Noncanonical Decapentaplegic Signaling Activates Matrix Metalloproteinase 1 To Restrict Hedgehog Activity and Limit Ectopic Eye Differentiation in Drosophila

Abstract

One of the pertinent issues associated with cellular plasticity is to understand how the delicate balance between the determined state of cells and the extent to which they can transdetermine is maintained. Employing the well-established model of generating ectopic eyes in developing wing discs of Drosophila by ectopic eyeless expression, we provide evidence for the genetic basis of this mechanism. By both loss-of-function and gain-of-function genetic analyses, we demonstrate that Matrix metalloproteinase 1 (Mmp1) plays an important role in regulating the extent of ectopic ommatidial differentiation. Transcriptional activation of ectopic Mmp1 by the morphogen Decapentaplegic (Dpp) is not triggered by its canonical signaling pathway which involves Mad. Rather, Dpp activates an alternate cascade involving dTak1 and JNK, to induce ectopic Mmp1 expression. Mutational analyses reveal that Mmp1 negatively regulates ectopic eye differentiation by restricting the rate of proliferation and the levels of expression of retinal-determining genes dachshund and eyes absent. This is primarily achieved by restricting the range of Hedgehog (Hh) signaling. Importantly, the increase in proliferation and upregulation of target retinal-determining genes, as observed upon attenuating Mmp1 activity, gets significantly rescued when ectopic eyes are generated in wing discs of hh heterozygous mutants. In conjunction with the previously established instructive and permissive roles of Dpp in facilitating ectopic eye differentiation in wing discs, the outcome of this study sheds light on a mechanism by which Dpp plays a dual role in modulating the delicate balance between the determined state of cells and the extent they can transdetermine.

IMAGINAL discs are sacs of epithelial cells present within the larvae of holometabolous insects that serve as anlagen for adult appendages such as wing, eye, antennae, genitalia, leg, etc. The study of imaginal discs in the fruit fly Drosophila melanogaster has contributed significantly to our understanding of the role of signaling molecules like morphogens in cell fate specification and pattern formation. For instance, the precise coordination between Hedgehog (Hh) and Decapentaplegic (Dpp) signaling pathways in regulating pattern formation in developing eye imaginal discs of Drosophila highlights the role of these morphogens in cell fate induction. Hh, secreted from the differentiating ommatidial clusters, is important not only in regulating cell growth and proliferation by promoting expression of Cyclin E and Cyclin D (Duman-Scheel et al. 2002), but is also essential for initiating photoreceptor differentiation by triggering the expression of the retinal-determining (RD) genes, eyes absent (eya) and dachshund (dac) (Curtiss and Mlodzik 2000; Pappu et al. 2003; Firth and Baker 2009). Furthermore, Hh triggers the expression of dpp along the morphogenetic furrow (Heberlein et al. 1993) which induces the precursors to enter a preproneural state (Greenwood and Struhl 1999). In the developing wing discs, on the other hand, Hh signaling from cells of the posterior compartment drives the expression of dpp in a stripe of anterior cells along the anterior/posterior (A/P) compartment boundary (Masucci et al. 1990; Basler and Struhl 1994). The Dpp protein in turn acts as a long-range morphogen to play a pivotal role in regulating growth and patterning of wing disc (Tanimoto et al. 2000; Affolter and Basler 2007).

Drosophila imaginal discs have also served as an excellent experimental model to understand several mechanisms underpinning cell-fate alteration. Regeneration studies, which primarily involved transplanting larval imaginal disc fragments, revealed that these fragments retain their ability to differentiate into appropriate adult structures upon metamorphosis (Gehring 1966). However, in some rare instances, these disc fragments undergo alteration in cell-fate specification by a process termed as transdetermination, to ultimately differentiate into disc-inappropriate adult structures (Hadorn 1968; Schubiger 1971). Studies from different laboratories demonstrated that transdetermination could also be achieved by genetic manipulations of developing imaginal discs. For instance, ectopic expression of wingless (wg) in developing leg imaginal discs leads to conversion of leg to wing (Maves and Schubiger 1995). Likewise, ectopic eyes can be generated on different parts of adult flies by inducing the expression of RD genes eyeless (ey), eya, and twin of eyeless in the developing imaginal discs (Halder et al. 1995; Salzer and Kumar 2010).

Given the importance of morphogens during patterning in imaginal discs, it is intuitively obvious that they would also play a pivotal role during transdetermination. Indeed, several studies have demonstrated that high level of Wg expression is specifically associated with cells undergoing leg-to-wing conversion during regeneration in surgically fragmented leg discs (Gibson and Schubiger 1999). Interestingly, during this process, ectopically expressed Wg acts synergistically with higher levels of Dpp to support leg-to-wing transdetermination (Maves and Schubiger 1998; Ing et al. 2013). Likewise, ectopically expressed ey can induce ectopic eye formation only when it collaborates with high levels of Dpp and Hh (Kango-Singh et al. 2003). It is generally perceived that during ectopic eye induction in wing discs, these two morphogens play roles identical to that observed during normal ommatidial differentiation in eye discs. However, since both of these morphogens are equally important for normal wing development, it is not yet known whether the interaction between these two morphogen signaling genes gets reorganized when ectopic eyes are induced in wing discs.

On a separate note, transcriptome analysis of cells undergoing leg-to-wing cell-fate alteration due to ectopic expression of Wg demonstrated a fourfold enrichment in the level of Drosophila Matrix metalloproteinase1 (Mmp1) (Klebes et al. 2005). As a genetic correlate, it was subsequently demonstrated that ectopic expression of wg in leg discs of larvae heterozygous mutant for Mmp1 not only led to significant elevation in the rate of leg-to-wing transdetermination, but was also associated with ectopic transdetermination in the nonblastema cells that did not generally undergo cell-fate change (McClure et al. 2008). Importantly, this phenomenon of transdetermination in multiple sites within the Mmp1 heterozygous mutant discs was attributed to the failure of these cells in submitting to cell-cycle arrest and a plausible increase in the level of Dpp signaling. Though these findings clearly implicated the importance of Mmp1 in regenerative proliferation and cellular plasticity, our understanding of the precise role played by Mmp1 in regulating the process of cell-fate alteration still remains elusive. Likewise, the signaling pathway that triggers ectopic expression of Mmp1 during transdetermination is yet to be elucidated.

Matrix metalloproteinases (Mmps) belong to the family of zinc-dependent extracellular endopeptidases responsible for tissue remodeling and degradation of extracellular matrix (ECM) across divergent phyla (Page-McCaw et al. 2007; Page-McCaw 2008; Marino and Funk 2012). The genome of Drosophila codes for two Mmp proteins, Mmp1 and Mmp2 (Llano et al. 2000, 2002), and one tissue inhibitor of metalloprotease (TIMP) (Wei et al. 2003). While Mmp1 is secretory in nature (Page-McCaw et al. 2003), Mmp2 has a predicted GPI membrane anchor (Llano et al. 2002). Importantly, Drosophila TIMP (dTIMP) can inhibit mammalian MMP function (Wei et al. 2003) and mammalian TIMPs can inhibit fly Mmps (Llano et al. 2000), thereby demonstrating that their functions are evolutionarily conserved. Although embryos double mutant for both Mmp1 and Mmp2 are viable (Page-McCaw et al. 2003), Mmp2 plays a crucial role in axon guidance of motor neurons during embryonic development (Miller et al. 2011). The role of both the Mmps, rather, seems to be more important during later stages of development in processes that demand tissue remodeling (Page-McCaw et al. 2003). While the activities of both the Mmps are required for dendritic remodeling (Kuo et al. 2005), as well as for dissociation of fat bodies during metamorphosis (Jia et al. 2014), they are also independently and reciprocally required for synaptogenesis along neuromuscular junctions (Dear et al. 2016). In addition, Drosophila Mmps also contribute significantly to reepithelization during wound healing (Stevens and Page-McCaw 2012), as well as to tumor invasiveness in flies (Uhlirova and Bohmann 2006; Beaucher et al. 2007). Since it has been evidenced that mammalian Mmps can actively participate in modifying signaling pathways by modulating the ECM proteins (Yu and Stamenkovic 2000; Hamano et al. 2003; Lee et al. 2005; Zhang et al. 2006), we sought to determine whether ectopic expression of Mmp in transdetermining tissue plays any role in controlling morphogen activity.

