Imaging and Analysis of Tissue Orientation and Growth Dynamics in the Developing Drosophila Epithelia During Pupal Stages

Summary

This protocol is designed for the imaging and analysis of the dynamics of cell orientation and tissue growth in the Drosophila abdominal epithelia as the fruit fly undergoes metamorphosis. The methodology described here can be applied to the study of different developmental stages, tissues, and subcellular structures in Drosophila or other model organisms.

Abstract

Within multicellular organisms, mature tissues and organs display high degrees of order in the spatial arrangements of their constituent cells. A remarkable example is given by sensory epithelia, where cells of the same or distinct identities are brought together via cell-cell adhesion showing highly organized planar patterns. Cells align to one another in the same direction and display equivalent polarity over large distances. This organization of the mature epithelia is established over the course of morphogenesis. To understand how the planar arrangement of the mature epithelia is achieved, it is crucial to track cell orientation and growth dynamics with high spatiotemporal fidelity during development in vivo. Robust analytical tools are also essential to identify and characterize local-to-global transitions. The Drosophila pupa is an ideal system to evaluate oriented cell shape changes underlying epithelial morphogenesis. The pupal developing epithelium constitutes the external surface of the immobile body, allowing long-term imaging of intact animals. The protocol described here is designed to image and analyze cell behaviors at both global and local levels in the pupal abdominal epidermis as it grows. The methodology described can be easily adapted to the imaging of cell behaviors at other developmental stages, tissues, subcellular structures, or model organisms.

Introduction

To achieve their roles, epithelial tissues fully rely on the spatial organization of their cellular components. In most epithelia, cells are not only packed against each other to create a precise cobblestone layer but they orient themselves relative to the body axes.

The functional importance of precise tissue organization is obvious in sensory epithelia, such as the vertebrate inner ear and retina. In the first case, hair and supporting cells align in a specific axial direction to efficiently sense mechanical inputs such as sound and motion^1,2. Similarly, photoreceptor cell spatial organization is essential for achieving optimal optical properties by the retina³. Spatial control of cell position and orientation is thus of particular relevance for proper physiological function.

Drosophila is a holometabolous insect that undergoes a complete transformation of its larval body structures through metamorphosis, giving rise to its adult tissues. The Drosophila pupa is an excellent model for the noninvasive live imaging of a variety of dynamic events, including developmental cell migration⁴, cell division and growth dynamics⁵, muscle contraction⁶, cell death⁷, wound repair⁸, and cell orientation⁹. In the adult Drosophila, the external epithelium shows a high degree of order. This is easily observed on the arrangements of trichomes (i.e., cell protrusions originating from single epithelial cells) and sensory bristles all over the fly's body surface¹⁰. Indeed, trichomes are aligned in parallel rows guiding airflow¹¹. The morphogenesis of the adult epithelia and the ordered arrangement of the individual cells starts during embryogenesis and culminates during pupal stages. While in embryos cell divisions, intercalations, and shape changes all decrease tissue order^12,13, this is reverted at later stages of development, especially at pupal stages, when the fly approaches maturity⁹.

The immobile Drosophila pupa provides an ideal system to evaluate cell shape and orientation changes. The pupal abdominal epidermis presents special advantages. While the precursors of the adult head, thorax, genitalia, and appendages grow and get patterned from larval stages, the histoblasts, which are integrated into the larval epidermis, start growing and differentiating only at pupariation¹⁴. This feature allows the tracking of all spatiotemporal events involved in the establishment of tissue order in its entirety⁹.

Histoblasts are specified during embryonic development at contralateral positions in each presumptive abdominal segment. The dorsal abdominal epidermis of the adult derives from dorsolaterally located histoblast nests present at the anterior and posterior compartments^15,16. As histoblasts expand, replacing the larval epithelial cells (LECs), the contralateral nests fuse at the dorsal midline forming a confluent sheet^17,18,19,20.

This work describes 1) a methodology for dissection, mounting, and long-term live imaging of the Drosophila pupae, and 2) analytical methods to study the dynamics of cellular orientation and growth at high spatiotemporal resolution. A detailed protocol is provided here, covering all the steps required from the initial pupae preparation (i.e., staging and imaging) to the extraction and quantification of directionality and orientation features. We also describe how to infer local tissue properties from the analysis of cell clones. All the steps described are minimally invasive and allow long-term live analyses. The methods described here can be easily adapted and applied to other developmental stages, tissues, or model organisms.

