Fatty Acids and Albumin are Transported by Distinct Mechanisms in the Proximal Tubule

Abstract

Under physiologic conditions, proximal tubules depend on basolateral fatty acid (FA) uptake for metabolism. In pathophysiologic conditions due to glomerular filtration barrier disruption, albumin-bound FA undergo filtration and proximal tubule reabsorption, which leads to lipotoxicity and tubular atrophy. Apical proximal tubule albumin uptake is accomplished by the megalin/cubilin complex and receptor-mediated endocytosis, whereas apical proximal tubule FA uptake is primarily mediated by apical fatty acid transport protein-2 (FATP2). However, a commonly proposed (but untested) alternative model is that intact albumin-FA complex is co-transported by megalin/cubilin-mediated endocytosis, similar to apolipoproteins. Microperfused mouse proximal tubules demonstrated divergent one- vs. two-phase albumin and FA uptake kinetics, with significantly faster albumin compared to FA uptake. LLC-PK1, HPCT and OK proximal tubule cell lines all expressed megalin, cubilin and FATP2 mRNA, though in varying amounts. LLC-PK1 cells showed similar one-phase kinetics of dual fluorescently-labeled albumin and FA uptake, whereas HPCT cells demonstrated one-phase albumin and two-phase FA uptake kinetics, with significantly faster albumin compared to FA uptake (similar to perfused proximal tubules). FATP2 inhibition blocked FA uptake, but had no effect on albumin uptake in LLC-PK1 and HPCT cells. Megalin and cubilin deletion in OK cells inhibited albumin uptake, but had no effect on FA uptake. We conclude that apical proximal tubule albumin and FA are transported by distinct mechanisms, implying that FAs dissociate from albumin within the proximal tubule lumen prior to uptake.

NEW & NOTEWORTHY

Reabsorption of aberrantly filtered albumin-bound fatty acids by the apical proximal tubule is important for chronic kidney disease progression. Whether fatty acids and albumin are taken up as intact complexes or dissociate within the lumen prior to uptake has been controversial. Data derived from in vitro and ex vivo models demonstrate separate albumin and fatty acid uptake kinetics, implying dissociation prior to uptake.

Graphical Abstract

INTRODUCTION

In the proximal tubule ATP generation is achieved primarily through β-oxidation of long-chain (C14-C20) fatty acids (FA) to acetyl CoA, which subsequently enter the Krebs cycle (1). The greater ATP yield from FA compared to glucose metabolism fulfills the high metabolic demand of the proximal tubule (2). Intracellular FA accumulate by de novo lipogenesis, lipolysis or uptake from extracellular fluid. There are three sources of circulating FA: unbound/free FA, FA covalently bound to glycerol as triglycerides, or most commonly, FA non-covalently bound to albumin (3). The concentration of free FA in the circulation is ~7 nM, which is 10⁵-fold lower than the bound concentration (4, 5). The extracellular release of FA from triglycerides is catalyzed by lipoprotein lipases (6), while FA dissociation from albumin is enzyme-independent, but still rapid (7).

Apical proximal tubule exposure to FA is infrequent, since albumin and triglycerides are too large to cross the glomerular filtration barrier. Even in glomerular injury associated with the nephrotic syndrome, the diameter of triglycerides exceeds the threshold for filtration (8), and are therefore unlikely to appear in the urine in large concentrations. However, glomerular filtration barrier injury permits albumin-bound FA passage into the ultrafiltrate, and apical proximal tubule FA reabsorption (9). The resulting apical plus constitutive basolateral proximal tubule FA uptake contributes to lipotoxicity-induced tubular atrophy (10).

In most tissues, FA uptake is achieved by membrane-localized transporters (11, 12). Several proximal tubule FA transporters have been described, including FA transport protein-2 (FATP2), kidney injury molecule-1 (KIM-1), CD36, and free FA receptor 1 (FFAR1) (13–16). In whole animal organ surveys, FATP2 is most abundantly expressed in kidney, and within kidney, FATP2 localizes exclusively to the apical proximal tubule membrane (13, 17). Apical perfusion of FATP2 gene (Slc27a2)-deleted proximal tubules resulted in significantly diminished FA uptake (13), implying that FATP2 is the major apical FA transporter.