To address this issue, we resorted to in vivo genetic analyses of Drosophila larval wing discs undergoing ectopic eye differentiation due to induced expression of the gene ey. Our results show that the morphogen, Dpp, triggers the expression of Mmp1 adjacent to the site of ectopic photoreceptor differentiation. However, instead of the canonical signaling pathway, Dpp employs an alternate signaling cascade involving Tak1 and JNK to activate Mmp1. Employing in vivo loss-of-function [both RNA interference (RNAi) and classical mutant] genetic analyses, we establish that enrichment of Mmp1 around the cells producing Hh is actually instrumental in restricting the range of Hh signaling into the zone of ectopic photoreceptor differentiation. In doing so, Mmp1 modulates the rate of proliferation and activation of the genes eya and dac, two members of the RD gene network, to limit ectopic photoreceptor differentiation.

Materials and Methods

Drosophila stocks and fly husbandry

All stocks were reared on regular corn-flour medium and experiments were carried out at 25°. For details of the fly stocks used, their sources, and the genotypes of the lines generated for this study, please refer to Supplemental Material, Table S1.

Generation of somatic clones

To generate somatic clones of Mmp1², hs-flp/hs-flp; FRT42D, Ubi-GFP, M flies were crossed to FRT42D, Mmp1² flies. Synchronized hatched larvae were administered heat-shock pulses at 37° during the second larval instar. All the larvae were finally reared at 25° till analyses. For generating control mock clones, FRT42D flies were crossed to hs-flp/hs-flp; FRT42D, Ubi-GFP, M flies and a similar procedure as mentioned above was followed.

Immunohistochemistry and confocal microscopy

All immunostaining experiments were performed at least three times (n = 15 per experiment) following the standard protocol. The primary antibodies rat anti-ELAV (7E8A10; 1:100), mouse anti-Dac (mAbdac2-3; 1:20), mouse anti-En (4D9; 1:2), mouse anti-Eya (10H6; 1:2), rat anti-Ci (2A1; 1:3), and anti-Mmp1 (14A3D2, 3A6B4, 3B8D12, 5H7B11 used in a 1:1:1:1 dilution) were obtained from the Developmental Study Hybridoma Bank, Iowa, and anti-β-galactosidase (Z3781, 1:100) was obtained from Promega, Madison, WI. The following secondary antibodies for fluorescent detection were used at a 1:600 dilution: Cy^TM3-conjugated goat anti-mouse, FITC-conjugated goat anti-mouse, FITC-conjugated donkey anti-rat (Jackson ImmunoResearch Laboratories), and Phalloidin Texas Red (Molecular Probes, Eugene, OR). Immunofluorescence images were captured using a Laser Scanning Confocal Microscope (LSM 780; Carl Zeiss, Thornwood, NY) and processed using ImageJ and Adobe Photoshop.

Cryosectioning of imaginal discs

After secondary antibody immunostaining, tissues were refixed with 4% formaldehyde for 30 min, washed, and incubated in 30% sucrose solution (prepared in 1× PBS) overnight at 4°. Wing imaginal discs were oriented in Tissue-Plus O.C.T (optimal cutting temperature) solution (no. 4585; Fisher Healthcare), frozen, and cut into sections of 25 µm by using a Leica Cryotome (CM3050) instrument. Sections were washed carefully with 1× PBS on glass slides, mounted in VECTASHIELD, and imaged.

Quantification of fluorescence intensity

Fluorescence intensity of expression was quantified in terms of Gray values of Dac- and Eya-expressing areas by using ImageJ. Each experiment was repeated three times along with an appropriate control and imaged at similar parameter settings. Ci activity was quantified in terms of pixel intensities within fixed area of wing discs imaged at the same settings by using the plot profile function of ImageJ. The average of the pixel intensities of five imaginal discs was taken to calculate the final Ci intensity profile.

Area measurement

We used Imaris x64 and ImageJ for calculating the area of Dac- and Eya-expressing cells in the wing discs. For each case the calculated areas were normalized with the total disc area. The results shown here are the outcomes of three independent experiments. The ommatidial number was also calculated using Imaris x64.

5-Ethynyl-2′-deoxyuridine incorporation assay

The 5-ethynyl-2′-deoxyuridine (EdU) incorporation assay was performed by using the Click-iT EdU Alexa Fluor 594 Imaging Kit (Molecular Probes). The imaginal discs were incubated in EdU for 40 min. After completion of the procedure, the discs were processed for cryosectioning and imaging. The numbers of EdU-positive cells were calculated in the Dac-expressing domain and represented as number of EdU-positive cells per unit area of Dac expression.

Quantitative analysis of transcripts

Transcript levels of Mmp1, hh, and kayak were analyzed by quantitative RT-PCR (qRT-PCR) on RNA isolated from 50 to 60 wing discs. In all cases, expression levels of these transcripts were normalized with respect to rp49 expression. Each experiment was repeated three times, using triplicates each time. The list of primers used is as follows:

• rp49-forward, 5′-CTAAGCTGTCGCACAAATGGC-3′.

• rp49-reverse, 3′-TTCTGCATGAGCAGGACCTC-5′.

• Mmp1-forward, 5′-GCGTGTGAAGAACCTCACCT-3′.

• Mmp1-reverse, 3′-TCTCCACGAACTTGATCTCG-5′.

• hh-forward, 5′-GCTCCGTCAAGTCAGATTCG-3′.

• hh-reverse, 3′-AGGTTGCGGTCCATGAAGAG-5′.

• kayak-forward, 5′-GACCGATACTTCAAGTGCCCA-3′.

• kayak-reverse, 3′-TTGAGGTATTCGCGTTGCTG-5′.

Live imaging of cytonemes

Live imaging of cytonemes in the wing discs was performed following the standard protocol (Ramirez-Weber and Kornberg 1999) with slight modifications. In brief, third instar larval wing discs were dissected in Schneider’s Insect Medium (S0146, Sigma Chemical, St. Louis, MO) supplemented with 2.5% Phalloidin from Amanita phalloides from a stock solution of 1 mg/ml (P2141, Sigma Chemical) to stabilize the F-actin-rich cytonemes emanating from the basal side of the wing imaginal disc. The unfixed imaginal discs were mounted immediately and imaged in a Carl Zeiss LSM 780 Confocal Microscope.

Statistical analyses

Data are expressed as mean ± SD of values from three independent experiments. Statistical analysis was performed using two-tailed Student’s t-test. The following P-values were accepted as statistically significant: * P < 0.01, ** P < 0.001, and *** P < 0.0001.

Data availability

The authors state that all data necessary for confirming the conclusions presented in the article are represented fully within the article. Reagents are available upon request.

Results

Mmp1 activity is required to limit differentiation of ectopic photoreceptors in larval wing imaginal discs

For our studies, we induced ectopic eyes in the wings of Drosophila by employing the binary Gal4-upstream activation sequence (UAS) system (Brand and Perrimon 1993). We used the Gal4 driver DPP-Gal4 to ectopically drive the expression of UAS-ey in developing wing imaginal discs. Dpp being expressed in the anterior compartment of cells adjacent to the A/P boundary of larval wing imaginal discs (Figure 1A) (Blackman et al. 1991), upon driving ectopic ey expression by Dpp-Gal4, we observed ectopic differentiation of photoreceptors along the A/P axis in late third instar larval wing discs, 136 hr after egg laying (AEL) (Figure 1B).

Figure 1.