Protocol

NOTE: This protocol is divided into five steps: (1) staging the pupae, (2) preparing the pupae for imaging, (3) live imaging of the growing abdominal epithelia, (4) generation of genetic mosaics, (5) data processing and analysis (including sections describing how to analyze cell orientation dynamics from cell junction outlines and growth dynamics from cell clones).

1. Staging of Drosophila pupae before imaging

1. Culture flies of the appropriate genotype on standard medium in plastic vials at 25 °C for 5 days (±12 h) after egg laying (AEL).
   NOTE: Metamorphosis starts within the confinement of the third-instar larvae into the pupal case at 120 h AEL until 0 h after puparium formation (APF). This transition is easily identifiable, because larvae stop feeding and moving and the opercular region is formed (Figure 1A) at 0-12 h APF. The puparium is initially soft and white, but progressively hardens and tans.
2. Transfer white prepupae (0 h APF) to a fresh plastic vial using a moisten paintbrush. Animals can be kept at different temperatures depending on the designed experiment until the desired age.
   NOTE: Pupa formation (i.e., pupation) occurs at 12 h APF, when the head and the appendages of the adult fly are totally everted (Figure 1A). By this time the pupal case is fully separated from the pupa, allowing its complete removal (Figure 1A).

2. Preparing pupae for live imaging

NOTE: After staging, the pupae are dissected and mounted as described below (see also Figure 1).

1. Remove staged pupae from the wall of the vial with the help of the forceps.
2. Glue the ventral side of each pupa on a glass slide covered with double-sided sticky tape. Gently tap on the head spiracles (i.e., opercular region) and the dorsal surface of the pupa with the tips of the forceps to assure the adhesion of the pupal case to the tape (Figure 1A,B).
   NOTE: The dorsal surface should face up to facilitate the dissection of the case and the recovery of the pupa.
3. Begin dissection under a stereomicroscope by gently removing the operculum from the puparium with the forceps (Figure 1C).
4. Insert one tip of the forceps in a shallow angle between the pupal case and the pupa surface through the opercular opening. Tear the case from head to tail laterally in one or more swings, avoiding pinching the pupa (Figure 1D). Fold back the cracked pupal case to the lateral sides as you keep proceeding to the posterior end (Figure 1E).
   NOTE: The pupal case is quite rigid and cracks easily. In case of high humidity or in particular genotypic backgrounds, the pupal case becomes softer and tearing is more difficult. In these cases, the cracking of the pupal case can be helped by pricking its free edges with both tips of the forceps.
5. Remove the pupa from the opened-up pupal case by carefully inserting the forceps under the animal and gently pulling up (Figure 1F). The pupa will stick to the tip of the forceps (Figure 1G−H).
6. Transfer the pupa with the help of the forceps to a glass-bottom dish and deposit it on top of a small drop of gas-permeable halocarbon oil (Figure 1I). Hold the pupa gently by the ventral side to avoid any possible tissue damage.
   NOTE: The drop of halocarbon oil has to be small, with a diameter approximately less than half the length of the pupa. Such amount is sufficient to adhere the pupa to the glass by capillarity and to correct the optics for oil immersion objectives.
7. Roll a piece of wet tissue paper at the edges of the dish to maintain humidity. Cover the dish to avoid dehydration of the pupae during imaging.
   NOTE: Both female and male pupae can be employed for imaging. We recommend employing the third abdominal segments (AIII) as a reference abdominal metamere since it is almost identical in both sexes in terms of size, shape, and patterning.

3. Live imaging of growing abdominal epithelia

NOTE: An inverted laser scanning confocal microscope equipped with a 40x/1.3 NA oil immersion objective was used to image pupae at different developmental stages.

1. Orient the pupa over the oil drop on the glass-bottom dish according to the domain and the process to be evaluated (e.g. dorsolaterally for long-term live imaging of the early expansion of the dorsal nests, or dorsally to image their late expansion and tissue remodeling). See Figure 1J,K and Figure 4.
2. Transfer the glass-bottom dish containing the mounted pupae to the microscope stage and focus on the surface of the abdominal area using the transmitted light.
   NOTE: Even if this protocol is optimized for imaging on inverted microscopes, it is also possible to perform imaging on an upright microscope. In that case, the sample is placed on the microscope stage with the glass-bottom surface facing up. The halocarbon oil holds each pupa on a meniscus.
3. Set the acquisition parameters: 1) the number of Z-slices are usually between 20−40 to allow appropriate two-dimensional (2D) reconstruction of the abdominal epidermis; 2) step size between each slice (e.g., 1 micron); 3) time interval for recording (a 5 min interval is suitable for high fidelity analyses of cell orientation dynamics); and 4) frame resolution (e.g., 1024 x 1024).
4. Turn on the appropriate lasers (i.e., 488 nm and 561 nm to visualize GFP and RFP fluorophores respectively) and adjust the laser power and gain/offset settings to visualize the marked cells. Use the lowest possible laser power (in the range 5%−20%) to minimize photobleaching and phototoxicity.
5. Manually set the position and the appropriate Z-stack limits for multiple pupae using the attached motorized stage and the microscope multiposition acquisition software.