Despite the existence of proximal tubule FA transporters, the precise mechanism of apical FA uptake is still debated. In one model, FA and albumin dissociate within the proximal tubule lumen. Free FA are then reabsorbed by apical FA transporters, and albumin by the megalin/cubilin complex and receptor-mediated endocytosis (18). This paradigm is consistent with models of FA-albumin dissociation prior to transport in liver, heart and in vitro models (3, 19–21). Furthermore, in diabetic mice Slc27a2 deletion reduced albuminuria (22), implying that albumin reabsorption increases in the setting of reduced FA uptake. However, a commonly cited alternative model persists, which proposes that intact albumin-bound FA are co-transported across the apical proximal tubule membrane, and then undergo intracellular dissociation within endosomes (23–30). In this scenario, the primacy of FA transporters is diminished, and albumin-FA co-transport by megalin/cubilin receptor-mediated endocytosis has been inferred, in part because the megalin/cubilin complex is associated with uptake of other FA-containing ligands (31), such as apolipoproteins and liver-type FA binding protein.

Using ex vivo microperfusion and proximal tubule cell culture models, we examined whether apical proximal tubule uptake of albumin-bound long-chain FA is mediated by separate transporters or a single co-transport mechanism. We conclude that proximal tubule FA and albumin transport are distinct mechanisms, implying extracellular albumin-FA dissociation prior to uptake.

MATERIALS AND METHODS

Proximal tubule microperfusion.

Mouse S2 proximal tubule segments were microdissected and apically perfused with iFluor 350-labeled FA-free BSA (0.2 mg/ml, Sigma Aldrich, St. Louis, MO) complexed with BODIPY-labeled FA (2.5 μM, Invitrogen, Carlsbad, CA) according to previously described methods (13). The fluorescence detection system was mounted on a Nikon TE2000 inverted microscope (Tokyo, Japan). Excitation λ = 345 and 490 nm pulses were delivered for iFluor/BODIPY, respectively, and emission λ = 450 and 510 nm were recorded at 6.7-sec intervals for up to five min. Images were recorded using a 40X immersion oil objective and a Coolsnap HQ digital camera (Photometrics, Tucson, AZ). Fluorescence measurements were recorded using Metafluor version 7 imaging software (Universal Imaging, Downingtown, PA). No differences were noted between tubules derived from male and female mice for any parameters (Figure S1). Therefore, equal numbers of male and female mice were used in each experiment. All studies were conducted in accordance with protocols approved by the Institutional Animal Care and Use Committee of Case Western Reserve University School of Medicine.

FA-albumin dissociation in solution.

Palmitate (2.5 μM) was labeled with BODIPY and then complexed to 0.2% albumin as previously described (13). Unlabeled BODIPY was removed with a 7 kDa cutoff Zeba Spin Desalting Column (Thermo Fisher, Waltham, MA). The albumin-palmitate complex was resuspended in 3 ml microperfusion buffer, and then dialyzed in tubing with 1 kDa cutoff (SpectrumLabs, Rancho Dominguez, CA), which was submerged in 1L microperfusion buffer. Aliquots (100 μl) from the dialysis bag were sampled hourly to determine loss of BODIPY fluorescence, as an index of dissociation. No unaccounted increase in fluid volume within the dialysis tubing was observed at any time point, indicating that the decline in sample fluorescence over time was not due to oncotic effects of 0.2% albumin, and water movement into the dialysis bag.

Proximal tubule cell lines.

Human HKCC cells were a gift from Dr. Lorraine Racusen (32), and are cultured on plastic in DMEM-F12 (Invitrogen) + 10% fetal bovine serum as previously described (33). Human proximal tubule cells (HPCT) were obtained from Dr. Ulrich Hopfer (34), and stably transfected with lentiviral vectors to overexpress human FATP2. HPCT cells were maintained on permeable supports in Keratinocyte-SFM (Thermo Fisher) + 1% FBS, then passaged to plastic wells, and grown to confluence for uptake studies. Porcine LLC-PK1 cells were purchased from ATCC (Manassas, VA) and maintained on plastic in DMEM-F12 (Invitrogen) + 10% FBS as previously described (35). Opossum kidney (OK-P subclone) parental and knockout (KO) proximal tubule cell lines were cultured on permeable supports in DMEM-F12 + 10% FBS as previously described (36), then passaged to plastic wells, and grown to confluence for uptake studies. CRISPR/Cas9 megalin (Lrp2) and cubilin (Cubn) KO OK clones were previously characterized (37).

