Iron (Fe) is an essential metal for life. Evolution has harnessed its favorable redox potential, and consequently, Fe is a key constituent of numerous essential proteins, including hemoglobin, cytochromes, and NADH-coenzyme Q reductase (Ponka, 1999; Williams, 2012). Organisms have evolved mechanisms that exercise fine control over Fe balance because Fe deficiency or excess result in morbidity and mortality. The discovery of Fe transporter proteins expressed in kidney nephron epithelial cells has led to the suggestion that Fe is reabsorbed by the kidney (Gunshin et al., 1997; Kozyraki et al., 2001; Wolff et al., 2011; Smith et al., 2019). Yet, a key piece of knowledge is lacking in support of this suggestion, namely, there are no direct measures of Fe in kidney tubular fluid or ultrafiltrate. Not surprisingly, in the absence of this fundamental piece of information, the prevailing dogma persists that healthy kidneys do not filter Fe because Fe is tightly bound to transferrin and in the absence of pathology transferrin is not filtered.

Since the discovery and characterization of the first mammalian Fe transporting protein DMT1 (Gunshin et al., 1997), we have worked to determine the role of this and other Fe transporting proteins in the kidney (Smith and Thévenod, 2009; Thévenod and Wolff, 2016). From our work and that of others, it is well established that the components necessary for transcellular Fe transport are expressed in mammalian proximal convoluted tubule (PCT) cells. Briefly, the membrane-bound divalent metal transporters DMT1, ZIP8, ZIP14, and ferroportin 1 are expressed in PCT cellular membranes (Abouhamed et al., 2006; Wolff et al., 2011; Ajjimaporn et al., 2012; Van Raaij et al., 2018). Furthermore, transferrin receptor 1 and the megalin/cubilin receptor complex, both suggested to mediate reabsorption of filtered protein-bound Fe, are expressed in the apical PCT membrane (Christensen et al., 2012; Smith et al., 2019).

Despite this relative wealth of knowledge, the key questions whether Fe is present in ultrafiltrate and if so whether Fe is reabsorbed by the kidney remain unanswered. In view of these gaps in our collective knowledge, the aim of this study was to determine the Fe concentration in PCT S2 rat ultrafiltrate, determine whether the proximal tubule reabsorbs Fe, and begin to characterize the pathway. To do this, we collected nanoliter samples of PCT fluid from anesthetized rats and measured Fe concentration in tubular fluid along the PCT. In addition, using in vivo microperfusion, we measured PCT unidirection flux of radiolabeled Fe (⁵⁵Fe) and assessed the effect of competition by cadmium, copper, manganese, or zinc on ⁵⁵Fe flux.

The aim of the current study was to determine the Fe concentration in ultrafiltrate collected from the lumen of rodent S2 proximal tubule and to determine whether Fe is reabsorbed by the PCT. To do this, we utilized in vivo renal micropuncture and measured Fe concentration in samples of ultrafiltrate collected from S2 nephron segment. We also performed divalent metal competition studies to begin to functionally characterize the Fe reabsorptive pathway.

Tubular fluid collected from rat PCT contained Fe. This finding suggests Fe is either filtered by the glomerulus or secreted into the tubular lumen. By comparing tubular Fe concentration to the distance, a collection was made from the glomerulus and we determined whether Fe was reabsorbed or secreted. Because 50–60% of the water in the ultrafiltrate is reabsorbed from the PCT, a solute that is not reabsorbed or one that is secreted would increase in concentration the further a collection was made from the glomerulus (Rector, 1983). In the case of isosmotic reabsorption, no change in concentration would be expected and for higher rates of reabsorption the concentration of an ion/solute along the PCT would decrease. By comparing collections made in the first and last accessible millimeter of the S2 PCT, it was very evident that there was no discernable increase in PCT Fe and the concentration remained relatively constant or if anything decreased slightly along the PCT. These findings when taken together strongly suggest Fe is not secreted into the tubule lumen, but that it is filtered by the glomerulus and reabsorbed along the proximal tubule.

Comparing Fe flux and water flux enabled us to determine whether reabsorption of Fe might be linked to water movement. We found a correlation between water and Fe flux, indicating that the two processes were linked. This was not unexpected given that movement of a solute across a membrane generates an osmotic gradient. On the other hand, movement of water out of the tubule serves to increase the effective concentration of a solute and hence increases the solute concentration gradient thus favoring transport out of the tubular lumen. It was not possible from our experiments to definitively determine which mechanism led to the observed coupled movement; however, the finding that PCT Fe and water movement were linked reinforces the conclusion that Fe is reabsorbed from the PCT.

