Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2017 Oct 30.
Published in final edited form as: Cell Rep. 2017 Oct 17;21(3):559–569. doi: 10.1016/j.celrep.2017.09.083

A novel cell migration tracking tool supports coupling of tissue rotation to elongation

Dong-Yuan Chen 1, Justin Crest 1, David Bilder 1,2,*
PMCID: PMC5661965  NIHMSID: NIHMS910098  PMID: 29045826

SUMMARY

Cell migration is indispensable to morphogenesis and homeostasis. Live imaging allows mechanistic insights, but long-term observation can alter normal biology, and tools to track movements in vivo without perturbation are lacking. We develop here a tool called M-TRAIL, which reveals migration histories in fixed tissues. Using clones that overexpress GFP-tagged ECM components, motility trajectories are mapped based on durable traces deposited onto basement membrane. We applied M-TRAIL to Drosophila follicle rotation, comparing in vivo and ex vivo migratory dynamics. The rate, trajectory, and cessation of rotation in WT follicles measured in vivo and ex vivo were identical, as was rotation failure in fat2 mutants. However, follicles carrying intracellularly truncated Fat2, previously reported to lack rotation ex vivo, in fact rotate in vivo at a reduced speed, thus revalidating the hypothesis that rotation is required for tissue elongation. The M-TRAIL approach could be applied to track and quantitate in vivo cell motility in other tissues and organisms.

ETOC BLURB

Chen et al. describe M-TRAIL, a new ECM-based tool for tracking cell migration in living tissues, and demonstrate its utility in resolving conflicting models linking migration and morphogenesis in Drosophila egg chambers.

graphic file with name nihms910098u1.jpg

INTRODUCTION

Cell migration is critical to metazoan biology. Developing organisms require massive morphogenetic movements to shape their organs, while tissues that encounter physical or chemical challenges induce cell motility during immune responses, wound healing and regenerative processes to maintain homeostasis. In cancer, cells’ acquisition of the ability to migrate and metastasize is a hallmark of progression. To understand these fundamental biological processes, tools to accurately track and quantitate cell motility in tissues are crucial.

Tracing movements of individual or collectively migrating cells in non-transparent systems in vivo remains a challenge. Advances in ex vivo preparations, allowing an explanted tissue to be imaged for a period of hours, allow motility to be studied at cellular resolution (Shamir and Ewald, 2014). Nevertheless, few current techniques fully recapitulate development or homeostasis over long periods of time. Furthermore, imaging per se can lead to artifacts that may alter biological processes, via heat production, phototoxicity, and ROS generation (Liu et al., 2015). Overall, tools allowing long-term tracing of cell migration routes and velocities in intact but opaque organisms are highly limited.

One biological system that exemplifies the above is the Drosophila follicle (egg chamber). This epithelium-encased organ develops in the adult female abdomen, and undergoes a whole-tissue collective cell migration that spans several days (Bilder and Haigo, 2012; Cetera and Horne-Badovinac, 2015; Spradling, 1993). This migration leads to rotation of the entire organ through multiple rounds of revolution, and has been proposed to drive its elongation from spherical to ellipsoid (Haigo and Bilder, 2011). Because of its long-term nature and internal site, follicle rotation and its dynamics have only been imaged and quantified through explants in ex vivo culture. Whether the dynamics of cell migration differ in vivo remains unknown.

In this report, we develop a tool called M-TRAIL that allows unperturbed tracking of cell migration in vivo, and show its applicability in the follicle and other Drosophila tissues. We further use M-TRAIL to critically evaluate results that have challenged the model coupling follicle rotation to elongation. We show that limitations of ex vivo culture account for the conflict, while the M-TRAIL-derived in vivo results are fully consistent with the model. These data demonstrate the value of M-TRAIL, which should be applicable to a range of other tissues and systems.

RESULTS

M-TRAIL as a tool to trace cell migrations in vivo

A valuable tool to track long-term cell migration in vivo would meet several criteria. First, it should minimally perturb normal physiology. Second, it should allow spatial mapping of past location at cellular resolution. Third, it should allow extraction of temporal information for quantifying migratory dynamics. We moved to develop such a tool, based on the interactions between cells and extracellular matrix (ECM). Cell-ECM interactions are central to migration, as cells in physiological conditions often crawl over an ECM substrate such as the basement membrane (BM); migrating cells can also remodel the local BM environment (Wolf and Friedl, 2011). On the timescale of most cell motility, BMs are relatively static and long lasting.

We exploited these properties in an approach called ‘Matrix-labelling Technique for Real-time And Inferred Location (M-TRAIL)’ (Fig. 1A). In M-TRAIL, migrating cells are engineered to inducibly and constitutively express tagged BM components. Following induction, these components will be secreted and deposited by the cell into the local BM. If an induced cell moves, a trail of relatively stable tagged ECM molecules will be deposited in the BM along the route of migration, and this durable trail can be imaged following fixation. Knowing the onset of BM component deposition allows one to infer average velocity as well as the trajectory of the migrating cell.

Figure 1. M-TRAIL, a novel tool to track cell migration in vivo.