Mmp1 negatively regulates ectopic photoreceptor differentiation in wing imaginal discs. In all wing discs, anterior compartment is to the left and the genotypes are as mentioned. The nuclei are marked with DAPI (blue). (A) Schematic representation of third instar larval wing disc showing the pattern of Dpp expression. (B) Ectopically differentiated photoreceptor clusters, marked by ELAV expression, adjacent to the A/P boundary in wing discs of late third instar larvae (136 hr AEL). (C) Limited number of ectopically differentiated photoreceptors is detected in wing discs of midthird instar larvae (124 hr AEL). (D and E) Robust increase in ectopic photoreceptor differentiation is seen upon knocking down Mmp1 expression as in D UAS-ey/UAS-Mmp1i(B); Dpp-Gal4 and (E) UAS-ey/UAS-Mmp1i(kk); Dpp-Gal4 midthird instar larval (124 hr AEL) wing discs as compared to UAS-ey; Dpp-Gal4 larval wing discs of similar age (C). (F and G) Similar increase in number of ectopic photoreceptors is observed upon attenuating the Mmp1 activity in (F) UAS-ey/Mmp1^Q273; Dpp-Gal4 and (G) UAS-ey; Dpp-Gal4/UAS-Timp midthird instar larval wing discs as compared to UAS-ey; Dpp-Gal4 larval wing discs of similar age (C). (H) Overexpression of Mmp1 leads to a drastic drop in the number of ectopic photoreceptors in midthird instar larval wing discs of UAS-ey; Dpp-Gal4/UAS-Mmp1 larvae. (I) Quantification of the average number of ectopic photoreceptors in the midthird instar larval wing discs of genotypes mentioned. Mean ± SD; *** P < 0.001. (J) The increase in the number of ectopic photoreceptors as observed in Mmp1^Q273 mutant background (F) gets drastically reduced upon overexpressing Mmp1. (K) Quantitative estimate of average number of ectopic photoreceptors in the early third instar larval wing discs of genotypes mentioned (mean ± SD; ns, not significant). Bars, 20 μm.

To investigate the role of Mmp1, if any, in modulating the induction of ectopic eyes in developing wing discs, we analyzed the number of differentiating photoreceptors when ectopic eyes were generated in wing discs either with compromised levels of Mmp1 activity or having Mmp1 overexpressed. To avoid the gross morphological distortions associated with late third instar wing discs of these genotypes undergoing ectopic eye differentiation, we used wing discs from larvae that were 124 hr old AEL for our analyses. As evident from Figure 1C, at this developmental time point the extent of ectopic photoreceptor differentiation in wing discs of UAS-ey; Dpp-Gal4 larvae was much less when compared to that observed in discs of late third instar larvae (compare Figure 1C with Figure 1B). Interestingly, compared to the control (Figure 1, C and I), we observed an almost three- to fourfold increase in the number of ectopic photoreceptors upon driving two independent UAS-Mmp1-RNAi constructs, obtained from two different sources, by Dpp-Gal4 (Figure 1, D, E, and I). Likewise, generating ectopic eyes in wing discs that were heterozygous mutant for Mmp1^Q273, a hypomorphic allele of Mmp1 (Page-McCaw et al. 2003), also yielded a robust increase in the number of ectopic photoreceptors (Figure 1, F and I). In fact, the increase in number of ectopic photoreceptors observed in these discs was relatively more than that observed upon knocking down Mmp1. Overexpression of the sole dTIMP by Dpp-Gal4 also resulted in an increase in the number of ectopic photoreceptors, comparable to that observed for knocking down Mmp1 (Figure 1, G and I). In tune with these results, we observed a drastic reduction in the number of ectopic photoreceptors upon overexpressing Mmp1 by Dpp-Gal4 (Figure 1, H and I). The involvement of Mmp1 in regulating ommatidial differentiation, however, appears to be specific for ectopic photoreceptors, as normal ommatidial differentiation remained unaffected in eye discs of Mmp1^Q273 homozygous mutant larvae (Figure S1, A and B, and File S1).

To establish that the observed increase in the number of ectopic photoreceptor differentiation was due to specific loss of Mmp1 activity, we were interested to determine whether the increased number of ectopic photoreceptors as observed in Mmp1^Q273 heterozygous mutant wing discs could be rescued by overexpression of Mmp1. Indeed, driving UAS-Mmp1 in the wing discs of UAS-ey/Mmp1^Q273; Dpp-Gal4 larvae drastically reduced the number of ectopic photoreceptors to a state that was comparable to that seen upon driving UAS-ey by Dpp-Gal4 (Figure 1, J and K). Together, these results clearly established that Mmp1 activity was required to limit the extent of ectopic photoreceptor differentiation.

Dpp triggers ectopic Mmp1 expression during wing-to-eye fate alteration

Next we wanted to study the pattern of Mmp1 expression in wing discs undergoing ectopic photoreceptor differentiation. Detectible levels of Mmp1 expression in wing discs of UAS-ey; Dpp-Gal4 larvae were first observed at 94 hr AEL (Figure 2B) in contrast to wild-type wing discs of same developmental stage (Figure 2A). Closer observation revealed the presence of high levels of Mmp1 along the cell boundaries, creating a web-like pattern (Figure 2B´). As time progressed, we observed an increase in the domain of Mmp1 expression in wing discs of UAS-ey; Dpp-Gal4 larvae. By 114 hr AEL, ectopic Mmp1 expression was observed in an extensive area spanning the A/P axis of UAS-ey; Dpp-Gal4 wing discs (Figure 2, D and D′). Strikingly, the control wing disc of comparable developmental stage was devoid of any such patterned expression of Mmp1 (Figure 2C). Even third instar larval eye discs did not exhibit any patterned expression of Mmp1 (Figure S2, A and A′)

Figure 2.

Mmp1 gets ectopically expressed in wing discs undergoing ectopic photoreceptor differentiation. In all wing discs, anterior compartment is to the left and the genotypes are as mentioned. (A) Expression of Mmp1 is not observed in wing imaginal discs of early third instar larvae (94 hr AEL). (B and B′) Ectopic induction of UAS-ey by Dpp-Gal4 leads to ectopic Mmp1 expression in wing imaginal discs of early third instar larvae (94 hr AEL). (B′) represents the magnified image of the portion marked with a box in (B). (C) Basal level of Mmp1 expression in wing imaginal discs of third instar larvae (114 hr AEL). (D and D′) Intense ectopic expression of Mmp1is seen along the A/P axis in wing imaginal discs of UAS-ey; Dpp-Gal4 third instar larvae (114 hr AEL). (D′) Zoomed-in image of the portion marked with a box in (D). (E and F) Ectopic expression pattern of Mmp1 with respect to high levels of Dpp expression in third instar larval wing discs undergoing ectopic photoreceptor differentiation. While no Mmp1 expression with respect to Dpp expression is detected along the apical region of disc proper cells (E); strong Mmp1 expression posterior to the domain of Dpp expression is seen in Z-sections that are located basally (F). (G–G″) Zoomed-in images of a portion of midthird instar larval wing imaginal disc of UAS-ey/UAS-RFP; Dpp-Gal4 larvae reveal ectopic expression of Mmp1 with respect to RFP expression in the disc proper cells located basally. Arrowheads in G′ and G″ mark the punctate Mmp1 expression in between Dpp-expressing cells and arrows in G″ mark high levels of Mmp1 expression around cells expressing Dpp. (H) Drastic reduction in ectopic Mmp1 expression is seen in wing discs upon coexpressing UAS-Mmp1i and UAS-ey by Dpp-Gal4. Bars, 20 μm.

To have better spatial information about this topological area of Mmp1 expression, we analyzed Mmp1 expression with respect to that of reporter Dpp-RFP expression along the A/P axis. As reported earlier (Aggarwal et al. 2016), we observed that ectopic expression of ey by Dpp-Gal4 led to higher levels of Dpp expression in a relatively wider domain (Figure S2C), as compared to their wild-type control wing discs (Figure S2B). Analyses of Dpp-RFP and Mmp1 expression through the Z-stacks of confocal images revealed a very interesting pattern of Mmp1 expression. While there was no detectable Mmp1 expression in the Z-sections along the apical surface of wing disc (Figure 2E), we observed Mmp1 expression in Z-sections that are located basally (Figure 2F). Importantly, the expression of Mmp1 was observed juxtaposed to the wide band of Dpp expression toward the posterior compartment of the wing disc (Figure 2F). When observed under higher magnification, we found that even though the maximum accumulation of Mmp1 was around the cells just next to those expressing Dpp, a fraction of Dpp-positive cells had membranous Mmp1 expression (arrows in Figure 2G″ and Figure S2D″). In addition, we also detected the presence of several punctate Mmp1 in between the Dpp-expressing cells (marked by arrowheads in Figure 2, G′ and G″ and Figure S2, D′ and D″). Taken together, these results indicated that the Dpp-expressing cells might be responsible for the release of Mmp1.