Quantitative PCR.

FATP2, megalin and cubilin mRNA expression were determined by quantitative PCR using a QuantStudio 3 machine (Applied Biosystems, Foster City, CA) according to established methods (38). Custom primers were purchased from Eurofins (Lancaster, PA). Primer sequences are listed in Table S1.

Fluorescent probe labeling.

FA-free bovine serum albumin (Sigma-Aldrich) in 0.1 M NaHCO₃ was slowly mixed with Texas Red-X succinimidyl ester (10 mg/mL in DMSO). The reaction was stopped with 1.5 M hydroxylamine incubation (1 hr, room temp), and Texas red-coupled albumin was eluted from Zeba Spin desalting columns (7K MWCO, Thermo Scientific) with 20 mM HEPES/HBSS washes. The Texas red-albumin was then complexed with BODIPY-conjugated dodecanoic acid containing a four-carbon linker (C16-FA, QBT fatty acid uptake assay kit containing a quenching reagent; Molecular Devices, Sunnyvale, CA, final concentration 2.5 μM in 0.2% albumin) according to previously described methods (13, 39). The fluorescence spectral properties of Texas red and BODIPY are shown in Table S2, and indicate similar luminosity ranges for both fluorophores. For microperfusion studies, identical methods were used to couple iFluor 350 succinimidyl ester (AAT Bioquest, Pleasanton, CA) to bovine serum albumin, which was then complexed with BODIPY-conjugated FA.

Uptake methods.

Cells were seeded in 96-well, black-walled, clear bottom polystyrene plates (Corning #3904, Corning, NY), and cultured to confluence over 24 h. Wells were washed with serum-free, phenol-free media for 2 h at 37° C. The co-labeled BODIPY-C16-FA/Texas red-albumin complex was robotically added at time = 0. Alternating excitation λ = 490 and 590 nm pulses were delivered for BODIPY/Texas red, respectively, and emission λ = 510 and 615 nm were recorded at 10-sec intervals for up to 10 min. Plates were imaged on the Synergy Neo2 HTX Multi-Mode Microplate reader (BioTek, Winooski, VT) and averaged from six fields captured from each well using Gen5 software (Albuquerque, NM). Uptake curves were analyzed using GraphPad Prism v7 software (La Jolla, CA). The better fit for one- vs. two-phase kinetics was determined according to the curve with the greater R² value. Furthermore, if the two-phase fit yielded the same fast and slow uptake half-life (T_1/2), then the curve was assumed to be a one-phase fit and a single fast phase T_1/2 and rate constant (K) are presented. IC₅₀ values for Lipofermata (MedChemExpress, Monmouth Junction, NJ) were also calculated using GraphPad Prism software.

Statistics.

Albumin and FA uptake curves were quantified according to uptake half-life (T_1/2) ± 95% confidence interval (CI) using GraphPad Prism software. Statistical significance between albumin and FA uptake curves is defined as the lack of 95% CI overlap. Data are expressed as mean ± SEM, unless otherwise noted in the figure legend.

RESULTS

As an initial screen for transport of FA and albumin, the relative uptake kinetics were tested in freshly isolated mouse proximal tubules, perfused with BODIPY-FA complexed to iFluor 350-labeled albumin (spectral characteristics are shown in Table S2. Figures 1A–1D are from a representative proximal tubule segment before and four minutes after perfusion. No difference in albumin or FA uptake was noted between male and female mice (Figure S1, DOI 10.5281/zenodo.16739355). Figure 1E and Table 1 correspond to curves that best fit two-phase kinetics for albumin, and one-phase kinetics for FA uptake. The initial uptake velocity was markedly faster for albumin [(K = 0.095 s⁻¹, T_1/2 = 7.3 s (95% CI 4.3–12.2s)] compared to FA [K = 0.012 s⁻¹, T_1/2 = 57 s (95% CI 56–61s)] (Table 1), whereas the slow phase of albumin uptake was comparable to FA uptake kinetics. The data indicate that apical FA and albumin are taken up by distinct mechanisms, rather than by co-transport, and suggest that FA-albumin dissociation occurs prior to uptake.