Figure 1

Open in a new tab

(A) Schematic of M-TRAIL strategy. A one-cell clone (red) overexpressing a tagged ECM protein (green) is generated through genetic mosaic techniques. Secreted ECM protein is deposited in local BM, leaving a trail if the cell migrates before fixation.

(B) Implementation of M-TRAIL. Early st. 9 follicle with M-TRAIL induced, 13 hrs phs. Single epithelial clones reveal trajectory of cell migration during follicle rotation. ‘Anti-GFP’ here and below indicates that anti-GFP immunohistochemistry was performed without cell permeabilization. Left panel shows maximum Z projection of confocal stacks. Right panels show rotated views of opacity projection. Scale bar:10 μm.

(C) Genetic components used to generate M-TRAIL in Figs. 1–5. A heat shock-inducible recombinase causes excision of a stop cassette to initiate GAL4 production. GAL4 in turn drives expression of a nuclear RFP and GFP-tagged chain of Collagen IV.

(D) Early st. 9 follicle with non-ECM secreted protein tagged with GFP, 12 hrs phs. No trails are seen.

(E, F) Intracellular GFP-Cg25c can be detected 6 hrs phs, but trails are not evident until 7–8 hrs phs.

(G) Quantitation of M-TRAIL length, measured by percent of follicle circumference. Data are represented as mean +/− SEM.

(H, I) Trails can be seen in follicles from st. 5 to st. 10.

We tested whether the M-TRAIL concept could be successfully implemented, by exploring cell migrations in Drosophila. In this organism, stochastic single-cell expression clones can be generated using heat-induced FLP recombinase (hsFLP) to drive expression of a GAL4 gene separated from a constitutive promoter by cis-paired FRT sites (‘FLPout GAL4’) (Blair, 2003). We paired hsFLP and FLPout GAL4 transgenes with a UAS-driven, GFP-tagged version of one of the two Drosophila Collagen IV chains (GFP-Cg25c; Flybase: Col4a1), along with UAS-Histone RFP (Fig. 1C). Histone-RFP expression identifies the GAL4-expressing clone, although we noted that high levels of intracellular GFP-Cg25c puncta, presumably within the secretory system, were also sufficient.

We first applied M-TRAIL to motile cells in the follicle of Drosophila ovaries. In this organ, epithelial cells migrate collectively, circumferentially around the anterior-posterior axis, in a morphogenetic process known as follicle rotation (Haigo and Bilder, 2011). We heat-shocked females carrying M-TRAIL transgenes and then dissected, fixed, and examined ovaries 13 hours later. At 13 hour post heat shock (phs), a clear trail of GFP in the BM contiguous with the GAL4-expressing clone could be detected, extending in a single direction away from it in a swath extending around most of the circumferential axis (Fig. 1B). This is the axis in which follicle cells migrate in ex vivo preparations, and in which a previous BM-based in vivo assay inferred that migration occurs (Haigo and Bilder, 2011). Trails on the BM were not seen when clones expressed a soluble GFP-tagged secreted protein (preproANF-EMD) rather than GFP-Cg25c (Fig. 1D).

To determine when M-TRAILs could first be detected and how long they persisted, we examined follicles at various time points phs. In st. 8 follicles, which have been undergoing rotation for at least 36 hours, Cg25c-GFP was detected inside the clone at 6 hours phs, but no trails were visible (Fig. 1E). Trails were first detected at 7 hours phs, and increased in length as well as strength with time (Fig. 1F, 1G). Weak intracellular puncta and trails can be seen as early as st.5 follicles within an ovariole 17 hours phs (Fig. 1H). Trails were difficult to detect in st. 4 or earlier follicles, possibly because of the weak efficacy of the act5c promoter during these stages. We explored the stability of follicle M-TRAILs by examining ovarioles at 25 hours phs. Robust trails were detected in st. 10 follicles, which were likely at st. 6 when trail deposition began (Fig. 1H). Overall, these results demonstrate the applicability of M-TRAIL to mapping cell migration.

Exploring criteria for M-TRAIL applicability

We evaluated the parameters within which M-TRAIL is useful for migration analysis by testing cells other than the follicle epithelium. The best-characterized motile cells in Drosophila are the border cells, a specialized subset of follicle cells that delaminate from the epithelium and migrate through the central germline towards the oocyte at st. 9 (Pocha and Montell, 2014). Interestingly, migrating border cells expressing M-TRAIL left only scant and transient trails along their route (Fig. 2A and A′). Since germ cells do not express BM components (Fig. 2B, C), this result suggests that a local BM may be necessary to provide a depot for the ECM components induced by M-TRAIL.

Figure 2. Criteria for M-TRAIL application.

Figure 2

Open in a new tab

(A) In contrast to follicle epithelial cells, migrating border cells expressing GFP-Cg25c do not leave trails.

(B, C) Germline cells lack expression of BM components Vkg and Perlecan.

(D) In larvae expressing GFP-Vkg in hemocytes, unwounded wing imaginal discs (expressing a red fluorophore in the pouch) show only low-level GFP signal in the BM.

(E) 7 hours following wounding, hemocytes (arrowhead) adhere close to wound sites (white dotted line) and are often found next to BM trails of GFP (grey arrow).