However, in absence of a strong overlap between the expressions of Dpp-RFP and Mmp1, to ascertain whether Mmp1 was actually secreted by Dpp-expressing cells, we resorted to genetic analysis. Employing Dpp-Gal4, we ectopically expressed UAS-Mmp1 RNAi in wing discs undergoing ectopic photoreceptor differentiation and then checked for ectopic Mmp1 expression. As shown in Figure 2H, this resulted in a drastic loss of Mmp1 expression, thereby demonstrating that the Dpp-expressing cells were indeed responsible for ectopic Mmp1 expression. This induction of Mmp1 expression by Dpp seems to be context dependent as mere overexpression of Dpp does not lead to ectopic Mmp1 expression in wing discs (Figure S2, E and F).

Ectopic Mmp1 expression is triggered by dTak1-JNK-mediated Dpp signaling

The canonical Dpp signaling pathway involves binding of Dpp to its receptors, Thickveins (Tkv) and Punt. Activated Tkv, in turn, phosphorylates Mad which forms a heterodimer with Medea (Med) and together they translocate to the nucleus to transcriptionally activate the downstream target genes (Figure 3A) (Hamaratoglu et al. 2014). A previous report from our laboratory has demonstrated that besides this pathway, specifically during ectopic eye differentiation, Dpp upon binding to its receptors can also trigger an alternate signaling cascade that involves TGF-β-activated kinase1 (Tak1), leading to activation of JNK (Figure 3A) (Aggarwal et al. 2016).

Figure 3.

Dpp signaling through dTak1-JNK pathway regulates ectopic Mmp1 expression during wing-to-eye fate alteration. In all wing discs, anterior compartment is to the left and the genotypes are as mentioned. The nuclei are marked with DAPI (blue). (A) Schematic representation of the possible Dpp-mediated signaling pathways that can lead to ectopic Mmp1 expression. (B and C) Sharp reduction in ectopic Mmp1 expression is observed either upon knocking down tkv expression (B) or when ectopic eyes are induced in wing discs of tkv⁷ heterozygous larvae (C). (D) Attenuation of Mad activity does not affect ectopic Mmp1 expression. (E and F) Drastic reduction in ectopic Mmp1 expression is detected when ectopic photoreceptors are either induced in wing discs of dtak1² mutants (E) or upon driving UAS-dTak1i in wing discs undergoing ectopic photoreceptor differentiation (F). (G and H) Similarly, impairing the JNK signaling pathway by expressing UAS-bsk^DN (G) or by driving UAS-Kayi (H) results in a significant drop in ectopic Mmp1 expression. (I) Quantitative analysis of the fold change in Mmp1 transcript reveals that the robust increase in the level of Mmp1 expression as observed in UAS-ey; Dpp-Gal4 wing discs gets significantly reduced upon coexpressing UAS-fos^DN along with UAS-ey by Dpp-Gal4. ** P < 0.001. (J–J″) Zoomed-in image of the region of wing disc showing reporter Mmp1-lacZ expression (J) that completely overlaps with ectopic Mmp1 expression (J′ and J″) as revealed by immunostaining. (K) Drastic decrease in ectopic Mmp1-lacZ expression is observed upon knocking down the expression of dTak1 in wing discs undergoing ectopic photoreceptor differentiation. Bars, 20 μm.

To determine the involvement of Dpp signaling in inducing ectopic Mmp1 expression, we therefore checked for the involvement of both the signaling cascades. To start with, we knocked down the expression of the receptor Tkv in these cells. As evident from Figure 3B, this led to a significant drop in ectopic Mmp1 expression. An analogous result was obtained upon generating ectopic eyes in wing discs heterozygous mutant for the tkv allele, tkv⁷ (Figure 3C). In contrast, ectopic Mmp1 expression remained unaltered when Mad activity was knocked down (Figure S3A) or when ectopic eyes were generated in wing discs heterozygous mutant for Mad¹² (Figure 3D). On the other hand, a significant reduction in ectopic Mmp1 expression was observed when ectopic eyes were generated in dTak1 mutant wing discs (Figure 3E), as well as upon knocking down the expression of dTak1 (Figure 3F). From these results we conclude that activation of Mmp1 expression by Dpp is mediated by dTak1, not by the canonical pathway involving Mad.

To further consolidate this finding, we checked for Mmp1 expression in wing discs undergoing ectopic eye differentiation with compromised levels of expression of members of the JNK pathway. As evident from Figure 3G, a dominant-negative form of basket (bsk; JNK in flies), when coexpressed with ey, resulted in a significant drop in the levels of Mmp1 expression. Likewise, inhibition of dFos activity (one of the transcriptional regulators activated by JNK) by knocking down kayak expression (the gene encoding dFos; Figure S3B) resulted in a considerable decrease in the level of ectopic Mmp1 expression (Figure 3H). Similar results were obtained upon coexpression of a dominant-negative form of dFos in the transdetermining tissue (Figure S3C). In tune with these observations, results of our qRT-PCR analyses also revealed that the increase in expression of Mmp1 transcripts as observed in wing discs of UAS-ey; Dpp-Gal4 larvae became significantly reduced upon downregulating the activity of dFos (Figure 3I).

The transcriptional regulator dFos is known to form a heterodimer with dJun to form the AP1 complex which binds to its specific DNA binding site and regulates transcriptional activity of the target gene (Kockel et al. 2001). We therefore employed a reporter Mmp1-lacZ line, which harbors a construct with reporter lacZ cloned downstream to the regulatory region of Mmp1 with three putative AP1 binding sites (Uhlirova and Bohmann 2006), to determine whether the ectopic Mmp1 expression observed was mediated by AP1. Upon generating ectopic eyes in wing discs of these Mmp1-lacZ larvae followed by co-immunostaining with antibodies against Mmp1 and β-galactosidase, we observed a significant overlap in their expression patterns (Figure 3, J–J″). Furthermore, we observed that knocking down dTak1 activity brought about a drastic drop in the expression of reporter Mmp1-lacZ in wing discs undergoing ectopic photoreceptor differentiation (Figure 3K). From all these results, we conclude that Dpp, involving a signaling cascade that includes dTak1 and JNK, triggered the transcription of ectopic Mmp1 during ectopic eye induction.

Ectopic Mmp1 expression gets spatially oriented toward the posterior compartment

The larval wing imaginal disc consists of two layers of epithelial cells. The disc proper is made up of a layer of columnar epithelial cells, and the apical surface of these cells is juxtaposed with the peripodial membrane made up of squamous epithelial cells. Gross morphological analysis, as assayed by phalloidin staining, revealed that, compared to their wild-type control discs, wing discs undergoing ectopic eye differentiation had a portion of the disc proper cells protruding outward. Interestingly, this protrusion was always observed toward the basal side. Transverse section of these wing discs further confirmed that the protrusion was indeed formed by the columnar epithelial disc proper cells, more or less midway along the A/P direction of the wing disc (arrowhead in Figure 4, B and B′). Similar sections through wild-type wing discs did not exhibit any such fold in the layer of disc proper cells (Figure 4, A and A′). To have better positional information about this protrusion, we next checked for the expression of Engrailed (En) and Dpp in these wing discs. While cells in the posterior arm of the protrusion were found to express En, a bona fide marker for the posterior compartment of wing disc (Figure 4, C and C′), we observed Dpp expression in the anterior arm of the protrusion (Figure 4D). Together, these results demonstrated that the disc proper cells indeed protruded outward basally, midway along the A/P axis, with the A/P boundary around its tip.

Figure 4.

