Mmp2 regulates basement membrane remodeling and dedifferentiation of the visceral musculature during Drosophila metamorphosis
Uwe Töpfer, Ina Dahlitz, Anne Holz
TL;DR
This study shows that Mmp2 helps reshape the basement membrane and allows muscle cells to change during fruit fly metamorphosis.
Contribution
The study identifies Mmp2 as a key regulator of basement membrane remodeling during Drosophila metamorphosis.
Findings
Mmp2 is localized during basement membrane degradation and is required for breaking down its main components.
Mmp2 is essential for survival and proper tissue transformation during metamorphosis.
Mmp2-mediated basement membrane remodeling is necessary for muscle dedifferentiation.
Abstract
The basement membrane (BM) is a specialized extracellular matrix that surrounds most tissues and organs. Remodeling of the BM is critical for morphogenesis and to control tissue homeostasis. During Drosophila metamorphosis, most tissues undergo apoptosis and become histolyzed to be replaced by progenitor cells to generate adult structures, but the visceral musculature trans-differentiates to give rise to new adult muscles. The molecular mechanisms of the BM remodeling during this extensive tissue reorganization are poorly understood. Here, we identified Matrix metalloprotease 2 (Mmp2) as a key regulator of BM remodeling in visceral musculature. We find that Mmp2 is localized when the BM is degraded and that Mmp2 is required for degradation of the major BM components. In addition, Mmp2 is important for survival and tissue metamorphosis. Our results suggests that Mmp2-mediated BM…
Genes, proteins, chemicals, diseases, species, mutations and cell lines named across the full text — each resolved to its canonical identifier and authoritative record.
Click any figure to enlarge with its caption.
Figure 1
Figure 2
Figure 3
Figure 4- —Justus-Liebig-Universität Gießen (3114)
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
Taxonomy
TopicsCellular Mechanics and Interactions · Cell Adhesion Molecules Research · Cellular transport and secretion
Introduction
The basement membrane (BM) is a specialized extracellular matrix (ECM) that underlies epithelia and surrounds tissues like the nervous system or muscles^1,2^. These sheet-like scaffolds, composed of laminin and collagen IV networks, proteoglycans such as perlecan, and glycoproteins like nidogen, are important for stabilizing tissues and providing biochemical and biophysical information that influences cell behavior such as cell migration, adhesion, polarization, or differentiation^3,4^. Irregular BM homeostasis have been associated with human disease like Alport syndrome or diabetes mellitus^5^. The BM must be remodeled during dynamic tissue growth, changes in tissue shape or tumor invasion, processes often associated with the function of matrix metalloproteases (Mmps)^6–9^. Mmps are a family of zinc proteases that degrade extracellular matrix components. This gene family of matrix proteases is upregulated in almost every form of cancer and are therefore thought to play a key role in cancer progression by influencing metastasis, cancer cell growth and migration^10^. Originally, Mmps have been identified as enzymes that degrade fibrillar collagen in tadpole tails during metamorphosis in amphibian tissues^11^. Subsequently, Mmps were classified as collagenases, gelatinases, matrilysins and stromelysins based on their specificity for extracellular matrix components. However, the range of Mmp substrates is still expanding, and Mmps are now grouped according to their structural properties and between secreted and membrane-type. In contrast to humans, where 23 Mmps are thought to be enzymatically redundant or compensatory^8,12^, the genetic model organism Drosophila encodes only two Mmps^13^, making it easier to identify the substrates of individual Mmps. Hereby, Mmp1 as well as Mmp2 has been shown to cleave Casein and Mmp2 can additionally cleave Gelatin^14^ and using GFP-tagged Collagen IV, Mmp1 and Mmp2 knockout have been shown to increase Collagen IV protein level^15,16^. Amorphic alleles of Mmp1 cause disruption of tracheal development in larval stages and problems with head eversion during pupal stages, whereas Mmp2 mutants have defects in larval tissue histolysis and epithelial fusion processes during metamorphosis^13^.
