1. Cell Biology
  2. Developmental Biology
Download icon

Relish plays a dynamic role in the niche to modulate Drosophila blood progenitor homeostasis in development and infection

  1. Parvathy Ramesh
  2. Nidhi Sharma Dey
  3. Aditya Kanwal
  4. Sudip Mandal
  5. Lolitika Mandal  Is a corresponding author
  1. Department of Biological Sciences, Indian Institute of Science Education and Research (IISER) Mohali, India
  2. Developmental Genetics Laboratory, IISER Mohali, India
  3. Molecular Cell and Developmental Biology Laboratory, IISER Mohali, India
Research Article
  • Cited 0
  • Views 1,650
  • Annotations
Cite this article as: eLife 2021;10:e67158 doi: 10.7554/eLife.67158

Abstract

Immune challenges demand the gearing up of basal hematopoiesis to combat infection. Little is known about how during development, this switch is achieved to take care of the insult. Here, we show that the hematopoietic niche of the larval lymph gland of Drosophila senses immune challenge and reacts to it quickly through the nuclear factor-κB (NF-κB), Relish, a component of the immune deficiency (Imd) pathway. During development, Relish is triggered by ecdysone signaling in the hematopoietic niche to maintain the blood progenitors. Loss of Relish causes an alteration in the cytoskeletal architecture of the niche cells in a Jun Kinase-dependent manner, resulting in the trapping of Hh implicated in progenitor maintenance. Notably, during infection, downregulation of Relish in the niche tilts the maintenance program toward precocious differentiation, thereby bolstering the cellular arm of the immune response.

Introduction

The larval blood-forming organ, the lymph gland, is the site for definitive hematopoiesis in Drosophila (Banerjee et al., 2019; Evans et al., 2003; Jung et al., 2005; Lanot et al., 2001; Mandal et al., 2004). Interestingly, there are noticeable similarities between the molecular mechanisms that regulate the lymph gland and those essential for progenitor-based hematopoiesis in vertebrates (Evans et al., 2003; Gold and Brückner, 2014). The lymph gland is formed in embryonic stages, and through various larval stages, it grows in size. The mature third-instar larval lymph gland is a multi-lobed structure with well-characterized anterior lobe/primary lobes with three distinct zones. The heterogeneous progenitor cells (Baldeosingh et al., 2018; Cho et al., 2020) are medially located and define the medullary zone (MZ), while the differentiated hemocytes populate the peripheral zone or cortical zone of the primary lobe (Jung et al., 2005). The innermost core progenitors are maintained by the adjacent cardiac cells that serve as niche (Destalminil-Letourneau et al., 2021), while the bulk of primed progenitors are maintained by the posterior signaling center (PSC) or the niche (Baldeosingh et al., 2018; Sharma et al., 2019). Except for one study that claims otherwise (Benmimoun et al., 2015), several studies demonstrate that PSC/niche maintains the homeostasis of the entire organ by positively regulating the maintenance of these progenitors (Figure 1A and B; Jung et al., 2005; Kaur et al., 2019; Krzemień et al., 2007; Mandal et al., 2007; Mondal et al., 2011; Sharma et al., 2019). During development, this organ is the site of proliferation, maintenance, and differentiation of hemocytes. Only with the onset of pupation do the lymph glands rupture to disperse the blood cells into circulation (Grigorian et al., 2011).

Figure 1 with 1 supplement see all
Relish expression and its function in hematopoietic niche of Drosophila larval lymph gland.

Genotypes are mentioned in relevant panels. Scale bar: 20 μm. (A) Schematic representation of Drosophila larval lymph gland with its different cell types. (B) Hematopoietic niche in larval lymph gland visualized by Antp-Gal4,UAS-GFP and Antennapedia (Antp) antibody. (C–D') Expression of Relish (antibody: red) in larval lymph gland. (C) Relish is expressed in the hematopoietic niche of lymph gland and in the progenitor population. (C') Zoomed in view of the niche showing the expression of Relish in control niche. (D–D') Relish expression is abrogated in the niche upon RNAi mediated downregulation. (E) Quantitation of Relish expression in the niche. Significant reduction in Relish expression was observed in niche (n=10, p-value=7.4 × 10−9, two-tailed unpaired Student’s t-test), whereas progenitor-specific expression remained unchanged (n=10, p-value=0.764 , two-tailed unpaired Student’s t-test). (F–G'') Effect of Relish loss from the niche on cell proliferation (F–F''), Antp expression marks the niche of wild-type lymph gland. (G–G'') Loss of Relish function from niche leads to increase in niche cell number. (H–I') Hematopoietic progenitors of larval lymph gland (red, reported by DE-Cadherin [Shg] immunostaining). Compared to control (H–H'), drastic reduction in progenitor pool was observed when Relish function was attenuated from niche (I–I'). (J) Quantitation of Shg-positive progenitor population upon Relish knockdown from the niche using Antp-GAL4 (n=10, p-value=8.47 × 10−6, two-tailed unpaired Student’s t-test). (K) Quantitation of niche cell number upon Relish knockdown from the niche using Antp-GAL4 (n=10, p-value=1.3 × 10−7, two-tailed unpaired Student’s t-test) and pcol85-GAL4 (n=11, p-value=1.2 × 10−12, two-tailed unpaired Student’s t-test). (L–M') Hematopoietic progenitors of larval lymph gland (red, reported by Ci155 immunostaining) (L–L'). Loss of Relish from the niche resulted in reduction in Ci155-positive progenitor pool (M–M'). (N–O') Compared to control (N–N'), increase in the amount of differentiated cell population (red, P1 immunostaining) was observed upon niche-specific downregulation of Relish (O–O'). (P) Quantitative analysis of (N–O') reveals significant increase in the amount of differentiated cells in comparison to control (n=10, p-value=2.3 × 10−9, two-tailed unpaired Student’s t-test). (Q–Q') Scheme based on our observation. The white dotted line mark whole of the lymph gland in all cases and niche in (F–G''). Yellow dotted lines mark the progenitor zone in (H–I') and (L–M'). In all panels, age of the larvae is 96 hr AEH. The nuclei are marked with DAPI (blue). Error bar: standard deviation (SD). Individual dots represent biological replicates. Data are mean ± SD. *p<0.05, **p<0.01, and ***p<0.001.

It is fascinating to note how this reserve population within the lymph gland is prevented from precociously responding to all of the environmental challenges during normal development. Interestingly, during infection, the lymph gland releases the differentiated hemocytes into circulation in larval stages (Khadilkar et al., 2017; Lanot et al., 2001; Louradour et al., 2017; Sorrentino et al., 2002).

The three Drosophila NF-κB factors – Dorsal, Dorsal-related immunity factor (DIF), and Relish – regulate the insect humoral immunity pathway that gets activated during infection (Govind, 1999; Hetru and Hoffmann, 2009; Louradour et al., 2017). Drosophila NF-κB signaling pathways show conspicuous similarity with vertebrates. The NF-κB family consists of five members – RelA (p65), RelB, c-Rel, p50/p105, and p52/p100 (Ganesan et al., 2010). In vertebrates, these factors are critical for producing cytokines, regulating cell death, and controlling cell cycle progression (Gilmore, 2006). In Drosophila, Dorsal and Dif activation happens during embryogenesis as well as during gram-positive bacterial and fungal infections. In both cases, it is triggered by the activation of the Toll pathway by cleaved cytokine Spatzle (Valanne et al., 2011).

On the other hand, gram-negative bacterial infections activate the Imd pathway. The diaminopimelic acid (DAP)-type peptidoglycan from the cell wall of the bacteria directly binds to the peptidoglycan recognition protein-LC (Choe et al., 2002; Gottar et al., 2002; Kaneko et al., 2006; Rämet et al., 2002) or peptidoglycan recognition protein-LE (PGRP-LC or PGRP-LE). This binding initiates a signaling cascade that elicits the cleavage, activation, and nuclear translocation of Relish with the subsequent transcription of antimicrobial peptide genes (Choe et al., 2002; Hedengren et al., 1999).

IMD pathway has been studied intensively in the context of immunity and inflammation, but far less is understood about the developmental function of this pathway. Accumulating evidence from studies, however, suggests that the IMD pathway may also have distinct roles in development. For example, in Drosophila, Relish and its target genes are activated during neurodegeneration and overexpression of Relish during development causes apoptosis in wing disc cells, neurons, photoreceptors (Cao et al., 2013; Chinchore et al., 2012; Katzenberger et al., 2013; Tavignot et al., 2017) and autophagy in salivary gland cell (Nandy et al., 2018). These studies point out to diverse developmental requirements of Relish beyond immunity in Drosophila. Since IMD is an evolutionarily conserved signaling cascade, Drosophila, therefore, turns out to be a great model to explore the diverse function of the components of this pathway.

Expression of Relish in the hematopoietic niche of the lymph gland during non-infectious conditions prompted us to investigate its role in developmental hematopoiesis. We found that Relish acts as an inhibitor of c-Jun Kinase Signaling (JNK) in the hematopoietic niche. During infection, Relish inhibits JNK signaling through tak1 in Drosophila (Park et al., 2004). Interestingly, we found similar crosstalk being adopted during development in the hematopoietic niche. Activation of JNK signaling in Drosophila is associated with alteration of the cytoskeletal architecture of cells during various developmental scenarios, including cell migration, dorsal closure, etc (Homsy et al., 2006; Jacinto et al., 2000; Kaltschmidt et al., 2002; Kockel et al., 2001; Rudrapatna et al., 2014). We found that upon Relish loss, JNK activation causes upregulation of actin remodelers, Enabled and Singed in the niche. The actin cytoskeletal remodeling, in turn, affects the formation of cytoneme-like filopodial projections leading to precocious differentiation at the expense of progenitors. These filopodial projections are proposed to facilitate the transporting of Hh from the niche to the adjoining progenitors (Mandal et al., 2007). We further show that perturbation in filopodial extensions via downregulation of Diaphanous affects Hh delivery and disrupts the communication between niche and progenitors. The hematopoietic niche maintains the delicate balance between the number of progenitors and differentiated cells of the lymph gland (Baldeosingh et al., 2018; Krzemień et al., 2007; Mandal et al., 2007; Sharma et al., 2019). During development, this organ accumulates hemocytes for post-larval requirements. However, during wasp infestation, this organ precociously releases the content into circulation (Lanot et al., 2001) due to the activation of the Toll pathway in the PSC/hematopoietic niche (Louradour et al., 2017). Therefore, a switch is essential to enable the transition from basal hematopoiesis toward the emergency mode to enable the organism to combat infection. The pathway identified in this study, critical for niche maintenance and developmental hematopoiesis, is also exploited during the immune challenge. The circuit engaged in niche maintenance and, therefore, crucial for developmental hematopoiesis gets disrupted during bacterial infection. We found that Relish in the niche serves as a joystick to achieve control between developmental and immune response.

Previous studies have demonstrated that Relish needs to be activated in the fat body to mount an immune response (Cha et al., 2003; Charroux and Royet, 2010). We show that to reinforce the cellular arm of the innate immune response, Relish needs to be downregulated in the niche during infection. Though the candidate that breaks the maintenance circuit remains to be identified, nonetheless, our study illustrates that the hematopoietic niche can sense the physiological state of an animal to facilitate a transition from normal to emergency hematopoiesis.

Results

The hematopoietic niche requires Relish during development

Drosophila NF-κB-like factor, Relish, has been studied extensively as a major contributor of humoral immune defense mechanism against gram-negative bacterial infections (Buchon et al., 2014; Ferrandon et al., 2007; Ganesan et al., 2010; Gottar et al., 2002; Kleino and Silverman, 2014). During larval development, Relish expresses in the hematopoietic niche (marked by Antp-GAL4>UAS-GFP, a validated reporter for niche cells; Figure 1C–C'). In addition to the niche, the hemocyte progenitor cells (MZ) also express Relish (arrow, MZ, Figure 1C). The niche-specific expression was further validated by the downregulation of Relish using UAS-Relish RNAi within the niche that resulted in complete loss of Relish protein therein (Figure 1D–D'). As evident from the quantitative analysis (Figure 1E) of the above data, the expression of Rel in the niche was drastically affected, while that of the MZ is comparable to the control. Whether this transcription factor executes any role in developmental hematopoiesis, beyond its known role in immune response, inspired us to carry out in vivo genetic analysis using Drosophila larval lymph gland.

We employed the TARGET system (McGuire et al., 2004) to investigate the role of Relish, if any, in the hematopoietic niche. Compared to the control, wherein the number of cells in the hematopoietic niche ranges from 40 to 45 (Figure 1F–F'' and K), a niche-specific downregulation of Relish results in a fourfold increase in the cell number (Figure 1G–G'' and K). A similar increase is evidenced upon downregulation of Relish by another independent niche-specific driver, collier-GAL4 (Krzemień et al., 2007; Figure 1—figure supplement 1A–B' and Figure 1K). To further validate the phenotype, the lymph gland from the classical loss of function of Relish (RelE20) was analyzed. Interestingly, compared to control, RelE20 niches exhibit a twofold increase in cell number (Figure 1—figure supplement 1C-D' and E'). Likewise, overexpression of Relish specifically, in the niche, causes a decline in the niche cell number (Figure 1—figure supplement 1F-G'' and H).

To investigate whether the hyperproliferative niche is still capable of performing its function of progenitor maintenance (Mandal et al., 2007), we assayed the status of the progenitors. Interestingly, compared to the control, the loss of Relish from the niche results in a drastic reduction in the number of the progenitor cells (visualized by DE-Cadherin: Shg Jung et al., 2005; Sharma et al., 2019; Figure 1H–I' and J) and Cubitus interruptus: Ci155 (Figure 1L–M') with a concomitant increment in the number of differentiated hemocytes (visualized by plasmatocyte marker by P1, Nimrod; Figure 1N–O'Asha et al., 2003; Jung et al., 2005; Kurucz et al., 2007). Quantitation of differentiation index in the genotype described above reveals a twofold increase in plasmatocyte number (Figure 1P). Moreover, in these lymph glands, the differentiated cells, instead of being spatially restricted in the CZ, are dispersed throughout (Figure 1N–O').

Although the differentiation index increases, there was no induction of lamellocytes (visualized by lamellocyte marker β-PS: myospheriod; Stofanko et al., 2008; Figure 1—figure supplement 1I–J'). The crystal cell numbers also remain unaltered (Figure 1—figure supplement 1K-L' and M), suggesting a tilt toward plasmatocyte fate upon Relish loss from the niche.

These results collectively indicate that Relish plays a critical role in determining the number of niche cells in the developing lymph gland (Figure 1Q–Q').

Relish loss from the hematopoietic niche induces proliferation

Our expression analysis throughout development reveals that around 45–48 hr AEH (after egg hatching), Relish can be detected in the niche as well as in the progenitors (Figure 2—figure supplement 1A-E'). The co-localization of Rel with validated markers of progenitors like TepIV (Dey et al., 2016; Irving et al., 2005; Kroeger et al., 2012; Shim et al., 2013) and Ance (Benmimoun et al., 2012; Sharma et al., 2019) further endorsed Rel’s progenitor-specific expression (Figure 2—figure supplement 1F-F''). On the other hand, co-labeling with Pxn-YFP, a differentiated cell marker (Nelson et al., 1994), reveals that Rel is downregulated from CZ (Figure 2—figure supplement 1G-G'). Therefore, we traced back to post second-instar stages to get a better insight into the phenotype caused by Relish loss from the niche. At 54–64 hr AEH, compared to wild type (Figure 2A–A'', C–C'', and I), downregulation of Rel by Antp-Gal4 results in an increase in EdU incorporation in the niche (Figure 2B–B'', D–D'', and I). In context to the niche, a definite proliferation pattern is observable during development. Compared to the rest of the lymph gland, niche cell proliferation decreases by 86 hr AEH (Figure 2E–E'' and I). Beyond this time point, EdU incorporation rarely occurs in the niche (Figure 2G–G'' and I). In sharp contrast to this, upon niche-specific downregulation of Relish, there is a failure in attaining the steady-state proliferative pattern by 86 hr AEH (Figure 2F–F'' and I). Quite strikingly, EdU incorporation continues even at 96 hr when the control niche cells have stopped proliferating (compare Figure 2G–G'' with H–H'' and Figure 2I). These proliferating niche cells are indeed mitotically active is evident by the increase in phospho histone H3 (PH3) incorporation compared to the control (Figure 2J–K'' and L). In addition to these snapshot techniques, in vivo cell proliferation assay of the niche was done employing the FUCCI system (fluorescent ubiquitination-based cell cycle indicator) (Zielke and Edgar, 2015). Fly-FUCCI relies on fluorochrome-tagged probes where the first one is a GFP fused to E2F protein, which is degraded at the S phase by Cdt2 (thus GFP marks cells in G2, M, and G1 phases). The second probe is an mRFP tagged to the CycB protein, which undergoes anaphase promoting complex/cyclosome-mediated degradation during mid-mitosis (thereby marking cells in S, G2, and M phases). While in control by 96 hr AEH, niche cells are mostly in G2-M (yellow), and in G1 state (green), in loss of Relish, abundance in S phase can be seen at the expense of G1 (Figure 2—figure supplement 1H-I'''' and J).