(F) The ECM protein SPARC-HA can be used in M-TRAIL, although it diffuses farther from cells than GFP-Cg25c.

We then asked whether only cells that endogenously produce large amounts of BM components are suitable for M-TRAIL analysis. While follicle cells secrete Collagen IV into surrounding BM, imaginal disc epithelia instead construct their local BM predominantly from soluble Collagen IV synthesized by the fat body (Isabella and Horne-Badovinac, 2015a; Pastor-Pareja and Xu, 2011). We tested the ability of discs to secrete M-TRAIL-induced Cg25c-GFP. Staining for extracellular proteins revealed incorporation of disc cell-derived Cg25c-GFP into the local BM (Fig. S1A). Although cells in imaginal discs do not obviously migrate, the M-TRAIL approach appears applicable to cells that do not professionally secrete BM.

We turned to a third tissue to ask whether M-TRAIL might be used for tracking motility in cells outside of the ovary. Drosophila hemocytes are macrophage-like cells that undergo cell migration in response to tissue damage (Pastor-Pareja et al., 2008). We expressed GFP-Vkg specifically in hemocytes and then compared control imaginal discs to those that had been physically wounded in situ. Consistent with the documented production of basement membrane by hemocytes (Lunstrum et al., 1988), unwounded discs displayed a low level of GFP accumulation at their basal surface (Fig. 2D). In wounded discs, linear trails of basal GFP were often associated with hemocytes at the site of damage (Fig. 2E). These results suggest that M-TRAIL can be implemented in additional motile cell types.

Biogenesis, secretion, and incorporation of mature Collagen IV trimers into BM involves a specialized set of proteins, raising the question of whether alternative BM resident proteins might work more efficiently for M-TRAIL (Saito et al., 2009). We compared Cg25c-GFP to SPARC-HA, a tagged version of another secreted protein that is incorporated into the BM, by inducing M-TRAIL clones expressing both (Martinek et al., 2002). In follicles, SPARC-HA appeared to be more efficiently secreted than Cg25c-GFP, as less protein was detected intracellularly, but it appeared to diffuse more from the cell trajectory (Fig. 2F). However, SPARC overexpression can modify local BM properties and alter follicle shape (Isabella and Horne-Badovinac, 2015b). By contrast, Cg25c represents only one component of the obligate heterotrimeric Collagen IV, suggesting that its overexpression would not structurally alter the BM (Van De Bor et al., 2015). Consistent with this hypothesis, overexpression of Cg25c in the entire epithelia does not lead to any evident perturbation of morphogenesis (Fig. S1B–D).

M-TRAIL analysis of WT follicle morphogenesis in vivo

Having established the general utility of M-TRAIL, we applied it to analyze follicle rotation in more detail, comparing the in vivo data to that inferred from other approaches. Rotation has previously been analyzed using ex vivo live imaging, from which linear and angular velocities (ω) have been measured (Cetera et al., 2014; D.-Y. Chen et al., 2016; X. J. Chen et al., 2014; Haigo and Bilder, 2011; Viktorinová and Dahmann, 2013). Current ex vivo culture protocols support follicle viability for a maximum of ~10 hours, representing only a fraction of the 42 hours through which rotation is thought to occur in vivo. We used M-TRAIL to compare in vivo rotation velocities and those ex vivo, analyzing the length of trails in fixed M-TRAIL-induced st. 8 follicles 9 hours phs by using the ‘unrolling’ algorithm ImSAnE (D.-Y. Chen et al., 2016) (Fig. 3A–C). Accounting for the 7 hour onset time of M-TRAIL induction, at st.8 in vivo ω equals 50°/hr, very similar to that measured ex vivo (Fig. 3D). These results demonstrate that in WT follicles, short-term ex vivo conditions recapitulate the speed of cell motility in vivo.

Figure 3. Characterization of follicle morphogenesis in vivo using M-TRAIL.

Figure 3

Open in a new tab

(A) Use of ImSAnE ‘flattening’ algorithm to project 3D follicle epithelium onto 2D plane.

(B, C) Flattened trails present in M-TRAIL follicles cultured for 2 hrs in vivo vs ex vivo.

(D) Rotational angular velocity (ω, calculated by trail length as a fraction of follicle circumference) reveals no difference ex vivo vs in vivo. Data are represented as mean +/− SEM.

(E) Anterior clones (arrow) leave trails that are circumferential during st. 7–8, but then shift toward the posterior after rotation ceases at st. 9. Posterior clones (arrowhead) show rotation but no A-P shift.

(F) St. 9 follicle with two clones showing long term tracing. Note the complete rotations evident; also note the greater posterior shift in the more anterior clone.

In addition to rotation from st. 2 to 8, follicles undergo a morphogenetic event variously termed ‘posterior migration’ or ‘anterior accommodation’ from st. 9 to 10, after rotation ceases (Kolahi et al., 2009; Spradling, 1993). During this process, the anterior follicle epithelium undergoes rapid reorganization: cells alter their morphology from cuboidal to squamous and concurrently elongate anisotropically towards the anterior pole (Brigaud et al., 2015). Interestingly, M-TRAIL captured this movement: in trails initiating at late st. 7, both cessation of rotation and a posteriorly-directed shift were seen in the squamous ‘stretch’ cells at the anterior, while only the former process was seen in the columnar cells at the posterior (Fig. 3E). This result confirms that the anterior ‘stretch’ cells change their position relative to the BM, while posterior cells are relatively static. Moreover, the lengthy (>24 hour) time courses that capture complete rotations as well as anterior morphogenesis (Fig. 3F) far exceed the limits of follicle viability under current ex vivo culture conditions; M-TRAIL uniquely allows their visualization.