During Drosophila metamorphosis most larval tissues are degraded and form anew from imaginal cells. Thereby, cells like the larval midgut epithelium undergo an autophagy-initiated cell death program^17^, most parts of the somatic musculature are histolysed and just a small group of cells get remodeled and persist to adulthood^18^ and the visceral and heart muscle cells transdifferentiate into adult muscle cells^19–22^. The visceral midgut musculature consists of circular muscles and overlying longitudinal muscles which shares their common BM with the underlying midgut epithelium during larval stages^20^. During metamorphosis, the multinucleated circular and longitudinal muscles dedifferentiate in the pupal stage by losing their myofibrils and forming secondary myoblasts, and then redifferentiate to form the adult midgut musculature^19,20^. This process coincides with the loss and reassembly of Perlecan^19^. To what extent the BM becomes degraded in such drastic tissue reorganization and which regulators are essential for remodeling is largely unknown.
Here we identified Mmp2 as a regulator of BM remodeling during the metamorphosis of the visceral midgut musculature. Knockdown of Mmp2 inhibits dedifferentiation of visceral muscle cells, blocks metamorphosis, and ultimately induces pupal death. We show that Mmp2 is expressed in the visceral musculature when the BM is degraded and disappears when the BM is rebuilt. Moreover, systematic analysis of the BM composition of Mmp2 knockdown guts, show a dependence of degradation of all BM core components on Mmp2 expression. Thus, Mmp2 emerges as main regulator of BM remodeling during Drosophila midgut metamorphosis, which is a key requirement for metamorphosis and survival.
Results
Mmp2 is required for midgut metamorphosis and survival
Metamorphosis in Drosophila requires the fundamental remodeling of larval organs and the formation of adult organs. Excessive remodeling of the BM is required to implement these changes. To identify regulators of BM remodeling during Drosophila metamorphosis, we screened for matrix proteases and found that knockdown of Mmp2 lead to impaired metamorphosis (Fig. 1A-F). Control pupae degrade larval organs, such as trachea and fat body, which are visible during pupal stage 3. Most tissues are rebuilt by progenitor cells thereafter, such as during the pupal stage 6, and these morphogenetic processes are finalized by the end of metamorphosis, as seen in the pupal stage 12. The result are late-stage pupae with adult organs and structures, including eyes, wings and bristles (Fig. 1A-C). In contrast, Mmp2 knockdown pupae stop development between stages 3–6, begin to collapse and form melanotic tumors (Fig. 1E, F, arrows). Survival tests with two independent RNAi lines show complete lethality at 29 degrees (Sup. Figure 1). To elucidate the consequences of Mmp2 knockdown on the tissue level, we studied the morphology of the midgut and the surrounding muscles using F-actin staining (Fig. 1G-J’). In control pupae, visceral muscles of the midgut start to shrink with the onset of metamorphosis and dedifferentiate until they are no longer detectable by F-actin staining at stage 6 while the larval midgut epithelium starts to compact and form the so-called ‘yellow-body’^19^(Fig. 1I’). In comparison, Mmp2 knockdown pupae of stage 6 still have detectable myofibrils by F-actin staining (see Fig. 1I and J), pointing to an arrest of visceral muscle dedifferentiation. The midgut of Mmp2 knockdown pupae do not condensate to the same extent like controls, leading to significant differences in midgut size in stage P3 and P6 (Fig. 1K) and do not shrink as controls (Fig. 1L).
Taken together, Mmp2 is required for midgut metamorphosis and survival during pupal development.
Fig. 1. Mmp2 is required for midgut metamorphosis.** (A-F)** Pupae morphology of control Actin5C-Gal4/+ (A-C) and Actin5C-Gal4/UAS-Mmp2-RNAi (D-F) genotypes at the indicated stages. Black arrows in E and F points to the formation of melanotic tumours in Mmp2 knockdown pupae. (G-J’) F-actin staining (white) on midguts of control Actin5C-Gal4/+ (G, G’, I, I’) and Actin5C-Gal4/UAS-Mmp2-RNAi (H, H’, J, J’) genotypes in top view to show visceral muscles (G, H, I, J) and cross-view (G’, H’, I’, J’) to show additionally the midgut epithelium. (K and L) Violin plots of sagittal midgut area (K) and midgut length (L). Mean ± SEM (central black point and range) and single data point (scatter plot) are shown. ∗∗∗p < 0.001, ∗p < 0.05 (Welsh’s two-sided t test). n = 7 midguts per stage and genotype. Scale bar = 100 μm.
