Maturing reticulocytes internalize plasma membrane in glycophorin A-containing vesicles that fuse with autophagosomes before exocytosis

Abstract

The erythrocyte is one of the best characterized human cells. However, studies of the process whereby human reticulocytes mature to erythrocytes have been hampered by the difficulty of obtaining sufficient numbers of cells for analysis. In the present study, we describe an in vitro culture system producing milliliter quantities of functional mature human adult reticulocytes from peripheral blood CD34(+) cells. We show that the final stage of reticulocyte maturation occurs by a previously undescribed mechanism in which large glycophorin A-containing vesicles forming at the cytosolic face of the plasma membrane are internalized and fuse with autophagosomes before expulsion of the autophagosomal contents by exocytosis. Early reticulocyte maturation is characterized by the selective elimination of unwanted plasma membrane proteins (CD71, CD98, and β1 integrin) through the endosome-exosome pathway. In contrast, late maturation is characterized by the generation of large glycophorin A-decorated vesicles of autophagic origin.

Figure 1

Maturation, proliferation, and enucleation of CD34⁺ cells from adult peripheral blood. (A) Differential counts and cytomicrographs show the morphology of the cells at different stages in 2 representative cultures using protocol B. (B) Packed pellet of filtered reticulocytes.

Figure 2

Scanning EM and cytospins of nascent reticulocytes before and after filtration. All cells had been cultured following protocol A. (A) Scanning electron microscopy image of day 18 reticulocytes pre-filtration. The arrowheads indicate free nuclei and/or enucleating cells (i). The panel on the right shows an enucleating reticulocyte (ii). (B) Scanning electron microscopy images of leukocyte filtered reticulocytes (i,ii,iv) and adult peripheral blood (iii). Arrows indicate some of the more mature reticulocytes. Panels ii and iv depict an R2 reticulocyte and an almost mature RBC, respectively. (C) Cytospin image of unfiltered (i) and leukocyte-filtered (ii) day 20 reticulocytes. Scale bar indicates 20 μm.

Figure 3

Confocal analysis of erythroblasts, enucleating erythroblasts, and reticulocytes with Abs to organelle and plasma membrane marker proteins. All cells were cultured following protocol A. Cells were fixed in 1% (wt/vol) paraformaldehyde and permeabilized with 0.05% (wt/vol) saponin. (A) Cells were harvested on day 12 and stained for the presence of the plasma membrane markers glycophorin A, glycophorin C, Rh polypeptides, β1 integrin, CD98, and CD147. (B) Cells were stained for the presence of the organelle and cytosolic markers LAMP1 (lysosomes), calreticulin (ER), CD63 (endosomes), giantin (Golgi), pericentrin (centrioles), and tubulin. (C) Cells were stained for the presence of filamentous F-actin. Scale bars indicate 5 μm.

Figure 4

Oxygen binding, deformability, membrane stability of nascent reticulocytes. (A) The oxygen association (left panel) and dissociation (right panel) curves are plotted for 20 μL of cord blood (mean p50 oxy = 14.26, deoxy = 14.57), cultured (following protocol B) and filtered reticulocytes (mean p50 oxy = 28.44, deoxy = 28.57), and adult peripheral blood (p50 adult oxy = 28.97, deoxy = 28.97). (B) Deformability curve of elongation index (EI) vs. shear stress for erythrocytes and cultured reticulocytes (i). Membrane stability curve of elongation index versus time for mature adult erythrocytes (red) and cultured reticulocytes (black) subjected to a continuous high shear stress of 57.5Pa (ii).