Ectopic Mmp1 expression during wing-to-eye fate alteration is oriented to the posterior compartment. In all transverse-section images of wing discs, anterior compartment is to the left and the genotypes are as mentioned. The nuclei are marked with DAPI (blue). (A and A′) Transverse section of wild-type wing discs in the A/P orientation showing the columnar epithelial disc proper cells (dp) and the squamous epithelial cells of the peripodial (pp) membrane. (B and B′) Transverse section through wing discs of UAS-ey; Dpp-Gal4 larvae in the A/P orientation reveals the formation of a protrusion of the disc proper cells in a direction opposite to that of the peripodial cells. (C and C′) Expression of En is observed in the posterior arm of the protrusion. (D) Cells of the anterior arm of the protrusion express Dpp. (E and E′) Expression of Mmp1 is detected along the cell boundaries of cells of the posterior arm of the protrusion that are adjacent to the Ci^act-expressing cells of the anterior compartment. (F–F″) Expression of UAS-GMA marking the cytonemes that are projected toward the anterior compartment in wing discs of UAS-GMA; Dpp-Gal4 larvae. (F′ and F″) Zoomed-in images of the regions marked in F and F′, respectively. (G–G″) Ectopic expression of ey in wing discs of UAS-GMA; UAS-ey; Dpp-Gal4 larvae leads to a change in the orientation of the cytonemes. As demarcated by the expression of UAS-GMA, cytonemes in these wing discs project toward the posterior compartment. (G′ and G″) Zoomed-in images of the regions marked in G and G′, respectively. Bars, 20 μm.

To determine the spatial orientation of ectopic Mmp1 expression with respect to these protrusions, we checked for Mmp1 expression in transverse sections, wherein the anterior compartment cells were marked with activated Ci (Ci^act) expression. Mmp1 expression was specifically evident around the cells of the posterior compartment adjacent to the posterior boundary of the Ci^act-expression domain (Figure 4, E and E′). Results of these expression studies, apart from providing topological information about ectopic Mmp1 expression, also explained the reason for observing Mmp1 expression specifically in Z-stacks of confocal images along the basal surface of wing discs as shown in Figure 2F. Importantly, this kind of directionality of Mmp1 secretion toward the posterior compartment appears to be specific for these wing discs, as it was not evidenced upon overexpression of Mmp1 by Dpp-Gal4 in the anterior compartment cells along the A/P boundary (Figure S4, A–A′′′). Rather, Mmp1 expression was observed both in the anterior and posterior compartments with relatively more in the anterior compartment compared to that in the posterior one.

To account for this directionality in Mmp1 secretion, we further investigated the role of cytonemes, if any. Cytonemes are thin, actin-rich cellular projections involved in signaling between cells, and have been evidenced to play a critical role in patterning in the developing wing discs (Roy and Kornberg 2015). For our analyses, we primarily focused on the cytonemes emanating from the Dpp-expressing cells, as our results indicated the involvement of these cells in secreting Mmp1. We marked the actin-rich cytonemes of the Dpp-expressing cells by driving UAS-GMA with Dpp-Gal4. The UAS-GMA is a chimeric construct that has the actin-binding region of Drosophila moesin fused to the C terminus of GFP cloned into the pUAST transformation vector (Bloor and Kiehart 2001). Expression of this construct within a cell allows long-term live imaging of morphogenetic events at a high spatial resolution without affecting the normal behavior of F-actin. In developing third instar larval wing discs, we observed the cytonemes originating from these cells are visibly oriented toward the anterior compartment (Figure 4, F–F″). In sharp contrast, these cytonemes get majorly oriented toward the posterior compartment in wing discs of UAS-ey; Dpp-Gal4 larvae (Figure 4, G–G″). However, this kind of change in the directionality of cytonemes was not observed upon overexpressing Mmp1 by Dpp-Gal4 (Figure S4, B–B″). Taken together, these results indicate the possible involvement of cytonemes in specifying the directionality of Mmp1 secretion in wing discs of UAS-ey; Dpp-Gal4 larvae.

Loss of Mmp1 activity augments ectopic Eya and Dac expression, resulting in an increase in ectopic photoreceptor differentiation

Differentiation of photoreceptors in Drosophila is triggered by the activation of a gene-regulatory network that primarily consists of ey, eya, sine oculis (so), and dac (Heberlein and Treisman 2000; Silver and Rebay 2005). As shown in Figure 5A, while all the members of this network regulate each other’s expression by transcriptional activation and feedback loops, the three proteins, Eya, So, and Dac are known to physically interact with one another (Chen et al. 1997; Pignoni et al. 1997). Moreover, it has also been established that dac, along with eya and so, is a downstream transcriptional target of Dpp signaling (Chen et al. 1999; Pappu et al. 2005). We therefore analyzed the role of Mmp1 in regulating ectopic photoreceptor differentiation in the light of Dac and Eya expression.

Figure 5.

Mmp1 expression is required to restrict the domain of ectopic Eya and dac expression during wing-to-eye fate alteration. In all transverse-section images of wing discs, anterior compartment is to the left and the genotypes are as mentioned. The nuclei are marked with DAPI (blue). (A) Schematic representation of the gene regulatory network involved in activation of the RD genes during normal photoreceptor differentiation. (B–B″) While all cells of the anterior arm of the protrusion express Dpp (B and B″), a subset of them toward the posterior ectopically express Dac (B′ and B″). (C–C″) Expression of Mmp1 (C and C″) is detected along the cell boundaries of cells of the posterior arm of the protrusion that are adjacent to the Dac-expressing cells (C′ and C″). (D) Schematic representation of protrusion showing the expression patterns of Dpp, Dac, and Mmp1. (E–G) Both the intensity and domain of ectopic Dac expression get enhanced upon knocking down Mmp1 expression (F) or in Mmp1^Q273 mutant background (G) compared to that observed in wing discs of UAS-ey; Dpp-Gal4 larvae (E). (H and I) Quantification of the increases in the domain of ectopic Dac expression (H) as well as in the intensity of ectopic Dac expression per unit area (I) in wing discs of the genotypes mentioned when compared to their control UAS-ey; Dpp-Gal4 larval wing discs. (J and K) Both the intensity and domain of ectopic Eya expression get enhanced upon generating ectopic eyes in Mmp1^Q273 mutant background (K) compared to that observed in wing discs of UAS-ey; Dpp-Gal4 larvae (J). (L and M) Quantification of the increases in the domain of ectopic Eya expression (L) as well as in the intensity of ectopic Eya expression per unit area (M) in wing discs of the genotype mentioned when compared to their control UAS-ey; Dpp-Gal4 larval wing discs. Bars, 20 μm.

To start with, we checked for the expression of Dac and Mmp1. While ectopic Dac expression was seen in a subset of cells positioned toward the posterior end of the Dpp-expressing domain (Figure 5, B–B″), we observed Mmp1 expression around the cells adjacent to the posterior boundary of the Dac-expression domain (Figure 5, C–C″ and D). Subsequently, we observed that knocking down Mmp1 expression in wing discs undergoing ectopic eye induction resulted in a significant increase in the domain of ectopic Dac-expressing cells (Figure 5F) compared to that observed in wing discs of UAS-ey; Dpp-Gal4 larvae (Figure 5E), with a concurrent increase in the number of ectopically differentiating photoreceptors (Figure S5, A and B′). An identical increase in the domain of Dac-expressing cells (Figure 5G) and number of ectopic photoreceptors (Figure S5, C and C′) was also observed when ectopic eyes were induced in wing discs heterozygous mutant for Mmp1^Q273. Quantitative analyses revealed that compromised activity of Mmp1 led to a twofold increase in the domain of Dac expression when compared to that observed in control UAS-ey; Dpp-Gal4 wing discs (Figure 5H). Interestingly, this increase in the domain of ectopic Dac expression was also associated with an increase in its level of expression. Quantification of fluorescence intensity in the Dac-expression domain revealed an almost 1.8-fold increase in the level of ectopic Dac expression per unit area upon attenuating Mmp1 when compared to that detected in control UAS-ey; Dpp-Gal4 wing discs (Figure 5I).