To begin to functionally characterize the proposed reabsorptive pathway, we performed a series of PCT microperfusion experiments. We perfused PCTs and measured the flux of ⁵⁵Fe out of the proximal tubule as a marker for Fe reabsorption. To ensure that the PCT was functioning normally, we simultaneously measured sodium and water flux and obtained values very comparable to those previously reported by our laboratory under similar perfusion conditions (Wareing and Green, 1994; Wareing et al., 1995). The mean recovery of Fe from PCT’s perfused with a solution containing 12μmol.l⁻¹ Fe was 66%. Therefore, on average, a third of the Fe presented to the PCT was taken up by the tubule. The rate of Fe transport measured was 105.2±12.7 fmol. mm⁻¹.min⁻¹, and since this is the first measurement of PCT Fe flux, it is not possible to compare it to other published data, and while the fluxes of other ions out of the PCT (e.g., Na, Ca, K, or Mg) have been measured by our group and others groups given the methodological differences, it is not possible to directly compare these values.

Adding unlabeled Fe to the perfusate decreased the apparent unidirectional ⁵⁵Fe flux. This was due to unlabeled Fe competing with ⁵⁵Fe, possibly for a fixed number of transporter carrier binding sites. Recalculating the molar activity of ⁵⁵Fe to take into account the addition of unlabeled Fe revealed that increasing the Fe concentration five-fold caused a three-fold increase in Fe flux. This outcome is important for two reasons: Firstly, it shows that our experimental protocol was capable of uncovering decreased or increased Fe flux; secondly, it showed that the transport pathway function was concentration dependent.

The membrane-bound metal transporter proteins DMT1, ZIP8, and ZIP14 are expressed in the mammalian PCT (Wang et al., 2012; Van Raaij et al., 2018). These proteins are selective for several divalent metals including Fe²⁺. With this in mind, our next aim was to determine the characteristics of PCT Fe flux. Firstly, we looked at the effect of including ascorbate in the perfusate, reasoning that DMT1, ZIP8, and ZIP14 mediate the transport of Fe²⁺ and the perfusate we used contained predominately Fe³⁺; thus, inclusion of ascorbate would be expected to increase Fe flux (Sparkman et al., 2011; Illing et al., 2012; Wang et al., 2012). We found ascorbate did not alter Fe flux suggesting that ascorbate was not necessary for PCT Fe translocation. Ferrous Fe²⁺ is promoted by inclusion of ascorbic acid; therefore, the finding that ascorbate did not increase flux argues against involvement of DMT1, ZIP8, or ZIP14 in transporting Fe out of the PCT. Interestingly, ascorbate caused water flux to increase marginally. This might reflect the transport activity of ascorbate by the Na+−coupled L-ascorbic acid transporter SVCT1. SVCT1 is expressed in the PCT, and like SGLT1 is likely to transport water and its cognate substrate (Castro et al., 2008; Mackenzie et al., 2008; Bürzle et al., 2013).

To further probe the reabsorptive pathway, we performed ion competition experiments. A prominent characteristic of DMT1, ZIP8, and ZIP14 is that in addition to Fe, these proteins transport other divalent metals: DMT1 has substrate selectivity for Cd²⁺, Co²⁺, Mn²⁺, and to a lesser extent Zn^2+; ZIP 8 has selectivity for Cd²⁺, Co²⁺, Zn²⁺, and to a lesser extent Mn²⁺ and ZIP 14 has the capacity to mediate transport of Cd²⁺, Mn²⁺, and Zn²⁺(Sparkman et al., 2011; Illing et al., 2012; Wang et al., 2012). In light of this, we microperfused tubules with solutions containing Cd, Zn, Cu, or Mn. Addition of competing metals did not significantly alter Fe, sodium, or water flux, although addition of CdSO₄ to the perfusate caused a marginal decrease in sodium and water flux that did not reach statistical significance. Taken together, these data suggest DMT1, ZIP8, and ZIP14 are unlikely major contributors to unidirectional Fe flux out of the PCT.

In our previous work, we proposed that DMT1 is responsible for transport of Fe across the apical membrane of the late loop of Henle and early distal tubule because we showed DMT1 is expressed in the apical plasma membrane of these segments and transport of Fe was reduced by inclusion of other divalent metals which are known ligands of DMT (Wareing et al., 2000; Ferguson et al., 2001). In contrast, DMT1 is unlikely to transport Fe across the apical PCT membrane based on the functional studies presented in this work and this is supported by robust immunolabelling studies showing DMT1 expressed in late endosomes and lysosomes of PCT cells in rodents and not in the apical plasma membrane (Abouhamed et al., 2006).