M-TRAIL analysis of fat2 mutant follicles

As M-TRAIL can be used to study WT follicle development, we then turned to mutant analysis. The most dramatic perturbation of follicle cell migration occurs in flies mutant for fat2, which encodes an atypical cadherin implicated in planar polarity and symmetry-breaking (D.-Y. Chen et al., 2016; Viktorinova et al., 2009). In fat2 null follicles, rotation ex vivo completely fails, never initiating at st. 2 and not measurably moving throughout st. 8 (Cetera et al., 2014; D.-Y. Chen et al., 2016; Viktorinová and Dahmann, 2013). We tested whether the ex vivo results accurately reflect the in vivo situation by inducing M-TRAIL in fat2 null follicles. Fat2 follicles produced ectopic Cg25c-GFP and deposited it into the BM comparable to WT. However, M-TRAIL clones induced either at st. 5–6 or st. 7–8 did not produce a measurable circumferential trail, even 17 hours phs (Fig. 4A, B). Importantly, posteriorly-directed trails in stretch cells at st. 9 were visible (Fig. S2A), consistent with the observation that acquisition of squamous morphology and anterior follicle morphogenesis are undisturbed by fat2 mutation.

Figure 4. fat2ΔICR follicles rotate in vivo, albeit at a slower rate.

Figure 4

Open in a new tab

(A–D) Maximum intensity Z and ImSAnE-flattened projections of M-TRAIL induced follicles carrying different fat2 alleles. (A) Follicles carrying a WT allele generate long trails; (B) homozygous fat2 follicles show no trails; (C) follicles carrying fat2ΔICR show short trails; (D) follicles carrying fat2ΔTMICR show no trails.

(E) Angular velocity calculated by trail length; fat2ΔICR follicles rotate at 57% the rate of those carrying a WT allele, while fat2ΔTMICR do not rotate at all. Data are represented as mean +/− SEM.

Thus, the lack of circumferential M-TRAILs in fat2 mutants rigorously establishes the absence of rotation in vivo, as previously seen ex vivo.

Follicles lacking the Fat2 intracellular region rotate in vivo

A recent paper interrogated the role of fat2 in rotation and elongation by designing and testing novel alleles, carried on fosmids in a fat2 null mutant background (Aurich and Dahmann, 2016). These authors found that deletion of either the intracellular region (ΔICR) alone or the transmembrane domain in conjunction with the ICR (ΔTMICR) prevented rotation ex vivo. Interestingly, follicles lacking ICR alone (fat2ΔICR) were able to elongate appropriately and achieve late-stage PCP cytoskeletal organization, despite their observed lack of PCP rotation. The authors proposed that this genotype, which apparently uncouples the two processes, demonstrates that PCP rotation is not required for follicle elongation and PCP morphogenesis.

We investigated this proposition using M-TRAIL to analyze rotation of the new fat2 alleles in vivo. We first confirmed that fat2ΔICR follicles elongate normally but fail to rotate ex vivo (Fig. S2B–E). As previously reported, no circumferential rotation could be detected under ex vivo live imaging during a 2 hour course; in this fat2ΔICR follicles resembled fat2 nulls (Fig. S2B). Surprisingly, when we utilized M-TRAIL to visualize in vivo migration, clear circumferentially oriented trails were visible in fat2ΔICR follicles but not fat2 null nor fat2ΔTMICR 13 hours phs (Fig. 4B, C, and D). Interestingly, trails appeared shorter than those seen in WT follicles (Fig. 4A). We therefore tested if fat2ΔICR-carrying follicles have altered rotation dynamics by directly comparing them with heterozygous sibling follicles under the same M-TRAIL induction conditions (Fig. 4E). While st.8 follicles carrying either a WT fat2 allele or a functional fat2-GFP fosmid have ω=44 /hr, follicles carrying only fat2ΔICR were ~50% slower, with ω=25°/hr. We confirmed these results using a second fat2 mutant background (Fig. S2F). This result demonstrates that fat2ΔICR follicles indeed rotate in vivo, although an undefined limitation of current culture conditions does not support ex vivo rotation. Importantly, it further establishes that this genotype does not in fact uncouple rotation from elongation.

