**4. Discussion**

Our goals in the present study were to elucidate the source of urinary Cd and to relate that source to the pathogenesis of Cd nephropathy. Data were obtained from clinically healthy Thai subjects residing in areas with low, moderate, or high exposure to Cd. In Mae Sot District, Tak Province, intensity of exposure was determined from the Cd content of rice grains [35,36]. Exposure in Bangkok was assumed to be low on the basis of food analyses and dietary histories [33].

Subjects in the three subsets were demographically dissimilar (Table 1). Both age and percentage of smokers rose with intensity of exposure. The percentage of women was particularly high in the moderate exposure group, and some women were of childbearing age. Because iron and Cd share a transporter in intestinal epithelium, menstruating women in this group may have absorbed Cd with exceptional avidity and incurred exceptional tubular toxicity secondarily (Table 1). In the high exposure subset, increased age may have conferred additional reasons for deterioration of GFR, and smoking may have provided a second environmental source of Cd. Whether smoking itself could have accelerated the progression of CKD is unresolved [40–42]. Neither age, nor smoking per se, nor the source of exogenous Cd obscured the significant relationship between eGFR and ECd/Ccr.

Table 1 shows that with the increasing intensity of exposure, ECd/Ccr rose and eGFR fell in stepwise fashion. In contrast, both ENAG/Ccr and <sup>E</sup>β2MG/Ccr were higher in the moderate than in the high exposure group. Although one might expect a direct relationship between NAG excretion and the number of intact nephrons, the higher ENAG/Ccr in the moderate group implies that the median excretion of NAG *per intact nephron* was also higher in these subjects. At the same time, the overlap of ENAG/Ccr between the moderate and high exposure groups was substantial (Figures 2 and 5), and

consistent relationships among ECd/Ccr, ENAG/Ccr, and eGFR were demonstrable at all intensities of exposure (Figures 1–3). We speculate that the number of menstruating women in the moderate exposure subset was sufficient to increase Cd absorption and tubular toxicity in the entire group, but insufficient to disrupt the statistical relationships seen in all groups among ECd/Ccr, ENAG/Ccr, and eGFR.

Analogous statements can be made about β2MG. As impaired reabsorption of this protein is an early sign of proximal tubular injury, it is not surprising that median <sup>E</sup>β2MG/Ccr tracked with median ENAG/Ccr (Table 1). As would be expected, <sup>E</sup>β2MG/Ccr varied directly with ENAG/Ccr in the moderate and high exposure groups, but also varied directly with ECd/Ccr, which increased progressively with the intensity of exposure (Figures 4 and 5).

Although CKD-EPI equations estimate GFR imprecisely [37,38], each exposure group differed significantly from the others with respect to eGFR (Table 1). Estimated GFR was inversely related to ECd/Ccr and ENAG/Ccr in the entire sample and each subset (Figures 1 and 2), and ENAG/Ccr varied directly with ECd/Ccr (Figure 3). For all comparisons in Figures 1–3, both linear and quadratic relationships were significant, and with one exception (Figure 1B), R<sup>2</sup> rose with the exposure intensity. Standardized β followed the same pattern. Despite the statistical significance of all comparisons, some R<sup>2</sup> values indicated that the fractional contribution of ECd/Ccr or ENAG/Ccr to eGFR was <10% (Figures 1B and 2A); simultaneously, however, standardized β indicated robust effects of changes in *x* on changes in *y*. Factors other than ECd/Ccr and ENAG/Ccr affected eGFR, but variation in each ratio was associated with substantial variation in eGFR.

Taken together, the graphs in Figures 1–3 imply that in each subset, GFR was inversely related to the severity of cellular injury per nephron (ENAG/Ccr), which in turn was associated with the amount of Cd excreted per nephron (ECd/Ccr). In addition, the quadratic relationships in Figures 1 and 2 sugges<sup>t</sup> that small increments in the most advanced injury were accompanied by disproportionate reductions in GFR. Slope analyses of curves in Figures 1D and 2D confirm this inference (Table 2).

Other investigators have described direct relationships of [NAG]u/[cr]u and eGFR to [Cd]u/[cr]u, but we have not found a synthesis of those relationships into a satisfactory pathophysiologic narrative [35,43–48]. A cogen<sup>t</sup> interpretation of Figures 1–3 must explain how eGFR and ENAG/Ccr—results of *cumulative* Cd sequestration—were related physiologically to ECd/Ccr, an indicator of Cd excretion *at the time of sampling*. Cd was excreted for two possible reasons; it was filtered and not reabsorbed, or it was released from tubular cells [32]. Although both processes may have occurred, we do not see how the first, excretion after filtration, could have produced a physiologic connection between ECd/Ccr and ENAG/Ccr (Figure 3). In contrast, if Cd was released from damaged tubules, then Cd and NAG emanated from the same source, and both ECd/Ccr and ENAG/Ccr measured cellular injury. This shared attribute of Cd and NAG explains the statistical association of the ratios.

Additional evidence for the tubular origin of excreted Cd is provided by demonstrated extrusions of MT into tubular lumens [49], documented correlations of ECd with renal tissue content of Cd [50–52], and direct relationships between ECd and GFR (number of intact nephrons) [53–55]. Experiments in rabbits demonstrated a high tubular maximum for reabsorption of CdMT that would preclude excretion of filtered Cd in the typically intoxicated human [56].

