In eukaryotic cells, macropinocytosis is the most efficient way to internalize extracellular fluid through plasma membrane-formed large vacuoles called macropinosomes (Racoosin and Swanson, 1993; Swanson and Watts, 1995; Lim and Gleeson, 2011). As an ancient cellular behavior, macropinocytosis is essential for many physiological and pathological processes, such as nutrients uptake, pathogen capture, antigen presentation, and tumorigenesis (Kerr and Teasdale, 2009; Diken et al., 2011; Liu and Roche, 2015; Bloomfield and Kay, 2016). Sharing similar intracellular mechanism, macropinocytosis is thought to be largely homologous to phagocytosis, neuronal bulk endocytosis and other actin-driven endocytosis (Bloomfield and Kay, 2016).

Macropinocytosis provides a non-selective route to internalize extracellular fluids. In cancer cells, macropinocytosis is utilized for nutrient uptake to support metabolic needs and promote tumor growth (Commisso et al., 2013). Several infectious pathogens, such as bacteria, virus and protozoa, opportunistically hijack macropinocytosis to invade host cells and evade immune responses (Haraga et al., 2008; Gobeil et al., 2013). Observation of macropinocytosis will provide insight into the underlying regulatory molecular mechanisms and enable the physiological control of macropinocytosis for drug delivery in anti-cancer or -infection therapies. However, most observations of macropinocytosis were obtained from in vitro experiments or unicellular organisms (Chubb et al., 2000; Chen et al., 2015), e.g., Dictyostelium amoebae, instead of naturalistic models that do not fully reflect the complexity of in vivo situations, limiting their application. Therefore, considerable gaps remain in the knowledge of the relevance of macropinocytosis, especially the lack of optical imaging approaches, in living organisms. It is essential to develop consistent and reliable methods for in vivo macropinocytosis studies.

Most Drosophila melanogaster (fruit fly) genes are evolutionarily conserved with human and other mammals (Reiter and Bier, 2002). With its short life cycle and genetic amenability, Drosophila provides attractive model systems for various researches (Brand and Perrimon, 1993). After removal of chorions, Drosophila embryos become transparent, but still tolerance to subsequent operations for live imaging, rendering this model feasible for in vivo cell behavioral and cell biological studies.

In the present study, we described a protocol for in vivo studies of macropinocytosis. By microinjection, fluorescence-labeled dextrans were delivered into Drosophila embryos. Engulfed by Drosophila hemocytes, which resemble mammalian macrophages, fluorescent dextrans were internalized with associated membrane, resulting in formation and subsequent transportation of macropinosomes. For microscopic methods, macropinocytosis was fluorescently visualized and monitored in live embryos. This method provides a novel way for observation of the organization and subsequent processing of macropinosomes in vivo, and an ideal model for revealing the underlying mechanisms of macropinocytosis.

Using our embryo microinjection method, we first sought to determine whether embryonic hemocytes were capable for evoking macropinocytosis. After induction by ATPγS, a non-hydrolysable ATP analog, macropinocytosis was observed using a confocal microscope. Five standard rules are followed to identify formed macropinosomes; (1) dextran-positive; (2) approximately pellet and ellipse-shaped; (3) surrounded by GFP-positive cytoplasm;(4) larger than 0.2 μm in diameter; (5) fluorescence intensities of macropinosomes are comparable with or higher than that of the extracellular space.

In our observation, responding to injected ATPγS, GFP-positive hemocytes efficiently engulfed large volumes of extracellular fluids containing fluorescent-labeled dextran with diverse molecular weights and different fluorophores (Figures 2A,E). However, macropinosomes were seen in almost 100% of embryos after 3-kDa dextran injection, whereas only 81.8, 60.0, and 38.1% of embryos generated macropinosomes after injection of 10, 40, and 70-kDa dextran, respectively (Figure 2B). To clarify whether hemocytes have equal ability to uptake dextrans with different sizes, we measured the number of macropinosomes in each hemocyte. Our result shows that about 2 macropinosomes were generated in each hemocyte and there were no statistically significant differences among hemocytes uptake 3-kDa to 70-kDa dextrans (Figure 2C). In addition, our result showed that dextran-positive pinosomes were heterogeneous in sizes, ranging from 0.2 to 10 μm, and approximately 90% of the macropinosomes were 1 to 4 μm in diameter (Figure 2D). Combined with genetic strategies, e.g., GAL4/UAS system, this model is appropriate for molecular studies, especially in endocytosis and its subsequent events. For instance, using hemocyte-specific promoter srp-Gal4 to drive GFP fused tubulin-associated protein tau (tau-GFP), microtubule-filaments could be visualized and further analyzed in formation and trafficking of nascent macropinosomes (Figure 2F).