Slow-rotating fat2ΔICR follicles have functionally softer but anisotropic BMs

Does the normal elongation of fat2ΔICR follicles establish that even a low rate of PCP follicle rotation is sufficient to generate the tissue organization and mechanical properties required for appropriate morphogenesis? We have recently shown that follicle shape is determined by mechanical properties provided by its fibril-like BM, with key parameters including BM PCP organization and A-P patterned levels (Crest et al., 2017). We analyzed both parameters in fat2ΔICR follicles with ImSAnE. In agreement with previous data, fat2ΔICR follicles generate PCP Collagen IV fibrils but these fibrils are shorter as well as less dense than those seen in WT (Fig. 5A–D) (Aurich and Dahmann, 2016). However, the anisotropic A-P expression gradient of Collagen IV resembled WT (Fig. 5E–F). The mechanical stiffness of the fat2ΔICR follicle BM was then tested, using both Atomic Force Microscopy (AFM) along the A-P and circumferential axes (Fig. 5G) and an osmotic swelling assay in which follicles are placed in distilled water (Fig. 5H). We have previously used these techniques to show that WT follicles display an anisotropic A-P stiffness gradient that directs elongation (Crest et al., 2017). AFM analysis on fat2ΔICR follicle BM showed the existence of an anisotropic A-P stiffness gradient, but revealed that it appears softer than WT (Fig. 5I). Indeed, when challenged with osmotic stress, fat2ΔICR follicles burst more frequently than WT (Fig. 5J). Nevertheless, bursting is restricted to poles (Fig. 5K), consistent with the anisotropic stiffness measurements. Thus, reduced fibril formation and a softer BM correlate with the slower albeit normally-oriented tissue rotation revealed by M-TRAIL. Overall, the results are consistent with the model that a degree of PCP migration of the follicle epithelium, generating PCP fibrils and consistent circumferential stiffness, is indeed essential for appropriate organ elongation.

Figure 5. fat2ΔICR follicles show altered mechanical properties.

Figure 5

Open in a new tab

(A–B) Confocal projections of Collagen IV-GFP in WT and fat2ΔICR follicles.

(C–D) Collagen IV fibril analysis quantitates shorter and less dense fibrils in fat2ΔICR vs control (fat2 heterozygous) follicles. Data are represented as mean +/− SEM.

(E–F) Anisotropic A-P distributions of Collagen IV-GFP are similar between the two genotypes.

(G) Schematic of Atomic Force Microscopy experiments. Follicles are probed at different regions along the A-P axis to measure BM (green) stiffness.

(H) Schematic of bursting assay. Follicles are placed in distilled water to swell tissue; timing and position of BM (green) rupture reflects BM mechanics.

(I) AFM measurements reveal that fat2ΔICR follicles maintain an AP stiffness gradient similar to WT, but display modest softening. Data are represented as mean +/− SEM.

(J) Following osmotic challenge, fat2ΔICR follicles burst more frequently than WT follicles, but less frequently than fat2ΔTMICR follicles. Data are represented as mean +/− SEM.

(K) Like WT, bursting in fat2ΔICR follicles does not occur in the central region, in contrast to fat2ΔTMICR follicles.

DISCUSSION

M-TRAIL tracking of cell migration in vivo

Ex vivo live imaging has led to important discoveries about cell motility during development and homeostasis. Whether explant culture systems fully recapitulate in vivo biology, however, remains an issue, especially for biological processes that span longer time scales. Here we provide a tool to track long-term cell motility in vivo, by quantitating ECM traces stably deposited by migrating cells. The genesis of this strategy lies in a previous experiment in the Drosophila follicle, where large clones endogenously expressing tagged Collagen IV were marked by mitotic recombination in trans and used to demonstrate that follicle rotation organizes BM in vivo (Haigo and Bilder, 2011). M-TRAIL uses cis-recombination to allow temporal control over ECM component secretion, as well as to free induction from cell division, and can be utilized even in natively non-ECM-producing cells. Importantly, the ability to generate single cell clones allows cellular resolution of migratory tracks. Overall, M-TRAIL enables long-term, quantitative measurements of cell migration in fixed tissues and can potentially track other cellular dynamics, such as cell division and rearrangements, under physiological conditions.

Several points are relevant to applying and interpreting M-TRAIL. First, cells must be able to secrete the M-TRAIL protein. Export of Collagen chains, including Collagen IV, can involve specialized machinery that may not be active in all cells (Saito et al., 2009). In our experiments, an alternative ECM protein, SPARC, was more efficiently secreted, but appeared to diffuse farther and also has the ability to alter ECM stiffness. Inactive mutant forms of SPARC or other ECM proteins may be appealing directions for future refinements (Morrissey et al., 2016). Second, M-TRAIL depends on the pre-existence and stability of the BM. In early animals, there may be insufficient BM for incorporation, and throughout development there are migrations that do not involve integrin-ECM interactions (Paluch et al., 2016). Moreover, trail stability is subject to BM remodeling, such as degradation by matrix metalloproteases. BMs can also be more dynamic than commonly appreciated (Loganathan et al., 2016; Morrissey and Sherwood, 2015); this should be considered when plotting speed and trajectory of the source cell relative to the BM and the rest of the organism. Third, it is necessary to determine the lag time to trail detection in the cells under analysis. This includes not just genetic induction but also microscopic visualization. Since the amount of ECM protein deposited will correlate with the cellular ‘dwell time’ over a region of BM, rapid migrations as well as those that occur over short periods of time will require more sensitive detection. Improved secretion, brighter fluorophores, and multiple epitope tags can benefit this. While we have demonstrated M-TRAIL’s value in Drosophila, the relationship of migrating cells to the ECM is conserved throughout metazoans (Brown, 2011), so M-TRAIL should be useful for a wide variety of biological systems.