BM degradation and Mmp2 localization during metamorphosis
Collagen IV and Perlecan have already been shown to become degraded and reassemble during midgut metamorphosis^19^. To study if the laminin network is remodeled in the same way and to identify the temporal-spatial localization of Mmp2 proteins, we used GFP-tagged versions of LamininB1 (LanB1)^23^ and Mmp2^24^ in combination with a Mhc-Gal4,* UAS-mCD8::RFP* to mark the musculature and generate cryosections of pupae. This method allowed us to study the localization of the matrix protease and compare its expression dynamics with BM remodeling (Fig. 2). At the onset of pupal development, a laminin layer covers the visceral and somatic muscles and fat body and is somewhat weakened in P3 (Fig. 2A-A’’, C-C’’), while Mmp2 is at P1 slightly visible at visceral muscles (Fig. 2B-B”), but show strong accumulation around visceral muscles as like midgut cells in P3 (Fig. 2D-D”). The protein distribution of LanB1 at stage P6 shows a complete loss of laminin layer at internal organs with one exception, the malpighian tubule show the only LanB1 signal we are able to detect in P6 (Fig. 2E-E’’). At this stage, Mmp2 is no longer detectable (Fig. 2F-F’’). Between stages P6 and P12, LanB1 reassembles and covers all internal organs (Fig. 2G-G”), while Mmp2 is still absent (Fig. 2F-H”). In the next step, we wanted to find out whether the detected protein localization of Mmp2 was due to tissue-specific gene expression in the midgut or the midgut musculature. To address this question, we used a Mmp2^CRIMIC^-Gal4 line^25^ and found strong expression in longitudinal visceral muscles and a weaker expression in circular visceral muscles in pupal stage 3 (Sup. Figure 2).
To conclude, we show that laminin, which is indicative of the inner part of the BM, is degraded and reassembled during Drosophila midgut metamorphosis. We also show that there is a time delay in the protein localization of Mmp2 compared to laminin. This time delay in protein expression may suggest that Mmp2 could be responsible for laminin degradation.
Fig. 2. Protein localization of LanB1 and Mmp2 during metamorphosis. (A-A’’, C-C’’, E-E’’, G-G’’) Cryosections of Pupae with indicated stages stained for LanB1::sfGFP to visualize laminin protein distribution (blue in A, C, E and G; white in A’, C’, E’ and G’), Mhc-Gal4,* UAS-mCD8::RFP* to visualize muscle cells (red) as well as DNA staining (orange). (B-B’’,** D-D’’, F-F’’, H-H’’)** Cryosections of pupae with indicated stages stained for Mmp2::eGFP to visualize Mmp2 protein distribution (violet in B, D, F and H; white in B’, D’, F’ and H’), Mhc-Gal4,* UAS-mCD8::RFP* to visualize muscle cells (red) as well as DNA staining (orange). Scale bar = 100 μm.
Mmp2 is required for BM degradation
In order to understand the dynamics of BM remodeling during metamorphosis of the midgut, we used two complementary approaches. First, we used histological analysis of cross-sections through pupae of different stages to analyze the morphology of the visceral musculature and the surrounding BM (Fig. 3). We note that in control as well as in Mmp2 knockdown animals, the BM is prominently visible surrounding the visceral muscles in stage 3 pupae (Fig. 3A, A’, D and D’). In P6 controls, the entire midgut becomes extremely flat and we were not able to identify a BM (Fig. 3B, B’), while a thin BM is restored until stage P12 (Fig. 3C, C’). In contrast, in Mmp2 knockdown pupae, the BM never disappears and a thick BM surrounds the visceral musculature even at stages P6 and P12 (Fig. 3E, E’, F, F’).