4. Generation of genetic mosaics to follow behaviors of cell clones

NOTE: We employ mitotic recombination to induce genetic mosaics in the abdominal epithelium via site-specific recombination (FLP/FRT system^21,22) (Figure 2).

1. Cross virgin females carrying a heat shock-inducible Flippase transgene (hs-FLP), a FLP recognition target (FRT) site at a specific genomic location (e.g., FRT site at position 40A at the L arm of chromosome 2), and a recognizable cellular marker (e.g., Ubi-RFP.nls or Ubi-GFP.nls) distal to the FRT site, to mutant males carrying an FRT site at the equivalent genomic location (Figure 2A−C).
   NOTE: Autonomous and nonautonomous effects within or outside clones for any gene loss of function could be studied employing specific recessive alleles distal to the FRT site.
2. Generate FLP/FRT somatic clones in the histoblasts by heat shock treatment at the third instar larval stage of the progeny of the cross. This is performed by submerging the plastic vials containing the animals in a water bath at 37 °C for 45 min to 1 h at the wandering larvae (LIII) stage.
   NOTE: The sensitive period for mitotic recombination is the G2 phase of the cell cycle. Histoblasts are arrested in G2 during the whole larval development.
3. Score twin clones for absence (i.e., mutant cells) or enhanced (i.e., wild type twin-spot cells) levels of the fluorescent protein marker from 16 h APF onward (Figure 2D).
   NOTE: On average, 45 min to 1 h at 37 °C renders approximately only 2−3 twin clones per region of interest (e.g., the abdominal hemisegment). Pupae showing too a high clone density (e.g., more than four twin clones per hemisegment) should be discarded from further quantitative analyses.
4. Upon clone identification, image living pupae at the desired stage and for the desired length of time as described in the previous section.

5. Data processing and analyses

NOTE: Data are processed using ImageJ (imagej.nih.gov/ij/).