Figure 2 with 1 supplement see all
Loss of Relish from the niche causes niche cell hyperplasia. Genotypes are mentioned in relevant panels.

Scale bar: 20 μm. Niche is visualized by Antp antibody expression. (A–H'') EdU or 5-ethynyl-2'-deoxyuridine marks the cells in S-phase of the cell cycle. EdU profiling at 54 hr AEH (A–B''), 64 hr AEH (C–D''), 86 hr AEH (E–F''), and 96 hr AEH (G–H'') displayed EdU incorporation in the niche (green) in control and upon Relish downregulation. Control niches showed scanty EdU incorporation beyond 84 hr (E–E'' and G–G''), whereas loss of Relish induced niche cells to proliferate more (F–F'' and H–H''). (I) Graph representing percentage of EdU incorporation in the niche during the course of development in control (black line) and Relish loss (red line). Significant increase in the niche cell number is observed with development in Relish loss scenario. (54 hr, n=6, p-value=0.294), (64 hr, n=6, p-value=1.3 × 10−3), (86 hr, n=6, p-value=2.9 × 10−2), (96 hr, n=6, p-value=5.9 × 10−3); two-tailed unpaired Student’s t-test. (J–K'') Significant increase in the number of mitotic cells (phospho-histone 3 [PH3], red) was observed upon Relish loss from the niche (K–K'') compared to the control (J–J''). (L) Quantitation of the mitotic index of wild-type and Relish loss niche (n=15, p-value=8.1 × 10−4; two-tailed unpaired Student’s t-test). The white dotted line marks whole of the lymph gland and the niches. In all panels, age of the larvae is 96 hr AEH, unless otherwise mentioned. The nuclei are marked with DAPI (blue). Individual dots represent biological replicates. Error bar: standard deviation (SD). Data are mean ± SD. *p<0.05, **p<0.01, and ***p<0.001.

Put together, these results implicate that Relish functions as the negative regulator of niche proliferation in the developing lymph gland.

Absence of Relish in the niche stimulates proliferation via upregulation of Wingless signaling

Previous studies have shown that the Wingless (Wg) pathway positively regulates niche cell number in addition to its role in the maintenance of the prohemocyte population in the MZ (Sinenko et al., 2009). Upon perturbation of Relish function, a drastic increase in the level of Wingless is evident (arrow, Figure 3B–B'') in the niche compared to the control (arrow, Figure 3A–A''). Quantitative analysis reveals a 1.6-fold increase in the fluorescence intensity of Wg per unit area in the niche where Rel function is attenuated compared to that of the control (Figure 3C). Tweaking of Wg in the background of Rel loss from the niche by RNAi constructs led to a decline in niche cell number compared to Rel loss from the niche (compare Figure 3G–G' with Figure 3E–E' and H), restores the hyperproliferative niche to a cell number comparable to the control (Figure 3D–D' and H).

Figure 3 with 1 supplement see all
Upregulated Wingless signaling leads to increase in niche cell number. The genotypes are mentioned in relevant panels.

Scale bar: 20 μm. (A–B'') Expression of Wingless (antibody) in the lymph gland. The hematopoietic niche is visualized by Antp-GAL4>UAS-GFP. (A'–A'') and (B'–B'') are higher magnifications of (A) and (B), respectively. In comparison to the wild-type niche (A–A''), Wingless protein levels were substantially high in Relish loss of function (B–B''). (C) Statistical analysis reveals elevated wingless expression upon Relish knockdown in niche (n=15; p-value=5.8 × 10−9, two-tailed unpaired Student’s t-test.) (D–G') The increased niche number observed upon Relish loss (E–E') is rescued upon reducing Wingless level by the wg RNAi (F–F') in Relish loss genetic background (G–G'). The rescued niche cell number is comparable to control (D–D'). (H) Statistical analysis of the data in (D–G') (n=10, p-value=1.1 × 10−11 for control versus Rel RNAiKK, p-value=3.15 × 10−10 for Rel RNAiKK versus Rel RNAiKK; wg RNAi, n=10, p-value=0.10 for control versus wg RNAi, n=10, p-value=0.29 for control versus Rel RNAiKK; wg RNAi; two-tailed unpaired Student's t-test). (J–M) Hematopoietic progenitors of larval lymph gland (red, reported by DE-Cadherin [Shg] immunostaining). Knocking down wingless function from the niche resulted in loss of Shg-positive progenitors (L). Downregulating wingless using wg RNAi in Relish loss genetic background was unable to restore the reduction in prohemocyte pool (M) observed in Relish loss (K) scenario in comparison to control (J). (N) Statistical analysis of the data in (J–M) (n=10, p-value=6.74 × 10−6 for control versus Rel RNAiKK, p-value=4.03 × 10−7 for control versus wg RNAi; Rel RNAiKK, p-value=3.42 × 10−8 for control versus wg RNAi; two-tailed unpaired Student’s t-test). The white dotted line marks whole of the lymph gland and the niches in (A–G') Yellow dotted lines mark the progenitor zone in (J–M). In all panels, age of the larvae is 96 hr AEH. The nuclei are marked with DAPI (blue). Individual dots represent biological replicates. Error bar: standard deviation (SD). Data are mean ± SD. *p<0.05, **p<0.01, and ***p<0.001.

Interestingly, although the niche cell number was restored in the above genotype, the defects in the maintenance of progenitors (Figure 3J–M and N) and differentiation (Figure 3—figure supplement 1A–D and E) observed upon Relish loss from the niche were still evident.

Similarly, reducing Wg by using a temperature-sensitive mutant allele wgts Bejsovec and Martinez Arias, 1991 following the scheme provided in Figure 3—figure supplement 1F, gave similar restoration of the hyperproliferative phenotype (Figure 3—figure supplement 1G–K). In this case also, there was a failure in rescuing the defects in progenitor maintenance (Figure 3—figure supplement 1L-O and P) as well as differentiation (Figure 3—figure supplement 1Q–T and U).

This set of experiments led us to infer that the upregulated Wg in Relish loss was responsible only for controlling the niche cell number.

In the absence of Relish, altered cytoskeletal architecture of the niche traps Hh

Various studies have established PSC as the niche for hematopoietic progenitors and have shown that it employs a morphogen Hedgehog for its maintenance. It has also been shown that niche expansion correlates to expansion in the progenitor population (Baldeosingh et al., 2018; Benmimoun et al., 2012; Mandal et al., 2007; Pennetier et al., 2012; Tokusumi et al., 2011). However, in contrast to the above studies, despite a threefold increment in niche cell number upon Relish downregulation, we observed a significant reduction in the progenitor pool (Figure 1L–M'). Moreover, restoration in the number of niche cells by modulating Wg levels in Relish knockdown condition failed to restore the differentiation defects observed upon downregulating Relish from the niche (Figure 3J–M and N, Figure 3—figure supplement 1A–D and E, Figure 3—figure supplement 1L-O and P, Figure 3—figure supplement 1Q–T and U). To understand this result, we assayed Hedgehog levels in the niche by using an antibody against Hh protein (Forbes et al., 1993). Interestingly, compared to that of the control, there is a substantial increase in Hh protein in the niche where the Relish function is abrogated (Figure 4A–B''). Quantitative analysis reveals an almost twofold increase in the level of Hh protein in the experimental niche (Figure 4C).

Figure 4 with 3 supplements see all
Hedgehog release from the niche is affected in Relish loss of function.

The genotypes are mentioned in relevant panels. Scale bar: 20 μm. (A–B'') Hedgehog (Hh) antibody staining in the lymph gland shows Hh enrichment in the niche. The hematopoietic niche in Relish loss of function (B–B'') exhibits higher level of Hh in comparison to the control (A–A''). (C) Statistical analysis of fluorescence intensity revealed more than 2.5-fold increase in Hh levels compared to control (n=15, p-value=2.5 × 10−17, two-tailed Students t-test). (D–E'') Progenitors in Relish loss of function exhibits lower level of Extracellular Hh (HhExtra) (E–E'') in comparison to those of control (D–D''). (E'' and D'') are zoomed in view of niche and the neighboring progenitor cells of (E' and D'), respectively. The yellow box denotes the area quantified in (F). (F) The intensity profile of HhExtra in progenitors (along the rectangle drawn from PSC to cortical zone housing differentiated cells in D' and E') reflects a stark decline in the level of HhExtra in Relish loss scenario compared to control. (G–I') Cellular filopodia emanating from the niche cells were stabilized by using untagged phalloidin. The filopodia in Relish loss of function niches were found to be smaller in length and fewer in number (H–H', I–I') as compared to control (G–G'). (J–K) Significant reduction in filopodial length (J, n=10, p-value=6.64 × 10−9, two-tailed Student’s t-test) and number (K, n=6, p-value=9.19 × 10−10, two-tailed Student’s t-test) were observed in Relish loss scenario compared to control. The white dotted line marks whole of the lymph gland and niches in A–B'', D-E'. In all panels, age of the larvae is 96 hr AEH. The nuclei are marked with DAPI (blue). Individual dots represent biological replicates. Error bar: standard deviation (SD). Data are mean ± SD. *p<0.05, **p<0.01, and ***p<0.001.

However, despite having a higher amount of Hh in the niche upon Relish downregulation, there was a decline in the amount of extracellular Hh (HhExt) in the prohemocytes compared to control (Figure 4D–E'' and F). This result is in sync with the observation that Rel loss from the niche leads to the reduction in the levels of Ci155 in the progenitors (Figure 1L–M'), suggesting that Hh produced by the niche is not sensed by the progenitors resulting in their precocious differentiation.

The alteration in extracellular Hh and decline in Ci155 level in the progenitors prompted us to speculate that loss of Relish from niche might have interfered with Hh delivery to the progenitor cells. Several reports in diverse tissues across model organisms have demonstrated filopodia mediated Hh delivery (Bischoff et al., 2013; González-Méndez et al., 2019). Although the filopodial extension has been documented in the case hematopoietic niche (Krzemień et al., 2007; Mandal et al., 2007), its role in Hh delivery is yet to be demonstrated. To check this possibility, we assayed the status of these actin-based cellular extensions emanating from the niche cells in freshly dissected unfixed tissue of control as well experimental. For this purpose, UAS-GMA (also known as UAS-moesin-GFP) that marks F-actin (Kiehart et al., 2000) was expressed in a niche-specific manner. Multiple cellular processes with variable length are detectable in control, while upon Relish knockdown, filopodial extensions are highly compromised (arrowheads, Figure 4G–I'). Quantitative analyses of the data reveal that both length (Figure 4J) and number (Figure 4K) are altered upon Rel loss from the niche. Intrigued with this finding, we independently downregulated Diaphanous (dia), an actin polymerase known to be important in filopodial formation, elongation and maintenance (Homem and Peifer, 2009; Nowotarski et al., 2014), from the niche. As expected, compared to control niches, dia loss resulted in compromised filopodial length and number (Figure 4—figure supplement 1A–B and C-D). Quite similar to Rel loss from the niche, these defects in filopodial in turn affected Hh delivery from the niche (Figure 4—figure supplement 1E–F' and G). As a consequence, there was a decline in the number of progenitors (Figure 4—figure supplement 1H–I and L) and a concomitant increase in the differentiated cells (Figure 4—figure supplement 1J–K and M) compared to control.

Additionally, compared to the control, F-actin (visualized by rhodamine-phalloidin) expression is significantly increased in the cell cortex upon Relish loss from the niche (Figure 4—figure supplement 2A–B'' and C). This accumulation of cortical F-actin intrigued us to further probe into F-actin associated proteins' status, Singed and Enabled in the niche cells upon loss of Relish. While Singed is the Drosophila homolog of Fascin and is involved in cross-linking actin filaments and actin-bundling (Cant et al., 1994; Tilney et al., 2000), Enabled is a cytoskeletal adaptor protein involved in actin polymerization (Gates et al., 2007; Lin et al., 2009). In comparison to the control, where there is a basal level of Singed or lack of Ena expression in the niche, a significant increase in the level of both of these actin-associated proteins occurs upon downregulation of Relish function (Singed: Figure 4—figure supplement 2D–E'' and F and Ena: Figure 4—figure supplement 2G–H'' and I).

Interestingly, co-expressing the RNAi construct of Ena and Rel in the niche partially rescued the defects in progenitor maintenance (Figure 4—figure supplement 3A–C and D) and differentiation (Figure 4—figure supplement 3E–G and H), which is otherwise seen upon Rel loss. This rescue in the phenotype can be attributed to the resurrection of the transport defects of Hh seen upon Rel loss from the niche (Figure 4—figure supplement 3I–K' and L-N).

These results demonstrate that loss of Relish from the niche induces cytoskeletal rearrangement, which disrupts the proper delivery of Hedgehog to the adjoining progenitors. These results further emphasize how aberrant cytoskeleton architecture might interfere with niche functionality by trapping Hh.

Ectopic JNK activation leads to precocious differentiation in relish loss from the niche

Next, we investigated how Relish loss causes alterations in cytoskeletal architecture within the niche. Studies across the taxa have shown mitogen-activated protein kinases (MAPKs) as a major regulator of cellular cytoskeleton dynamics (Densham et al., 2009; Pichon et al., 2004; Reszka et al., 1995; Šamaj et al., 2004). The c-Jun-NH2-terminal kinase (JNK) or so-called stress-activated protein kinases, which belong to the MAPK superfamily, are one such key modulator of actin dynamics in a cell. Whether the cytoskeletal remodeling of the niche in the absence of Relish is an outcome of JNK activation was next explored. Compared to the control where there is a negligible level of activation of JNK signaling in the niche, visualized by TRE-GFP: a transcriptional reporter of JNK (Chatterjee and Bohmann, 2012), a robust increase in the expression occurs in the niche where the function of Relish is abrogated (Figure 5A–B' and C). This result implicates that during development, Relish inhibits JNK activation in the hematopoietic niche.

Figure 5 with 3 supplements see all
Loss of Relish from the niche activated JNK causing niche hyperplasia.

The genotypes are mentioned in relevant panels. Scale bar: 20 μm. (A–B') Upregulation of JNK signaling visualized by its reporter TRE-GFP (green) in Relish knockdown (B–B') compared with WT niche (A–A'). (C) Statistical analysis of fluorescence intensity (A–B') revealed a significant increase in TRE-GFP levels compared to control (n=15, p-value=4.2 × 10−19, two-tailed Student’s t-test). (D–G') Upon niche-specific simultaneous knockdown of Rel and JNK, the niche hyperplasia observed upon loss of Relish (E–E') is rescued (G–G') and is comparable to control (D–D') whereas loss of bsk from the niche does not alter niche cell number (F–F'). (H) Statistical analysis of the data in (D–G') (n=10, p-value=5.6 × 10−8 for control versus Rel RNAi, p-value=8.0 × 10−7 for bsk DN; Rel RNAi versus Rel RNAi, p-value=0.10 control versus for bsk DN; two-tailed unpaired Student's t-test). (I–N) Cellular filopodia from the niche cells in Rel loss of function is found to be smaller in length and fewer in numbers (J and M–N). Simultaneous loss of both JNK using bsk DN and Relish (L and M–N) rescued the stunted, scanty filopodia to control state (I and M–N), whereas loss of JNK did not affect filopodia formation (K and M–N). (M–N) Statistical analysis of the data in (I–L) (Filopodia number: n=10, p=6.96 × 10−8 for control versus Rel RNAi, p-value=8.11 × 10−7 for bsk DN; Rel RNAi versus Rel RNAi, p-value=0.153 for bsk DN versus control. Filopodia length: n=6, p-value=2.78 × 10−16 for control versus Rel RNAi, p-value=1.84 × 10−6 for bsk DN; Rel RNAi versus Rel RNAi, p-value=0.22 for bsk DN vs control; two-tailed unpaired Student’s t-test). (O–R) Knocking down JNK function from the niche did not have any effect on progenitors (visualized by Shg) (Q). Downregulating bsk function in Rel loss genetic background was able to restore the reduction in prohemocyte pool (R) observed in Relish loss (P) scenario in comparison to control (O). (S) Statistical analysis of the data in (O–R) (n=10, p-value=2.26 × 10−6 for control versus Rel RNAi, p-value=1.94 × 10−7 for bsk DN; Rel RNAi versus Rel RNAi, p-value=0.521 for control versus bsk DN; two-tailed unpaired Student's t-test) The white dotted line marks whole of the lymph gland in all cases and niches in (A–G'). Yellow dotted lines mark the progenitor zone in (O–R). In all panels, age of the larvae is 96 hr AEH. The nuclei are marked with DAPI (blue). Individual dots represent biological replicates. Error bar: standard deviation (SD). Data are mean ± SD. *p<0.05, **p<0.01, and ***p<0.001.