Fig. 3. Histological analysis of BM remodeling during midgut metamorphosis.** (A-F’)** Toluidine blue-stained semi-thin cross sections of pupae with control Actin5C-Gal4/+ (A-C’) and Actin5C-Gal4/UAS-Mmp2-RNAi (D-F’) genotypes at the indicated stages. Orange rectangles indicate area of close-ups (A’, B’, C’, D’, E’, F’). Semi-transparent blue color annotate the BM in close-ups. The arrows indicate the absence of BM surrounding the midgut in stage P6 of the control genotype (B’), as well as the presence of thickened BM in the midguts of Mmp2 knockdown pupae in stage P6 (E’). Abbreviation: Basement membrane (BM), visceral musculature (VM), somatic musculature (SM), midgut epithelium (ME), yellow body (YB). Scale bar = 100 μm.
In the second approach, we compared the presence of major BM components in control and Mmp2 knockdown pupae at stages P3, a time point when dedifferentiation of visceral midgut muscles begins, and P6, when wild-type pupae no longer show laminin protein localization (Fig. 2E) and we were no longer able to detect a BM in histological sections (Fig. 3B, B´). We analyzed the main BM components starting with LamininA, LamininB1, and Nidogen which we detected with specific antibodies, while the other BM proteins Perlecan and Collagen IV were detected with GFP tagged protein variants (Fig. 4A-T’). During pupal stage P3 all main BM components are enriched in the midgut BM (Fig. 4A-E’), although LamininA is more prominent in the circular visceral muscles (Fig. 4A, A’). Likewise, the entire BM is degraded in the mid of pupal development in stage P6 (Fig. 4F-J’). Mmp2 knockdown pupae show similar expression of all BM components in stage P3 in comparison to the controls (Fig. 4K-O’) but in the knockdown condition all analyzed BM components are not degraded and are still present (Fig. 4P-T’).
Taken together, BM degradation between pupal stages 3–6 depends on the expression of Mmp2.
Fig. 4. Mmp2 is required for degradation of main BM components.** (A-T’)** Midguts of pupal stage 3 (A-E’, K-O’) and stage 6 (F-J’, P-T’) with Actin5C-Gal4 as control genotype (A-J’) and additional UAS-Mmp2^dsRNA^ (K-T’) stained for LanA (A, A’, F, F’ K, K’, P, P’), LanB1 (B, B’, G, G’, L, L’, Q, Q’), Nidogen (C, C’, H, H’, M, M’, R, R’), trol::GFP (encodes Perlecan) (D, D’, I, I’, N, N’, S, S’) and vkg::GFP (encodes Col IVα2) (E, E’, J, J’, O, P’, T, T’) (pseudocolor violet to orange). Scale bar = 100 μm.
Discussion
Mmps contribute to morphogenesis through ECM remodeling and act on critical processes like tumor progression and wound repair^7,8,10,26^. The work described here, provides a mechanistic model of how BM remodeling may be controlled by a Mmp to control morphogenesis. Specifically, we show that Mmp2, one of two Drosophila Mmps, is critical for the degradation of major BM components in the visceral muscle BM. The degradation and reassembly of the BM and the protein localization of Mmp2 show a spatially and temporally opposing pattern, suggesting a dynamic regulation of BM remodeling by Mmp2. Furthermore, in Mmp2 knockdown pupae dedifferentiation of visceral midgut muscles is blocked and pupae subsequently die.