Figure 1. Albumin and FA uptake in microperfused mouse proximal tubules.

Microdissected mouse proximal tubules were perfused with iFluor 350-labeled bovine serum albumin complexed with BODIPY-labeled fatty acids as described in Methods. (A) baseline iFluor fluorescence (emission λ = 450 nm) at time = 0, (B) iFluor fluorescence at 4 minutes, (C) baseline BODIPY fluorescence (emission λ = 510 nm) at time = 0, (D) BODIPY fluorescence at 4 minutes. (E), normalized iFluor-labeled albumin and BODIPY-labeled FA uptake over time. Results are mean from six experiments from three male and three female mice (see Figure S1 for sex-specific uptake graphs). Error bars are omitted for clarity. Uptake kinetics half-lives are significantly different between the two curves.

Table 1.

Albumin and fatty acid uptake kinetics in perfused mouse proximal tubules.

	FA	Albumin
R2	0.995	0.981
% fast	100	35.8
Kfast (s−1)	0.012	0.095
Kslow (s−1)	N/A	0.008
Fast T1/2 (s) (95% CI)	57.1 (53/61)	7.3 (4.3/12.2)
Slow T1/2 (s) (95% CI)	N/A	91.5 (62/213)

FA, fatty acids; K, uptake rate constant; T_1/2, uptake half-life; N/A, not applicable

To explore the possibility of reagent instability as a non-physiological cause of FA-albumin dissociation, BODIPY-labeled FA coupled to albumin was reconstituted in perfusion buffer, and loss of fluorescence across a dialysis membrane with a 1 kDa cut-off over six hours was measured as an index of dissociation. In contrast to the rapid kinetics of fatty acid and albumin uptake by perfused tubules, Figure S2 demonstrates very slow decline in fluorescence (K = 0.51 hr⁻¹, T_1/2 = 1.35 hrs). Taken together, Figures 1 and S2 reflect transporter-facilitated FA-albumin dissociation mechanisms, rather than dissociation due to FA-albumin instability in solution.

To test for consistency between perfused tubules and in vitro models, we conducted experiments in LLC-PK1 cells, which are well-characterized for albumin and FA uptake (13, 40–42). LLC-PK1 cells express FATP2, megalin and cubilin mRNA (Table S3). Figure 2A shows that FA and albumin uptake curves for LLC-PK1 cells most accurately fit a one-phase model. The initial rates of both FA and albumin uptake were rapid, and not significantly different (Table 2). For comparison, the same experiments were conducted in a human proximal tubule cell line (34), which expresses endogenous megalin and cubilin and was stably transfected to express FATP2 (Table S3). In these cells, albumin uptake obeyed one phase kinetics, whereas FA conformed to a two phase model (Figure 2B, Table 3), further suggesting that FA and albumin uptake are mediated by separate mechanisms in this cell line.

Figure 2. Albumin and FA uptake in cell lines.

LLC-PK1 cells (A) and HPCT cells with stable FATP2 expression (B) were incubated with Texas red-labeled bovine serum albumin complexed with BODIPY-labeled fatty acids. A and B represent normalized uptake of fluorophore-labeled albumin and FA, which was measured every 10 seconds, as described in Methods. Each data point represents the mean from five experiments. Error bars are omitted for clarity. Uptake kinetics half-lives are not significantly different between the two curves in A, whereas the two curves in B are significantly different.

Table 2.

Albumin and fatty acid uptake kinetics in LLC-PK1 cells.

	FA	Albumin
R2	0.991	0. 740
K (s−1)	0.030	0.028
T1/2 (s) (95% CI)	22.9 (21.3/24.8)	24.7 (12.9/75.4)

FA, fatty acids; K, uptake rate constant; T_1/2, uptake half-life

Table 3.

Albumin and fatty acid uptake kinetics in HPCT cells.