Immunostaining of wing discs of Dpp-Gal4; UAS-ey larvae with antibodies against the other RD protein, Eya, revealed its expression in the anterior half of the protrusion of disc proper cells (Figure 5J) in a pattern similar to that observed for ectopic Dac expression. Ectopic eyes, when generated in wing discs of larvae that were heterozygous mutant for Mmp1^Q273, resulted in a significant increase in the domain of ectopic Eya expression (Figure 5, K and L) in a manner identical to that observed for ectopic Dac expression. This increase in the domain of ectopic Eya expression was accompanied with an almost 1.5-fold increase in the intensity of Eya expression (Figure 5M). In sum, these results clearly demonstrated that the observed increase in ectopic photoreceptor differentiation in Mmp1 mutant background or upon attenuating Mmp1 expression was actually due to an increase in the intensity as well as in the domain of ectopic Dac and Eya expression.

Loss of Mmp1 leads to increased proliferation of Dac-expressing cells

To ascertain whether the expansion in the Dac-expression domain was an outcome of increased cell proliferation, we assessed cell-cycle progression of the Dac-expressing cells by the incorporation of EdU. EdU is a thymidine homolog that gets incorporated in the newly synthesized strand of DNA and thereby marks the cells that are in the S phase of cell division. Uniform incorporation of EdU was observed in both the anterior and posterior compartments of the wild-type wing disc (Figure 6, A–A″; the posterior compartment being marked by En expression). However, the pattern of EdU incorporation in wing discs of UAS-ey; Dpp-Gal4 larvae was quite distinct. Although a uniform distribution of EdU incorporation was observed in the transverse section, quite strikingly, a portion in the anterior part of the protrusion exhibited scanty EdU-positive cells (marked by dotted lines in Figure 6B). Interestingly, co-immunostaining of these sections with anti-Dac antibody revealed an exact overlap of this region to the domain of ectopic Dac expression (Figure 6, B′ and B″), thereby indicating limited proliferation of the Dac-expressing cells. However, knocking down Mmp1 activity or generating ectopic eyes in a Mmp1^Q273 mutant background led to a significant increase in the number of EdU-positive cells in the Dac-expression domain (Figure 6, C–D″). Quantification of the above result revealed a twofold increase in the number of EdU-positive cells per unit area of the Dac-expression domain in wing discs with attenuated Mmp1 activity when compared to the control (Figure 6E). Together, these results clearly established that the increase in ectopic Dac- and Eya-expression domain, as observed upon compromising Mmp1 activity, was essentially due to an increase in proliferation within this domain.

Figure 6.

Mmp1 regulates cell proliferation to limit the domain of ectopic Dac-expressing cells. In all transverse-section images of wing discs, anterior compartment is to the left and the genotypes are as mentioned. The nuclei are marked with DAPI (blue) and the domain of ectopic Dac expression is denoted by dotted line. (A–A″) Transverse section of wild-type wing discs in the A/P orientation showing uniform EdU incorporation in both the anterior and the posterior compartments [posterior compartment is marked by En expression in (A′ and A″)]. (B–B″) Transverse section through wing discs of UAS-ey; Dpp-Gal4 larvae in the A/P orientation reveals EdU incorporation in the protrusion (B and B″). While uniform EdU incorporation is seen in most of the region, very few EdU-positive cells are detected in the domain expressing ectopic Dac (B′ and B″). (C–C″) Knocking down Mmp1 expression results in increased level of EdU incorporation (C and C″) in cells of the ectopic Dac-expressing domain (C′ and C″). (D–D″) Generating ectopic eyes in wing discs of Mmp1^Q273 heterozygotes also leads to an increase in EdU incorporation (D and D″) in cells of the ectopic Dac-expressing domain (D′ and D″). (E) Quantification of the increase in EdU-positive cells per unit are of the ectopic Dac-expressing domain in the wing discs of the genotypes mentioned when compared to their control UAS-ey; Dpp-Gal4 larval wing discs. Bars, 20 μm.

Ectopic expression of Mmp1 defines the range of Hh signaling in wing discs undergoing ectopic eye differentiation

Previous studies have demonstrated that, during normal eye development, the differentiating clusters of photoreceptors in the larval eye imaginal disc express the secretory morphogen Hh, which coordinates proliferation of the eye primordial cells, and initiates differentiation of photoreceptors, apart from inducing the expression of Dpp in cells ahead of the differentiating clusters (Curtiss and Mlodzik 2000; Duman-Scheel et al. 2002; Pappu et al. 2003; Firth and Baker 2009). Hh is also expressed by cells of the posterior compartment of wing disc and is known to induce the expression of Dpp in the anterior compartment (Basler and Struhl 1994). Moreover, ectopic photoreceptor differentiation always takes place in Dac-expressing cells adjacent to the posterior compartment. We were intrigued to determine whether the increased proliferation in the domain of ectopic Dac expression, as well as the increase in intensity of Dac and Eya expression, as observed upon attenuating Mmp1 activity in wing discs of UAS-ey; Dpp-Gal4 larvae, was an outcome of increased Hh activity.

We therefore analyzed the expression pattern of Ci^act, the transcriptional activator molecule of the Hh signaling pathway, in these wing discs (Huangfu and Anderson 2006). In control wing discs, we observed Ci^act expression in cells of the anterior compartment with maximum expression along the A/P border which gradually formed a decreasing gradient toward the periphery (Figure 7, A and B). In contrast to this, we detected a modest increase in the level of Ci^act expression in the anterior compartment of wing discs of UAS-ey; Dpp-Gal4 larvae (Figure 7, C and D). Interestingly, knocking down Mmp1 expression in wing discs of UAS-ey; Dpp-Gal4 larvae resulted in a robust increase in the level of Ci^act expression (Figure 7, E and G). Not only was the intensity of Ci^act expression along the A/P boundary higher, but even the gradient of Ci^act expression in the anterior compartment exhibited a significant level of increase. Analogous results were observed in wing discs of Dpp-Gal4; UAS-ey/Mmp1^Q273 larvae (Figure 7, F and G). Together, these results clearly demonstrated that loss of Mmp1 activity led to increased Hh signaling in the anterior compartment of wing discs undergoing ectopic eye induction.

Figure 7.

Ectopic Mmp1 expression regulates cell proliferation and levels of ectopic Dac and Eya expression by restricting the diffusion of Hh from the posterior compartment. In all images of wing discs and the transverse sections, anterior compartment is to the left and the genotypes are as mentioned. The nuclei are marked with DAPI (blue). (A) Expression pattern of Ci^act in wild-type wing discs. (B) Average intensity profile of Ci^act expression in wild-type wing discs. (C) Modest increase in the level of Ci^act expression is observed in wing discs of UAS-ey; Dpp-Gal4 larvae. (D) Average intensity profile of Ci^act expression in wing discs of UAS-ey; Dpp-Gal4 larvae compared to that observed in wild-type wing discs. (E and F) The level of Ci^act expression gets drastically enhanced either upon knocking down Mmp1 (E) or upon generating ectopic eyes in wing discs of Mmp1^Q273 heterozygotes (F). (G) Average intensity profiles of Ci^act expression in wing discs of larvae of the genotypes mentioned compared to that observed in wild-type wing discs. (H) Quantitative analysis of the fold change in hh transcript reveals that compared to wild- type wing discs, the level of hh expression remain unaltered in UAS-ey; Dpp-Gal4 and UAS-ey/Mmp1^Q273; Dpp-Gal4 wing discs. (I and I′) Expression of Ci^act in wing discs with wild-type mock clones. Absence of GFP in I′ marks the clone. (J and J′) A marginal increase in the level of Ci^act is observed in somatic clones of Mmp1². The clonal area is marked by the absence of GFP in (J′). (K) Average intensity profiles of Ci^act expression in clones of Mmp1² compared to that observed in wild-type clones. (L) The level of hh expression gets reduced in wing discs of UAS-ey/Mmp1^Q273; Dpp-Gal4/hh^AC larvae. (M) Compared to that observed in UAS-ey/Mmp1^Q273; Dpp-Gal4 wing discs (F), a drop in the level of Ci^act expression is observed in wing discs of UAS-ey/Mmp1^Q273; Dpp-Gal4/hh^AC larvae. (N) Average intensity profiles of Ci^act expression in wing discs of larvae of the genotypes mentioned compared to that observed in wild-type wing discs. (O–O″) Amount of EdU incorporation in the domain of ectopic Dac expression gets reduced in wing discs of UAS-ey/Mmp1^Q273; Dpp-Gal4/hh^AC larvae. (P) Quantification of the relative change in EdU-positive cells per unit area of the ectopic Dac-expressing domain in wing discs of the genotypes mentioned when compared to their control UAS-ey; Dpp-Gal4 wing discs. ** P < 0.001. (Q) Transverse section through the wing discs of UAS-ey/Mmp1^Q273; Dpp-Gal4/hh^AC larvae demonstrating reduction in both the intensity and domain of ectopic Dac expression. (R and S) Quantification of the decrease in the domain of ectopic Dac expression (R) as well as in the intensity of ectopic Dac expression per unit area within the domain (S) in wing discs of UAS-ey/Mmp1^Q273; Dpp-Gal4/hh^AC larvae when compared to their control UAS-ey/ Mmp1^Q273; Dpp-Gal4 wing discs. ** P < 0.001. (T) Transverse section through the wing discs of UAS-ey/Mmp1^Q273; Dpp-Gal4/hh^AC larvae demonstrating reduction in both the intensity and domain of ectopic Eya expression. (U and V) Quantification of the decrease in the domain of ectopic Eya expression (U) as well as in the intensity of ectopic Eya expression per unit area within the domain (V) in wing discs of UAS-ey/Mmp1^Q273; Dpp-Gal4/hh^AC larvae when compared to their control UAS-ey/ Mmp1^Q273; Dpp-Gal4 wing discs. ** P < 0.001. (W) Reduction in the number of ectopic photoreceptors as evidenced in the wing discs of UAS-ey/Mmp1^Q273; Dpp-Gal4/hh^AC larvae. (X) Quantification of the drop in the number of ectopic photoreceptors in wing discs of UAS-ey/Mmp1^Q273; Dpp-Gal4/hh^AC larvae when compared to their control UAS-ey/ Mmp1^Q273; Dpp-Gal4 wing discs. Mean ± SD; *** P < 0.0001. (Y) Schematic representation of the regulation of Hh activity by Dpp in wing discs undergoing ectopic eye differentiation. Transcriptional activation of Mmp1 by Dpp signaling through a pathway involving Tkv, Tak1, and JNK leads to release of Mmp1 outside the cell. Modification of the ECM by Mmp1 restricts the amount of Hh molecules that diffuse to the anterior compartment. Expression of the target genes of Hh signaling that include cycE, dpp, dac, and eya depend upon the level of Ci^act generated in the anterior compartment cells. Bars, 20 μm.