Interestingly, activation of JNK alone (expression of Hepact) in the niche can recapitulate the phenotypes associated with Relish loss to a large extent, for example, hyperproliferative niche (visualized by Antp, Figure 5—figure supplement 1A–B'), ectopic differentiation (visualized by Nimrod P1, Figure 5—figure supplement 1D–E' and F), and upregulated cytoskeletal elements (visualized by Enabled, Figure 5—figure supplement 1G–H'' and I). Moreover, downregulating wg function in the same genetic background restores the cell number within the niche. These results further validate the epistatic relation of JNK and Wg in context to the hematopoietic niche (Figure 5—figure supplement 1J–M and N).

To further understand the relationship of Relish-JNK in the context of niche cell proliferation and functionality, a double knockdown of both JNK and Relish from the niche was analyzed. The concurrent loss of JNK and Relish rescues the increase in niche cell proliferation, seen upon Relish loss (Figure 5D–G' and H). Moreover, downregulating JNK in conjunction with Relish loss from the niche restores the abrogated filopodial extension (Figure 5I–L). The quantitative analyses further reveal the restoration of filopodial length (Figure 5M) and number (Figure 5N) in the above genotype. The rescue, in turn, restored the progenitor pool (Figure 5O–R and S) and the differentiation defect (Figure 5—figure supplement 2A–D and E) noted in the lymph gland upon Relish loss from the niche. The rescue in ectopic differentiation coupled with the resurrection of the filopodial extension suggests a re-establishment of the communication process between the niche and the progenitors.

To have a functional insight into this result, we checked the extracellular Hh level (HhExt) in the same genetic background. We found that in the double knockdown of JNK and Relish, the level of HhExt present in the progenitors is similar to that of the control (Figure 5—figure supplement 2F–H' and I-K). Therefore, the downregulation of the elevated JNK in Relish loss restores niche cell number, as well as the proper communication between niche cells and progenitors, which is mandatory for the maintenance of the latter.

Collectively, these results indicate that Relish functions in the niche to repress JNK signaling during development. In the absence of this regulation, upregulated JNK causes cytoskeletal re-arrangements within the niche and disrupts Hh delivery to the progenitors. The morphogen trapped within the niche is unable to reach the progenitors, thereby affecting their maintenance.

Relish inhibits JNK signaling by restricting tak1 activity in the niche during development

It is essential to understand how the repression of JNK by Relish is brought about in a developmental scenario. Several in vitro and in vivo studies in vertebrates have shown the inhibitory role of NF-κB signaling over JNK during various developmental and immune responses (Clark and Coopersmith, 2007; Nakano, 2004; Tang et al., 2001; Volk et al., 2014). In Drosophila, mammalian MAP3 kinase homolog TAK1 activates both the JNK and NF-κB pathways following immune stimulation (Boutros et al., 2002; Kaneko et al., 2006; Vidal et al., 2001). Interestingly, during bacterial infection, Relish, once activated, leads to proteasomal degradation of TAK1, thereby limiting JNK signaling to prevent hyper-immune activation (Park et al., 2004). It is intriguing to speculate that a similar circuit is engaged in the niche to curtail JNK signaling during development. If this is the case, then the loss of tak1 should restore the elevated TRE-GFP expression in a niche where Relish is downregulated. Indeed, upon genetic removal of one copy of tak1 in conjunction with Relish loss from the niche, a drastic decrease in TRE-GFP expression is noted (Figure 5—figure supplement 3A–D). Furthermore, we found a significant reduction in cell number; analogous to what we observe when JNK and Relish activity is simultaneously downregulated from the niche (Figure 5—figure supplement 3E–H and I). It is interesting to note that there is a restoration in the progenitors (Figure 5—figure supplement 3j–M and N) along with the rescue of the precocious differentiation (Figure 5—figure supplement 3O–R and S) observed upon Relish loss from the niche, which is comparable to the control state in the above genotype.

These results led us to infer that Relish restricts the activation of JNK signaling in the hematopoietic niche via tak1 during development. The restraint on JNK activity is essential for proper communication between niche cells and progenitor cells, which is necessary for maintaining the latter.

Ecdysone-dependent activation of Relish in the niche is a developmental requirement

Cleavage, activation, and nuclear translocation of Relish during bacterial infection are brought about by binding of the cell wall component of gram-negative bacteria to membrane-bound receptor PGRP-LC (Kaneko et al., 2006; Leulier et al., 2003). We wondered whether the niche is employing a similar mechanism to regulate Relish activation during development by engaging the endogenous microbiota. To explore this possibility, we checked the status of the hematopoietic niche in the germ-free/axenic larvae (which were devoid of commensal microflora, Figure 6—figure supplement 1A–A' and B). We found no significant change in the niche cell number in an axenic condition (Figure 6A–B' and D) compared to the control. Additionally, JNK signaling (visualized by TRE-GFP) is not active in the hematopoietic niche of the axenic larva (Figure 6—figure supplement 1C-C'), neither the ectopic differentiation (visualized by Hemolectin, green) of the progenitors was evident (Figure 6—figure supplement 1D-D'). Furthermore, we employed a deletion mutant allele of PGRP-LB (PGRP-LB delta). This gene codes for an amide that specifically degrades gram-negative bacterial peptidoglycan (PGN) (Paredes et al., 2011; Zaidman-Rémy et al., 2006). Even in this scenario, where the systemic PGN level is known to be elevated, there is no increase in the niche cell number (Figure 6C–C' and D). The above results demonstrate that during development, Relish expression and activation in the hematopoietic niche are independent of the commensal microflora.

Figure 6 with 2 supplements see all
Ecdysone regulates Relish expression and functionality in the niche.

The genotypes are mentioned in relevant panels. Scale bar: 20 μm. (A–C') Niche number remains comparable to control (A–A') both in axenic larval lymph gland (B–B') and in PGRP-LB mutant where there is upregulation in systemic peptidoglycan levels (C–C'). (D) Statistical analysis of the data in (A–C') (n=9; p-value = 0.262 for control versus germ free and 0.392 for control versus PGRP-LB mutant; two-tailed unpaired Student's t-test). (E–G') Compared to that of control (E–E') Rel expression is significantly downregulated both in EcR loss (G–G') as well as in Rel loss from the niche (F–F'). (H) Statistical analysis of the data in (E–G') (n=10, p-value=7.81 × 10−12 for control versus Rel RNAi loss and p-value = 3.76 × 10−10 for control versus EcR-DN; two-tailed unpaired Student's t-test). (I–K') Similar to Rel loss from the niche (J–J'), EcR loss also results in increase in niche cell numbers (K–K') compared to that of control (I–I'). (L) Statistical analysis of the data in I-K' (n=10, p-value=6.6 × 10−5 for control versus EcR-DN and p-value = 3.1x10−5 for control versus Rel RNAi; two-tailed unpaired Student’s t-test). (M–O') Compared to control (M–M'), both loss of Rel (N–N') and EcR (O–O') from the niche results in increase in differentiation. (P) Statistical analysis of the data in (M–O') (n=10, p-value=4.3 × 10−5 for control versus Rel RNAi and p-value=2.2 × 10−6 for control versus EcR-DN; two-tailed unpaired Student’s t-test). (Q–T') Increase in niche cell numbers observed upon EcR loss from the niche (R–R') is rescued to control levels (Q–Q') when Relish was overexpressed in an EcR loss genetic background (T–T'). Overexpression of Relish in the niche reduced the cell number compared to control (compare S–S' and Q–Q'). (U) Statistical analysis of the data in (Q–T') (n=10; p-value=1.7×10−9 for control versus EcR-DN, p-value=7.8 × 10−11 for Ecr-DN versus UAS-Rel 68kD; EcR-DN, p-value=3.63 × 10−6 for control versus UAS-Rel 68kD; two-tailed unpaired Student’s t-test). The white dotted line marks whole of the lymph gland and niches in all the cases. In all panels, age of the larvae is 96 hr AEH. The nuclei are marked with DAPI (blue). Individual dots represent biological replicates. Error bar: standard deviation (SD). Data are mean ± SD. *p<0.05, **p<0.01, and ***p<0.001.

Interestingly, activation of the IMD pathway components PGRP-LC and Relish is transcriptionally regulated by steroid hormone 20-hydroxyecdysone signaling during bacterial infection (Rus et al., 2013). Moreover, a recent study also reveals that the activation of Relish and IMD-dependent genes is mediated via ecdysone signaling in the Malpighian tubules during development (Verma and Tapadia, 2015). Strong expression of the Ecdysone receptor in the hematopoietic niche (Figure 6—figure supplement 1E–E'') prompted us to check the possibility of ecdysone-dependent regulation of Relish expression and activation in the niche. Upon expression of a dominant-negative allele of the receptor EcR in the niche, a drastic reduction in the amount of Relish protein is evident (Figure 6E–G'). Intensity analysis reveals a threefold decrease in Relish expression upon blocking ecdysone signaling compared to the control niches (Figure 6H). Since transcriptional regulation of Relish through ecdysone signaling has been previously reported (Rus et al., 2013), we decided to explore whether this holds in case of the hematopoietic niche. Fluorescent in situ hybridization (FISH) analysis reveals the presence of Rel transcript in the lymph gland as well as in the salivary gland of control third-instar larvae (Figure 6—figure supplement 2A–A'' and C–C'). Due to increase in differentiation, the number of Rel expressing progenitors is less compared to control (Figure 6—figure supplement 2B–B'). The sense probe was used as the negative control (Figure 6—figure supplement 2D–E).

To probe the status of Rel transcripts specifically in the niche, we performed whole-mount immunofluorescence (IF) along with FISH on the third-instar lymph gland. Drastic reduction of the Rel transcript is noticeable in the niche from where EcR expression was downregulated compared to the control (Figure 6—figure supplement 2F–G'' and H), implicating that Rel is transcriptionally regulated through ecdysone signaling.

This observation indicates that the phenotypes observed upon EcR loss from the niche should be analogous to Rel loss. Attenuation of ecdysone signaling indeed leads to a significant increase in niche cell proliferation compared to the control (Figure 6I–K' and L). Furthermore, to understand whether the functionality of the niche is also compromised in the above genotype, we checked the differentiation status. Similar to Relish loss, downregulation of ecdysone signaling from the niche results in precocious differentiation (Figure 6M–O' and P). Niche-specific overexpression of Rel in conjunction with EcR loss can restore the cell number of the niche (Figure 6Q–T' and U) as well as its functionality (Figure 6—figure supplement 2I–L and M).

These results, therefore, collectively suggest that ecdysone signaling regulates the expression and activation of Relish in the hematopoietic niche during development (Figure 6—figure supplement 2N). These results also underscore the requirement of a hormonal signal in regulating Relish during developmental hematopoiesis.

During bacterial infection Relish in the niche is downregulated to facilitate immune response

In Drosophila, ecdysone-mediated immune potentiation has shown to have a greater impact on the development of immunity in embryos (Tan et al., 2014) as well as the survival of flies during bacterial infection (Flatt et al., 2008; Rus et al., 2013; Tan et al., 2014; Verma and Tapadia, 2015; Xiong et al., 2016). Interestingly, we found a fourfold decrease in Relish expression from the hematopoietic niche during bacterial infection compared to uninfected larvae (compared Figure 7A–A' with C -C' and quantitated in Figure 7D). To rule out the possible effect of injection on Rel expression, we compared the infected with sham control. There was a 2.6-fold decrease in the intensity of Rel expression within the niche of infected larvae compared to the sham control (compare Figure 7B–B' with C–C', quantitated in Figure 7D). In contrast, upon bacterial infection, we could see the nuclear expression of Relish in the fat body cells as previously reported (Figure 7E–G; Cha et al., 2003; Kim et al., 2006). Interestingly, niche-specific overexpression of the N-terminal domain of Relish (UAS-Rel68kD), which is known to translocate to the nucleus and induce target gene expression (Stöven et al., 2000), is unable to sustain Relish expression post-infection (Figure 7H–H'), implicating the post-transcriptional regulation on Relish during bacterial infection. Relish activity is modulated through proteasomal degradation in Drosophila and Bombyx mori (Khush et al., 2002; Ma et al., 2015).

Figure 7 with 1 supplement see all
Niche-specific expression and function of Relish is susceptible to pathophysiological state of the organism.

The genotypes are mentioned in relevant panels. Scale bar: 20 μm. (A–C') Compare to uninfected conditions (A–A') and sham (B–B'), significant reduction in Relish expression was observed in the hematopoietic niche 4 hr post-infection (C–C'). (D) Statistical analysis of the data in (A–C') (n=15; p=6.62×10−18 for unpricked versus infected, p=2.5×10−7 for sham versus infected, two-tailed unpaired Student’s t-test). (E–G) Nuclear expression of Relish was observed in infected (G) fat body cells 4 hr post in contrast to uninfected (E) and sham (F) larval fat body. (H–H') Overexpressing Relish N-terminus (UAS-Rel-68kD) could not rescue loss of Relish expression post-infection. (I–J) Compared to sham (I), significant reduction in Shg-positive progenitors (red) were observed in infected lymph glands (J). (K) Statistical analysis of the data in (I–J) (n=10; p-value=5.2 × 10−6 for sham versus infected, two-tailed unpaired Student's t-test). (L–M) Drastic increase in differentiation (visualized by Pxn-YFP, green) was observed in infected lymph glands (M) compared to sham (L). (N) Statistical analysis of the data in (L-M) (n=10; p-value = 4.65×10−6 for sham versus infected, two-tailed unpaired Student's t-test). The white dotted line mark whole of the lymph gland in all cases. Yellow dotted line marks the niche in (A– C' and H–H') and the boundary between CZ and MZ in (L–M). In all panels, age of the larvae is 96 hr AEH. The nuclei are marked with DAPI (Blue). Individual dots represent biological replicates. Error bar: standard deviation (SD). Data are mean ± SD. *p<0.05, **p<0.01, and ***p<0.001.

More importantly, we also found that compared to control, 4 hr post-bacterial challenge, the progenitor pool declines (Figure 7I–K), accompanied by a concomitant precocious differentiation (Figure 7L–N). These phenotypes show a remarkable similarity to the ones seen on the loss of Relish from the niche (Figure 1H–P). As a response to systemic bacterial infection, upregulation of JNK is detected throughout the lymph gland, including the niche compared to sham control (Figure 7—figure supplement 1A–B'). The short duration of systemic infection adopted in our study induced proliferation in the otherwise quiescent niche cells (Figure 7—figure supplement 1C–E). Based on these studies, we speculate that Relish, in this case, might also undergo ubiquitin-mediated degradation (by Factor X, Figure 7—figure supplement 1F) that overrides the developmental signal (Figure 6—figure supplement 2M) during bacterial infection.

These data collectively elucidate that a differential regulation on Relish is mandatory during bacterial infection to boost immune response.

Discussion

Our study unravels the molecular genetic basis of the hormonal control on Relish expression in the hematopoietic niche essential for maintaining the hemocyte progenitors of the lymph gland during development. Hemocytes present in the lymph gland are not actively involved in immune surveillance under healthy conditions. Within this organ, the hemocytes proliferate to create a pool of progenitors and differentiated cells. However, with its content, this organ takes care of all post-larval hematopoiesis and therefore is not precociously engaged. Our study illustrates how the hematopoietic niche recruits neuroendocrine-immunity (Ecdysone–Relish) axis to maintain the progenitors of the lymph gland during larval development (Figure 6—figure supplement 2M). The loss of Ecdysone/Relish, therefore, results in precocious maturation of the progenitors. The mechanism underlying the control of niche state and function by Relish involves repression of the Jun Kinase signaling. Interestingly, Relish during infection is known to inhibit JNK activation in response to gram-negative bacterial infection in Drosophila (Park et al., 2004). We found that this antagonistic relation of Relish and JNK, essential for innate immunity, is also relevant during development to facilitate the functioning of the hematopoietic niche. Our results suggest that two independent events occur in the niche if JNK is activated (Figure 6—figure supplement 2M). Firstly, the activation of JNK leads to supernumerary niche cells due to an increase in Wingless expression. Secondly, the JNK pathway negatively regulates the actin-based cytoskeletal architecture essential for the release of Hh from the niche cells.

Though perceived as a pro-apoptotic signal, a large body of work has evidenced the role of the JNK pathway to induce proliferation in diverse developmental scenarios (Kaur et al., 2019; Ohsawa et al., 2012; Pérez-Garijo et al., 2009; Pinal et al., 2019; Wu et al., 2010). The JNK pathway is also known for its ability to release proliferative signals that can stimulate the growth of the tissue (Pinal et al., 2019). For instance, during compensatory proliferation in the developing larval wing disc, JNK triggers wingless to stimulate the proliferation of the non-dead cells (Ryoo et al., 2004). Moreover, wingless signaling has been reported as a mitogenic signal for stem cells in diverse contexts (Deb et al., 2008; Lin et al., 2008; Song and Xie, 2003), and aberrant activation of this pathway contributes to various blood cell disorders and cancers (Grainger and Willert, 2018; Klaus and Birchmeier, 2008; Lento et al., 2013; Reya and Clevers, 2005). Drosophila hematopoietic niche is known to positively rely upon wingless (Wg) signaling for its proliferation during larval development. Downregulation of the signaling by expressing a dominant-negative form of its receptor Frizzled results in a reduction in niche cell numbers (Sinenko et al., 2009). We believe, to prevent hyperproliferation of the niche cells, Relish is reining in Wingless by inhibiting JNK signaling during development.