Coupling of follicle elongation to tissue rotation

The original discovery of tissue rotation in WT follicles, and its failure in non-elongating mutants, led to the proposal that rotation is required for elongation, by creating PCP organization of cytoskeletal and BM elements to form the previously hypothesized ‘molecular corset’ (Gutzeit et al., 1991; Haigo and Bilder, 2011). This initial model has guided numerous subsequent studies, whose data are largely consistent with its formulation. The recent report that fat2ΔICR follicles uncouple rotation and elongation thus presents an important challenge to the model (Aurich and Dahmann, 2016). Here we resolve this discrepancy by showing that fat2ΔICR follicles indeed rotate in vivo, although they do not in ex vivo culture for reasons currently unknown. At this time, we are unaware of any non-rotating follicle genotype that is capable of elongation, consistent with the proposition that rotation is a morphogenetic mechanism that generates the mechanical environment necessary to shape the Drosophila egg.

Importantly, although fat2ΔICR follicles show appropriate aspect ratios, they are not fully normal in several parameters. Migration is slowed, the BM fibrils that are produced are attenuated, and the result is a functionally softer BM than in WT follicles. Although these phenotypes could point to a direct role of the fat2ΔICR in BM secretion, an attractive alternative hypothesis is that they instead reflect a role for tissue rotation rates in constructing BM networks and their mechanical properties. Integrin-mediated traction forces are known to drive a positive feedback loop linking ECM stiffness to cell migration speeds in many systems (Parsons et al., 2010); this conserved molecular mechanism, amplified by the continuous topology of the follicles’ axis of rotation, could drive the construction of a robust BM that instructs elongation.

The demonstration that fat2ΔICR follicles undergo PCP cell migration sheds further light on the activity of Fat2 itself. Fat2 shows PCP localization in rotating follicles and binds to actin polymerization-stimulating WAVE complex proteins; both of these properties are mediated by the ICR (Barlan et al., 2017; Squarr et al., 2016; Viktorinová and Dahmann, 2013). The ability of follicles lacking the Fat2 ICR to rotate demonstrates that a direct Fat2-WRC interaction is not essential for collective cell motility, placing more emphasis on the Fat2 extracellular domain, perhaps in a ligand-like role (see also (Barlan et al., 2017)). This ability also emphasizes that it is PCP rotation, rather than Fat2 localization, that is correlated with proper organization of cytoskeleton and ECM components. Recent work identifies an initial and essential requirement for Fat2 in regulating symmetry-breaking microtubule polarity in the germarium, in a manner independent of actin organization (D.-Y. Chen et al., 2016). The localization of Fat2 during these stages is not known but the ability of fat2ΔICR follicles to break chiral symmetry is consistent with an ECD-mediated role. In general, the use of M-TRAIL to both reveal unexpected migration and quantitate its rate in this genotype, as well as in future manipulations, will guide testable hypotheses for mechanisms that couple rotation of a tissue to the mechanical properties that drive its elongation.

While M-TRAIL uncovered one limitation of ex vivo follicle imaging, reconciling contradictory results, it is important to emphasize that in other aspects it was concordant with prior analyses. In particular, the migration rate, trajectory, and cessation of rotation in WT follicles measured in vivo by M-TRAIL were identical to those observed ex vivo, as was the failure of rotation in fat2 mutants. This validates the general ex vivo approach to study follicle rotation, but points out that live imaging of some mutant genotypes may not accurately reflect what occurs in the animal. The creation of more genetically convenient as well as more sensitive M-TRAIL variants will expand the range of situations in which this new tool can be used.

EXPERIMENTAL PROCEDURES

All fluorescent images were acquired on a Zeiss LSM700 confocal microscope. M-TRAIL follicle clones were induced by 15 or 20 minute heat shocks of 3-day old, fed adult females in a 37°C water bath. Ovaries were dissected and stained after hours post heat shock (phs) indicated on the figures. Over 50% follicles show multiple single clones along the A-P arc and ~10% have non-overlapping single clones. Non-permeablized immunostains were performed without detergent. ImSAnE unrolling and analyses were executed as previously described on full stacks of confocal sections of follicles stained with Alexa Fluor647-conjugated phalloidin (D.-Y. Chen et al., 2016). The phalloidin channel was used to generate point clouds for unrolling. Conditions for ex vivo follicle culture were as previously described (D.-Y. Chen et al., 2016). AFM measurements, bursting assays, and Collagen IV-GFP fibril intensity measurements were done as described (Crest et al., 2017). Fibril density and length were measured as in (Isabella and Horne-Badovinac, 2015b) on the central region of ImSAnE-flattened images. Error bars in charts represents SEM and statistical significance was assessed by Welch’s unequal variances t-test. Additional details are described in Supplemental Experimental Procedures.

Supplementary Material

Figure S1
Figure S2
Supplemental information

HIGHLIGHTS.