To account for the increase in the gradient of Ci^act expression in wing discs of UAS-ey; Dpp-Gal4 larvae, we then checked for Hh expression. However, our qRT-PCR analyses did not reveal any detectable increase in the level of hh transcripts in wing discs of UAS-ey; Dpp-Gal4 larvae when compared to that observed in control wing discs (Figure 7H). Likewise, we also found that the level of hh expression remained unaltered even in wing discs of UAS-ey/Mmp1^Q273; Dpp-Gal4 larvae (Figure 7H), which demonstrated a robust increase in the gradient of Ci^act expression in the anterior compartment (Figure 7, F and G). Therefore, our results demonstrated that the increase in the gradient of Ci^act was not an outcome of increased hh expression. Rather, these results suggested that loss of Mmp1 might have allowed a greater amount of Hh to diffuse into the anterior compartment of the wing discs of UAS-ey/Mmp1^Q273; Dpp-Gal4 larvae to account for the increase in the level of Ci^act. These results suggested that ectopic Mmp1 expression in wing discs of UAS-ey; Dpp-Gal4 larvae might be playing a role in restricting the diffusion of Hh molecules in the anterior compartment of the wing disc.

To ascertain whether Mmp1 plays a similar role in restricting Hh signaling in the anterior compartment during normal development, we generated somatic clones of Mmp1² [a null allele of Mmp1 (Page-McCaw et al. 2003)] in the wing disc by employing a flp-FRT technique to check for Ci^act expression. As shown in Figure 7, J and J′, we observed a marginal increase in Ci^act expression in somatic clones of Mmp1², when compared to that in wild-type mock clones (Figure 7, I and I′). This increase is further evident in the plot profile analysis, as shown in Figure 7K. A similar increase in the level of Ci^act expression in the anterior compartment was also observed when we knocked down the expression of Mmp1 along the A/P axis by Dpp-Gal4 (Figure S6, A–C). From these results, we conclude that, though marginally, Mmp1 is instrumental in limiting Hh activity in the anterior compartment of wild-type wing discs.

We then argued that if this holds true, then we should be able to rescue all the phenotypes observed in wing discs of UAS-ey/Mmp1^Q273; Dpp-Gal4 larvae by genetically manipulating the amount of Hh molecules being produced. For that purpose, we adopted genetic means to reduce the amount of Hh expression in the posterior compartment of wing discs by employing heterozygous mutants for hh^AC, a null allele of hh (Lee et al. 1992). As evident from Figure 7L, wing discs of UAS-ey/Mmp1^Q273; hh^AC/Dpp-Gal4 larvae demonstrated a 30% reduction in the level of hh expression as compared to the UAS-ey/Mmp1^Q273; Dpp-Gal4 control. Consistent with this, when these wing discs were checked for the expression of Ci^act, they demonstrated a drop in expression along the A/P border as well as in its gradient in the anterior compartment (Figure 7M). Quantitative analysis of the intensity profile revealed the level of Ci^act expression was comparable to that observed in wing discs of UAS-ey; Dpp-Gal4 larvae (Figure 7N).

To determine whether this drop in Hh signaling was able to rescue the increase in cell proliferation of the Dac-expression domain of UAS-ey/Mmp1^Q273; Dpp-Gal4 larvae, we analyzed the pattern of EdU incorporation in wing discs of hh^AC/Dpp-Gal4; UAS-ey/Mmp1^Q273 larvae. As evident from Figure 7, O–O″, an appreciable drop in the level of EdU incorporation was detected, suggesting that the cells in the Dac-expressing domain were indeed dividing less. Quantification of the number of EdU-positive cells per unit area of the domain of Dac expression in these wing discs exhibited a significant drop when compared to that observed in wing discs of UAS-ey/Mmp1^Q273; Dpp-Gal4 larvae (Figure 7P). We subsequently checked for the expression of the RD genes Dac and Eya. For each of them, we observed a significant drop both in their domain as well as in their level of expression (Figure 7, Q–V). Finally, as a functional correlate of these results, we observed that reducing the amount of Hh signaling resulted in a significant drop in the number of ectopic photoreceptors in wing discs of UAS-ey/Mmp1^Q273; hh^AC/Dpp-Gal4 larvae (Figure 7, W and X). In sum, results of our expression studies, in conjunction with that of in vivo genetic analyses, corroborated the hypothesis that the altered phenotypes observed in UAS-ey/Mmp1^Q273; hh^AC/Dpp-Gal4 wing discs was an outcome of decreased Hh signaling in the anterior compartment. These results further established that in wing discs of UAS-ey; Dpp-Gal4 larvae, Mmp1 was playing an important role in limiting the range of Hh signaling in the anterior compartment.

Discussion

Our results unravel an otherwise unknown mechanism by which Dpp signaling regulates the range of Hh signaling (Figure 7Y) during ectopic eye differentiation in wing discs of Drosophila. We show that by employing a signaling cascade that involves Tak1 and JNK, Dpp activates the expression of Mmp1. The role of secreted Mmp1 seems to regulate the diffusion of Hh molecules into the anterior compartment and, in the process, restricts proliferation and the level of ectopic dac and eya expression. In doing so, Mmp1 limits the amount of ectopic photoreceptor differentiation. With these findings we have been successful not only in unraveling the developmental signaling cascade leading to ectopic Mmp1 expression, but also in providing evidence to establish the role played by Mmp1 during ectopic photoreceptor differentiation. From a broader perspective, this entire regulatory mechanism (Figure 7Y) unearths a unique phenomenon where one morphogen, Dpp, regulates the activity domain of another morphogen, Hh, by employing Mmp1, a known modulator of the ECM.