Figure 5

TEM of cytoplasmic and membrane remodeling in erythroblasts and reticulocytes and detection of GPA-positive vesicles by confocal microscopy. (A) Representative TEM images of day 23 filtered reticulocytes containing large autophagic compartments (autophagosomes and amphisomes). (i-ii) Fields of R2 reticulocytes, some retaining large autophagic vacuoles (arrowheads) containing poorly degraded organelles; identical compartments were also observed undergoing exocytosis (arrow). (iii) Higher magnification image of the autophagic vacuole outlined in panel ii. (iv-v) Examples of exocytic events in reticulocytes. (B) Evidence for plasma membrane blebbing during human in vitro erythropoiesis. (i-ii) Plasma membrane blebs on the surface of erythroblasts. (iii-v) Plasma membrane blebs showing electron-dense constrictions at their bases (arrowheads). (vi) Example field of a reticulocyte culture containing cellular fragments (arrows) of comparable size to the plasma membrane blebs observed in erythroblasts. (C) Analysis of reticulocytes isolated after 7 days in final stage medium and analyzed at days 7, 11, and 14 (numbers of cells analyzed: 54 at day 7, 37 at day 11, and 43 at day 14). Black bars represent the mean. *P ≤ .05; **P ≤ .01; ***P ≤ .001. (D) GPA-stained cells before and after filtration (left and right panels, respectively) show the presence of vesicles (arrows; i). (ii-iv) GPA (red) dual-stained aquaporin-1 (Aq-1), Glut-1, and band 3 (all green), respectively. (E) Filtered reticulocytes were dual stained with GPA (green) and Abs to organelle markers (red) for autophagy (LC3; i), ER (calreticulin; ii), Golgi (giantin; iii), lysosomes (LAMP1; iv), and mitochondria (MitoTracker; v). Arrows highlight vesicles fusing with the plasma membrane. All cells were cultured following protocol A. (F) Presence of GPA-positive vesicles in vivo. GPA-positive vesicles (red) containing organelle markers (green) for autophagy (LC3; i) and Golgi (giantin; ii) and GPA-positive vesicles (green) with mitochondrial probe MitoTracker (red) found in cells from peripheral blood (iii). Arrows highlight colocalization. Scale bars indicate 5 μm.

Figure 6

Endocytosis and exocytosis of GPA-positive vesicles in maturing reticulocytes. (A) Unfiltered reticulocytes cultured using protocol B were labeled with BRIC256 at 10°C, followed by incubation at 37°C for 0, 10, 20, 40, and 60 minutes. Incubation with rabbit anti–mouse Fab before permeabilization ensured that any external GPA stained red (Alexa Fluor 546) and internal GPA stained green (Alexa Fluor 488). (B) After a BRIC256 (green) internalization assay for 20 and 60 minutes, reticulocytes were dual stained with anti-LC-3 (red). Scale bars indicate 5 μm. (C) Scanning electron microscopy of filtered reticulocytes showing membrane blebs. Scale bars indicate 5 μm. (D) Live-cell imaging of a reticulocyte showing membrane blebbing (left: phase contrast; right: labeled with anti–GPA-FITC). Scale bars indicate 5 μm. (E) Immunoblot of vesicles purified from filtered reticulocytes (i, iii, v, vii, ix, and xi) and from membranes made from these reticulocytes (ii,iv,vi,viii,x,xii) using rabbit anti-glycophorin A (i-ii), mouse monoclonal anti–band 3 (BRIC170; iii-iv), rabbit anti-GLUT1 (v-vi), mouse monoclonal anti–β actin (clone AC-15; vii-viii), rabbit anti-ankyrin (with cross-reactivity for both α and β spectrin; ix-x), and mouse monoclonal anti α spectrin (BRIC174; xi-xii).

Figure 7

Model for the involvement of GPA-labeled endosomes in the exocytosis of autophagocytosed cytoplasmic content during reticulocyte maturation. (A) Autophagosomes derive from isolation membranes (i) that expand to engulf organelles and other cytoplasmic content before sealing (ii). Fusion of autophagosomes with late endosomes generates amphisomes (iii). As a prelude to fusion, autophagosomes lose LC3 from their outer membrane (a process known as delipidation), meaning that LC3 is found only on the inside of the resultant amphisome. During reticulocyte maturation, active endocytosis of GPA is observed (v), and these endosomes (vi) converge with LC3-positive amphisomes to generate an LC3/GPA–positive hybrid organelle (vii) that has the capacity to fuse with the plasma membrane (viii). Exocytosis may occur at sites of weakened underlying skeleton and is predicted to release cytoplasmic content concomitant with delivering GPA back on to the plasma membrane. (B) The above model predicts that plasma membrane surface area would increase as a result of internal vesicle maturation and exocytosis. We postulate, therefore, that the exocytic event is coupled with a process of plasma membrane blebbing, facilitated in vivo by passage through the spleen. By coupling these events, integral membrane proteins of the exocytic vesicle (including GPA) would be incorporated stochastically into the nascent bud, thereby effecting the shedding of redundant material that had previously been enriched in endosomes.