ER-mediated phagocytosis: myth or reality?

Dear Editor, In a manuscript by Touret et al. [1], published in the present issue, the authors oppose a so-called “conventional model” of phagocytosis involving the plasma membrane (PM) and early endosomes to an “endoplasmic reticulum (ER)-mediated” model, where the ER is the sole source of initial membrane for the formation of phagosomes. We argue that none of these models integrate accurately the sum of knowledge acquired in the last few years on membrane trafficking events during phagocytosis. Moreover, the ER-mediated model shown certainly does not represent our current view. It would have been more appropriate to present as a conventional model the original concept derived from Elie Metchnikoff's observations, where the PM is the sole source of membrane, and the alternative view is that other organelles, including various endocytic organelles and ER, also contribute membrane, as proposed recently [2]. Indeed, the seminal work of Metchnikoff at the end of the 19th century led to the proposal that phagosomes were formed by the invagination of the PM. For over a century, this prevailing view was unchallenged. However, it appeared to many scientists that the use of the cell surface might waste an important membrane to form a compartment destined for degradation and that in some conditions, the membrane needed to form phagosomes was exceeding the actual cell surface, implying that membranes from endovacuolar organelles might participate directly in the formation of phagosomes. It is only in the last few years that evidence for the contribution of several organelles in the formation of phagosomes at the cell surface, including recycling and late endosomes, was provided in the Touret et al. [1] paper. More recently, we and others [3, 4] presented a significant body of evidence indicating that ER is also directly involved in the formation of nascent phagosomes. Although the functional advantage of using recycling or early endosome membranes for the formation of phagosomes is still unclear, the contribution of ER led to the proposal that ER-mediated phagocytosis must confer functional properties, enabling phagosomes to play a direct role in the processing and presentation of exogenous peptides on major histocompatibility complex class I molecules, a process referred to as cross-presentation [2]. Following this proposal, three independent studies published simultaneously confirmed that phagosomes, in macrophages and dendritic cells, were competent organelles for cross-presentation [5–7]. In the Touret et al. [1] manuscript, the authors challenge the idea that ER contributes to the formation of phagosomes and their competence in cross-presentation. Although five major studies have contributed, so far, to the concept of ER-mediated phagocytosis and its significance in cross-presentation, Touret and collaborators [1] chose to criticize mainly the articles of Gagnon et al. [3] and Houde et al. [7]. They base their arguments on claims of their inability to reproduce some of our results, although no detail is provided about the methods used to generate their data and series of unpublished observations or data not shown. In this context, it is quite difficult to try and understand what could be the reason for these apparent discrepancies. As they cited some of our work incorrectly, it is important to put back our results in their correct context to allow a fair discussion of the significance of ER-mediated phagocytosis. The proposal that ER participates in phagocytosis and constitutes part of the phagosome membrane is supported by various experimental approaches based around biochemical analysis, morphological observations, and functional assays. Nevertheless, Touret and collaborators [1] went on to suggest that ER is unlikely to constitute even a small part of the phagosome membrane and propose an explanation as to how a series of artifacts generated by these various approaches might have misled us to conclude that ER was a genuine constituent of phagosomes. According to them, all the biochemical data, including proteomics and Western blot analyses, showing the presence of ER proteins on phagosomes, are questionable as a result of a contamination of the phagosome preparations by ER vesicles. It is possible that some ER elements can be copurified during the flotation procedure used to isolate latex bead-containing phagosomes. However, various observations suggest that the ER molecules detected on the phagosome are directly associated to that compartment. When phagosomes are isolated at different time-points after their formation (to study phagosome maturation), ER proteins are not observed at all time-points. Indeed, Western blot analyses indicated that ER proteins are present on phagosomes early after phagocytosis, and they disappear and reappear in cycles throughout phagolysosome biogenesis (see Fig. 2C in ref. [3]). Thus, this brings the unlikely concept that if ER is contaminating the phagosome preparations, it does so at