Several studies have outlined the regulation of Dpp signaling by Hh in both eye and wing imaginal disc development (Heberlein et al. 1993; Basler and Struhl 1994). However, regulation of Hh activity by Dpp has not yet been evidenced (Shen and Dahmann 2005). We provide the first in vivo genetic evidence for the regulation of Hh signaling by Dpp. The mechanism involved does not lead to any transcriptional regulation of Hh signaling. Rather, Dpp achieves this regulation through activation of the JNK signaling cascade that elicits ectopic expression of Mmp1. Mmp1, in turn, limits the range of Hh signaling. Importantly, our results indicate that this kind of restriction of Hh diffusion by Mmp1 also happens, although marginally, in developing wild-type wing discs, as revealed by an increase in the level of Ci^act in anterior compartment cells that are homozygous mutant for Mmp1, as well as upon knocking down the expression of Mmp1 along the A/P boundary. Since the glypican, Dally-like protein (Dlp), is a potential target of Mmp1 and Dlp facilitates the transmission of Hh signaling by acting as coreceptor for Patch in the Hh receiving cells of the anterior compartment (Gallet et al. 2008), it is quite possible that in these cases the increase in Ci^act in anterior compartment cells might be an outcome of an increased level of Dlp. However, even though the molecular basis of regulation of Hh by Mmp1 during ectopic eye differentiation is yet to be made evident, we believe it to be independent of Dlp, primarily because we observe ectopic Mmp1 expression around the Hh-releasing cells of the posterior compartment. Since no increase in Mmp1 expression is observed in cells of the anterior compartment, it is quite evident that during this process, Mmp1 does not control Hh signaling in the receiving cells by modulating Dlps. Rather, Mmp1 actually modulates the ECM around the Hh-producing cells to regulate the diffusion range and accessibility of Hh molecules to its cognate receptors present on cells of the anterior compartment by an alternate mechanism yet to be made clear.

Our results clearly demonstrate that even though the domain of Dpp expression gets significantly increased in wing discs of UAS-ey; Dpp-Gal4 larvae, ectopic expression of Dac and Eya is restricted only to the posterior part of the Dpp-expression domain. Consistent with this, ectopic photoreceptor differentiation is also observed in this domain with ectopic dac and eya expression. Despite Ey and Dpp being expressed in the entire Dpp-expressing domain, this restriction of ectopic dac and eya expression toward the posterior part of this domain highlights the requirement of Hh signaling in ectopic eye differentiation. Interestingly, restriction of ectopic dac and eya expression in cells juxtaposed to the A/P boundary which have a high expression of Ci^act, suggests threshold-dependent activation of these genes by Hh. This notion becomes further endorsed as we find that increase in expression of Ci^act in Mmp1 mutant wing discs leads to a concomitant increase in ectopic dac and eya expression and subsequent photoreceptor differentiation. Importantly, this increase in the level of ectopic dac and eya expression can be rescued by lowering the amount of Hh signaling from the posterior compartment. Likewise, Hh signaling from the posterior compartment also regulates proliferation in the Dac-expressing domain. Lowering Mmp1 levels allows more Hh signaling to induce the Dac-expressing cells, which were otherwise nonproliferating, to proliferate further. Interestingly, the increase in proliferation occurs toward the posterior part of the Dac-positive region with the cell-cycle arrest still persisting anteriorly. To a certain extent, this bears similarity to the progression of the morphogenetic furrow by Hh signaling, as seen during ommatidial differentiation in developing eye discs (Borod and Heberlein 1998). With increased Hh signaling, the region of cell-cycle arrest moves anteriorly and the cells entering S phase in the posterior part of the Dac-positive region might resemble the second mitotic wave. As a consequence, new rows of photoreceptors differentiate anteriorly.

A striking phenotype associated with ectopic ommatidial differentiation is the protrusion formed by disc proper cells around the region undergoing cell-fate alteration in wing discs of UAS-ey; Dpp-Gal4 larvae. This process appears to be independent of Mmp1 activity, as the formation of the protrusion is not affected in Mmp1 mutant or in an Mmp1 knock-down background. Within this protrusion, ectopic Mmp1 expression is specifically directed toward the posterior compartment. Based on our results, we speculate that altered orientation of the cytonemes emanating from the Dpp-expressing cells, resulting from prominent remodeling of disc proper cells, might be responsible for this directionality. Since this kind of change in the orientation of cytonemes as well as directionality in Mmp1 secretion is not seen when Mmp1 is overexpressed or upon ectopically expressing another RD gene, eya (results not shown), in all likelihood, ectopic expression of ey is responsible for this structural remodeling by a mechanism yet to be revealed. The expansion in the domain of Dpp expression in anterior compartment cells observed in these wing discs might also result from more Hh molecules diffusing from posterior to anterior compartment as a result of local alteration in ECM due to these structural changes. However, since the cells of the wing imaginal discs are in a determined state, it is quite possible that activation of Mmp1 by Dpp signaling might be a measure adopted by these cells to limit Hh diffusion, so as to counteract (or outcompete) the extent of fate alteration, and in the process maintain their determined state. In this context, it is important to note that previous studies have demonstrated that Dpp signaling elicits both instructive and permissive roles to regulate the ectopic expression of RD genes critical for ectopic photoreceptor differentiation (Chen et al. 1999; Curtiss and Mlodzik 2000; Pappu et al. 2005; Aggarwal et al. 2016). The outcome of this current study, in conjunction with those results, clearly establishes a dual role of Dpp signaling in cellular plasticity. While on one hand it facilitates cell fate alteration by creating a microenvironment conducive for the process, simultaneously it also restricts the proliferation and differentiation potential of the cells undergoing fate alteration. By balancing these two events, Dpp plays a pivotal role in maintaining tissue homeostasis.

Studies in the recent past, primarily involving in vitro analyses of cancer cells, have implicated the involvement of JNK, Wnt, NF-κB, and TGF-β signaling pathways in transcriptional activation or repression of Mmp and TIMP genes (Yan and Boyd 2007; Fanjul-Fernandez et al. 2010). Transcriptional regulation of MMPs in cell lines by TGF-β signaling, however, is rather complex. While binding of Smad to the TGF-β inhibitory element negatively regulates transcription of MMP-1 and MMP-7 (Yuan and Varga 2001), Smads play an important role in altering MMP expression (Verrecchia et al. 2001) by interacting with AP1, the transcriptional regulator of JNK pathway. Although the involvement of JNK signaling has also been evidenced in transcriptional activation of Mmp1 in flies (Uhlirova and Bohmann 2006), the involvement of TGF-β signaling in triggering Mmp1 expression was not known. Here we present in vivo genetic evidence of an otherwise complex regulatory mechanism that involves cross talk between Dpp (a member of the TGF-β family of proteins) and JNK signaling in the transcriptional activation of Mmp1 in a Mad- (the fly homolog of Smad) independent manner. Of note, involvement of the TGF-β-TAK1 signaling cascade in activating mammalian MMP-2 has also been demonstrated in preneoplastic MCF10A human breast epithelial cells (Sano et al. 1999; Kim et al. 2007), providing evidence for the evolutionary conservation of this signaling module.

Aberrant activation of the Sonic hedgehog (Shh; the vertebrate counterpart of Hh) signaling pathway has been implicated in initiation, progression, and metastatic invasion of multiple cancer types (Evangelista et al. 2006; Song et al. 2016). While deregulated Shh signaling upregulates levels of different components of cell-cycle machinery leading to unrestricted proliferation of cancer cells (Kenney and Rowitch 2000; Mille et al. 2014), Shh signaling also promotes metastasis through induced expression of MMP-2 and MMP-9 (Wang et al. 2015). MMPs, on the other hand, play a dual role in tumor growth and metastasis processes. Although numerous studies have demonstrated tumor-specific upregulation of MMPs together with their involvement in tumor invasion and metastasis in various cancer types, growing evidence suggests that some MMPs also have tumor-inhibiting functions in a context-dependent manner (Kessenbrock et al. 2010; Gialeli et al. 2011). It would be really intriguing to find, in the future, whether mammalian MMPs play any role in limiting Shh signaling, as demonstrated in this study, to limit uncontrolled proliferation and cancer progression.

Supplementary Material

Supplemental material is available online at www.genetics.org/lookup/suppl/doi:10.1534/genetics.117.201053/-/DC1.