Several studies have shown that actin-based cellular extensions or cytonemes (Bischoff et al., 2013; González-Méndez et al., 2019; Gradilla et al., 2014; Kornberg and Roy, 2014; Portela et al., 2019) play a crucial role in transporting Hh from the source to several cell diameter distances (Rojas-Ríos et al., 2012), thereby contributing in the establishment of Hh gradient. Coincidently, Drosophila hematopoietic niche cells are also known to emanate cytoneme-like filopodial projections to the nearby progenitor cells (Mandal et al., 2007; Pennetier et al., 2012; Tokusumi et al., 2011). We demonstrate that perturbation of this filopodial extension disrupts the transportation of Hh from the niche. The current study is in sync with the understanding that these cellular extensions are required to maintain the undifferentiated cell population by facilitating the crosstalk between niche and hematopoietic progenitors (Krzemień et al., 2007; Tokusumi et al., 2011). Here, we show that upon Relish loss from the niche, filopodial formation gets impaired in a JNK- dependent manner. Ectopic activation of JNK signaling leads to altered expression of cytoskeletal elements that disrupt the process of filopodial formation. Consequently, the morphogen Hh gets trapped within the niche cells, thereby hamper the proper communication between the niche and the progenitor cells of the lymph gland (Figure 8). Previous studies have demonstrated that activation of Relish leads to the disruption of cytoskeletal architecture in S2 cells to bring about the necessary changes associated with cell shapes for the proper immune response (Foley and O'Farrell, 2004). However, the underlying mechanism of the modulation of cytoskeletal elements by Relish was not evident. Here we provide in vivo genetic evidence for the process by which Relish loss causes alteration of the cytoskeletal elements of the niche cells by ectopic JNK activation.

Developmental requirement of Relish in the niche for progenitor maintenance.

Scheme describing how loss of Relish from the niche alters cytoskeletal elements of the cells. The change in cytoskeletal architecture affects cytoneme-like filopodial formation thereby trapping Hedgehog within the niche. The failure of Hh delivery in turn interferes with progenitor maintenances and pushes them toward differentiation.

Another enthralling finding of our study is identifying 20-hydroxyecdysone signaling as a regulator of Drosophila developmental hematopoiesis. The underlying reasons for this hormonal control on Relish seem to be intriguing. The need for this regulatory network during development may be related to the various microbial threats commonly confronted and dealt with by the circulating hemocytes of the larvae. While the circulating hemocytes cater to this need, the blood cells in the lymph gland proliferate and undergo maturation, creating a reservoir of hemocytes dedicated to deal with the post-larval requirements. Therefore, to safeguard the reserve population from responding to all of the common threats faced during development, the niche employs the ecdysone–Relish axis to prevent the disruption in definitive hematopoiesis. However, during a high infection load, the lymph gland ruptures, suggesting a break in this circuit. This notion gets endorsed when the niche is analyzed post-infection. A previous study demonstrated that the septate junction in the niche is dismantled during infection, leading to the disbursing of differentiation signals that facilitated the maturation of the hemocytes (Khadilkar et al., 2017). We demonstrate that bacterial infection results in downregulation of Rel from the niche, which alters cytoskeletal architecture and traps the maintenance signal. As a consequence, precocious differentiation sets in the lymph gland, while in the case of the earlier study, seeping out of too many differentiation signals leads to ectopic differentiation underlining the fact that maintenance and differentiation are both sides of the same coin.

Quite intriguingly, the downregulation of Relish in the niche during bacterial infection and the response of the lymph gland mimic the genetic loss of Relish from the niche. These observations confirm that the developmental pathway gets tweaked in the hematopoietic niche to combat high bacterial infection (Figure 7—figure supplement 1F).

During bacterial infection, the activation of Relish by ecdysone signaling in the fat body results in the production of antimicrobial peptides (Rus et al., 2013). In contrast to this, we show that upon infection, Relish needs to be downregulated in the niche to bolster the cellular immune response. This downregulation of Relish facilitates the release of a large pool of macrophages from the lymph gland to augment the circulating hemocytes to combat infection. The lymph gland hemocytes do not participate in immune surveillance during development. However, during wasp infection, activation of the Toll/NF-κB signaling occurs in the niche to recruit lymph gland hemocytes to encapsulate wasp eggs (Louradour et al., 2017). We show that during bacterial infections, Relish, another member of the NF-κB pathway, is downregulated in the niche to disperse the lymph gland hemocytes into circulation. It is intriguing to see that the contrasting regulation of NF-κB components by the hematopoietic niche is essential for mounting an adequate immune response.

Interestingly, de novo production of neutrophils occurs in the bone marrow in response to systemic bacterial infection (Zhao and Baltimore, 2015). In mouse, ‘emergency granulopoiesis’ demands the activation of the TLR (Toll-like receptors)/NF-κB pathway via TLR4 in the vascular niche (Boettcher et al., 2014). It will be important to investigate whether this differential regulation on NF-κB members is evident in vertebrate bone marrow niches during infection.

For an organism to combat an infection successfully, a quick shift of the ongoing hematopoiesis toward emergency mode is absolutely necessary. We show that the hematopoietic niche is the sensor that gauges the physiological state of the animal and diverts the basal hematopoiesis toward the emergency hematopoiesis.

In conclusion, the present work reveals an unexpected role of Relish in developmental hematopoiesis. Furthermore, it unravels the systemic regulation of the hematopoietic niche by the neuroendocrine system. Also, it sheds light on how, during infection, this pathway gets suppressed to reinforce the cellular arm of the innate immune response.

Materials and methods

Key resources table
Reagent type
(species)
or resource
DesignationSource or
reference
IdentifiersAdditional
information
Gene (Drosophila melanogaster)AntpFlybase:FB2020_01FLYB:FBgn0260642
Gene (Drosophila melanogaster)HmlFlybase:FB2020_01FLYB:FBgn
0029167
Gene (Drosophila melanogaster)Collier/knFlybase:FB2020_01FLYB:FBgn0001319
Gene (Drosophila melanogaster)wgFlybase:FB2020_01FLYB:
FBgn0284084
Gene (Drosophila melanogaster)hepFlybase:FB2020_01FLYB:FBgn0010303
Gene (Drosophila melanogaster)EcRFlybase:FB2020_01FLYB:FBgn0000546
Gene (Drosophila melanogaster)PGRP-LBFlybase:FB2020_01FLYB:FBgn0037906
Gene (Drosophila melanogaster)Tak1Flybase:FB2020_01FLYB:FBgn0026323
Gene (Drosophila melanogaster)bskFlybase:FB2020_01FLYB:FBgn
0000229
Gene (Drosophila melanogaster)EnaFlybase:FB2020_01FBgn0000578
Gene (Drosophila melanogaster)HhFlybase:FB2020_01FBgn0004644
Gene (Drosophila melanogaster)DiaFlybase:FB2020_01FBgn0011202
Genetic reagent (D. melanogaster)Antp-Gal4Emerald and Cohen, 2004FLYB:FBal0155891FlyBase symbol: GAL4Antp-21
Genetic reagent (D. melanogaster)P(col5-cDNA)/CyO-TM6B, TbKrzemień et al., 2007FLYB:FBti0077825FlyBase symbol: P{GAL4}col85
Genetic reagent (D. melanogaster)Hml-GAL4.ΔSinenko and Mathey-Prevot, 2004FLYB:FBtp0040877FlyBase symbol:P{Hml-GAL4.Δ}
Genetic reagent (D. melanogaster)UAS-Rel RNAiKKVienna Drosophila Resource CenterVDRC:v108469;
FLYB:FBti0116709;
RRID:FlyBase_FBst0477227
FlyBase symbol: P{KK100935}VIE-260B
Genetic reagent (D. melanogaster)w[1118]Bloomington Drosophila Stock CenterBDSC:3605; FLYB:FBal0018186;RRID:BDSC_3605FlyBase symbol: w1118
Genetic reagent (D. melanogaster)UAS-Rel RNAiBloomington Drosophila Stock CenterBDSC:33661; FLYB:FBti0140134;RRID:BDSC33661FlyBase symbol: P{TRiP.HMS00070}attP
Genetic reagent (D. melanogaster)UAS-wg RNAiBloomington Drosophila Stock CenterBDSC:33902; FLYB:FBal0263076; RRID:BDSC_33902FlyBase symbol: P{TRiP.HMS00844}attP2
Genetic reagent (D. melanogaster)UAS-dia RNAiBloomington Drosophila Stock CenterBDSC:35479;
FLYB:FBtp0068562; RRID:BDSC_35479
FlyBase symbol: P{TRiP.GL00408}
Genetic reagent (D. melanogaster)UAS-hep.ActBloomington Drosophila Stock CenterBDSC:9305; FLYB:FBti0074410; RRID:BDSC_9305FlyBase symbol: P{UAS-Hep.Act}1
Genetic reagent (D. melanogaster)UAS-FUCCIBloomington Drosophila Stock CenterBDSC:55121; RRID:BDSC_55121FlyBase symbol: P{UAS-GFP.E2f1.1–230}32; P{UAS-mRFP1.NLS.CycB.1–266}19
Genetic reagent (D. melanogaster)TRE-GFPBloomington Drosophila Stock CenterBDSC:59010; FLYB:FBti0147634; RRID:BDSC_59010FlyBase symbol: P{TRE-EGFP}attP16
Genetic reagent (D. melanogaster)Pxn-YFPKyoto Stock Centerkyoto:115452;
FLYB: FBti0143571;
RRID:FlyBase_FBst0325439
FlyBase symbol: PBac{802 .P.SVS-2}PxnCPTI003897
Genetic reagent (D. melanogaster)hhF4f-GFPTokusumi et al., 2012FBtp0070210FlyBase symbol:P{hhF4f-GFP}
Genetic reagent (D. melanogaster)UAS-GMABloomington Drosophila Stock CenterBDSC:31774; FLYB:FBti0131130; RRID:BDSC_31774FlyBase symbol:P{UAS-GMA}1
Genetic reagent (D. melanogaster)UAS-Rel 68kDBloomington Drosophila Stock CenterBDSC:55778; FLYB:FBti0160486; RRID:BDSC_55778FlyBase symbol: P{UAS-FLAG-Rel.68}i21-B
Genetic reagent (D. melanogaster)UAS-Rel 68kDBloomington Drosophila Stock CenterBDSC:55777;
FLYB:FBti0160484
RRID:BDSC_55777
FlyBase symbol:
P{UAS-FLAG-Rel.68}
Genetic reagent (D. melanogaster)UAS-EcR.B1ΔBloomington Drosophila Stock CenterBDSC:6872; FLYB:FBti0026963; RRID:BDSC_6872FlyBase symbol: P{UAS-EcR.B1-ΔC655.W650A}TP1-9
Genetic reagent (D. melanogaster)PGRP-LB[Delta]Bloomington Drosophila Stock CenterBDSC:55715; FLYB:FBti0180381; RRID:BDSC_55715FlyBase symbol: TI{TI}PGRP-LBΔ
Genetic reagent (D. melanogaster)wgl-12 cn1 bw1/CyOBloomington Drosophila Stock CenterBDSC:7000; FLYB:FBal0018504; RRID:BDSC_7000FlyBase symbol: wgl-12
Genetic reagent (D. melanogaster)Tak1(2)Bloomington Drosophila Stock CenterBDSC:26272; FLYB:FBal0131420; RRID:BDSC_26272FlyBase symbol: dTak12
Gene (Drosophila melanogaster)RelE20Flybase:FB2020_01FLYB:FBgn0014018
Genetic reagent (D. melanogaster)UAS-bsk[DN]Bloomington Drosophila Stock CenterBDSC:6409; FLYB:FBti0021048; RRID:BDSC_6409FlyBase symbol: P{UAS-bsk.DN}2
Genetic reagent (D. melanogaster)UAS-ena RNAiKKVienna Drosophila Resource CenterVDRC: v106484
FBst0478308;
RRID:v106484
FlyBase symbol: P{KK107752}VIE-260B
Genetic reagent (D. melanogaster)UAS-mCD8: RFPBloomington Drosophila Stock CenterBDSC:27400; FLYB:FBti0115747; RRID:BDSC_27400FlyBase symbol: P{UAS-mCD8.mRFP.LG}28a
Genetic reagent (D. melanogaster)tubGAL80[ts20]Bloomington Drosophila Stock CenterBDSC:7109; FLYB:FBti0027796; RRID:BDSC_7109FlyBase symbol: P{tubP-GAL80ts}20
AntibodyAnti-P1 (Mouse monoclonal)Kurucz et al., 2007Cat# NimC1, RRID:AB_2568423IF(1:50)
AntibodyAnti-c Rel (Mouse monoclonal)Stöven et al., 2000Cat#21F3,
RRID:AB_1552772
IF (1:50)
AntibodyAnti-Ci155 (Rat polyclonal)Developmental Studies Hybridoma BankCat# 2A1,
RRID:AB_2109711
IF(1:2)
AntibodyAnti-Wg (Mouse monoclonal)Developmental Studies Hybridoma BankCat#4D4
RRID:AB_528512
IF(1:3)
AntibodyAnti-Singed
(Mouse monoclonal)
Developmental Studies Hybridoma BankCat# sn 7C
RRID:AB_528239
IF(1:20)
AntibodyAnti-Enabled
(Mouse monoclonal)
Developmental Studies Hybridoma BankCat#5G2
RRID:AB_528220
IF(1:30)
AntibodyAnti-PH3(Rabbit monoclonal)Cell signaling
Technology
Cat# 3642S
RRID:AB_10694226
IF(1:150)
AntibodyAnti-Hh (Rabbit monoclonal)Forbes et al., 1993IF(1:500)
AntibodyAnti-Hnt
(Mouse monoclonal)
Developmental Studies Hybridoma BankCat#1G9
RRID:AB_528278
IF(1:5)
AntibodyAnti-EcR common
(Mouse monoclonal)
Developmental Studies Hybridoma BankCat#DDA2.7
RRID:AB_10683834
IF(1:20)
AntibodyAnti-Ance (rabbit monoclonal)Hurst et al., 2003IF(1:500)
AntibodyAnti-GFP
(rabbit polyclonal)
Cell signaling
Technology
Cat#2555IF(1:100)
AntibodyAnti-shg
(rat monoclonal)
Developmental Studies Hybridoma BankCat#DCAD2
RRID:AB_528120
IF(1:50)
AntibodyAnti-β-PS
(mouse monoclonal)
Developmental Studies Hybridoma BankCat#CF.6G11
RRID:AB_528310
IF(1:3)
AntibodyAnti-DIG-POD (sheep polyclonal)Sigma-AldrichCat#11207733910IF(1:1000)
Chemical compound, drugPhalloidin from Amanita phalloidesSigma-AldrichCat#P2141IF(1:500)
Chemical compound, drugRhodamine PhalloidinThermo ScientificCat# R415
RRID:AB_2572408
IF(1:500)
Sequence-based reagentRelish cDNA cloneDGRCClone id: GH01881
FLYB: FBcl0110737
Sequence-based reagentActin_FElgart et al., 2016PCR primersGGAAACCACGCAAATTCTCAGT
Sequence-based reagentActin_RElgart et al., 2016PCR primersCGACAACCAGAGCAGCAACTT
sequence-based reagentAceto_FElgart et al., 2016PCR primersTAGTGGCGGACGGGTGAGTA
Sequence-based reagentAceto_RElgart et al., 2016PCR primersAATCAAACGCAGGCTCCTCC
Sequence-based reagentLacto_FElgart et al., 2016PCR primersAGGTAACGGCTCACCATGGC
Sequence-based reagentLacto_RElgart et al., 2016PCR primersATTCCCTACTGCTGCCTCCC
Software, algorithmFijiFijiRRID:SCR_002285
Software, algorithmPhotoshop CCAdobeRRID:SCR_014199
Software, algorithmImarisBitplaneRRID:SCR_007370
Commercial assay or kitClick-iTEdU plus (DNA replication kit)InvitrogenCat# C10639
Commercial assay or kitAlexa Fluor 594 Tyramide ReagentThermo FischerCat# B40957

Fly stocks

Request a detailed protocol

In this study, the following Drosophila strains were used: Antp-Gal4 (S. Cohen, University of Copenhagen, Denmark), PCol85-Gal4 (M. Crozatier, Université de Toulouse, France), RelE20 (B. Lemaitre, École polytechnique fédérale de Lausanne, Switzerland), and hhF4f-GFP (R. Schulz, University of Notre Dame, USA). Hml-GAL4.Δ (S. Sinenko, Russian Academy of Sciences, Moscow), UAS-Rel RNAi (II), Pxn-YFP, and UAS-ena RNAi (II) were from the Vienna Drosophila Resource Center. The following stocks were procured from Bloomington Drosophila Stock Center: w1118, UAS-Rel RNAi, UAS-Rel 68kD (I), UAS-Rel 68kD (II), UAS-EcR.B1Δ, PGRP-LBΔ, UAS-wg RNAi, UAS-dia RNAi, TRE-GFP, UAS-bsk DN, UAS-mCD8-RFP, UAS-Hepact, wgts/cyo, UAS-GMA, UAS-FUCCI, tubGAL80ts20. Detailed genotype of the fly lines used for the current work is listed in Key Resources Table.

