Rab5 and Rab11 maintain hematopoietic homeostasis by restricting multiple signaling pathways in Drosophila

Essential revisions:

1) Rab5/Rab11 and autophagy

A general concern is at the end of the study where autophagy is interrogated. Unlike the previous sections of the manuscript, this section appears more preliminary. A knowledge gap remains: what is the functional role of autophagy in hematopoiesis? Does the autophagy induction from Rab5/11 perturbation require functional lysosomes? If lysosome function is perturbed, does this suppress the increased autophagy phenotype? An alternative model posits that autophagy merely sequesters signaling molecules away from recycling endosome-cell surface cycling. If so, fusion of autophagsomes with lysosomes may not be necessary for the role of autophagy in lamellocyte formation. Can autophagosomal SNARES be knocked down or perturbed to see the impact on formation? Evidence is presented that Rab5/11 knockdown increases autophagy, and this is correlated with lamellocyte formation/JNK activation. This conclusion could be tuned-down or directly tested (using for instance Hep or Bsk RNAi).

We thank the reviewers for the suggestions to further clarify the role of autophagy in lamellocyte formation. Syntaxin 17 (Syx17) is an autophagosomal SNARE that is important for the fusion of lysosomes and autophagosomes (Takáts et al., 2013). Inhibiting Syx17 disrupts the fusion of lysosomes and autophagosomes and subsequently blocks autolysosomal degradation. To examine whether autophagy induction from Rab5/Rab11 perturbation requires functional lysosomes, we knocked down Syx17 in Hml>UAS-Rab5/11^DN larvae. We found that increased LysoTracker signals were decreased in Hml>UAS-Rab5/11^DN>Syx 17 RNAi circulating hemocytes (Figure 9Q-T), showing that the increased lysosomal activity was blocked. In addition, the lamellocyte count was decreased in Hml>UAS-Rab5/11^DN>Syx 17 RNAi circulating hemocytes (Figure 9U-X, BB). The above data indicated that lysosomal activity is critical to the role of autophagy in lamellocyte formation upon the inhibition of Rab5/Rab11.

To further validate that JNK is required for autophagy activation in Hml>UAS-Rab5/11^DN hemocytes, we repressed JNK signaling by using UAS-bsk RNAi or UAS-bsk^DN flies. Accordingly, the increased LysoTracker signals were restored in Hml>UAS-Rab5/11^DN>UAS-bsk RNAi and Hml>UAS-Rab5/11^DN>UAS-bsk^DN hemocytes (Figure 9J-P), suggesting that this process is JNK-dependent.

Regarding the functional role of autophagy in hematopoiesis, 20 autophagy-related genes have been identified to date (https://flybase.org/reports/FBgg0000076), but only Atg6 has been shown to affect hematopoiesis (Shravage et al., 2013). Autophagy is associated with many biological processes, including the immune response, aging and neurodegeneration. We are also interested in the detailed role of autophagy in hematopoiesis and will further focus on the mechanisms underlying how autophagy regulates hematopoiesis.

2) An increase in the number of circulating hemocytes

The origin of the increased number of circulating hemocytes should be clarified. The authors do not observe an increase in proliferation in circulating hemocytes at the L3 stage (in contrast with the situation in the lymph gland). Increased proliferation may have occurred earlier. Alternatively, it is possible that the increase is merely due to a disruption of the sessile hemocyte compartment. These 2 possibilities could be easily assessed.

To examine whether increased proliferation occurred at an earlier stage, we stained hemocytes from Hml>UAS-GFP>UAS-Rab5/11 RNAi 2^nd and 3^rd instar larvae with anti-PH3 antibodies. When compared with the Hml>UAS-GFP>WT hemocytes, the PH3-positive hemocyte count was increased in Hml>UAS-GFP>UAS-Rab5/11 RNAi hemocytes at the L2 stage (Figure 1—figure supplement 1G-I, Q), whereas we did not observe significantly increased proliferation at the L3 stage (Figure 1—figure supplement 1J-L, Q). In addition, we also examined sessile hemocytes from Hml>GFP>UAS-Rab5/11^DN larvae, but the sessile hemocyte count was not decreased compared with that in Hml>UAS-GFP>WT larvae (Figure 1—figure supplement 1M-O). These data indicated that increased circulating hemocytes upon the inhibition of Rab5/Rab11 resulted from the high proliferation of hemocytes and not from the decreased sessile hemocyte count.

3) Rab5/Rab11-gal4 expression

Different phenotypes driven by Rab5/Rab11-gal4 require additional experiments to address why it leads to the discrepancy when compare to hemocyte-specific drivers. This could simply be due to different levels of overexpression (and a threshold effect to observe lamellocyte induction in response to JNK activation). Or it is possible that Rab5/11-gal4 is expressed in other tissues causing systemic effects. This needs to be assessed either by looking directly at Bsk accumulation or with a UAS-destablizedGFP. At present, the conclusion that bsk needs to be expressed "under endosomal control" to induce lamellocyte seems weak and the concept still ill-defined.