some time-points only. Further evidence, incompatible with a simple contamination of phagosomes by ER, is the observation that the distribution of calnexin in maturing phagosomes follows a pattern compatible with the proteolytic degradation of its intralumenal domain with time, a process that could occur in phagosomes but not in ER (see Fig. 2B in ref. [3]). Observation at the electron microscope (EM) indicates that the flotation procedure used to isolate latex bead-containing phagosomes generates highly enriched preparations (compare with a total cell lysate), in which pieces of membrane or vesicles of unidentified origin are also present (see arrows in Fig. 1A). The nature of these structures is unknown. However, when phagosome isolation is done with cells that preinternalized horseradish peroxidase (HRP) by endocytosis (HRP is present in endosomes and phagosomes in this case), several of the apparent contaminating vesicles contain HRP [8]. This indicates that a significant part of the vesicles or membranes isolated with phagosomes includes endosomes or vesicles originating from broken phagosomes. When isolated phagosomes are incubated to reveal the ER molecule calnexin by a pre-embedding method using high-resolution electron microscopy, a high labeling for calnexin is observed directly on the phagosome membrane, and the contaminating membranes are not labeled [3]. Similar results were obtained for the transporter associated with antigen presentation, another ER molecule [6]. In both studies, nothing suggested the presence of an additional layer of membrane coming from collapsed vesicles, as argued by Touret and collaborators [1]. To show that contaminating membranes could be labeled if they were of ER origin, we deliberately added ER vesicles to our phagosome preparations and showed that in these conditions, calnexin is observed on phagosomes and the added vesicles (Fig. 1B). It is such an image from our work that is presented in Figure 2B of the Touret et al. [1] paper, without mentioning that the labeled vesicles, which could be mistaken for a contaminant, were added to the preparation as a positive control. Also note that contrary to what their legend indicates, there is no labeling for LAMP-1 in the panel taken from our published data. Moreover, there is no indication that what they present as foreign membranes in their pictures in Figure 2, E and F, is of ER or even foreign origin. In fact, in their Figure 2F, the latex bead shown does not appear to be surrounded by a membrane, suggesting that the membrane bits in contact with the naked bead could be the broken phagosome membrane. A second EM method based on the localization of glucose-6-phosphate (G-6-P) activity, an enzyme specific for the ER, was used to visualize the possible contribution of ER to phagosome formation. This long-established method of detection is highly specific. In cells without phagosomes, only the ER and the nuclear envelope, which is of ER origin, could be labeled using this technique [3]. In no case did we observe labeling on any other cellular organelles, including the PM, mitochondria, or endosome-like structures. In cells where endosomes were preloaded with bovine serum albumin (BSA)-gold particles to facilitate their identification, G-6-P activity was not detected in compartments containing BSA-gold (unpublished data). It is true that it has been reported that in some cases, the product of reaction detected by this method could come from the activity of acid or alkaline phosphatase. However, the use of a specific inhibitor for acid phosphatase had no effect on our results (unpublished results). The contribution of alkaline phosphatase can practically be ruled out, as this enzyme is highly present at the surface of cells, which display no reactivity in our conditions. When applied to cells containing phagosomes formed by the internalization of various types of particles (latex beads, Leishmania parasites, or red blood cells), G-6-P activity was observed in a high proportion of these compartments. Touret and collaborators [1] wrongly claim that the presence of G-6-P activity in these phagosomes could be a result of the use of drugs that alter membrane trafficking, such as phosphatidylinositol-3 kinase inhibitors. In fact, none of our experiments with G-6-P were performed in the presence of drugs, except for one case, which had nothing to do with the internalization of any of the particles studied, where a proton pump ATPase inhibitor was used. In these conditions, we do not see how they could propose that a re-examination of the location of G-6-P is justified. In fact, the G-6-P approach was shown to be a good technique to visualize the recruitment of ER during phagocytosis in a well-established system. Indeed, the study of the intracellular trafficking of Brucella has shown that this bacteria is internalized in phagosomes derived from the cell surface and that it is delivered to ER-like compartments few hours after its internalization. As expected, the group of Gorvel [9] showed