In addition to addressing the likely source of excreted Cd, we must also ask why declining eGFR, the result of continuous loss of intact nephrons over time, was associated in the present study with parameters of *current* cellular injury, ECd/Ccr and ENAG/Ccr. To address this paradox, we propose that eGFR, ECd/Ccr, and ENAG/Ccr were simultaneous consequences of the tubular content of Cd. As the content rose, cellular injury per nephron and the rate of nephron loss increased proportionately; at any moment in a subject's exposure history, the three variables were quantitatively associated because they were traceable to the same burden of sequestered Cd.

β2MG, a small protein made by nucleated cells, is almost completely filtered by glomeruli [29]. Ordinarily, the proximal tubule reabsorbs and degrades over 99% of filtered β2MG [30]. Because tubulopathies increase <sup>E</sup>β2MG [31], excessive <sup>E</sup>β2MG is conventionally interpreted as evidence of reabsorptive dysfunction [35,57,58]. This interpretation is understandable, but we suspect that it is an oversimplification. One reason is that endogenous production of β2MG may be increased by chronic inflammatory conditions, solid tumors, lymphatic malignancies, and multiple myeloma [59]. If tubular degradation (TDβ2MG) remains constant as production rises, <sup>E</sup>β2MG also rises even though TDβ2MG has not fallen (SM). Moreover, if both β2MG production and TDβ2MG per volume of filtrate (TDβ2MG/Ccr) remain constant as GFR falls, TDβ2MG also falls, and <sup>E</sup>β2MG rises (SM). Although these inferences are unproven, it seems likely that a combination of reabsorptive dysfunction and reduced GFR caused associations of <sup>E</sup>β2MG/Ccr with ECd/Ccr and ENAG/Ccr (Figures 4 and 5).

An additional observation requires explanation. In the low exposure group, a cluster of subjects exhibited exceptionally low <sup>E</sup>β2MG/Ccr. At a fixed rate of β2MG filtration (equal to endogenous production) and a fixed value of TDβ2MG/Ccr, the rate of β2MG reabsorption increases with the number of intact nephrons, and <sup>E</sup>β2MG decreases simultaneously. In the isolated cluster, mean eGFR was 105.3 mL/min/1.73 m2; in the remainder of subjects in the study, it was 89.3 mL/min/1.73 m2. We suspect that extremely low <sup>E</sup>β2MG/Ccr in the cluster was the result of high Ccr, high TDβ2MG, and secondarily reduced <sup>E</sup>β2MG.

In the present study, we have continued the recently introduced practice of normalizing excretion of Cd, NAG, and β2MG to creatinine clearance instead of creatinine excretion [26]. Because the resulting ratios express ECd, ENAG, and <sup>E</sup>β2MG as functions of intact nephron mass, they nullify sources of imprecision that accompany normalization to Ecr or [cr]u. At any GFR, Ecr is primarily a function of muscle mass [60]; consequently, at a given ECd (for example), [Cd]u/[cr]u may vary by a multiple over the range of human body size. Moreover, multiple groups have reported *direct* rather than inverse relationships between GFR and [Cd]u/[cr]u after Cd exposure [5,53–55]. If the nephron number determines ECd at a given cellular burden of the metal, then [Cd]u/[cr]u may exaggerate the burden at normal GFR and underestimate it at reduced GFR. Normalization of ECd to Ccr—that is, calculation of [Cd]u[cr]p/[cr]u—eliminates the confounding effects of both muscle mass and nephron number on [Cd]u/[cr]u. In addition, because the required measurements are made in aliquots of urine and serum, the calculation quantifies amounts of Cd (or other substances) excreted per volume of filtrate while eliminating the need for timed urine collections and direct determinations of GFR. We plan to address optimal expression of excretion rates relevant to Cd nephropathy in a separate publication.

In summary, we draw the following conclusions from the significant regressions described herein. ENAG/Ccr varied directly with ECd/Ccr because sequestered Cd induced the release of NAG and Cd from tubular cells into filtrate. Estimated GFR varied inversely with both ratios because all three parameters reflected the extent of tubular Cd accumulation. ENAG/Ccr and ECd/Ccr quantified ongoing cellular injury, and eGFR quantified the loss of intact nephrons. We suspect that the significant regressions of <sup>E</sup>β2MG/Ccr on ECd/Ccr and ENAG/Ccr resulted from effects of Cd on both tubular reabsorption and nephron number.

**Supplementary Materials:** The following are available online at http://www.mdpi.com/2305-6304/7/4/55/s1, The equation for normalization to creatinine clearance.

**Author Contributions:** S.S., D.A.V., W.R., M.N., and G.C.G. formulated the study designs and protocols. S.S. and W.R. obtained ethical institutional clearances for research on human subjects and supervised the collection of biologic specimens in Thailand. S.S. organized and analyzed the data, created the tables and figures, and revised the manuscript for important intellectual content. K.R.P. proposed normalization of excretion rates to creatinine clearance, provided logical data interpretation, and was the primary author of the manuscript.

**Funding:** This research received no external funding.

**Acknowledgments:** This work was partially supported by the Commission for High Education, Thailand Ministry of Education, and the National Science and Technology Development Agency (NSTDA). Additionally, it was supported with resources of the Stratton Veterans' Affairs Medical Center, Albany, NY, USA, and was made possible by facilities at that institution. Opinions expressed in this paper are those of the authors and do not represent the official position of the United States Department of Veterans' Affairs.

**Conflicts of Interest:** The authors have no potential conflicts of interest to declare.

#### *Toxics* **2019**, *7*, 55