Given that we observed lamellocytes in Rab5/Rab11>UAS-bsk lymph glands but not in da>UAS-bsk lymph glands, we crossed da-Gal4 and Rab5/Rab11-Gal4 with UAS-GFP to determine where there was a difference in expression in the whole larvae. As shown in the figure, the overall fluorescence intensity in Rab5/Rab11-Gal4 was comparable with that in da>UAS-GFP, however; the fluorescence intensity was stronger in salivary glands from Rab5/Rab11-Gal4 larvae (Author response image 1A-C). In addition, based on gene expression data from FlyBase, Rab5 or Rab11 is highly expressed in the gut, fat bodies and salivary glands (Rab5: https://flybase.org/reports/FBgn0014010; Rab11: https://flybase.org/reports/FBgn0015790), while the daughterless gene has a mild expression pattern in these tissues (da: https://flybase.org/reports/FBgn0267821). These data suggest that Rab5-Gal4 and Rab11-Gal4 are stronger drivers in some tissues.

Author response image 1

Download asset Open asset

Moreover, to examine whether systemic effects exist, we crossed Hml>UAS-bsk with NP3084-Gal4 (midgut-specific driver), ppl-Gal4 (fat body-specific driver) and elav-Gal4 (nervous system-specific driver) and stained hemocytes with anti-L1. We observed lamellocyte formation in Hml>NP3084>UAS-bsk and Hml>ppl>UAS-bsk hemocytes (Figure 4—figure supplement 2A-D), indicating that the interaction between hemocytes and fat bodies or midguts may play a role in aberrant lamellocyte differentiation. The above results showed that the different phenotypes in Hml>UAS-bsk and Rab5/Rab11>UAS-bsk may result from systemic effects, so our conclusion that bsk expression under Rab5/Rab11 control is crucial for lamellocyte formation is not correct. We have deleted the previous conclusion “These data indicated that activation of JNK under endosomal control can activate downstream transcription factors in the JNK pathway, resulting in massive lamellocyte production.” from our paper.

4) Mixture of lymph gland- and circulating hemocyte phenotypes

Lymph gland- and circulating hemocytes are intermingled, and some of the pathways, including the Toll pathway and autophagy, are validated only by circulating hemocytes. The mixture of two cell types throughout the manuscript makes it hard to follow the authors' rational. Moreover, hemocyte differentiation/proliferation mechanisms in two tissues may differ. The authors need to better align the two populations with more caution and confirm all the pathways in both cell types.

We previously confirmed the Toll pathway in lamellocyte differentiation in lymph glands, showing that Dif played a more important role in Rab5/Rab11^DN-induced lamellocyte formation (Figure 7I-P). To examine autophagy in lymph glands, we starved Hml>UAS-GFP larvae for 4 h, 6 h and 8 h and stained lymph glands with anti-Atg8. Compared with lymph glands from fed larvae, we could not see significant autophagosome formation after starvation (Author response image 2A-D). In contrast, after starvation for 4 h, 6 h or 8 h, we observed obvious autophagosome formation in hemocytes (Author response image 2E-H), showing that hemocytes are more sensitive to starvation and are probably a better model for studying the relationship between autophagy and hematopoiesis. In addition, we also confirmed the JNK and Ras/EGFR pathways in lamellocyte formation in circulating hemocytes (Figure 3—figure supplement 2; Figure 6—figure supplement 1).

Author response image 2

Download asset Open asset

5) The non-cell-autonomous phenotype in the lymph gland

In Figure 2H-K, the authors showed that the expression of RNAi against Rab5 or Rab11 in the CZ decreases the MZ in a non-cell-autonomous manner. How do the authors explain this phenotype by the current model? This should be shown with additional experiments.

Based on the current model, CZ-to-MZ progenitor maintenance is mainly regulated by PVR/STAT92E/adenosine signaling (Mondal et al., 2011). To examine whether the non-cell-autonomous regulation from the CZ in our study is dependent on the PVR/STAT92E/adenosine signal, we crossed UAS-STAT92E and UAS-Adgf-A with Hml>UAS-Rab5/11^DN and stained the lymph gland with anti-P1. However, the increased P1-positive area in Hml>UAS-Rab5/11^DN lymph glands was not restored by UAS-STAT92E and UAS-Adgf-A, indicating that this non-cell-autonomous regulation from the CZ is independent of the PVR/STAT92E/adenosine signal (Figure 2—figure supplement 2). The underlying mechanism warrants future investigation.

Physiological- or cell biological relevance of Rab5/11-mediated function

What is the physiological or cell biological relevance of the Rab5/Rab11 regulation shown in this study? It will be important to show whether the endocytic control indeed is critical for the immune response/hemocyte development to manage all the associated signals.

