of Standards, Washington, DC, USA) was used for quality assurance and control purposes. The low limit of detection of Cd quantitation, defined as 3 times the standard deviation of blank measurements, was 0.06 µg/L. None of the urine samples contained [Cd] u below the detection limit.

Estimated Glomerular Filtration Rates and CKD Stratified by the KDIGO Categories
The GFR is the product of nephron number and mean single nephron GFR, and, in theory, GFR is indicative of nephron function [19,20,49]. In practice, the GFR is estimated from established Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) equations and is reported as eGFR [20,49].
Male [cr] u = urine creatinine concentration (mg/dL). E x /C cr was expressed as the excretion of x per volume of filtrate [25].

Statistical Analysis
Data were analyzed with IBM SPSS Statistics 21 (IBM Inc., New York, NY, USA). The one-sample Kolmogorov-Smirnov test was used to identify departures of continuous variables from a normal distribution, and a logarithmic transformation was applied to variables that showed rightward skewing before they were subjected to parametric statistical analysis. The Kruskal-Wallis test was used to assess differences in means among three exposure groups. The chi-square test was used to determine differences in percentage and prevalence data. To determine strength of association of excretion rates of total protein and eGFR with independent variables, the multiple linear regression model analysis was used. The multivariable logistic regression analysis was used to determine the prevalence odds ratio (POR) for albuminuria and CKD in relation to six independent variables: age, gender, diabetes, smoking, hypertension, and Cd exposure, measured as Cd excretion (E Cd ).

Conclusions
The conventional method for adjusting excretion rates of Cd and proteins of low and high molecular weight, namely, β 2 M, α 1 M, albumin, and total protein, to creatinine excretion understate the severity of the nephrotoxicity of Cd. The body burden of Cd among residents of an area affected by Cd contamination is indistinguishable when the urinary Cd excretion levels are adjusted to creatinine excretion, thereby nullifying a dose-response relationship analysis. In comparison, normalization of excreted Cd, β 2 M, α 1 M, albumin, and total protein to creatinine clearance enables a dose-effect analysis and provides, for the first time, evidence that proteinuria after chronic Cd exposure is a manifestation of Cdinduced tubulopathy and Cd-induced nephron loss. The results of a dose-effect analysis also provide evidence that albumin is preferentially reabsorbed.