Phosphoinositide 3 kinase type III and dynamin are critical for apical endocytic recycling in kidney proximal tubular cells

Kidney proximal tubular cells (PTC) use the multigand, highly expressed and rapidly recycling tandem apical endocytic receptors, megalin and cubilin, to efficiently recapture ultrafiltrated proteins. Whereas megalin is a single-pass type I transmembrane receptor equipped with 3 NPXY and other cytoplasmic interaction motives, cubilin has no cytoplasmic domain. Megalin is a member of the LDL receptor-related (LRP) family, in which the scavenger receptor/LRP-1 exerts both endocytotic and signalling roles. By analogy to LRP-1, megalin/LRP-2 endocytosis was reported to be regulated by PI3Kinase (PI3K) class I. However, this conclusion was largely based on the broad-spectrum inhibitors wortmannin (irreversible) and LY294002 (reversible), which should also inhibit Vps34p (PI3K class III), a key downstream effector of the rate-limiting regulator of apical endosomal function, Rab5. I first re-addressed the issue of which PI3K isoform is a key regulator of apical receptor-mediated endocytosis (ARME) in the PTC-derived OK cell line: PI3K-I (internalization step) versus PI3K-III (endosomal maturation), using a substracting pharmacological approach combined with overexpression of a dominant-negative specific interacting partner of PI3K-III, GFP-(FYVE)2x. I next defined the role of PI3K at recycling endosomes by exploiting the reversibity of LY294002 inhibition for vital imaging. As reported, global PI3K inhibitors slowed down apical internalization of megalin ligands by ~2 fold, but they also decreased surface binding to the same extent, implying that internalization efficiency of bound ligands could be unaffected. Surface biotinylation confirmed a reversible ~2 fold decrease of megalin surface pool after short LY294002 treatment, suggesting reversible intracellular sequestration. Upon global PI3K inhibition, the Rab5 effector, EEA1, was fully released from sedimentable membranes into the high-speed supernatant, demonstrating uncoupling of the Rab5-dependent machinery. PI3K inhibitors or sequestration of the PI3K product, PI(3)P, by GFP-(FYVE)x2 overexpression, induced a major vacuolisation of apical endosomes without affecting late-endosomes/lysosomes. None of these effects were observed using ZSTK474, which efficiently inhibits PI3K class-I and -II but not class-III, and indeed suppressed insulin signalling in OK cells. Vital imaging revealed that LY294002 removal led to rapid shrinkage of the apical vacuoles concomitant with a burst of long radiating and branched tubular extensions labelled by endocytic cargoes, Rab5-, Rab11-, (FYVE)2x- and dynamin-GFP. Two dynamin-GTPase inhibitors, Dynasore and Dyngo4, abrogated tubule formation confirming dynamin dependence. Inhibition of actin polymerisation by latrunculin B amplified endosomal vacuolation induced by PI3K inhibitors. Inhibition of microbubule stability by nocodazole at the time of LY294002 removal abrogated directionality of recycling tubules. I conclude that (i) class I PI3K play no significant role in ARME where class III PI3K/Vps34 is the key partner (as expected); (ii) instead of acting mainly on fusion, Vps34 unexpectedly controls fission of sorting endosomes into centrifugal recycling endosomes; (iii) recycling depends on dynamin for tubulation and microtubules for directionality. A better understanding of the remarkably efficient (thus vulnerable) apical receptor recycling machinery may shed more light on genetic kidney Fanconi syndromes, including Dent’s and Lowe diseases.(SBIM 3) -- UCL, 201