Uncovering specific substrates and functions of Mmps remains challenging, especially in vertebrates, were 23 Mmps show genetic redundancy and compensation and act on more complex ECMs compared to the simple and comparatively reduced BM of invertebrates like Drosophila. Indeed, little is known about specific substrates of single Mmps. Even in Drosophila, with only two Mmps, the molecular targets are barely described. What we know is, that in substrate gel zymographic and enzymatic assays, both Drosophila Mmps cleave Casein and Mmp2, but not Mmp1, can cleave Gelatin^14^. In addition, Mmp1 has been shown to influence Collagen IV turnover during embryogenesis^27^ and both Mmps can act on Collagen IV, such as in in the heart tube, where mutations of Mmps lead to an increased Collagen IV signal^16^ and in the fat body, where ectopic expression of both Mmps leads to a loss of Collagen IV^15^. Interestingly, in this study, Mmp1 has also been implicated to act on cell junctions. During metamorphosis, both Mmps are required for cell dissociation in the fat body. Likewise to our findings, in the fat body knockdown of Mmp2 results in loss of the the BM component Collagen IV, while Mmp1 is maybe required to cleave the cell junction protein E-Cadherin^15^. Similarly, Mmp1 processes the cell adhesion protein Ninjurin A in the trachea^28^ and in vertebrates two Mmps (MMP-9 and MMP-12) cleave N-Cadherin in vascular smooth muscle cells to influence ß-Catenin signaling and proliferation^29^. Proteomic approaches to identify cleaved substrates of proteases, called degradomics, reveal the broad capability of Mmps to cleave ECM, growth factors, cytokines and cell surface-associated adhesion and signaling receptors^30–32^. However, the investigation of temporal-spatial expression of Mmp genes as well as the in vivo validation of target substrates remain open challenges to understand the mechanisms how Mmps are involved in tissue remodeling and homeostasis. Previous work^15,16^ and our result in Drosophila may indicate functional diversification, with different Mmps are maybe required for the degradation substrates in a hierarchical, sequential mechanism required for tissue remodeling.
Here we identified Mmp2 as a crucial regulator of BM remodeling during midgut metamorphosis. We propose a model in which the degradation of the visceral musculature BM at the onset of metamorphosis depends on Mmp2 expression to allow tissue remodeling. Without this process, tissues become stuck in development and eventually die, leading to pupal death.
Methods
Fly stocks and staging
The fly stocks used were Actin5C-Gal4 [Bloomington Drosophila Stock Center (BDSC), 4414], Mhc-Gal4 (BDSC, 55133), UAS-Mmp2^dsRNA^ I (BDSC, 31371), Mmp2::eGFP (BDSC, 60512), Mmp2^k00604^ (BDSC, 10358), Mmp2^CR01124−TG4^ (BDSC, 81180), UAS-mCD8::RFP (BDSC, 32219), UAS-Mmp2^dsRNA^ II [Vienna Drosophila Resource Center (VDRC), 330203], LanB1::sfGFP (VDRC, 318180), vkg::GFP^G00454^ [Drosophila Genetic Resource Center (DGRC), 110692], trol::GFP^ZCL1700^ (DGRC, 110807). Larvae were raised at 18 °C until day 4, sorted for genotype and shifted to 29 °C. Siblings of control and Mmp2 knockdown, which transformed into pupal stage 1 were timed together. Stages were identified due to the morphology of controls^33^ and knockdown pupae were treated like the controls with the same incubation time.
Immunohistochemistry
Midguts were dissected in PBS and fixed with 4% formaldehyde in PBS for 20 min. After washing four times with PBS-Tween (0.1%), samples were incubated with primary antibody overnight at 4 °C. The unbound antibody was washed out four times with PBS-Tween. Then the samples were incubated with secondary antibody labeled with fluorescent dye for two hours at 25 °C. The following primary antibodies were used: guinea pig anti-LamininA^34^(1:1000), rabbit anti-LamininB1^35,36^(1:400), rabbit anti-Nidogen^37,38^(1:500), rabbit anti-GFP (Abcam, ab290, 1:500) and rat anti-RFP (Chromotek, 5F8, 1:1000). Alexa Cy-coupled secondary antibodies were purchased from Dianova (goat anti guinea pig, 106–165-003, 1:100), Jackson ImmunoResearch (goat anti-rat, 112–165-167, 1:200) and Vector Laboratories (goat anti rabbit, DI-1488-1.5.5, 1:2000; goat anti rat, 1:500). Phalloidin-TRITC (Sigma Aldrich, P1951, 1:2000) was used to visualize F-actin. Hoechst 33,342 (Sigma Aldrich, 1:2000) or TOPRO3 (Invitrogen, 1:1000) were used to visualize DNA. Samples were embedded in Fluoromount-G (Southern Biotech) before visualization. Images were acquired using a laser scanning microscope Leica TCS SP2, Leica SP8 Falcon or Olympus FV1000. Brightfield images were taken with an Olympus BX 51.
