Intraendosomal flow cytometry: A novel approach to analyze the protein composition of antigen‐loaded endosomes

Adaptive immunity is induced when DCs internalize antigens for processing and presentation on MHC molecules. While antigens targeted toward lysosomes are degraded by lysosomal proteases and loaded onto MHC class II (MHC II), antigens internalized by the mannose receptor (MR) are routed toward a distinct endosome subpopulation, which retains characteristics of early endosomes [1]. From these endosomes, MR internalized antigens are processed for loading onto MHC class I (MHC I) [1], a process termed as cross-presentation. Such processing includes antigen transport into the cytoplasm for proteasomal degradation [2,3]. Despite intensive investigations, the key proteins regulating the export of antigen into the cytoplasm or the endosomal loading of MHC I molecules remain unknown. Hence, the protein content of such antigen-bearing endosomes is of high interest. While purification and analysis of phagosomes have been established quite extensively [4], the analysis of endosomes containing soluble antigens remains difficult. In this comment, we describe the establishment of a novel method to analyze the protein content of such endosomes by flow cytometry. Importantly, particular care was taken to avoid unwanted clustering of individual endosomes, which can occur during centrifugation, as this severely affects the reliability of the flow cytometric analysis. To analyze endosomes specifically supplied by the MR, we generated MR expressing HEK293 T cells (HEK-MR), which internalize high amounts of OVA (Fig. 1A). We then disrupted HEK-MR cells, which had been incubated with fluorochrome-labeled OVA, and isolated endosome preparations. Flow cytometric analysis of these preparations revealed OVA-containing particles in the endosomal preparations of HEKMR cells but not of control cells (Fig. 1B, Supporting Information Fig. 1 and 2). This indicated that these structures might indeed be endosomes. To test this hypothesis, we compared them with calibrated size beads, which revealed that the OVAcontaining structures have a size of around 500 nm (Fig. 1C), which corresponds to the size of endosomes supplied by the MR [5]. To ensure that the observed structures are indeed intact organelles, we treated them with different detergents before analysis (Supporting Information Fig. 3), which resulted in a release of OVA, demonstrating that these structures are indeed membrane-bound vesicles. Subsequently, we tested if such analysis could be used to analyze the protein content of individual endosomes. To this end, we mixed HEK-MR cells that were pretreated with either Alexa Fluor 488or Alexa Fluor 647-conjugated OVA, isolated endosomes from this cell mixture, and analyzed them by flow cytometry (Supporting Information Fig. 4A). Importantly, a clear population of endosomes, which were positive for both Alexa Fluor 488 and Alexa Fluor 647, was detected, indicating undesired clustering or fusion of endosomes during the preparation. Such clustering was not observed directly after cell disruption, showing that it might have been caused by the centrifugation of the endosome preparation. Similar observations were observed after isolation of endosomes from OVA-treated wild-type BM-DCs or from untreated MHC II-GFP expressing BM-DCs (Supporting Fig. 4B). If these endosomes were mixed before centrifugation, a clear population positive for both GFP and OVA was observed. To test whether the presence of double positive endosomes was caused by centrifugation, we centrifuged our preparations at different velocities, which revealed a close correlation between the presence of double positive endosomes and centrifugation speed (Supporting Information Fig. 5). Since the occurrence of such double positive endosomes is an unwanted artifact, which falsifies a flow cytometric analysis, we aimed to avoid such clustering by fixing the endosomes before centrifugation. Subsequent flow cytometric analysis revealed indeed that after fixation, no double positive endosomes were detected (Supporting Information Fig. 5), indicating that the inclusion of a fixation step allows reliable analysis of individual endosomes. Next, we stained the endosomal preparations with an antibody against Rab5, which colocalizes with MR-containing endosomes as shown by immune fluorescence microscopy [1], and against the cytoplasmic region of the MR. Importantly, we fixed the endosomes before centrifugation to prevent endosome clustering as shown in Supporting Information Fig. 5. Flow cytometric analysis showed that nearly all OVA-positive structures were positive for Rab5 and the MR (Fig. 1D and E), demonstrating that these structures indeed are endosomes and that this method can be used to detect proteins at the surface of endosomes. The Rab5and MR-specific