  • M-TRAIL allows retrospective tracking of unperturbed cell migration in vivo

  • Fluorescent proteins incorporate into BM, allowing visualization and quantitation

  • In vivo rotation of fly follicle can differ from ex vivo

  • Elongating follicles, even those with aberrant mechanics, rotate in vivo

Acknowledgments

We thank Christian Dahmann, Stephane Noselli, Eduardo Moreno, Katja Brückner, Iswar Hariharan, and the Bloomington Stock Center for sending reagents, Geert de Vreede for assistance in discs wounding experiments, and are grateful to the Bilder lab for helpful discussions. This work was supported by NIH RO1s GM068675 and GM111111.

Footnotes

AUTHOR CONTRIBUTIONS

D.Y.C. and J.D. conducted the experiments; D.Y.C and D.B. designed the experiments and wrote the paper.

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

References

  1. Aurich F, Dahmann C. A Mutation in fat2 Uncouples Tissue Elongation from Global Tissue Rotation. Cell Reports. 2016 doi: 10.1016/j.celrep.2016.02.044. [DOI] [PubMed] [Google Scholar]
  2. Barlan K, Cetera M, Horne-Badovinac S. Fat2 and Lar Define a Basally Localized Planar Signaling System Controlling Collective Cell Migration. Dev Cell. 2017;40:467–477. e5. doi: 10.1016/j.devcel.2017.02.003. [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Bilder D, Haigo SL. Expanding the morphogenetic repertoire: perspectives from the Drosophila egg. Dev Cell. 2012;22:12–23. doi: 10.1016/j.devcel.2011.12.003. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Blair SS. Genetic mosaic techniques for studying Drosophila development. Development. 2003;130:5065–5072. doi: 10.1242/dev.00774. [DOI] [PubMed] [Google Scholar]
  5. Brigaud I, Duteyrat J-L, Chlasta J, Le Bail S, Couderc J-L, Grammont M. Transforming Growth Factor β/activin signalling induces epithelial cell flattening during Drosophila oogenesis. Biology Open. 2015:4. doi: 10.1242/bio.201410785. [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Brown NH. Extracellular matrix in development: insights from mechanisms conserved between invertebrates and vertebrates. Csh Perspect Biol. 2011:3. doi: 10.1101/cshperspect.a005082. [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Cetera M, Horne-Badovinac S. Round and round gets you somewhere: collective cell migration and planar polarity in elongating Drosophila egg chambers. Curr Opin Genet Dev. 2015;32:10–15. doi: 10.1016/j.gde.2015.01.003. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Cetera M, Ramirez-San Juan GR, Oakes PW, Lewellyn L, Fairchild MJ, Tanentzapf G, Gardel ML, Horne-Badovinac S. Epithelial rotation promotes the global alignment of contractile actin bundles during Drosophila egg chamber elongation. Nature Communications. 2014;5:5511. doi: 10.1038/ncomms6511. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Chen DY, Lipari KR, Dehghan Y, Streichan SJ, Bilder D. Symmetry Breaking in an Edgeless Epithelium by Fat2-Regulated Microtubule Polarity. Cell Reports. 2016;15:1125–1133. doi: 10.1016/j.celrep.2016.04.014. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Chen XJ, Squarr AJ, Stephan R, Chen B, Higgins TE, Barry DJ, Martin MC, Rosen MK, Bogdan S, Way M. Ena/VASP proteins cooperate with the WAVE complex to regulate the actin cytoskeleton. Dev Cell. 2014;30:569–584. doi: 10.1016/j.devcel.2014.08.001. [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Crest J, Diz-Munoz A, Chen DY, Fletcher DA, Bilder D. Organ sculpting by patterned extracellular matrix stiffness. eLife. 2017;6:e24958. doi: 10.7554/eLife.24958. [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Gutzeit HO, Eberhardt W, Gratwohl E. Laminin and basement membrane-associated microfilaments in wild-type and mutant Drosophila ovarian follicles. journal of cell Science. 1991;100( Pt 4):781–788. doi: 10.1242/jcs.100.4.781. [DOI] [PubMed] [Google Scholar]
  13. Haigo SL, Bilder D. Global tissue revolutions in a morphogenetic movement controlling elongation. Science. 2011;331:1071–1074. doi: 10.1126/science.1199424. [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Isabella AJ, Horne-Badovinac S. Building from the Ground up: Basement Membranes in Drosophila Development. Current Topics in Membranes. 2015a;76:305–337. doi: 10.1016/bs.ctm.2015.07.001. [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Isabella AJ, Horne-Badovinac S. Dynamic regulation of basement membrane protein levels promotes egg chamber elongation in Drosophila. Dev Biol. 2015b doi: 10.1016/j.ydbio.2015.08.018. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Kolahi KS, White PF, Shreter DM, Classen AK, Bilder D, Mofrad MR. Quantitative analysis of epithelial morphogenesis in Drosophila oogenesis: New insights based on morphometric analysis and mechanical modeling. Dev Biol. 2009;331:129–139. doi: 10.1016/j.ydbio.2009.04.028. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Liu Z, Lavis LD, Betzig E. Imaging live-cell dynamics and structure at the single-molecule level. Mol Cell. 2015;58:644–659. doi: 10.1016/j.molcel.2015.02.033. [DOI] [PubMed] [Google Scholar]