	FA	Albumin
R2	0.999	0.967
% fast	16.1	56.3
Kfast (s−1)	0.031	0.015
Kslow (s−1)	0.007	N/A
Fast T1/2 (s) (95% CI)	22.2 (10.6/45.1)	46.5 (40.0/54.6)
Slow T1/2 (s) (95% CI)	102 (86.7/222)	N/A

FA, fatty acids; K, uptake rate constant; T_1/2, uptake half-life; N/A, not applicable

FATP2 mediates apical proximal tubule FA transport (13, 22, 43). To address the possibility that FATP2 co-transports albumin, FA and albumin uptake studies were conducted in the presence of the FATP2 inhibitor, Lipofermata (39, 44). Lipofermata inhibited FA uptake, but had no effect on albumin uptake in LLC-PK1 (Figure 3A) or HPCT cells (Figure 3B). These data suggest that FATP2 does not co-transport albumin and FA.

Figure 3. Effect of FATP2 inhibition on albumin and FA uptake.

LLC-PK1 cells (A) and FATP2-expressing HPCT cells (B) were pre-incubated with or without the FATP2 inhibitor Lipofermata (50 μM, 1 hr, 37^o C). Albumin and FA uptake were then measured as in Figure 1, and data are shown without normalization. Each data point represents the mean from five experiments. Error bars are omitted for clarity.

To further investigate whether FATP2 co-transports FA and albumin, Lipofermata dose-response curves were generated for FA and albumin uptake in proximal tubule cell lines. Figure 4A demonstrates Lipofermata IC₅₀ = 6 μM for FA uptake in LLC-PK1 cells, which is consistent with prior studies in other proximal tubule and liver epithelial cells lines (39, 44). In contrast, the IC₅₀ for albumin uptake was >50 μM (Figure 4A). In HPCT cells, Lipofermata IC₅₀ for FA uptake was 5 μM, and >50 μM for albumin uptake (Figure 4B). Taken together, Figures 3 and 4 strongly support that FATP2 transports FA, but does not co-transport albumin.

Figure 4. Lipofermata dose-response curves for albumin and FA uptake.

LLC-PK1 cells (A) and FATP2-expressing HPCT cells (B) were pre-incubated with the FATP2 inhibitor Lipofermata at concentrations ranging from 0.01–50 μM. Uptake for Texas red-labeled albumin (TR-albumin) and BODIPY-labeled FA (BODIPY-FA) are measured as in Figure 1. Results are means from three experiments. Each data point represents mean static value at 60 sec, which correlates with maximum uptake velocity. Error bars are omitted for clarity.

To then address whether the megalin-cubilin complex co-transports albumin and FA, dual fluorescently-labeled FA plus albumin uptake was tested in OK proximal tubule cells with CRISPR-mediated megalin or cubilin deletion (37). Wild-type OK cells are well-characterized for albumin uptake, and express abundant endogenous megalin, cubilin, as well as FATP2 mRNA at relatively low levels (36, 45)(Table S3). Figure 5A reveals that the rate of Texas red-labeled albumin uptake in wild-type OK cells was similar to LLC-PK1 cells (Figure 2A). Albumin uptake was diminished in cubilin-deleted cells, and to a greater extent in megalin knockout cells (Figure 5A), consistent with the role of both proteins in proximal tubule albumin uptake. Figure 5B shows BODIPY-labeled FA uptake in wild-type OK cells, which was not inhibited in megalin or cubilin knockout OK cells. The Figure 5A and 5B data indicate that the megalin-cubilin complex mediates albumin uptake, but does not co-transport FA.

Figure 5. Effect of megalin and cubilin deletion on albumin and FA uptake.

OK cells (green), and OK cells that underwent CRISPR-Cas9 deletion of cubilin (red) or megalin (blue) were assayed for albumin uptake (A) or FA uptake (B) as in Figure 1. Each data point represents the mean from three to six experiments. Error bars are omitted for clarity.

