Protocol for lattice light-sheet time-lapse imaging of early post-implantation mouse embryos

Summary

Lattice light-sheet microscopy offers unprecedented spatial and temporal resolution for visualizing morphogenetic and physiological processes, while minimizing photodamage. Here, we present a detailed protocol for time-lapse imaging of post-implantation mouse embryos, using lattice light-sheet microscopy. We describe steps for embryo isolation, mounting and culture, setting up of imaging parameters, and pipelines for processing the data generated in preparation for downstream analyses. This approach is also suitable for stem cell-derived embryo models, organoids, and small organ explants.

For complete details on the use and execution of this protocol, please refer to Thowfeequ et al.¹ and Stower et al.²

Graphical abstract

Highlights

• •Steps for dissecting post-implantation mouse embryos
• •Mounting post-implantation mouse embryos for lattice light-sheet imaging
• •Lattice light-sheet live imaging of cellular behavior in developing mouse embryos
• •Post-processing lattice light-sheet imaging data

Publisher’s note: Undertaking any experimental protocol requires adherence to local institutional guidelines for laboratory safety and ethics.

Lattice light-sheet microscopy offers unprecedented spatial and temporal resolution for visualizing morphogenetic and physiological processes, while minimizing photodamage. Here, we present a detailed protocol for time-lapse imaging of post-implantation mouse embryos, using lattice light-sheet microscopy. We describe steps for embryo isolation, mounting and culture, setting up of imaging parameters, and pipelines for processing the data generated in preparation for downstream analyses. This approach is also suitable for stem cell-derived embryo models, organoids, and small organ explants.

Before you begin

During mammalian embryonic development, the first few days after implantation are marked by key morphogenetic events during which the anatomical axes, the definitive germ layers and the first organs are established.¹ These events are driven by dynamic cellular behaviors and subcellular processes, which need to be visualized and characterized in four-dimensions including the three spatial and the time domain. Understanding their mechanistic underpinning requires perturbing candidate molecular and mechanical regulators and recording how these behaviors are affected. Although developmental processes have been visualized at high spatiotemporal resolution in many other model systems such as C. elegans, Drosophila or zebrafish,^3,4,5 the mouse post-implantation embryo lags behind, due to its in utero development, relatively low transparency and high sensitivity to photodamage.⁶ While advances in culture conditions,^7,8 reporter alleles^9,10,11 and imaging technology¹² has made imaging mouse embryogenesis possible, many of these approaches are still limited by the low number of embryos that can be imaged in one experiment and photodamage to the embryo during imaging. In particular, the latter severely limits our ability to capture fast events at high resolution, occurring through a large three-dimensional volume of the embryo over a long duration of time.

Here we present a protocol to image post-implantation mouse embryos using Lattice light-sheet microscopy (LLSM). We utilize the ZEISS LLSM L7, that generates a thin light-sheet derived from two-dimensional optical lattices of interfering Bessel beams.¹³ Owing to the selective plane illumination, light-sheet imaging in general results in lower illumination overhead leading to decreased photobleaching and increased viability when compared to confocal microscopy or epi-illumination widefield approaches.¹⁴ LLSM provides a resolution advantage over conventional (Gaussian) light-sheet approaches owing to reduced sheet thickness. This, combined with fast dual-camera based acquisition, makes LLSM well suited to imaging highly dynamic and photo-sensitive processes with high spatial and temporal resolution (see Methods videos S1 and S2). This protocol also offers the advantage of concurrently imaging several embryos, with relatively simple mounting approaches. This is particularly important given the biological variability between embryos at these early stages,¹⁵ which necessitates building an ‘average’ cellular behavior map for embryos at any given stage for effective downstream analyses. For example, we have previously used this protocol to image the migration of anterior visceral endoderm cells necessary for establishing the anterior-posterior axis and characterized how basal projections produced by the migrating cells and overall migratory behavior are affected in mutants of the semaphorin-plexin pathway.¹ This protocol offers a blueprint for live imaging the mouse embryo using LLSM but can also be easily extended with minor optimization steps, to visualize sub-cellular events such as actomyosin-, membrane- and chromosomal-dynamics as well as trafficking and signal propagation in specific regions or cells of the embryo [e.g., Stower et al., 2024²; see Methods video S2] or multicellular embryo-like models.