Following genotypes were recombined for the current study:

  1. Antp-Gal4.UAS-mCD8-RFP/Tb

  2. TRE-GFP/TRE-GFP; Antp-Gal4.UAS-mCD8-RFP/Tb

  3. UAS-bsk DN/UAS-bsk DN; +/+; UAS-Relish RNAi/UAS-Relish RNAi

  4. UAS-GMA/UAS-GMA; tubgal80ts/ tubgal80ts; Antp-Gal4 /Tb

  5. w; pcol85-Gal4/UAS-2XeGFP; tub-Gal 80ts

  6. UAS-Relish RNAiKK/UAS-Relish RNAiKK; UAS-Wg RNAi/ UAS-Wg RNAi

  7. tubgal80ts/ tubgal80ts; Antp-Gal4.UAS-2XeGFP/TM2

  8. UAS-Relish /UAS-Relish; UAS-EcR-DN/ UAS-ECR-DN.

  9. UAS-ena RNAiKK/cyo; UAS-Relish RNAi/Tb

  10. UAS-hepact/FM7RFP;+/+; UAS-wg RNAi/Tb

All stocks were maintained at 25°C on standard media. For GAL80ts experiments, crosses were initially maintained at 18°C (permissive temperature) for 2 days AEL to surpass the embryonic development and then shifted to 29°C till dissection.

For time series experiments, synchronization of larvae was done. Flies were allowed to lay eggs for about 4 hr. Newly hatched larvae within 1 hr intervals were collected and transferred onto food plates and kept at 29°C till dissection.

Immunohistochemistry

Request a detailed protocol

Immunostaining and dissection (unless said otherwise) were performed using protocols described in Jung et al., 2005; Mandal et al., 2007; Mondal et al., 2011 using primary antibodies: mouse anti-c-Rel (1:50, a gift from N.Silverman; Stöven et al., 2000), mouse anti Relish (1:50, 21F3, DSHB), mouse anti-Antp (1:10, 8C11, DSHB), mouse anti-Wg (1:3, 4D4, DSHB), mouse anti-P1 (1:40, a gift from I. Ando), rabbit anti-Ance (1:500, a gift from A. D. Shirras; Hurst et al., 2003), rat anti-Ci (1:5, 2A1, DSHB), mouse anti-singed (1:20, Sn7C, DSHB), mouse anti-enabled (1:30, 5G2, DSHB), rabbit anti-PH3 (1:150, Cell Signaling), rabbit anti-Hh (1:500, a gift from P. Ingham; Forbes et al., 1993), mouse anti-Hindsight (1:5, 1G9, DSHB), mouse anti-EcR common (1:20, DDA2.7, DSHB), mouse anti-β-PS (1:3, CF.6G11, DSHB), rabbit-anti-GFP(1:100, 2555, Cell Signalling), and rat anti-shg (1:50, DCAD2, DSHB). Secondary antibodies used in this study are as follows: mouse Cy3, mouse FITC, mouse Dylight 649, rabbit Cy3, (1:500), and rabbit-FITC (1:200) (Jackson Immuno-research Laboratories).

Tissues were mounted in Vectashield (Vector Laboratories) and then followed by confocal microscopy (LSM, 780, FV10i, LSM 900).

EdU incorporation assay

Request a detailed protocol

Click-iT EdU (5-ethynyl-2’-deoxyuridine, a thymidine analog) kit from Life Technologies was used to perform DNA replication assay (Milton et al., 2014). Larval tissue was quickly pulled out in 1× PBS on ice (dissection time not more than 25 min and fat body and salivary gland needs to be cleared from the tissue of interest). Incubation of the dissected tissue was done in EdU solution, Component A (1:1000) in 1× PBS on shaker at room temperature for 30–35 min followed by fixation in 4% paraformaldehyde (prepared in 1× PBS). Post-fixation tissues were washed with 0.3% PBS-Triton four times at 10 min interval followed by 30–35 min of blocking in 10% NGS in 0.3% PBS-Triton. EdU staining solution as per manufacturer’s instruction (for 50 μl staining solution, 43 μl 1× EdU buffer, 2 μl CuSO4 solution, 5 μl 1× EdU buffer additive, 0.12 μl Alexa solution) was used to stain the sample for 30 min at room temperature. Two quick washes with 0.3% PBS-Triton was followed by a quick wash in 1× PBS. If no further antibody staining was required, nuclear staining by DAPI was done in 1× PBS and then mounted in Vectashield.

Extracellular Hh staining and quantitation

Request a detailed protocol

For extracellular Hh staining, a detergent-free staining protocol was used. Lymph glands were dissected in ice-cold Schneider’s media (Gibco 21720024), rinsed with cold PBS twice, and fixed with 4% formaldehyde overnight at 4°C (Sharma et al., 2019). Subsequent processing of the samples was the same as mentioned above in the Immunohistochemistry section, except that no detergent was used.

Protocol described by Ayers et al., 2010 was used as a reference to perform quantitation. A rectangle (500 × 150 pixels) was drawn, spanning from the niche to the cortical zone diagonally with the medullary zone in the middle, as shown in Figure 4F. An extracellular Hedgehog profile was made using the ‘Plot Profile’ tool of ImageJ. The Plot profile tool displays a ‘column average plot’, wherein the x-axis represents the horizontal distance through the selection and the y-axis the vertically averaged pixel intensity, which in this analysis is formed by extracellular Hedgehog staining.

Filopodial detection and quantitation

Request a detailed protocol

UAS-GMA was used to label the filopodia using a niche-specific driver, Antp-GAL4. Lymph glands of the desired genotype were dissected in Schneider’s media (Gibco 21720024) and incubated in a solution containing Schneider’s media supplemented with 1% Phalloidin from Amanita phalloides (P2141 SIGMA) for 15 min in order to stabilize the filopodia. These tissues were then mounted and imaged directly under the confocal microscope.

The intact PSC cells are often scattered when we carry out a live analysis. This is mainly due to imaging requirements that demand a coverslip to be placed on the sample. The coverslip creates a pressure on the unfixed/live sample leading to the scattering of the cells.

Filopodial quantitation was done using ImageJ. The number of filopodia emanating from the niche in all the Z-stacks was counted manually per sample. The average number of filopodia emanating per sample was plotted using GraphPad Prism for different biological replicates. For filopodial lengths, the ‘Freehand line’ tool was used to mark the entire filopodial lengths, and the ‘Measure’ tool was employed to get values in µM. Filopodial lengths in all samples were then plotted collectively as individual points in GraphPad Prism.

Phalloidin staining

Request a detailed protocol

Lymph glands dissected were fixed and incubated in rhodamine-phalloidin (1:100 in PBS) (Molecular Probes) for 1 hr. The samples were then washed thrice for 10 min in PBS followed by mounting in DAPI Vectashield before imaging.

Quantification of intensity analysis of phalloidin

Request a detailed protocol

Membranous intensity of Phalloidin was measured using line function in Image J/Fiji. Mean intensity was taken in a similar manner as mentioned in Shim et al., 2012.

p-values of <0.05, <0.01, and <0.001, mentioned as *, **, and ***, respectively, are considered as statistically significant.

Imaging and statistical analyses

Request a detailed protocol

All images were captured as Z sections in Zeiss LSM 780 confocal microscope and Olympus Fluoview FV10i (Panel 7). Same settings were used for each set of experiments. All the experiments were repeated at least thrice to ensure reproducibility. Mostly, 10 lymph glands were analyzed per genotype for quantification analysis. Data expressed as mean ± standard deviation of values from three sets of independent experiments. At least 10 images of the lymph gland/niche were analyzed per genotype, and statistical analyses performed employed two-tailed Student’s t-test. p-values of <0.05, <0.01, and <0.001, mentioned as *, **, and ***, respectively, are considered as statistically significant. All quantitative analysis was plotted using GraphPad.

Quantitative analysis of cell types in lymph gland

PSC cell counting

Request a detailed protocol

Antp-positive cells were counted using the spot function in imaris software (Sharma et al., 2019). Data from three independent experiments are plotted in GraphPad prism as mean ± standard deviation of the values. All statistical analyses performed employing two-tailed Student’s t-test. (http://www.bitplane.com/download/manuals/QuickStartTutorials5_7_0.pdf).

Quantification of intensity analysis

Request a detailed protocol

Intensity analysis of Hh, TRE-GFP, Wg, Singed, Enabled, Relish antibody, and Rel transcript in different genotypes was done using protocol mentioned in http://sciencetechblog.files.wordpress.com/2011/05/measuring-cell-fluorescence-using-imagej.pdf. For each genotype, in about 10 biological samples, at least five ROIs were quantified. Data is expressed as mean ± standard deviation of values and are plotted in GraphPad prism. All statistical analyses performed employing a two-tailed Student’s t-test.

Differentiation index calculation

Request a detailed protocol

To calculate the differentiation index, middle most stacks from confocal Z sections were merged into a single stack for each lymph gland lobe using ImageJ/Fiji (NIH) software as described earlier (Shim et al., 2013). P1-positive area was marked by using Free hand tool. The size was measured using the Measure tool (Analyse–Measure). In similar way, DAPI area was also measured. The differentiation index was estimated by dividing the size of the P1-positive area by the total size of the lobe (DAPI area). For each genotype, mostly 10 lymph gland lobes were used, and statistical analysis was performed using two-tailed Student’s t-test.

Fucci cell cycle analysis

Request a detailed protocol

UAS-GFP-E2f11-230 UAS-mRFP1NLS-CycB1-266 (Zielke and Edgar, 2015) fly line depends on GFP- and RFP-tagged degrons from E2F1 and Cyclin B proteins. Both E2F1 and Cyclin B gets degraded by APC/C and CRL4cdt2 ubiquitin E3 ligases once they enter S and G2-M phase of cell cycle, respectively. Due to accumulation of GFP-E2f11-230, G1 phase will show green fluorescence, and due to accumulation of mRFP1NLS-CycB1-266, S phase will show red fluorescence. Since both GFP-E2f11-230 and mRFP1NLS-CycB1-266 are present in G2 and M phases, the cells will show yellow fluorescence. UAS-GFP-E2f11-230 UAS-mRFP1NLS-CycB1-266 fly stock was recombined with Antp-Gal4 and was crossed to UAS-Relish RNAi and w1118 to ascertain the cell cycle status niche cells. All flies were kept at 25°C, and larvae were dissected 96 hr AEH.

Generation of axenic batches

Request a detailed protocol

Germ-free batches were generated following the ethanol-based dechorination method provided in Elgart et al., 2016. According to this method, embryos were collected and washed using autoclaved distilled water to get rid of residual food particles. Embryos were further dechorinated for 2–3 min in 4% sodium hypochlorite solution. Once this is done, embryos were washed with autoclaved distilled water and were transferred to the sterile hood. Further manipulations were done inside the hood in order to avoid cross-contamination. Embryos were further washed twice with sterile water and were transferred into standard cornmeal food supplemented with tetracycline (50 µg/ml).

Bacterial plating experiment

Request a detailed protocol

For plating experiments, three to five late third-instar larvae were washed in 70% ethanol twice for 2 min. Furthermore, the larvae were washed using sterile H2O twice for 2 min. After this surface sterilization, the larvae were transferred into LB media and were crushed thoroughly using a pestle. Once crushed, the homogenates were spread on LB agar media and was incubated for 3–4 days at 25°C.

Measuring of bacterial content by qPCR

Request a detailed protocol

To measure bacterial composition in the gut, 12–15 third-instar larval guts were dissected and pooled, and DNA was isolated manually using the protocol provided by VDRC (https://stockcenter.vdrc.at/images/downloads/GoodQualityGenomicDNA.pdf) and followed by PCR analysis using species-specific primers. Drosophila actin was used as a control.

S. no.Gene/species namePrimer sequence
1Actin5′-GGAAACCACGCAAATTCTCAGT-3′
5′-CGACAACCAGAGCAGCAACTT-3′
2Acetobacter5′-TAGTGGCGGACGGGTGAGTA-3′
5′-AATCAAACGCAGGCTCCTCC-3′
3Lactobacillus5′-AGGTAACGGCTCACCATGGC-3’
5′-ATTCCCTACTGCTGCCTCCC-3′

Infection experiments

Request a detailed protocol

The following bacterial strains were used for infection: E. coli (OD600:100). For larval infection, bacterial cultures were concentrated by centrifugation; the pellet formed was resuspended in phosphate-buffered saline (PBS) to appropriate OD value. Synchronized third-instar larval batches were used for all analyses. Third-instar larvae were washed three times with sterile ddH2O and pricked using a fine insect pin dipped in bacterial suspension at the postero-lateral part. Mock injections were done using PBS dipped pins. Complete penetration was confirmed while dissection by looking at the melanization spots at the larval epithelial surface. Once infected, larval batched were transferred to food plates and incubated at 25°Celsius till dissection. All observations were made 4 hr post-infection.

IF-fluorescence in situ hybridization

Request a detailed protocol

The protocol we followed was modified from Toledano et al., 2012.

Probe preparation

Request a detailed protocol

Rel clone was procured from DGRC. Following plasmid linearization and restriction digestion using EcoRV and Xho1, the DNA fragments were loaded in agarose gel for electrophoresis. Furthermore, the desired DNA fragments were purified using PCI (phenol:chloroform:isoamyl alcohol) based gel purification and DIG-labeled RNA anti-sense, and sense probe was prepared using Sp6 and T7 polymerase enzyme, respectively. Following DNase treatment, the probes were precipitated using LiCl and ethanol. The RNA pellet was dried resuspended in RNase-free dH20 and stored at −80°C till further use.

Dual labeling of mRNA and protein in the hematopoietic niche

Request a detailed protocol

For IF-FISH, we followed the Part B of the Tissue preparation and fixation section of Toledano et al., 2012. Followed by quick dissection, the larval tissues (make sure of having minimum fat body cells since it can hinder the fixation and hybridization) were fixed for 30 min in 4% formaldehyde prepared in RNase-free PBS and further washed in PBTH (PBS containing 0.1% Tween 20 and 250 μg/ml yeast tRNA) for thrice, 10 min each. Samples were blocked using 5% BSA prepared in PBTHR (PBTH containing 0.2 U ml −1 RNase inhibitor and 1 mM DTT). Furthermore, tissues were incubated in rabbit anti-GFP (1:100, prepared in PBTHR) for 18 hr at 4°C. Tissues were washed using PBTH three times 10 min each, followed by blocking for 30 min using 5% BSA prepared in PBTHR. The tissues were then incubated in fluorescent-labeled secondary antibody (rabbit-FITC 1:100) for 4 hr at room temperature in a shaker. Following this, three washes of PBTH, 10 min each, tissues were fixed using 10% formaldehyde for 30 min. Post-fixation, tissues were washed thrice, 5 min each and rinsed with 0.5 ml of prewarmed hybridization buffer (HB) for 10 min in a 65°C in a hybridization chamber. Tissued were then blocked with PHB (HB mixed with tRNA [10 mg/ml]) for 1 hr in 65°C. Following blocking, tissues were transferred to preheated RNA probe prepared in PHB (2 μg/ml) and incubated at 65°C for 18 hr. Post-hybridization, stringent washes were given using 0.1% PBT: HB mix as mentioned in Toledano et al., 2012. The issues were then blocked in TNB buffer for 1 hr prior to incubation anti-DIG-POD (1:1000) for 18 hr at 4°C. Post-primary antibody incubation, tissues were washed using 0.1% PBT. For signal detection and amplification, Alexa Fluor 594 Tyramide Reagent was used. Tyramide amplification solution was prepared as mentioned in the user guide. Tissues were incubated in TSA working solution for 8 min. Following this, an equal amount of Reaction stop reagent solution was added and further incubated for 1 min. Post-TSA reaction, tissues were PBS rinsed thrice for 5 min and mounted in Vectashield.

Data availability

All data generated or analysed during this study are included in the manuscript and supporting files. Source data files have been provided for Figures 1- 7 and their respective figure supplements.

References

    1. Forbes A
    2. Nakano Y
    3. Taylor A
    4. Ingham P
    (1993)
    Genetic analysis of hedgehog signalling in the Drosophila embryo
    Development 119:115–124.
    1. Grainger S
    2. Willert K
    (2018) Mechanisms of wnt signaling and control
    Wiley Interdisciplinary Reviews: Systems Biology and Medicine 10:e1422.
    https://doi.org/10.1002/wsbm.1422
    1. Šamaj J
    2. Baluška F
    3. Hirt H
    (2004)
    From signal to cell polarity: mitogen‐activated protein kinases as sensors and effectors of cytoskeleton dynamicity
    Journal of experimental botany 55:189–198.