  18. Loganathan R, Rongish BJ, Smith CM, Filla MB, Czirok A, Bénazéraf B, Little CD. Extracellular matrix motion and early morphogenesis. Development. 2016;143:2056–2065. doi: 10.1242/dev.127886. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Lunstrum GP, Bachinger H-P, Fessler LI, Duncan KG, Nelson RE, Fessler JH. Drosophila basement membrane procollagen IV. I. Protein characterization and distribution. J Biol Chem. 1988;263:18318–18327. [PubMed] [Google Scholar]
  20. Martinek N, Zou R, Berg M, Sodek J, Ringuette M. Evolutionary conservation and association of SPARC with the basal lamina in Drosophila. Dev Genes Evol. 2002;212:124–133. doi: 10.1007/s00427-002-0220-9. [DOI] [PubMed] [Google Scholar]
  21. Morrissey MA, Jayadev R, Miley GR, Blebea CA, Chi Q, Ihara S, Sherwood DR. SPARC Promotes Cell Invasion In Vivo by Decreasing Type IV Collagen Levels in the Basement Membrane. PLoS Genet. 2016;12:e1005905. doi: 10.1371/journal.pgen.1005905. [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Morrissey MA, Sherwood DR. An active role for basement membrane assembly and modification in tissue sculpting. journal of cell Science. 2015;128:1661–1668. doi: 10.1242/jcs.168021. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Paluch EK, Aspalter IM, Sixt M. Focal Adhesion-Independent Cell Migration. Annu Rev Cell Dev Biol. 2016;32:469–490. doi: 10.1146/annurev-cellbio-111315-125341. [DOI] [PubMed] [Google Scholar]
  24. Parsons JT, Horwitz AR, Schwartz MA. Cell adhesion: integrating cytoskeletal dynamics and cellular tension. Nature Reviews Molecular Cell Biology. 2010;11:633–643. doi: 10.1038/nrm2957. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Pastor-Pareja JC, Wu M, Xu T. An innate immune response of blood cells to tumors and tissue damage in Drosophila. Dis Model Mech. 2008;1:144–154. doi: 10.1242/dmm.000950. [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Pastor-Pareja JC, Xu T. Shaping cells and organs in Drosophila by opposing roles of fat body-secreted Collagen IV and perlecan. Dev Cell. 2011;21:245–256. doi: 10.1016/j.devcel.2011.06.026. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Pocha SM, Montell DJ. Cellular and molecular mechanisms of single and collective cell migrations in Drosophila: themes and variations. Annu Rev Genet. 2014;48:295–318. doi: 10.1146/annurev-genet-120213-092218. [DOI] [PubMed] [Google Scholar]
  28. Saito K, Chen M, Bard F, Chen S, Zhou H, Woodley D, Polischuk R, Schekman R, Malhotra V. TANGO1 facilitates cargo loading at endoplasmic reticulum exit sites. Cell. 2009;136:891–902. doi: 10.1016/j.cell.2008.12.025. [DOI] [PubMed] [Google Scholar]
  29. Shamir ER, Ewald AJ. Three-dimensional organotypic culture: experimental models of mammalian biology and disease. Nature Reviews Molecular Cell Biology. 2014;15:647–664. doi: 10.1038/nrm3873. [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Spradling AC. Developmental genetics of oogenesis. In: Bate M, Martinez Arias A, editors. The Development of Drosophila Melanogaster. CSHL Press; New York: 1993. pp. 1–70. [Google Scholar]
  31. Squarr AJ, Brinkmann K, Chen B, Steinbacher T, Ebnet K, Rosen MK, Bogdan S. Fat2 acts through the WAVE regulatory complex to drive collective cell migration during tissue rotation. J Cell Biol. 2016;212:591–603. doi: 10.1083/jcb.201508081. [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Van De Bor V, Zimniak G, Papone L, Cerezo D, Malbouyres M, Juan T, Ruggiero F, Noselli S. Companion Blood Cells Control Ovarian Stem Cell Niche Microenvironment and Homeostasis. Cell Reports. 2015;13:546–560. doi: 10.1016/j.celrep.2015.09.008. [DOI] [PubMed] [Google Scholar]
  33. Viktorinova I, Konig T, Schlichting K, Dahmann C. The cadherin Fat2 is required for planar cell polarity in the Drosophila ovary. Development. 2009;136:4123–4132. doi: 10.1242/dev.039099. [DOI] [PubMed] [Google Scholar]
  34. Viktorinová I, Dahmann C. Microtubule polarity predicts direction of egg chamber rotation in Drosophila. Current Biology. 2013;23:1472–1477. doi: 10.1016/j.cub.2013.06.014. [DOI] [PubMed] [Google Scholar]
  35. Wolf K, Friedl P. Extracellular matrix determinants of proteolytic and non-proteolytic cell migration. Trends Cell Biol. 2011;21:736–744. doi: 10.1016/j.tcb.2011.09.006. [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Figure S1
NIHMS910098-supplement-Figure_S1.pdf (630.9KB, pdf)
Figure S2
NIHMS910098-supplement-Figure_S2.pdf (2.8MB, pdf)
Supplemental information
NIHMS910098-supplement-Supplemental_information.pdf (139.1KB, pdf)

RESOURCES