1. Distinguish cell orientation dynamics from cell junction outlines.
   1. Project the Z-stack slices acquired by confocal microscopy in 2D using the Maximum Intensity Projection (MIP) function of ImageJ.
      NOTE: The number of slices per stack should be kept to a minimum to avoid the out of focus noise generated by macrophages patrolling under the epidermis.
   2. Set a planar coordinate system identifying reliable tissue landmarks (e.g., A/P compartment boundaries) for analysis of each dataset (Figure 3A).
      NOTE: Employing the same planar references for each data set will allow multiple measurements to be compared.
   3. Use the OrientationJ plugin (bigwww.epfl.ch/demo/orientation/) of ImageJ^23,24 on local cell edges to obtain qualitative and quantitative orientation values. The plugin renders color-coded overlays on input images based on local orientations and provides numerical values when used in quantitative mode (Figure 3B−B'' and Figure C−C'''').
      NOTE: The OrientationJ plugin is based on structure tensors, 2 x 2 matrix of eigenvalues derived from gradient and directional derivatives (See references^23,24 for a detailed description).
   4. Use the OrientationJ Distribution option to color-code cell edge orientations relative to the set planar coordinate system (i.e., cell edge orientation maps) (Figure 3B). The Distribution option is found under the plugin menu in ImageJ. The settings to employ are: Gaussian window sigma = 1 pixel; Cubic Spline = Gradient; Minimum Coherency = 20%; Minimum Energy = 1%. Cell edge orientations are displayed as a color-coded image employing the Color Survey option of the plugin (Hue = orientation; Saturation = coherency; and Brightness = input image).
      NOTE: Areas that contain background fluorescence do not provide any directional information and must be manually excluded from the analysis. High threshold settings reduce the pixels considered in the processed images.
   5. Use the OrientationJ Measure option to quantify cellular orientations and directional cell-cell alignment (i.e., coherency) (Figure 3C). The Measure option is found under the plugin menu in ImageJ. Generate small adjacent non-overlapping regions of interest (ROIs) of uniform weight (64 x 64 pixels, about 20 µm x 20 µm) within the area occupied by the histoblasts (Figure 3C).
   6. Calculate the dominant local orientation (i.e., the averaged orientation between neighboring cells-averaged cell edge orientation map) and coherency from the ROIs. The software computes the predominant orientation and local coherency within each ROI (Figure 3C).
      NOTE: The largest and smallest eigenvalues of the structure tensor estimated by OrientationJ are employed to calculate coherency as the ratio between their difference and their sum. Coherency is bounded between values of 0−1. A value of 1 indicates full alignment uniformity, while a value of 0 indicates isotropic areas with no alignment.
   7. Statistically analyze the calculated orientation and coherency values from multiple images using free software packages such as PAST²⁵. Axial data such as orientation values (in degrees) are appropriately described by directional statistics.
   8. Through the software, calculate the mean direction and circular variance for each set of orientation distributions. The statistical significance of the difference between the distribution of the orientations between different genotypes or conditions is determined using the nonparametric Mardia-Watson-Wheeler test (W-test) for equal distributions.
   9. Calculate the statistical significance of the difference in coherency applying the non-parametric Kolmogorov-Smirnov test (K-S test). Display data graphically as desired (polar plots, bar charts, box plots, etc.).
2. Growth dynamics from cell clones
   NOTE: The following steps allow retrieval of geometrical and shape parameters for cells from 2D MIP images containing the clone(s) of interest. For comparisons between multiple clones, images must be acquired with the same settings.
   1. Segment clone areas by drawing their contours with the ImageJ Freehand Selection tool.
   2. Calculate geometrical and shape parameters by using the Set Measurements tool of ImageJ under the Analyze menu. Activate the Area, Perimeter, Fit Ellipse, and Shape Descriptor options.
      NOTE: This will allow retrieval of diverse geometrical parameters including the area (sum of the pixels within the clone), the perimeter (sum of the pixel of the clone border), the aspect ratio (AR, the ratio between the major and minor axes of the best-fit Legendre ellipse inscribed into the clone border), and the angle of orientation (i.e, the angle of the major axis of the clone relative to the anteroposterior boundary).
   3. Calculating non-dimensional ratios from these measurements retrieves shape parameters. These include roundness (4 x [area]/π x [major axis]²), roughness (solidity - area/convex area), and circularity (4π x [area]/[perimeter]²).
      NOTE: Each of the shape parameters represents the degree of deviation of the clones from ideal shapes such as a circle or a bounded convex hull, and they are all bounded between values of 0 and 1. Values equal to 1 indicate maximum symmetry (i.e., minimal complexity).
   4. Statistically analyze geometrical and shape parameters between different genotypes or conditions using Microsoft Excel and/or PAST. Statistical significance of the difference is determined using an unpaired two-tailed Student's t-test for equal mean or the non-parametric Kolmogorov-Smirnov K-S-test for equal distributions between conditions. Data can be graphically displayed as desired (e.g., bar charts, box plots, etc.). See Figure 5.

Results

The protocol described above covers the preparation of Drosophila pupae for long-term live imaging and the procedures for the analysis of cell orientation and growth dynamics of the abdominal epidermis. By applying this methodology it is possible to generate high-resolution movies of the developing pupae for periods of up to 48 h without significant photobleaching or phototoxicity. Snapshots depicting the abdominal epidermis (e.g., histoblasts and LECs) at different time points a...

Discussion

Long-range order is an essential characteristic of most functional physiological units. During morphogenesis, order is achieved through the integration of complex instructions implemented with high temporal and spatial precision. Multiple and multilevel constrains are integrated into stereotyped tissue arrangements.

Polarity and directionality are critical to ordered spatial arrangement during development. Polarity implies symmetry breaking during development. The achievement of asymmetry is n...

Materials

Name	Company	Catalog Number	Comments
Analysis Software	-	ImageJ	Analyzing data
Drosophila	Atpa::GFP	-	Strains employed for data collection
Drosophila	hsflp1.22;FRT40A/FRT40A Ubi.RFP.nls	-	Strains employed for data collection
Dumont 5 Forceps	FST	11251-20	1.5 mm diameter for dissection
Glass Bottom Plates	Mat Tek	P35G-0.170-14-C	Mounting pupae for data collection
Halocarbon Oil 27	Sigma-Aldrich	9002-83-9	mounting pupae
Inverted Confocal microscope	Zeiss	LSM700	Data collection
Stereomicroscope	Leica	DFC365FX	Visualization of the pupae during dissection