that during Brucella infection in macrophages, this bacteria is first observed in a compartment devoid of labeling for G-6-P activity, which is observed only 4–24 h after infection, as this pathogen actively recruits ER to build its niche. Membrane fusion is regulated by soluble N-ethylmaleimide-sensitive factor attachment protein receptor (SNARE) molecules. Rothman and collaborators [10] originally uncovered the pairing of the plasmalemma SNARE Sso1/Sec9c with the ER SNARE Sec22, a result that was not further discussed, as it was unthinkable then to propose that the ER could fuse with the PM. However, this peculiar pairing, originally viewed as a grain of sand in the SNARE hypothesis, was recently shown by Rothman's group [11] to be involved in the regulation of ER fusion with the PM during phagocytosis. Thus, as it is the case for endosome-PM fusion, the pairing of specific SNARE molecules controls the fusion of ER with the PM. It is interesting to mention that proteomics analyses of phagosomes at different time-points after their formation and Western blot analysis indicate that ERS-24/Sec22b is present only on early phagosomes (unpublished data). Tourette and collaborators [1] also mention that they were unable to reproduce some of our data linking phagosomes to cross-presentation. For example, they claim that they did not observe OVA in the cytoplasm after phagocytosis but rather in membrane vesicles (no method is provided to see how they performed their experiments). Based on our results, we are quite sure that the apparent discrepancy between their results and ours is simply a result of the fact that the same experiments were not compared. Indeed, in control cells, we observe similar vesicular structures containing OVA and labeled with LAMP-1 (see Fig. 1C and inset). The diffusion of OVA to the cytoplasm was observed only in the presence of proteasome inhibitors, indicating that translocated OVA is likely to be degraded rapidly by the proteasome once in the cytoplasm (Fig. 1, D and E). Furthermore, the fact that we could detect, by mass spectrometry, the association to phagosomes of ubiquitinated OVA (a process that takes place in the cytoplasm) confirms that translocation of peptides occurred. It is clear to us that the PM, the ER, as well as various endocytic organelles contribute a significant part of the initial phagosome membrane as proposed in a recent model [2]. In that context, we do not see how the presence of ER would interfere with the ability of phagosomes to acidify their lumen, as argued by Tourette and collaborators [1]. The maturation of early phagosomes into phagolysosomes would indeed favor the acquisition of properties displayed by late endocytic organelles, such as the ability to acidify their lumen. In fact, we were the first to show that phagosome maturation is accompanied by the sequential appearance and disappearance of the small GTPases rab5 and rab7 (originally observed on early and late endosomes) [8] and that this process functionally correlates with the first direct demonstration that phagosomes fuse sequentially with early and late endosomes and lysosomes [12]. Our results showing that ER molecules are more abundant on early phagosomes than late phagosomes are not in conflict with the maturation model of phagosomes. Obviously, more data are required to understand the complex mechanisms and the dynamic membrane trafficking events involved in phagocytosis and phagolysosome biogenesis. However, a growing body of evidence already indicates that the active involvement of ER in this process confers functional advantages to phagosomes. The extent of ER participation as well as other endovacuolar organelles remains to be determined. ER-mediated phagocytosis: morphological and functional properties. (A) Pre-embedding immunogold labeling of phagosomes with anticalnexin (cytoplamic domain) reveals that this protein is present directly in the phagosome membrane. The picture presented was selected to show an area of a phagosome preparation, where “contaminating” membranes (see arrows) are present, and to show that these are devoid of calnexin. (B) To show that ER vesicles could be labeled if they were present, purified microsomes (ER vesicles; see arrows) were added to the phagosome preparation. In that case, they display a significant labeling for calnexin. (C) Following phagocytosis of latex beads covered with fluorescent ovalbumin (OVA), this tracer is observed in phagosomes, as well as in lysosome-associated membrane protein 1 (LAMP-1)-positive vesicles (in a similar way as shown in the Touret et al. [1] paper). (D and E) In the presence of MG-132, a proteasome inhibitor, the fluorescent tracer can then be detected in the cytoplasm, suggesting that proteasomes rapidly degrade the tracer after its translocation.

Récupéré en direct depuis OpenAlex et désinversé. Les résumés ne sont pas conservés dans cette base de données : les index inversés représentent 8,6 Go des 9,3 Go de texte de la base, et le serveur dispose de 13 Go libres.