DISCUSSION

Proximal tubule reabsorption of aberrantly filtered albumin and non-covalently bound FA is critical for the pathophysiology of chronic kidney diseases associated with disruption of the glomerular filtration barrier. The accumulation of intracellular FA and FA metabolites may then lead to lipotoxic cell death and tubular atrophy. Although albumin and FA transport mechanisms have been well described, there is still considerable controversy regarding the mechanism of apical proximal tubule uptake. On the one hand, FA may dissociate from albumin within the proximal tubule lumen, and then be taken up by separate transporters. The alternative hypothesis is that FA-bound albumin is co-transported as an intact complex, by an uncharacterized mechanism, with subsequent intracellular dissociation (presumably within endosomes) (23–30). Using a combination of ex vivo and in vitro approaches, as well as genetic deletion and small molecule inhibition, we demonstrate that FA and albumin are taken up by the apical proximal tubule through separate transport mechanisms. The implication then is that FA and albumin undergo dissociation within the luminal (extracellular) space prior to reabsorption by the proximal tubule.

If albumin and FA were co-transported and dissociation was intracellular, e.g., within endosomes or lysosomes, one would expect equivalent kinetics of albumin and FA uptake, which was not observed in ex vivo or most in vitro experiments. Furthermore, albumin-FA dissociation in the absence of cells was very slow, and inconsistent with the rapid uptake in the experimental models. While changes in ultrafiltrate pH due to NHE3 activity might exert a modest effect on FA-albumin dissociation, osmolality is unlikely to contribute to dissociation. We propose that an equilibrium exists between albumin-bound and unbound fatty acids within the circulation, as well as proximal tubule lumen; when fatty acids are taken up by transporters, this creates a gradient, which stimulates extracellular dissociation.

Few studies have examined kinetics of albumin or FA uptake. For albumin, the rate constant in cultured endothelial cells = 0.04 s⁻¹ and t_1/2 = 17 sec (46), which is consistent with our data in cultured and perfused proximal tubules. For long-chain FA (oleate) uptake in lipid vesicles, the first-order rate constant was 19 s⁻¹ and t_1/2 0.037 sec, and the rate-limiting step for FA uptake was FA dissociation from the albumin donor (21). Our data demonstrate relatively slower FA uptake kinetics, which may reflect the different systems (vesicles vs. intact cells). Although the rate constants and half-lives between albumin and FA uptake were not as divergent as in these two reports (21, 46), the differences between albumin and FA kinetics were significantly different, especially for microperfused tubules and HPCT cells. Moreover, FATP2 inhibition did not affect albumin uptake, and megalin and cubilin deletion did not affect FA uptake, consistent with megalin and cubilin siRNA knockdown inhibiting albumin uptake, but not palmitate-induced reactive oxygen species generation (47).

Kinetics measurements have previously been conducted in proximal tubules perfused with radiolabeled albumin for 90–120 min (48, 49), which likely correlate with the plateau, rather than the initial rapid phase of albumin uptake. That our method detected this initial phase is due to the capacity for frequent (every 6.7s) fluorescence measurements. Albumin uptake by clathrin-mediated endocytosis is complex, requiring adaptor protein recruitment from the cytoplasm and extensive cytoskeletal and plasma membrane rearrangement (50). This process is extremely active in the proximal tubule, with the velocity of membrane internalization corresponding to turnover of the entire surface of apical invaginations within 78 seconds (51). Our initial phase albumin uptake data are consistent with these rapid morphologic changes. Nevertheless, because the uptake T_1/2 by transporters is typically measured in seconds, compared to minutes for many types of receptor-mediated endocytosis (52), it is surprising that the initial phase of albumin uptake was significantly faster than FA transport (Figures 1E and 2B).

Much of the data was generated in cultured proximal tubule cell lines, which have been viewed with caution for albumin uptake (41, 53). The major issue is loss of brush border in cultured cells, which renders a slower apical albumin uptake rate (41). Although the kinetic models differed slightly between HPCT cells and perfused tubules, both yielded relatively faster albumin compared to FA uptake rates. In contrast, LLC-PK1 cells, which have been extensively utilized for albumin uptake measurements (40–42), demonstrated similar one-phase kinetics for albumin and FA. Kinetics studies were not conducted in OK cells, but megalin/cubilin-mediated albumin and FATP2-mediated FA uptake were observed (Figures 4 and 5), consistent with RNA sequencing data showing robust megalin and cubilin, and detectable FATP2 mRNA expression (45)(Table S3). If the perfused tubules are viewed as the standard, we would endorse HPCT and OK as reliable cell lines for albumin and FA uptake, whereas LLC-PK1 albumin and FA uptake kinetics are quite discrepant compared to perfused tubules. Furthermore, the similar kinetics of albumin and FA uptake in LLC-PK1 cells perpetuates the notion of co-transport, which was not observed in any of the other systems.