Decision letter

  1. Bruno Lemaître
    Reviewing Editor; École Polytechnique Fédérale de Lausanne, Switzerland
  2. Utpal Banerjee
    Senior Editor; University of California, Los Angeles, United States

Our editorial process produces two outputs: i) public reviews designed to be posted alongside the preprint for the benefit of readers; ii) feedback on the manuscript for the authors, including requests for revisions, shown below. We also include an acceptance summary that explains what the editors found interesting or important about the work.

Acceptance summary:

Relish is one of the three NK–kB factors in Drosophila of which functions are well–known in the insect immunity. This study identified how Relish in the Drosophila larval hematopoietic niche contributes to hematopoiesis during normal growing conditions and how it achieves a switch to immune responses upon bacterial infections. The primary claims are well–supported, and this study will highly contribute to insect hematopoiesis and immunity.

Decision letter after peer review:

[Editors’ note: the authors submitted for reconsideration following the decision after peer review. What follows is the decision letter after the first round of review.]

Thank you for submitting your work entitled "The dynamic role of Relish in the niche to modulate Drosophila blood progenitor homeostasis in development and infection" for consideration by eLife. Your article has been reviewed by 2 peer reviewers, and the evaluation has been overseen by a Reviewing Editor and a Senior Editor. The reviewers have opted to remain anonymous.

Our decision has been reached after consultation between the reviewers. Based on these discussions and the individual reviews below, we regret to inform you that your work will not be considered further for publication in eLife. Nevetherless, you still has the possibility to re–submit a revised version but this will be considered as a new submission.

The two reviewers were quite enthusiast about the manuscript but they also find that there are too many issues to solve before acceptation that cannot be addressed in the two month time frame. Note that you still have the opportunity to re–submit a revised manuscript to eLife that will be considered as a new submission and likely be reviewed by the same reviewers. Unfortunately, we did not get answer from a third reviewer. In the eventually of a resubmission, we will contact an additional reviewer. At this stage you should decide if you wish to submit a revised manuscript that address the point or move your paper to another journal.

Essential revisions:

Reviewer #1:

In this manuscript, Mandal and colleagues identified a novel role for Relish in the hematopoietic niche development and its coordinative function with immune responses. The authors found that Relish is enriched in the PSC and loss of which causes a dramatic reduction of hematopoietic progenitors. Interestingly, the loss of Relish in the PSC causes the proliferation of PSC cells through activation of Wg and alters the actin cytoskeletal structure of the PSC, in turn trapping Hh in the PSC. The authors also observed that these phenotypes are led by ectopic activation of JNK through TAK1. The authors moved on to find the developmental and physiological relevance of this phenomenon and discovered that EcR signaling activates Rel in the PSC during normal development while innate immunity attenuates the Rel expression in the PSC.

This study first describes the novel role of Relish in blood development and its association with ecdysone signaling as well as with innate immunity. The study is interesting, well–designed and the mechanism shown is novel enough to merit the journal. But the last part, where the biological significance of Rel in the PSC is described, is rather weak compared to the other genetic interactions and require further assessments or better description.

1. Although the authors focus on the expression of Relish in the PSC, Relish is also expressed in progenitors, as shown in Figures 1C–D and Figure 2–supplement 1A–E. Related, loss of Rel in the PSC seems to attenuate Rel in both the medullary zone and the PSC (Figure 1D). Is the Relish expression coordinated in both zones? Does Rel RNAi lead to similar phenotypes when driven in the progenitors? Is the Rel function shown cell–autonomous?

2. Compared to the genetic assessments, paragraphs describing the developmental and immunological relevance of Relish are rather weak and overstated. The only genetic basis shown for the Relish and ecdysone interaction is the expression of UAS–EcR_DN. Does UAS–relish rescue the EcR_DN phenotypes? If indeed ecdysone modulates the Relish level, is there any oscillation in the level of Relish in the PSC during larval development? How does Rel–TAK1JNK–Wg or Hh axis change according to the ecdysone level during normal development? Are Relish and its downstream pathway up or downregulated in the PSC when ecdysone is additionally given?

3. Related to #2, the authors proposed that proteasomal degradation (Factor X) will downregulate Relish during bacterial infection. Have the authors verified that the other components in the pathway upon infection? Or is it already known? Is overexpression of UAS–Relish_WT sufficient to block the precocious differentiation of lymph gland cells upon infection? Does ecdysone level change during infection? Without showing direct evidence, lines 475–489 should be moved to the discussion, and the results need to be toned down.

Reviewer #2:

In this manuscript the authors investigate the role of Relish in the Drosophila lymph gland (LG). They establish that relish is expressed in PSC cells and that reducing its expression in these cells (by expressing relish RNAi with a PSC–gal4 driver) leads to an enlarged PSC, increased plasmatocyte differentiation, no effect on crystal cell numbers, and fewer progenitors in the medullary zone (MZ). In the PSC, Relish controls Wingless levels that in turn control PSC cell proliferation and thus PSC size. This study also establishes that the knock down of relish in the PSC leads to increased levels of several actin binding proteins, reduced filopodia formation in PSC cells and a decrease in Hh (HhExt) release from the PSC. In addition, relish knock–down in the PSC leads to the activation of the JNK pathway in the PSC. Epistasis experiments establish that JNK acts downstream of Relish to control filopodia formation and HhExt. Under normal conditions, Relish levels in the PSC are under the control of ecdysone. Finally, in response to an E. coli infection, a decrease in Relish levels in the PSC is observed together with increased plasmatocyte differentiation.

This is an important study describing a yet unknown regulation of Drosophila LG hematopoiesis. However, I have some concerns with the current version of the manuscript.

In the introduction: the authors do not state that the role of the PSC in the LG under normal conditions is under debate, and papers relative to this, even if they diverge from the dogma, must be cited. It is established in the literature that the MZ progenitors are heterogeneous; this information together with the corresponding papers should be introduced. In the Results section, my first concern is about the models proposed in Figure 7: many epistasis experiments are lacking, and thus at this stage of the analysis it is impossible to propose such a model (see comments below). Furthermore, several controls (pictures and quantifications) are lacking (see below).

The second problem concerns the relationships between relish knock–down in the PSC, the absence of filopodia and the decrease of HhExt. These are important and novel data and should be given in the main figures. However, they need to be consolidated (see comments below).

The third point relates to the role played by Relish in the PSC in response to bacterial infection. In the current version of the manuscript, the data presented are too preliminary to prove that Relish is required and to propose how it is involved in the control of LG stress hematopoiesis.

Figure 1: In the PSC>rel RNAi context, the consequence on blood progenitors must be analyzed in more detail Indeed, it is possible that an increase in the mature plasmatocytes numbers results from increased CZ (cortical zone) precursor maturation, without an input from the MZ pool. A systematic analysis of MZ markers (and quantifications) should be performed in all the genetic contexts analyzed (see below for other figures). Additional markers for MZ cells, other than Ci, must be tested since Ci is part of the Hh signaling pathway which is impaired in this context. Why is the crystal cell index unchanged in the PSC >rel RNAi context? Is there lamellocyte differentiation?

Figure 3 : What about the contribution of PSC Wg in the control of LG homeostasis? What about LG homeostasis (MZ and differentiated blood cells) under conditions where the PSC size is rescued (i.e.: PSC>rel RNAi, Wgts)? In Figure 3 D–G, the control Wgts (picture and quantification) is missing.

Figure 3B–B': Why is Wg staining stronger in PSC cells that express lower levels of GFP?

Figure 3 sup1 E: Is there a significant difference in PSC cell numbers between antp>wgRNAi and control LGs? This should be indicated. In Figure 3 sup 1 F–I: What about crystal cell differentiation and MZ progenitors in PSC>wgRNAi and PSC>WgRNAi, relRNAi?

The control i.e. PSC>wg RNAi (both pictures and quantifications) is missing. Is there any difference in LG size in this context as previously reported by Sinenko et al., 09?

Figure 4 : Are there LG defects when the levels of actin binding proteins (F actin, singed,..) are modified in the PSC ? Is there a functional relationship between Relish and actin binding protein levels in the PSC?

Figure 4 A–B : Is Hh expressed in a subset of CZ cells? Is there basal expression of Hh also in the MZ (see A' and B")?

Figure4 sup1 A–E: Data relative to filopodia formation and HhExt must to be given in the main figures since they are important results. However, they still need to be consolidated. Quantifications for HhExt (quantity, dispersion) and filopodia (numbers and size) are necessary.

Pictures A–A": Is HhExt, as shown with a comet–like pattern outside the PSC, always observed in controls, or is it specific to this picture?

Pictures C–C': These is a problem here: the cells that are shown become detached from the PSC and migrate. In such conditions there are more cytoplasmic extensions that reflect the migrating status of the cells. Therefore this is not a correct illustration for PSC cell filopodia.

Pictures F–F' '(as well as in Figure 4G–L): There are discrepancies between pictures and quantifications (a 6 fold increase in the PSC>rel RNAi context compared to the control is not observed in the pictures).

Figure 5 H–L: the control, corresponding to PSC> bskDN (pictures and quantifications) is missing. Figure F5 O–Q, the control, tak1/+ (pictures and quantifications) is missing.

Figure 5 L–L', should be replaced (blurry picture and migrating PSC cells).

What about MZ and crystal cell differentiation in these different conditions ( i.e: UAS–bskDN, UAS bskDN, Rel RNAi; tak1/+ and tak1 , rel RNAi)?

–Epistasis between the JNK pathway and Hh, and the JNK pathway and wg must be analysed. This is essential to support the models given in Figure 7.

Figure 6: The demonstration that Ecdysone controls relish transcription in the PSC is lacking Analysis of Relish protein levels is not enough, since post transitional regulation has been described for Relish. relish transcription must be analyzed in PSC>EcRDN LGs. Furthermore, epistasis experiments are required to firmly establish a functional link between Relish and ecdysone.

Figure 7: The link between Relish in the PSC and its role in LG hematopoiesis in response to bacterial infection is not yet clearly assessed. That an increase in blood cell differentiation occurs in the LG in response to bacterial infection has already been described (Khadilkar et al;, 2017). This reference should be cited in the manuscript. Whether the bacteria–induced decrease of Relish in the PSC is involved in the LG stress hematopoiesis is not established here since there are many discrepancies between LG phenotypes in PSC>rel RNAi larvae and in response to bacterial infection and rescue experiments are lacking. While in both cases there is an increase in plasmatocyte differentiation, there is no difference in PSC size in response to bacterial infection (which differs from PSC>rel RNAi LGs), and there is no LG disruption in PSC>rel RNAi larvae (which differs from bacterial infection). Furthermore, the LG disruption in response to bacterial infection is a novel data, but unfortunately it is not properly documented in this manuscript. Finally, in bacterial infection experiments, no data relative to a role of Hh or JNK pathway in PSC cells have ever been described so far.

Is the transcription of relish or the stability of the protein that is affected in response to bacterial infection? In Figure 7 D: >UAS –Rel 68KD, without infection, is Relish protein detected by immunostaining in PSC cells? This would clarify whether Relish protein stability is dependent on bacterial infection.

Figure 7 B–D: Both quantifications of PSC cell numbers and Relish levels are missing.

Figure 7 H–J': Plasmatocyte quantification is missing. Analyses and quantifications of crystal cells and MZ progenitors must also be provided.

Models in Figure A and K are not correct, since many epistasis experiments are missing.

Furthermore, an arrow between Wg in the PSC and MZ progenitors is missing, since it has been previously established that the Wg pathway in the PSC controls LG hemocyte differentiation (please see Sinenko et al., 2009).

Figure 8 : Are there any functional links between filopodia, the actin skeleton and Hh signaling in the LG ? This point should be further addressed.

For the PSC>rel RNAi context, both in the summary and in point e in the legend Figure 8, we read "precocious differentiation of the blood progenitors". This point should be better established. Indeed, systematic analyses of MZ markers in the different genetic contexts are required (see comments above).

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

Thank you for submitting your article "Relish plays a dynamic role in the niche to modulate Drosophila blood progenitor homeostasis in development and infection" for consideration by eLife. Your article has been reviewed by 2 peer reviewers, and the evaluation has been overseen by a Reviewing Editor and Utpal Banerjee as the Senior Editor. The reviewers have opted to remain anonymous.

The reviewers have discussed their reviews with one another, and the Reviewing Editor has drafted this to help you prepare a revised submission.

Essential revisions:

General comments by the editor following the discussion with the reviewers

Both reviewers feel that this paper warrants publication in eLife.However , they think that additional new data are needed to fully support the conclusion on the role of PSC fillopodia in signaling. At the very least, the authors must establish whether there is a functional relationship between Relish and actin binding protein levels in the PSC. Rescue experiments of relish knock down, by decreasing "actin protein" and looking at lymph gland homeostasis, fillopodia formation and Hh diffusion, have to be provided.

No functional relationship has been established for the role of Relish in the PSC in response to bacterial infection Thus, the authors should be more cautious, and both in the manuscript and the summary the text relative to this point has to be modulated. Furthermore, the authors should discuss the discrepancy between their model, where in response to bacterial infection Hh is trapped within the niche and unable to diffuse, versus the model proposed by (Khadilkar et al., 2017) where bacterial infection causes the breakdown of the permeability barrier in PSC cells.

To minimize the second round of revision, the reviewers suggested that the authors could also tone down the statements on the filopodia–Hh interaction and the role of Rel in the bacterial infection while adding possible additional works.

1) As I already suggested in the previous comment, the authors should impair PSC fillopodia formation by modifying actin binding proteins (F actin, singed,dia,…) and analysing both lymph gland defects and Hh diffusion. It is also necessary to establish whether there is a functional relationship between Relish and actin binding protein levels in the PSC. Rescue experiments of relish knock down, by decreasing "actin protein" and looking at lymph gland defect, fillopodia formation and Hh diffusion are missing.

In Figure 4 Sup1: In PSC>UAS Dia RNAi, what about fillopodia formation and Hh diffusion ? Does decreasing dia in PSC cells in relish knocked down condition restore fillopodia formation, Hh diffusion and lymph gland cell homeostasis? This is an important issue than remains to be addressed.

Figure 4A–C: when relish is knocked down in the PSC, is Hh the increase in PSC cells due to a defect of Hh diffusion and/or a decrease in Hh transcription in those cells? This has to be clarified.

2) The role of Relish in the PSC in response to bacterial infection is based on correlation, since rescue experiments are missing and thus no functional relationship has been established yet. Furthermore, whereas bacterial infection leads to the increase in crystal cell differentiation (Khadilkar et al., 2017), relish knock down in PSC cells has no impact on crystal cell number. Thus, the authors should be more cautious, and both in the manuscript and the summary the text relative to this point has to be modulated. Finally, the authors should discuss the discrepancy between their model where in response to bacterial infection Hh is trapped within the niche and unable to diffuse, versus the model proposed by (Khadilkar et al., 2017) where the bacterial infection causes the breakdown of the permeability barrier in PSC cells.

https://doi.org/10.7554/eLife.67158.sa1

Author response

[Editors’ note: the authors resubmitted a revised version of the paper for consideration. What follows is the authors’ response to the first round of review.]

Essential revisions:

Reviewer #1:

This study first describes the novel role of Relish in blood development and its association with ecdysone signaling as well as with innate immunity. The study is interesting, well–designed and the mechanism shown is novel enough to merit the journal.

Thank you very much for your comments. We are happy to note that our effort is well appreciated.

But the last part, where the biological significance of Rel in the PSC is described, is rather weak compared to the other genetic interactions and require further assessments or better description.

Thanks for your input. We have addressed all the concerns related to the biological significance of Rel in the hematopoietic niche/PSC. All the experiments suggested from your end have been performed to add strength to this part of our manuscript. Thanks a lot for enriching our study.

1. Although the authors focus on the expression of Relish in the PSC, Relish is also expressed in progenitors, as shown in Figures 1C–D and Figure 2–supplement 1A–E. Related, loss of Rel in the PSC seems to attenuate Rel in both the medullary zone and the PSC (Figure 1D). Is the Relish expression coordinated in both zones? Does Rel RNAi lead to similar phenotypes when driven in the progenitors? Is the Rel function shown cell–autonomous?

a. Loss of Relish expression in progenitor cells of MZ is due to the drastic decline in their number. Therefore the loss of Rel expression in the MZ is not per cell instead due to fewer progenitor cells.

b and d. To check whether the expression and function of Relish in the niche and Medullary zone are mutually independent, we checked the intensity of Relish protein expression in the niches and progenitors of both control and Rel RNAi lymph glands. Quantitative analysis revealed more than a 2-fold reduction in Relish expression in the Rel RNAi niches compared to control, whereas progenitor specific expression remained unaltered (please refer to Figure 1 C-E). The above observation itself indicates that the Rel function described in our study is cell-autonomous.

c. The Rel function within the progenitors is entirely independent to one of the niche. It is a part of an ongoing investigation in the laboratory and is beyond the scope of the current manuscript.