Rab proteins are important for signal transduction. The inhibition of Rab5/Rab11 leads to aberrant signal transduction and eventually activates multiple innate immunity-associated signaling pathways. Activation of multiple pathways will result in abnormal hematopoiesis, including overproliferation and overdifferentiation of hemocytes. We believe that endocytic control is important for immune response/hemocyte development based on the following:

1) In addition to the UAS-Rab5 RNAi and UAS-Rab11 RNAi lines, we utilized the UAS-Rab5^DN and UAS-Rab11^DN lines, which are defective for GTPase activity (Zhang et al., 2007). In our study, when we overexpressed UAS-Rab5/Rab11^DN in the blood system by using Hml-Gal4, we observed aberrant hematopoiesis, including increased hemocyte counts, an expanded CZ area and massive lamellocyte formation; we confirmed this phenotype using Rab RNAi lines. These results showed that not only the expression level of Rab5/Rab11 but also Rab5/Rab11 GTPase activity play crucial roles in maintaining hematopoietic homeostasis.

2) Many studies have shown that the inhibition of Rab5/Rab11, essential genes in endocytosis, can activate innate immunity-related signaling pathways, including the JNK, BMP, and Ras/EGFR pathways (Tang et al., 2017; Takino et al., 2014; Bhuin and Roy, 2010). A previous study showed that knocking down ARF1 or inhibiting ARF1 GTPase resulted in an increased Hrs-positive endosome count and the accumulation of Notch signaling in Hrs-positive endosomes, which eventually activated the Notch pathway (Khadilkar et al., 2014). Similarly, we also observed an increased Hrs-positive endosome count in Hml>UAS-Rab5/11^DN, which means that defective Rab5/Rab11 GTPase led to abnormal endocytic activity. We also showed higher degrees of colocalization between Hrs and p-JNK and observed activation of p-JNK signaling. These data indicated that the endocytic control of Rab5/Rab11 is crucial for maintaining the homeostasis of signaling pathways.

Tang, Y., Geng, Q., Chen, D., Zhao, S., Liu, X., Wang, Z., 2017. Germline Proliferation Is Regulated by Somatic Endocytic Genes via JNK and BMP Signaling in Drosophila. Genetics 206(1), 189-197.

Bhuin, T., Roy, J.K., 2010. Rab11 regulates JNK and Raf/MAPK-ERK signalling pathways during Drosophila wing development. Cell Biol. Int. 34(11), 1113-1118.

Zhang, J., Schulze, K.L., Hiesinger, P.R., Suyama, K., Wang, S., Fish, M., Acar, M., Hoskins, R.A., Bellen, H.J., Scott, M.P., 2007. Thirty-One Flavors of Drosophila Rab Proteins. Genetics 176(2), 1307-1322.

[Editors' note: further revisions were suggested prior to acceptance, as described below.]

Reviewer #1:

Yu and colleagues supplemented a large number of data and addressed most of the concerns raised by the reviewers. New data clearly resolved concerns raised and explained how circulating hemocytes are increased upon loss of Rab5/11 and why Rab5/11-gal4 differs from Hml-gal4. Despite the clarification, the biological relevance of Rab5/Rab11 signaling in hematopoiesis is still unclear. It is well-known that Rab5/Rab11 endosomal pathway controls multiple proteins involved in cellular signalings, including EGFR/Ras, Eiger/JNK, and Toll, and thus, in my opinion, it will be important to understanding why Rab5/Rab11-mediated control is significant in the blood.

1) Biological relevance

Although the authors extensively investigated the functions of Rab5/Rab11 in the blood cell signaling, it is still ambiguous how the endosomal control is maintained during development or immunity and why it is significant in hematopoiesis. For example, does Rab5/Rab11 control the proliferation of hemocytes in a specific developmental stage by repressing or activating Ras/Erk? Is Rab5/Rab11 inhibited by immune challenges to directly stimulate the lamellocyte differentiation? In which context Rab5/Rab11 gets modified and manipulates target signals similar to phenotypes derived by Rab5DN/Rab11DN or Rab5/Rab11 overexpression?

We thank the reviewer for these points. We believe that there could be an interaction between Rab5/Rab11 and immunity (the lamellocyte formation upon wasp challenge). We will focus on these issues in our future study.

2) Schematic diagram

A schematic diagram shown in Figure 10B does not fully represent the genetic pathways identified in the results. For example, there is no evidence showing a direct activation of pERK by Rab5/Rab11. And even though Rasv12 alone could not activate the lamellocyte differentiation, active Ras and Bsk together can induce the lamellocyte. The authors introduced dFOXO as a target of JNK and showed that loss of dFOXO in Hml>Rab5/11DN rescues the lamellocyte phenotype. These points are missing. Or the diagram could be simplified to help understand the major points.

Given that there is no evidence showing loss of Rab5/Rab11 could directly activate p-Erk, we used the dotted line and one question mark in our revised version to display the relationship between Rab5/Rab11 and p-Erk (Figure 10B). In addition, we showed dFOXO as the target of the JNK pathway to facilitate lamellocyte formation (Figure 10B). We also simplified the schematic diagram to make it clearer.