Gene expression patterns also differ between cells lines (45, 54)(Table S3), which may confer differences in function. To address this limitation, we employed loss of function strategies that focus on megalin/cubilin and FATP2 as the predominant regulators of albumin and FA uptake, respectively. The partial inhibition of albumin uptake in megalin and cubilin knockout OK cells (Figure 5A) is consistent with reports suggesting that both receptors are required for maximal uptake (37, 55). This point is reinforced by studies in Nphs2^−/− mice with albuminuria, in which superimposed proximal tubule Lrp2 and Cubn deletion resulted in minimal increases in albuminuria, and no change in serum albumin (56).

FA uptake is also incompletely inhibited in perfused tubules from FATP2 knockout mice (13), implying the presence of additional apical FA transporters. Among other FA transporters, CD36 has been implicated as an albumin-FA co-transporter (15). However, proximal tubule CD36 expression is not conserved across species [(13, 57, 58), https://esbl.nhlbi.nih.gov/helixweb/Database/NephronRNAseq/All_transcripts.html, https://humphreyslab.com/SingleCell/], which casts doubt on the physiologic relevance. Moreover, in contrast to the importance of FATP2 in kidney pathophysiology (13, 22), proximal tubule-specific CD36 transgene expression had no effect on kidney disease (58). Expression of the G protein-coupled FA transporter, FFAR1, has also been detected by RT-PCR in proximal tubule (16, 59). However, in situ hybridization in kidney cortex revealed FFAR1 mRNA primarily in collecting duct (59), and neither FFAR1 mRNA by single cell RNAseq, nor protein expression by immunohistochemistry or proteomics were observed in proximal tubules from multiple species [https://cello.shinyapps.io/kidneycellexplorer/, https://humphreyslab.com/SingleCell/, https://esbl.nhlbi.nih.gov/Databases/KSBP2/, (13)]. Furthermore, the FFAR1 agonist PBI-4050 did not affect FA uptake in HKCC cells (39). So, although megalin/cubilin and FATP2 may not be singular mechanisms of apical proximal tubule albumin and FA uptake, no other pathways have emerged to challenge the preponderance of evidence that megalin/cubilin and FATP2 are the major mechanisms. Therefore, the use of megalin/cubilin and FATP2 loss of function strategies to clarify the plausibility of co-transport seem justified.

The current studies are focused on apical proximal tubule membrane functions, since uptake was not measured from the basolateral bath in the perfusion experiments, and in vitro, basal membrane adherance to plastic would limit substrate access. We previously demonstrated that FA uptake was approximately five times more rapid from the apical compared to basolateral surface, and FATP2 gene deletion had no effect on basolateral FA uptake (13). These observations are consistent with the magnitude of Lipofermata inhibition (Figure 3), and implies that the residual FA uptake in Lipofermata-treated cells is from basolateral transport. More detailed investigation would be facilitated by identification of the basolateral FA transporter(s), which are currently unknown.

The presence of separate apical proximal tubule albumin and FA transport systems logically suggests that uptake mechanisms for albumin and FA are distinct and not due to co-transport. However, this hypothesis had not previously been tested. In this report, we use ex vivo microperfusion and in vitro cell culture strategies to demonstrate that albumin and FA obey significantly different uptake kinetics. Furthermore, megalin/cubilin deletion does not affect FA uptake, and FATP2 inhibition does not affect albumin uptake. We conclude that albumin and FA are transported by distinct mechanisms, whereby the albumin-FA complex dissociates in the proximal tubule lumen prior to uptake by megalin/cubilin and FATP2.

Supplementary Material

Supplementary Tables 1–3 and Figures 1 and 2: https://zenodo.org/records/16739355.