2. Compared to the genetic assessments, paragraphs describing the developmental and immunological relevance of Relish are rather weak and overstated. The only genetic basis shown for the Relish and ecdysone interaction is the expression of UAS–EcR_DN. Does UAS–relish rescue the EcR_DN phenotypes? If indeed ecdysone modulates the Relish level, is there any oscillation in the level of Relish in the PSC during larval development?

Thanks for raising this issue. In the current version of the manuscript, these missing pieces of information have been included.

Upon niche-specific overexpression of Relish in conjunction with EcR loss, we found a rescue in niche proliferation (Figure 6 Q-U) and differentiation defects (Figure 6 figure supplement 2A-E), which are otherwise associated with EcR loss (Figure 6 I-L and M-P). These are also mentioned in Main text Line: 407-413.

As far as Rel expression is concerned, we assayed the Rel protein at two different timepoints that 60 and 72hrs AEH which corresponds to Ecdysone crest and trough respectively. We do not see any significant difference in the level reflecting oscillation. We think that the basal level of Ecdysone present throughout the larval stages (Hodgetts et al., 1976; Kraminsky et al., 1980, Handler, 1982 Riddiford, L. M., 1994) is sufficient to sustain Rel expression in the niche. Please refer to Author response image1A-C below.

Author response image 1
A-C: No significant change in Relish expression in the niche was observed at 72 hours (AEH) (B-B') compared to 60 hours (AEH) ( A-A').

Statistical analysis of the data from A-B' (n=29 P-value =.297 ; two tailed Students t-test).

How does Rel–TAK1–JNK–Wg or Hh axis change according to the ecdysone level during normal development? Are Relish and its downstream pathway up or downregulated in the PSC when ecdysone is additionally given?

Our study reveals that during development, ecdysone-mediated Relish expression represses JNK activity in the niche. Therefore, in the absence of Rel, there is an ectopic activation of JNK signaling (Figure 5A-C). Thus, it can be speculated that an exogenous supply of ecdysone will increase the Relish level in the niche, strengthening the inhibitory arm on JNK signaling. Based on the existing literature (Karim et al.,1991, A.J Andres A.J and Thummel C S, 1994 ), we incubated lymph glands in 20E (5uM). A slight increase in Rel expression was consistently seen compared to mock treatment in these ex vivo experiments (See below, Figure 2 A-D). In sync with our result, JNK activation within the niche was not seen even in this scenario. This observation further

endorsed our claim that Rel expression can prevent JNK activation (See Author response image 2E-F).

Author response image 2
A-C': Post 20E incubation, slight increase in Relish expression was observed compared (B-C') to mock incubated samples (A-A').

D. Statistical analysis of the data from A-B' (n=24 P-value=6.19 x10-7; two tailed Students t-test). E-F. JNK expression remained unaltered in 20E incubated (F) and mock (E) incubated samples.

3. Related to #2, the authors proposed that proteasomal degradation (Factor X) will downregulate Relish during bacterial infection. Have the authors verified that the other components in the pathway upon infection? Or is it already known? Is overexpression of UAS–Relish_WT sufficient to block the precocious differentiation of lymph gland cells upon infection? Does ecdysone level change during infection? Without showing direct evidence, lines 475–489 should be moved to the discussion, and the results need to be toned down.

In the uninfected scenario, overexpressing Relish in the niche decreased niche cell number and progenitor differentiation (Figure 1 figure supplement 1F-H and Figure 6 figure supplement 2C and E). However, during infection, Relish gets degraded even in overexpression scenario compared to sham (Figure 7H-H'). Therefore, we can conclude that post-transcriptional regulation is affecting Relish stability during infection.

Based on EcR common expression (validated reporter for Ecdysone activity, Hogness DS et al.,1993, Segraves WA et al.,1999, Matunis EL et al.,2014), it is evident that the level of Ecdysone signaling remains unaltered during infection. However, subsequent loss of Relish expression points out to a regulation evoked during infection to override the developmental input (See Author response image 3A-C).

Author response image 3
A-B: No significant change in EcR expression in the niche was observed upon infection and sham.

C. Statistical analysis of the data from A-B (n= 12 P-value=.364).

Reviewer #2:

In this manuscript the authors investigate the role of Relish in the Drosophila lymph gland (LG). They establish that relish is expressed in PSC cells and that reducing its expression in these cells (by expressing relish RNAi with a PSC–gal4 driver) leads to an enlarged PSC, increased plasmatocyte differentiation, no effect on crystal cell numbers, and fewer progenitors in the medullary zone (MZ). In the PSC, Relish controls Wingless levels that in turn control PSC cell proliferation and thus PSC size. This study also establishes that the knock down of relish in the PSC leads to increased levels of several actin binding proteins, reduced filopodia formation in PSC cells and a decrease in Hh (HhExt) release from the PSC. In addition, relish knock–down in the PSC leads to the activation of the JNK pathway in the PSC. Epistasis experiments establish that JNK acts downstream of Relish to control filopodia formation and HhExt. Under normal conditions, Relish levels in the PSC are under the control of ecdysone. Finally, in response to an E. coli infection, a decrease in Relish levels in the PSC is observed together with increased plasmatocyte differentiation.

This is an important study describing a yet unknown regulation of Drosophila LG hematopoiesis.

Thanks so much for your comments and appreciation. We are happy to note that we have been able to convey our findings to you.

However, I have some concerns with the current version of the manuscript.

In the introduction: the authors do not state that the role of the PSC in the LG under normal conditions is under debate, and papers relative to this, even if they diverge from the dogma, must be cited. It is established in the literature that the MZ progenitors are heterogeneous; this information together with the corresponding papers should be introduced.

We have included the relevant references in the introduction section.

In the Results section, my first concern is about the models proposed in Figure 7: many epistasis experiments are lacking, and thus at this stage of the analysis it is impossible to propose such a model (see comments below).

Thanks for your input. Request you to kindly go through the current manuscript, which provides the necessary genetic and epistatic analysis to establish Ecdysone-Rel-JNK axis in the niche. (Line No: 401-413 and Figure 6Q-U, Figure 6 figure supplement 2A-E and Figure 6 figure supplement 2K-L'' ).

Furthermore, several controls (pictures and quantifications) are lacking (see below).

We want to mention that controls and their quantitation were done in all the cases. Due to lack of space, we did not include all the controls and their quantitation in the previous version. However, as per your suggestion in the revised version, all control and quantitation have been included in the main or Supplementary panels.

The second problem concerns the relationships between relish knock–down in the PSC, the absence of filopodia and the decrease of HhExt. These are important and novel data and should be given in the main figures. However, they need to be consolidated (see comments below).

Thanks for your input. We have now shifted these figures to the Main panel (details given below in). As advised, we have now included proper quantitation of HhExtra and that of the filopodia (numbers and length). We profusely thank the Reviewer for this input, as the outcome has further consolidated our claim.

The third point relates to the role played by Relish in the PSC in response to bacterial infection. In the current version of the manuscript, the data presented are too preliminary to prove that Relish is required and to propose how it is involved in the control of LG stress hematopoiesis.

Based on the input of both of the reviewers, we have strengthened this section of our manuscript. Kindly see the response below.

Figure 1: In the PSC>rel RNAi context, the consequence on blood progenitors must be analyzed in more detail Indeed, it is possible that an increase in the mature plasmatocytes numbers results from increased CZ (cortical zone) precursor maturation, without an input from the MZ pool. A systematic analysis of MZ markers (and quantifications) should be performed in all the genetic contexts analyzed (see below for other figures). Additional markers for MZ cells, other than Ci, must be tested since Ci is part of the Hh signaling pathway which is impaired in this context. Why is the crystal cell index unchanged in the PSC >rel RNAi context? Is there lamellocyte differentiation?

Regarding MZ markers, we have now included DE-Cad/Shotgun (Shg) immunostaining followed by the quantification of the same (Figure 1H-I' and J). Regarding crystal cell number, other than few outliers there is no significant change in the number. Lamellocyte differentiation was not observed in Rel loss scenario (βPS, Figure1 figure supplement 1I-J'). Thus, the ectopic differentiation seen upon Rel loss from niche is biased towards plasmatocyte fate.

Figure 3 : What about the contribution of PSC Wg in the control of LG homeostasis? What about LG homeostasis (MZ and differentiated blood cells) under conditions where the PSC size is rescued (i.e.: PSC>rel RNAi, Wgts)? In Figure 3 D–G, the control Wgts (picture and quantification) is missing.

a. Sinenko et al., 2010 has shown that downregulation of Wg signaling (by overexpression of Fz2DN) from the niche caused reduction in the progenitor pool. In sync with this finding, our result (employing UAS-wg-RNAi as well as a temperature-sensitive allele of wg) shows a decline in progenitor number (Shg, Figure 3I-M) and increased differentiation (P1, Figure 3 figure supplement 1B-F).

b. Curtailing wingless signaling in conjunction with the Rel downregulation from the niche rescues hyperproliferative niche but not the precocious differentiation of the progenitors (Figure 3 figure supplement 1G-J, L-P, and Q-U). Since maintenance of the progenitors is also dependent on Hh signaling from the niche, which is majorly affected upon Rel loss, is the reason for the failure in progenitor maintenance (Figure 3I-M and Figure 3 figure supplement 1L-P).

c. We have included the picture and the quantitation (Figure 3E-E' and H).

Figure 3B–B': Why is Wg staining stronger in PSC cells that express lower levels of GFP?

All niche cells do not express similar levels of Wingless protein.

Figure 3 sup1 E: Is there a significant difference in PSC cell numbers between antp>wgRNAi and control LGs? This should be indicated. In Figure 3 sup 1 F–I: What about crystal cell differentiation and MZ progenitors in PSC>wgRNAi and PSC>WgRNAi, relRNAi?

There was a significant decrease in niche cell proliferation in wgts (Figure 3E-E' and H), while AntpGal4>UASwgRNAi was almost comparable to control (Figure 3 figure supplement 1 I and K). However, upon co-expression with UAS-RelRNAi, the niche number was significantly restored (Figure 3G-G' and H). Interestingly, downregulating wg function in Rel loss genetic background could not rescue the ectopic differentiation of progenitors (Shg, Figure 3 figure supplement 1O and P, Figure 3 figure supplement 1T and U). The number of progenitors as reflected by Shg in this genotype is less compared to control.

Upon loss of wg from the niche, there was a slight decrease in crystal cell index; however, in conjunction with Rel loss from the niche, no significant difference with the control was observed (see Author response image 4).

Author response image 4
A-D.

Slight decrease in Crystal cell index was observed in wg-RNAi (C) compared to control (C) whereas no significant change was observed in Rel RNAi(KK) (B) and Rel RNAi(KK), wg RNAi (D). E. Statistical analysis of the data from A-D (n=11 P–value=3.4x10-2 for control versus wg-RNAi).

The control i.e. PSC>wg RNAi (both pictures and quantifications) is missing. Is there any difference in LG size in this context as previously reported by Sinenko et al., 09?

Loss of wg function through wgts, we do find a significant decrease in the size of the lymph gland as compared to control (see Author response image 5). Please note that we have used a temperature sensitive allele of wg whereas Sinenko et al., had downregulated wg function by overexpressing Dfz2DN from the PSC.

Author response image 5
Significant decrease in LG area was observed in wgts compared to control (n=10, P-value = 2.

3x10-4 two tailed students t-test).

Figure 4 : Are there LG defects when the levels of actin binding proteins (F actin, singed,..) are modified in the PSC ? Is there a functional relationship between Relish and actin binding protein levels in the PSC?

Upon niche-specific downregulation of Diaphanous (an actin polymerase), a significant increase in the differentiation compared to control (Figure 4 figure supplement 1A-B' and C) is observed.

Based on our genetic data, we would like to infer that the actin remodeling observed in Rel loss is JNK dependent.

Figure 4 A–B : Is Hh expressed in a subset of CZ cells? Is there basal expression of Hh also in the MZ (see A' and B")?

CZ specific expression of Hh maps to the subset of crystal cells since it colocalizes with Lozenge (a validated marker for crystal cell). The basal level is the extracellular Hh diffused from the source (niche) and is sensed by its receptor patched expressed in the MZ cells.

Author response image 6
Crystal cells marked by Lz>GFP also expresses Hh (red).

Figure4 sup1 A–E: Data relative to filopodia formation and HhExt must to be given in the main figures since they are important results. However, they still need to be consolidated. Quantifications for HhExt (quantity, dispersion) and filopodia (numbers and size) are necessary.

Thanks you for your suggestions. We do understand your point. We have done the necessary inclusions. Please refer to Main text: Line No 254-271.

We have also provided the Intensity analysis of HhExt Figure 4F, and quantitative analysis of filopodial number and length in the current version (Figure 4J-K).

Pictures A–A": Is HhExt, as shown with a comet–like pattern outside the PSC, always observed in controls, or is it specific to this picture?

Sorry for the confusion created. The comet-like pattern is generated due to the inclusion of the multiple stacks. We have now processed the few confocal stacks of the same image to get a better representation. Please look at the new figure (Figure 4 D-D'') in the revised manuscript. For your reference, the old and the new are shown below.

Pictures C–C': These is a problem here: the cells that are shown become detached from the PSC and migrate. In such conditions there are more cytoplasmic extensions that reflect the migrating status of the cells. Therefore this is not a correct illustration for PSC cell filopodia.

We want to draw the Reviewer's attention towards the fact that the imaging has been done on live tissues to preserve and document the dynamic and fragile filopodial extensions. As soon as we put the coverslip in live conditions, the pressure created tends to detach the niche cells from each other in the control lymph glands. However, the niche cells upon Relish loss remain attached to each other. This observation can be explained by our results, which demonstrate that loss of Rel from the niche cells increases cortical actin accumulation, forcing the cells to remain adhered to each other.

Pictures F–F' '(as well as in Figure 4G–L): There are discrepancies between pictures and quantifications (a 6 fold increase in the PSC>rel RNAi context compared to the control is not observed in the pictures).

We want to draw the Reviewer's attention towards the fact that the imaging has been done on live tissues to preserve and document the dynamic and fragile filopodial extensions. As soon as we put the coverslip in live conditions, the pressure created tends to detach the niche cells from each other in the control lymph glands. However, the niche cells upon Relish loss remain attached to each other. This observation can be explained by our results, which demonstrate that loss of Rel from the niche cells increases cortical actin accumulation, forcing the cells to remain adhered to each other.

Figure 5 H–L: the control, corresponding to PSC> bskDN (pictures and quantifications) is missing. Figure F5 O–Q, the control, tak1/+ (pictures and quantifications) is missing.

In the previous version, we had not included these pictures due to scarcity of space in the respective panels. In the current version, we have added few more supplementary figures to accommodate the figures into main panel. In the current manuscript, PSC>bskDN and tak1/+ images and quantitation has been included in Figure 5 (Figure 5F-F' and H for bskDN and Figure 5U and W for tak1/+).

Figure 5 L–L', should be replaced (blurry picture and migrating PSC cells).

This is a rescue experiment where the downregulation of excess cortical actin accumulation has enabled normal niche cell behavior. Kindly refer to the previous response towards migrating PSC.

What about MZ and crystal cell differentiation in these different conditions ( i.e: UAS–bskDN, UAS bskDN, Rel RNAi; tak1/+ and tak1 , rel RNAi)?

We have included the status of MZ in these two genotypes in the Figure 5 figure supplement 2. As evident from the figures, tak1 (classical loss of function) and JNK loss from the niche has a subtle effect on the progenitor number assayed by Shg.

Since Rel loss from the niche does not affect crystal cell index, this aspect has no direct relationship with the data presented. However, we made an effort to address this keeping your suggestion in mind. The results related to crystal cell index are presented below. Both tak1 (classical loss of function, Author response image 7) and bsk loss (Author response image 8) from the niche slightly decreases the crystal cell number.

Author response image 7
Slight decrease in Crystal cell index was observed in tak12 (C) and tak12; Rel RNAi (D) compared to control (A) whereas no significant change was observed in Rel RNAi (B).

E. Statistical analysis of the data from A-D (n=10 P-value=1.5 x10-2 for control versus tak12 and P-value=7.4x10-2 control versus tak12; Rel RNAi, two tailed students t-test).

Author response image 8
Slight decrease in Crystal cell index was observed in bskDN (C) and bskDN; Rel RNAi (D) compared to control (A) whereas no significant change was observed in Rel RNAi (B).

E. Statistical analysis of the data from A-D (n=10 P-value=8.5 x10-2 for control versus bskDN and P-value=6.5x10-2 for control versus bskDN; Rel RNAi, two tailed students t-test).

–Epistasis between the JNK pathway and Hh, and the JNK pathway and wg must be analysed. This is essential to support the models given in Figure 7.

We have done the epistatic analysis for Wingless and JNK. We have down-regulated wg function in UAS-hep genetic background and found a rescue in niche proliferation (Figure 5 figure supplement 1J-N).