Midgut size quantification
Midgut length and area were quantified using Fiji^39^. To quantify midgut length, a freehand line was drawn from the most anterior to the most posterior end (Sup. Figure 3).
Cryosectioning
Pupae were collected from the tubes with a brush and assigned to the appropriate stage under stereomicroscopic observation. Pupae of stages P1 to P4^33^ were incised transversely at the anterior end with fine eye scissors so that only the retracted mouthparts were hit. For pupae from stage P5 onwards, the anterior part of the pupal case around the head region was carefully removed using fine forceps. The pupae were fixed with 4% formaldehyde in PBS and 0.1% Triton (F-PBT) for one hour at room temperature and the residual pupal case was then carefully removed using eye scissors without damaging the tissue, by starting from the anterior and carefully cutting lengthwise towards posterior. The pupae were again fixed in F-PBT for one hour at room temperature. The fixed pupae were washed three times for 15 min each and then incubated with an ascending gradient in sucrose. The incubation time per step was at least four hours or overnight at room temperature. After sucrose infiltration, the pupae were transferred to the optimal cutting temperature solution (OCT - Tissue-Tek) and incubated for at least four hours. For freezing, the samples were filled halfway with OCT in cryomolds (Tissue-Tek) and frozen at −20 °C for five minutes. The cryomolds were returned to room temperature and the pupae were transferred from the infiltration solution to the solid OCT layer in the cryomolds. Under stereomicroscopic observation, the pupae were oriented according to the desired main anatomical direction of the subsequent sections. The cryomold with the contained pupa was filled with OCT. The sample was left at room temperature until the bottom layer became slightly viscous. The two OCT layers should merge into each other to prevent the sample from breaking later during cutting. The sample was then frozen to −20 °C and subsequently stored at −80 °C at least overnight. Prior to sectioning, slides were silanized by incubation in 5% HCl overnight, in 2% silane in acetone, twice in 2% silane in ddH2O and finally in ddH2O for five minutes each. The cryomicrotome (Leica CM 1950) was switched on the day before in order to ensure a suitable pre-cooling period. The room temperature was between 18 and 35 °C. The cooling of the object head was set to −16 °C and the cooling of the cryostat chamber to −20 °C. The section thickness was set to 12 μm to enable the best possible detection with antibodies and to obtain enough cell material to analyze the tissue. The samples were removed from the crymolds and oriented on the previously cooled sample bases. The OCT was melted by adding a little water to the bottom of the sample. The samples were placed on the Peltier elements of the cryomicrotome and refrozen. The frozen samples were anchored on the object head, brought close to the blade and trimmed in steps of 20–30 μm. Once the tissue was reached, sections were made for analysis. The sections were applied by advancing the silanized slide, which was kept at room temperature during the cutting process. Due to the temperature gradient, the cooled tissue sample attached itself to the warm slide without the sample being affected by touching it. The slides were then dehydrated in ice-cold isopropanol for ten minutes before being dried at room temperature for two hours.
Section and histology
For histological sections, pupae were fixed with 4% formaldehyde in PBS, three times washed with PBS, followed by dehydration in a raising ethanol series and embedded in Technovit 8100 (Kulzer, Hanau, GER). Pupae were cut sagittal to 5 μm sections on a rotary microtome (Leitz, Wetzlar, GER) using a diamond knife and stained in toluidine blue (0.1% toluidine blue and 0.1% sodium tetraborate in distilled water) for 2 min, differentiated in distilled water, dehydrated in xylene, dried, and mounted in Entellan (Merck, Darmstadt, GER).
Statistical analyses
Statistical significance was calculated using the Welsh two-sided t-test using R.
Survival rate
Larvae were raised at 18 °C until day 4, sorted for genotype and shifted to 29 °C. Batches of 10–20 larvae of respective genotypes were transferred to petri dishes with fly food. After one week, hatched flies were counted to quantify the survival rate.
Supplementary Information
Below is the link to the electronic supplementary material.
Supplementary Material 1
The reference list from the paper itself. Each links out to its DOI / PubMed record.