To clarify whether ATPγS was necessary for hemocytes to induce macropinosomes in our system. We injected 3 or 70-kD dextrans without ATPγS into embryonic hemocyte and observed spontaneous macropinocytosis in GFP positive hemocytes. Compared with ATPγS induced macropinosomes, spontaneous macropinosomes were much smaller (Figure 3), raising the difficulty for in vivo observation. Therefore, we recommend ATPγS induction to promote macropinocytosis in this model.

To further test whether this method is suitable for monitoring the cellular and subcellular events of macropinocytosis, we injected fluorescence-labeled dextrans together with ATPγS into the embryos with GFP-labeled hemocytes. As shown in Figure 4A, GFP-hemocyte extracellular dextrans were internalized along with cell surface ruffling and generated TRD70S-positive macropinosomes. Using low molecular weight TRITC-dextrans, TRD3S, we monitored the intracellular events of ATPγS-induced macropinosomes. During a 15 min observation period, dextran-containing macropinosomes transported in the cell body of a migrating hemocytes, indicating the method is reliable for cellular events and cell behavioral recordings (Figure 4B).

Adenosine triphosphate, or ATP, is the principal molecule for intracellular energy transfer in cells. Extracellular ATP is also an essential messenger for several physiologic and pathological processes (Lim and Gleeson, 2011; Li et al., 2013; Cisneros-Mejorado et al., 2015). Sensed by purinergic receptors, extracellular ATP activates intracellular signaling pathway and induces membrane ruffling and macropinocytosis (Grimmer et al., 2002; Li et al., 2013).

Here, we present an ATP analog microinjection-based system for in vivo observation of macropinocytosis and its subsequent cellular events. In our observation, ATPγS are capable to induce larger macropinosomes engulfed by Drosophila embryonic hemocytes using different molecular weights of dextrans (Figures 2, 3). In addition, small dextrans are more efficient for labeling macropinosomes. Moreover, this method is feasible to investigate cellular events of macropinosomes, such as generation, fusion, and trafficking (Figure 4).

Although macropinocytosis is a nonselective process in cell culture systems, natural structures in living tissues may filter molecules and cause different diffusion properties in different size dextrans. Therefore, we tried to find out which size of dextran would be suitable for our system. In our experiment, dextrans ranged from 3-kDa to 70-kDa were all capable of labeling macropinosomes. However, small dextran (TRD3S) labeled macropinosomes could be seen in almost 100% of injected embryos, whereas large dextran (TRD70S) could only successfully label macropinosomes in about 40% of embryos under same conditions (Figure 2B). These data suggest that small molecules are more efficient for labeling macropinosomes in our system. In addition, although induced by different-sized dextrans, each hemocyte generated approximate numbers of macropinosomes. Taken together, application of larger dextran will reduce the success rate of induction without changing the uptake ability of hemocytes. It also suggests a potential filter effect of extracellular matrix in live tissues.

There are some difference of macropinocytosis between in vivo and in vitro systems. In previously studies, cultured macrophages and microglia exhibit impressive capabilities in internalization of extracellular fluids by macropinosomes (Racoosin and Swanson, 1993; Chen et al., 2015; Canton et al., 2016; Fu et al., 2016), which could generate within 1 min followed by centripetally migration and rapid shrink (Racoosin and Swanson, 1993; Lee and Knecht, 2002). It is a relatively short window to observe each phase. By contrast, in vivo macropinocytosis takes longer to generate macropinosomes in about 30 min (Figure 3A), providing enough time for observation. In addition, primary macrophages and the Drosophila hemocyte S2 cell line generate macropinosomes with high density (Gupta et al., 2009; Canton et al., 2016), raising the difficulty for observation of each vesicle. In contrast, no more than 5 macropinosomes were observed in our in vivo model. The sparse labeling provides a convenient approach to distinguish and monitor macropinosomes.

Combined with genetic tools and other strategies, this method is suitable to investigate molecular functions in macropinocytosis. For instance, using srp-GAL4;UAS-tau-GFP transgenic Drosophila, microtubules were visualized around ATPγS induced macropinosomes (Figure 2F). Together with previous in vitro studies (Gilberti and Knecht, 2015), our results provide essential in vivo cues supporting that microtubule-associated structures may regulate macropinosomes formation and subsequent processes.

To expend applications, the method could be used to: (1) reveal the underlying mechanism in macropinocytosis; (2) uncover the way by which macrophages/microglia used for pathogens internalization; (3) develop new approaches for drug delivery via macropinocytosis. Taken together, this method provides novel sights for in vivo investigation of macropinocytosis and associated processes.

This study was carried out in accordance with the recommendations of the Guidance for the Care and Use of Laboratory Animals at Zhejiang University. The protocol was approved by the Institutional Animal Care and Use Committee at Zhejiang University.

LC and YL designed the research. LC, DC, JC, TZ, ZD, and YL performed the experiments together; LC, JC, and YL wrote the paper; LZ and HL helped with animal raising and experimental preparation; LC with JC analyzed the data; YL supervised the entire study. All authors discussed the results and commented on the manuscript.