Our earlier genetic analysis has established that JNK directly affects actin accumulation in the niche (UAShepact and Ectopic activation of JNK upon Rel loss from the niche). The outcome of upregulated actin leads to trapping of Hh within the niche, thereby causing progenitors to undergo precocious differentiation.

Figure 6: The demonstration that Ecdysone controls relish transcription in the PSC is lacking Analysis of Relish protein levels is not enough, since post transitional regulation has been described for Relish. relish transcription must be analyzed in PSC>EcRDN LGs. Furthermore, epistasis experiments are required to firmly establish a functional link between Relish and ecdysone.

We have done rescue experiments in which we over-expressed Relish in EcR loss genetic background and found a rescue in niche proliferation (Figure 6 Q-U) and differentiation defects (Figure 6 figure supplement 2A-E) associated with EcR loss. Further, we also did the whole mount IF along with FISH and found that compared to control; there is a reduction in Rel transcripts level when EcR is downregulated from the niche (Figure 6 figure supplement 2K-L'').

Thanks for your input. With newly included epistatic analysis and dual antibody-FISH results, have strengthened our manuscript to a large extent.

Figure 7: The link between Relish in the PSC and its role in LG hematopoiesis in response to bacterial infection is not yet clearly assessed. That an increase in blood cell differentiation occurs in the LG in response to bacterial infection has already been described (Khadilkar et al;, 2017). This reference should be cited in the manuscript. Whether the bacteria–induced decrease of Relish in the PSC is involved in the LG stress hematopoiesis is not established here since there are many discrepancies between LG phenotypes in PSC>rel RNAi larvae and in response to bacterial infection and rescue experiments are lacking. While in both cases there is an increase in plasmatocyte differentiation, there is no difference in PSC size in response to bacterial infection (which differs from PSC>rel RNAi LGs), and there is no LG disruption in PSC>rel RNAi larvae (which differs from bacterial infection). Furthermore, the LG disruption in response to bacterial infection is a novel data, but unfortunately it is not properly documented in this manuscript. Finally, in bacterial infection experiments, no data relative to a role of Hh or JNK pathway in PSC cells have ever been described so far.

Thanks for your inputs. We have included the references in the main text of the current version.

However, we disagree with the statement that there are discrepancies in the phenotype of Rel loss data compared to infection. Kindly note the following reasons:

1. Similar to loss of Rel from the Niche, JNK signaling is evoked during infection (please refer to Figure 7 figure supplement 1A-B').

2. In context to niche cell number, the lymph glands in the infection experiments were analyzed 4 hours post insult, a short time window to expect a similar degree of proliferation compared to that of Rel knockdown from the niche. However, even in this short window of insult, we encounter a significant and consistent increase in the niche cell numbers in infected samples compared to mock (Figure 7 figure supplement 1C-E).

3. In infected samples, we do see lymph gland disruption beyond 4 hours. Upon loss of Rel from the niche, we see a similar response in few cases where there is extreme differentiation accompanied by peeling off the LG around 96 AEH (Please see Author response image 9). Since most LGs were intact at 96AEH, we did not include the LG with peeled off phenotype as the representative image in the current manuscript. However, we are open to suggestions from your end and, if required, will include this in the current version.

Author response image 9
A-D.

Compared to control (A) ectopic differentiation and peeling off of lymph glands was observed in infected (B) as well as Rel loss samples (C).

Is the transcription of relish or the stability of the protein that is affected in response to bacterial infection? In Figure 7 D: >UAS –Rel 68KD, without infection, is Relish protein detected by immunostaining in PSC cells? This would clarify whether Relish protein stability is dependent on bacterial infection.

Thank you for your suggestion. Antp-GFP >UAS-Rel68KD was capable of enriching Rel expression in the niche (See results in Author response image 10). However, on infecting the same genotype, we encountered a drop in Rel expression Please refer to Figure 7H-H' of the current manuscript. Thus, we can infer that forced expression of Rel in the niche was not capable of withstanding the immune challenge. These results indicate that the stability of the protein is at stake during infection

Author response image 10
A-B: Rel expression in the niche was observed in control as well as in UAS-Rel68KD tissues.

Figure 7 B–D: Both quantifications of PSC cell numbers and Relish levels are missing.

Thanks for your suggestions. We have included these missing information’s in the current version.

The niche cell proliferation data (Figure 7 figure supplement 1C-E) and Relish levels (Figure 7A-D) has been included in the modified manuscript.

Figure 7 H–J': Plasmatocyte quantification is missing. Analyses and quantifications of crystal cells and MZ progenitors must also be provided.

We have included the quantitation of progenitors and plasmatocytes in the current version. Compared to control, there was significant decrease in DE-Cad (Shg) positive progenitor cells (Figure 7I-K) and a concomitant increase in the plasmatocyte differentiation (Figure 7L-N) was observed in infected lymph gland compared to mock. The crystal cell index is provided in Author response image 11. As previously described (Khadilkar et al., 2017), we also see an increase in their number upon infection.

Author response image 11
Significant increase in crystal cell index was observed in infected samples compared to sham (n=9, P-value =1.

8x10-3, two tailed students t-test).

Models in Figure A and K are not correct, since many epistasis experiments are missing.

Furthermore, an arrow between Wg in the PSC and MZ progenitors is missing, since it has been previously established that the Wg pathway in the PSC controls LG hemocyte differentiation (please see Sinenko et al., 2009).

We have enriched our findings with multiple epistatic analyses and did the necessary modifications in the current version.

Figure 8 : Are there any functional links between filopodia, the actin skeleton and Hh signaling in the LG ? This point should be further addressed.

As mentioned in the manuscript, there are multiple reports regarding the role played by cellular cytoneme like filopodia in Hh delivery from the niche to progenitor cells. In addition to our data, works from the laboratory of Prof Schulz (Tokusumi et al., 2012, Tokusumi et al., 2010) have demonstrated the functional link between filopodia and Hh delivery in the LG.

For the PSC>rel RNAi context, both in the summary and in point e in the legend Figure 8, we read "precocious differentiation of the blood progenitors". This point should be better established. Indeed, systematic analyses of MZ markers in the different genetic contexts are required (see comments above).

We have done Shg immunostaining and found a significant decrease in the number of progenitor cells compared to the control (Figure 1H-J). Additionally, integrin β-PS, another progenitor marker expression was significantly lower (Figure 1 figure supplement 1I-J') in loss of Rel loss from the niche compared to control.

[Editors’ note: what follows is the authors’ response to the second round of review.]

Essential revisions:

General comments by the editor following the discussion with the reviewers

Both reviewers feel that this paper warrants publication in eLife.However , they think that additional new data are needed to fully support the conclusion on the role of PSC fillopodia in signaling. At the very least, the authors must establish whether there is a functional relationship between Relish and actin binding protein levels in the PSC. Rescue experiments of relish knock down, by decreasing "actin protein" and looking at lymph gland homeostasis, fillopodia formation and Hh diffusion, have to be provided.

No functional relationship has been established for the role of Relish in the PSC in response to bacterial infection Thus, the authors should be more cautious, and both in the manuscript and the summary the text relative to this point has to be modulated. Furthermore, the authors should discuss the discrepancy between their model, where in response to bacterial infection Hh is trapped within the niche and unable to diffuse, versus the model proposed by (Khadilkar et al., 2017) where bacterial infection causes the breakdown of the permeability barrier in PSC cells.

To minimize the second round of revision, the reviewers suggested that the authors could also tone down the statements on the filopodia–Hh interaction and the role of Rel in the bacterial infection while adding possible additional works.

Essential revisions:

1) As I already suggested in the previous comment, the authors should impair PSC fillopodia formation by modifying actin binding proteins (F actin, singed,dia,…) and analysing both lymph gland defects and Hh diffusion.

We have downregulated dia from the PSC/niche, and the lymph gland defects observed has been included in the revised version Figure 4 figure supplement 1. Since the filopodial length and number were significantly compromised (A-D), we observed a defect in the transport of extracellular Hh (E-G). As a consequence we see an increased differentiation (J-K and M) and decline in progenitors index (H-I in L). We have also included this in the main text of the revised version. Please refer to page 12 and Line 276-284.

It is also necessary to establish whether there is a functional relationship between Relish and actin binding protein levels in the PSC. Rescue experiments of relish knock down, by decreasing "actin protein" and looking at lymph gland defect, fillopodia formation and Hh diffusion are missing.

Thanks for this suggestion. Since a robust upregulation of “actin binding protein” Ena is observed upon Rel loss from the PSC, we decided to down-regulate Ena in conjunction with Rel loss from PSC to address their relationship.

Loss of both Rel and Ena from the niche led to a partial rescue in the defects related to filopodia, Hh dispersal, and differentiation, which are otherwise observed upon Rel loss. Just by modulating one “actin binding protein”, getting a partial rescue indicates a functional relationship between Rel and actin-binding proteins.

Please refer to Figure4 figure supplement 3 and Page: 13 and Lines: 296-300 in the main text of the revised manuscript.

In Figure 4 Sup1: In PSC>UAS Dia RNAi, what about fillopodia formation and Hh diffusion ? Does decreasing dia in PSC cells in relish knocked down condition restore fillopodia formation, Hh diffusion and lymph gland cell homeostasis? This is an important issue than remains to be addressed.

As mentioned above we have performed this set of experiments. Please refer to the revised Figure 4 Figure supplement 1. As expected loss of Dia from PSC affects filopodia formation. As a result both defects in Hh transport and progenitor maintenance is evident upon dia loss. Based on the result we can speculate that loss of dia along with Rel would enhance the phenotype.

Figure 4A–C: when relish is knocked down in the PSC, is Hh the increase in PSC cells due to a defect of Hh diffusion and/or a decrease in Hh transcription in those cells? This has to be clarified.

We have now looked at the transcription readout of Hh: hhF4fGFP expression, which seems to be also elevated in niches where Rel is downregulated. Please refer Author response image 12:

Author response image 12
A-B'.

Loss of Relish function from the niche resulted in upregulation of hh transcription (A- A') (marked by hh-F4f GFP) compared to control niches (B-B')C. Statistical analysis of the data provided in A-B' (n=15, P-value= 2.2x10-11, two tailed Students t-test).

2) The role of Relish in the PSC in response to bacterial infection is based on correlation, since rescue experiments are missing and thus no functional relationship has been established yet.

In order to perform a rescue experiment, a bacterial infection was performed on individuals where UAS-Rel68kD was overexpressed in the niche. However, much to our surprise, the upregulated Rel gets degraded at the protein level during bacterial infection. This itself unraveled a niche specific-regulation on Rel during infection. Therefore, the conventional rescue experiment is not possible. Though the candidate that breaks the maintenance circuit remains to be identified, nonetheless, our study illustrates that the hematopoietic niche can sense the physiological state of an animal to facilitate a transition from normal to emergency hematopoiesis via Rel.

Furthermore, whereas bacterial infection leads to the increase in crystal cell differentiation (Khadilkar et al., 2017), relish knock down in PSC cells has no impact on crystal cell number. Thus, the authors should be more cautious, and both in the manuscript and the summary the text relative to this point has to be modulated.

Bacterial infection is a systemic challenge that can have multiple impacts on the lymph gland. One such impact is what we have shown through our study where a septic injury resulted in the loss of Relish from the hematopoietic niche. Hence, knocking down the function of just one gene (in our case Relish) cannot mimic all the phenotypes associated with the systemic immune challenge.

Moreover, we also speculate that in Relish loss from the niche, there can be loss of multiple positive as well as negative regulators involved in crystal cell differentiation, thereby dampening individual outputs.

Finally, the authors should discuss the discrepancy between their model where in response to bacterial infection Hh is trapped within the niche and unable to diffuse, versus the model proposed by (Khadilkar et al., 2017) where the bacterial infection causes the breakdown of the permeability barrier in PSC cells.

We want to draw the reviewer's attention towards the fact that comparison with the mentioned paper is not possible. This is because Khadilkar et al. has not assayed the status of Hh in their case; instead, their case emulates a niche that has decanted signals that can be both of maintenance and differentiation.

While in our case, cytoskeletal rearrangements in the niche prevent the dispersion of Hh or the dispersion of signals from the niche. We show that the trapped Hh is not sensed by the progenitors, which fails to maintain themselves and thus differentiates.

In the cited paper, seeping off too many differentiation signals (Please refer to the Discussion section of Khadilkar et al.) leads to ectopic differentiation.

If maintenance and differentiation are both sides of a coin: Khadlikar et al., describes scenario elevated differentiation signaling that pushes progenitor towards differentiation. On the other hand, we illustrate how the failure of maintenance can also be a reason for differentiation. Moreover, at this time point, we want to draw your attention to that signals for differentiation are not only elicited from Niche, the progenitors also generate them. As advised by the reviewer, we have discussed Khadilkar et al. findings in the Discussion section of the revised manuscript.

Please refer to page: 22 and Lines: 530-537.

https://doi.org/10.7554/eLife.67158.sa2

Article and author information

Author details

  1. Parvathy Ramesh

    1. Department of Biological Sciences, Indian Institute of Science Education and Research (IISER) Mohali, Knowledge City, India
    2. Developmental Genetics Laboratory, IISER Mohali, SAS Nagar, Punjab, India
    Contribution
    Conceptualization, Data curation, Formal analysis, Validation, Investigation, Visualization, Methodology, Writing - original draft, Writing - review and editing
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0001-7273-5848
  2. Nidhi Sharma Dey

    1. Department of Biological Sciences, Indian Institute of Science Education and Research (IISER) Mohali, Knowledge City, India
    2. Developmental Genetics Laboratory, IISER Mohali, SAS Nagar, Punjab, India
    Present address
    York Biomedical Research Institute, Hull York Medical School, University of York, York, United Kingdom
    Contribution
    Formal analysis, Validation, Investigation, Visualization, Methodology
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0001-9432-0221
  3. Aditya Kanwal

    1. Department of Biological Sciences, Indian Institute of Science Education and Research (IISER) Mohali, Knowledge City, India
    2. Developmental Genetics Laboratory, IISER Mohali, SAS Nagar, Punjab, India
    Contribution
    Conceptualization, Validation, Investigation, Visualization, Methodology, Writing - original draft
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0002-8066-6798
  4. Sudip Mandal

    1. Department of Biological Sciences, Indian Institute of Science Education and Research (IISER) Mohali, Knowledge City, India
    2. Molecular Cell and Developmental Biology Laboratory, IISER Mohali, SAS Nagar, Punjab, India
    Contribution
    Formal analysis, Writing - original draft, Writing - review and editing
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0002-2211-483X
  5. Lolitika Mandal

    1. Department of Biological Sciences, Indian Institute of Science Education and Research (IISER) Mohali, Knowledge City, India
    2. Developmental Genetics Laboratory, IISER Mohali, SAS Nagar, Punjab, India
    Contribution
    Conceptualization, Data curation, Formal analysis, Supervision, Funding acquisition, Validation, Investigation, Visualization, Methodology, Writing - original draft, Project administration, Writing - review and editing
    For correspondence
    lolitika@iisermohali.ac.in
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0002-7711-6090

Funding

The Wellcome Trust DBT India Alliance (IA/S/17/503100)

  • Lolitika Mandal

The funders had no role in study design, data collection and interpretation, or the decision to submit the work for publication.

Acknowledgements

We thank B Lemaitre, N Silverman, I Ando, U Banerjee, S Cohen, P Ingham, M Crozatier, A D Shirras, R Schulz and NG Prasad for reagents. Special thanks to Sushmit Ghosh and Kaustuv Ghosh for assisting with FISH and germ-free experiments. We thank all members of the two laboratories for their valuable inputs. We thank IISER Mohali’s Confocal Facility, Bloomington Drosophila Stock Center, at Indiana University, VDRC (Vienna) and Developmental Studies Hybridoma Bank, University of Iowa for flies and antibodies. Models ‘Created with BioRender.com’. DBT Wellcome-Trust India Alliance Senior Fellowship [IA/S/17/1/503100] to LM and Institutional support to PR, NSD, SM, and DBT-fellowship funding to AK for this study duly acknowledged.

Senior Editor

  1. Utpal Banerjee, University of California, Los Angeles, United States

Reviewing Editor

  1. Bruno Lemaître, École Polytechnique Fédérale de Lausanne, Switzerland

Publication history

  1. Received: February 2, 2021
  2. Preprint posted: April 5, 2021 (view preprint)
  3. Accepted: July 14, 2021
  4. Accepted Manuscript published: July 22, 2021 (version 1)
  5. Version of Record published: August 13, 2021 (version 2)

Copyright

© 2021, Ramesh et al.

This article is distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use and redistribution provided that the original author and source are credited.

Metrics

  • 1,650
    Page views
  • 202
    Downloads
  • 0
    Citations

Article citation count generated by polling the highest count across the following sources: Crossref, PubMed Central, Scopus.

Download links

A two-part list of links to download the article, or parts of the article, in various formats.

Downloads (link to download the article as PDF)

Download citations (links to download the citations from this article in formats compatible with various reference manager tools)

Open citations (links to open the citations from this article in various online reference manager services)