If not DMT1, then could ZIP8 or ZIP14 be responsible for PCT Fe flux? In order to fulfil this role, these transporters would have to be expressed in the apical PCT membrane. Unfortunately, to date, it is not possible to discern with any degree of certainty whether these proteins are expressed in this membrane or like DMT1 whether they are expressed in a subcellular compartment. Furthermore, in the current study none of the competing divalent metals we trialed which are known ligands of ZIP8 or ZIP14 inhibited Fe transport indicating that in the absence of pathology, neither of these proteins mediate transport of Fe in the segments we investigated (mainly S2). In support of this tentative conclusion, ZIP8 has negligible transport activity at pH 6.5, and the prevailing pH in the S2 segment is ~6.7 (Dubose et al., 1979; Wang et al., 2012). In comparison, ZIP14 shows some activity at pH6.5 (Sparkman et al., 2011); however, the lack of effect of zinc on Fe transport argues strongly against the proposed involvement of this protein in PCT apical membrane Fe flux (Liuzzi et al., 2006; Fujishiro et al., 2012).

Other potential candidate mechanisms are apparent in the PCT. The megalin/cubilin receptor complex and TFR1 are both expressed in the apical PCT membrane are likely to be responsible for reabsorbing filtered transferrin bound Fe (Christensen et al., 2012; Smith et al., 2019). Our micropuncture experiments simply demonstrated that Fe was present in ultrafiltrate and that the Fe concentration profile along the proximal tubule strongly indicated reabsorption; however, we did not assess what species of Fe was present. The tubular perfusion experiments we performed used a perfusate which was protein-free and contained Fe in its unbound state – sometimes referred to as non-transferrin bound Fe. We specifically designed the experiments to enable us to functionally test for the presence of DMT1, ZIP8, and ZIP14 since these transporters translocate unbound Fe. Clearly, future experiments should aim to examine the role of receptor-mediated mechanisms while bearing in mind the observations of the current study that indicate the presence of transport of elemental unbound- Fe.

Overall based on the data presented in this work, what is the contribution of renal Fe reabsorption to Fe balance? We propose that in addition to its role as a source of EPO, the kidney is critical for maintaining Fe balance due to the following reasoning: Within serum, the concentration of Fe measured using the TCA precipitation method was 50.5±3.9μmol.l⁻¹. The majority of this represents Fe bound to transferrin, and some of this passes the glomerular filtration barrier and becomes a component of the ultrafiltrate. The plasma volume of a 250g rat is 9.8 mls (Bijsterbosch et al., 1981); therefore, the total amount of Fe in plasma is 0.49μmol. In vitro we measured the ultrafiltered concentration of Fe to be 4.5μmol.l⁻¹ and in vivo the mean PCT concentration in the S2 segment to be 3.0μmol.l⁻¹. Assuming a mean glomerular filtration rate of a 250g rat is 2.25l.day⁻¹, then between 6.9 and 10.1μmol Fe would be filtered per day. In terms of maintaining Fe balance, the implication of this estimate is that the total amount of Fe in plasma would be lost in less than two hours if the kidney did not reabsorb filtered Fe. From the available data, we constructed a model of renal Fe handling (see Figure 6). It can be seen that the bulk of Fe is reabsorbed in the PCT, by a non-DMT1-dependent pathway that is not shared by copper, manganese, or zinc. There is significant DMT1-mediated transport of Fe by the loop of Henle and the collecting duct system which in comparison may also mediate transport of copper and manganese and to some extent zinc.

Micropuncture is not possible in human subjects, and imaging methods are not sufficiently refined to resolve nephron distribution of Fe; however, recent data suggest that transporter proteins necessary to affect epithelial Fe transport are present in the human kidney and may in some cases be implicated in deleterious renal Fe accumulation (Van Raaij et al., 2018). And although it can be inferred that Fe is filtered by human kidneys, definitive localization studies to pinpoint the nephron expression of transporter proteins are lacking and should ideally be the focus of future studies.

In conclusion, we have demonstrated that Fe is present in ultrafiltrate in the S2 PCT. The Fe concentration profile along the PCT is indicative of PCT Fe reabsorption. The mechanism responsible for the observed flux remains obscure; however, DMT1, ZIP8, or ZIP14 are unlikely to majorly contribute. We propose that the rat kidney filters and reabsorbs Fe and this process is potentially of critical importance to Fe homeostasis.

The original contributions presented in the study are included in the article/supplementary material, and further inquiries can be directed to the corresponding author.

All authors listed have made a substantial, direct and intellectual contribution to the work, and approved it for publication.

We gratefully acknowledge the financial support of the Wellcome Trust Grant no. 043322/Z/94 (CPS) and the Royal Society (CPS).

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.