Methods video S1. Slow time-lapse imaging of dynamic cellular behavior within the embryo, related to expected outcomes and Figure 6C

A 201-minute excerpt from a movie of maximum intensity projections of a gastrulating 7.5 dpc mouse embryo ubiquitously expressing membrane-tdTomato (in gray) with mosaic labeling of cells with membrane-GFP (in color) using a tamoxifen-inducible ROSA-CreERT2, imaged for 6 hours at 3-minute intervals.

Methods video S2. Fast time-lapse imaging of cellular events within the embryo, related to expected outcomes and Figure 6F

A 350 ms excerpt from a movie of the anterior region of an 8.0 dpc mouse embryo ubiquitously expressing the Ca²⁺ reporter GCaMP6f (in color) and membrane-tdTomato (in gray), imaged for 10 s at 5 ms intervals, showing the propagation of a Ca²⁺ wave laterally across the cardiac field and the subsequent contraction and relaxation of the tissue.

Institutional permissions

The necessary permissions and approvals need to be obtained from the relevant institutions before performing any experiments involving regulated species. All animal experimentation procedures outlined here were in full accordance with the UK Animals (Scientific Procedures) Act 1986, approved by Oxford University’s Biological Services Ethical Review Process and were performed under UK Home Office project licenses by personal license holders.

Setting up timed mating

Timing: 7–10 days

• 1.Plan experiments well in advance to make sure enough mice with the desired reporter alleles are available to set up for timed mating.

Note: Females from outbred mouse strains such as CD1, with genetically diverse backgrounds, can be crossed with males homozygous for desired fluorescent reporters to ensure large litters, so that healthy undamaged embryos can be chosen for live imaging experiments. Alternatively, wildtype embryos can be used with various live dyes that allow membranes or nuclei to be labeled for live imaging (see expected outcomes).

• 2.Pair stud males and females from the desired strains and genetic backgrounds close to the start of the dark period of the facility where the mice are housed.

Note: Choosing females that are in proestrus (i.e., exhibiting a swollen, moist, pink vaginal opening) for setting up plugs, will increase the chance of a successful pregnancy.¹⁶ Female mice between 6–18 weeks of age and male mice between 8–32 weeks should be used to set up mating pairs for optimal litter sizes.

• 3.Check the females early next morning for the presence of a vaginal plug.

Note: If no plug is present the female can be left with the male and checked every morning until a plug is seen. Noon of the day a vaginal plug is recorded is assumed to be 0.5 days post coitum (dpc) from which future embryonic stages can be calculated. All the embryonic ages mentioned in this protocol are based off this rule but variation in developmental stage can occur between litters due to the exact timing of the plug or even within litters.

Preparation and equilibration of dissection and culture medium

Timing: 1.5 h

Note: The amount of culture medium described here is more than sufficient for one experiment to image up to fifteen 5.5 dpc embryos (fewer for later stages) in a single imaging chamber for up to 8 hours. The surplus can be used to wash the embryos and fill the other wells to minimize evaporation and help maintain humidity within the sample carrier.

CRITICAL: Since no antibiotics are used, extra care must be taken to ensure that infection risks are minimized. Embryo Culture Medium should be prepared in a laminar flow hood or biosafety cabinet and any equipment used should be cleaned with 70% ethanol.

• 4.
  Prepare 4 mL of Embryo Culture Medium fresh for each experiment.

  • a.
    Clean the hood and all the components needed with 70% ethanol, before transferring them into the hood.
  • b.
    Add 2 mL of CMRL to a 15 mL tube with a screw cap.
  • c.
    Add 2 mL of Knock Out serum.
  • d.
    Add 42 μl of 200 mM L-Glutamine.

    Note: L-glutamine can be aliquoted and frozen. Before using a new aliquot, defrost it in a 37°C water bath or hybridization oven, and vortex the tube, to make sure that all crystals are in solution. Otherwise, L-glutamine crystals can cause imaging artifacts due to light refraction and scattering.

    Note: For later stage embryos (6.5+ dpc) other Embryo Culture Medium composition can be used (e.g., 50% DMEM-FluoroBrite with 50% rat serum).

  • e.
    Vortex to mix all the components together.
  • f.
    Leave the tube in a humidified incubator at 37°C and 5% CO₂ for at least 1 hour before use, with the screw cap loosened to allow gas exchange and adequate equilibration.
• 5.Remove an aliquot of the M2 dissection medium (∼8 mL per litter) from the fridge and warm it up to room temperature (22°C).

Assembly and pre-equilibration of the imaging chamber

Timing: 20 min and 1 h (for pre-equilibration)

CRITICAL: Wear the necessary protective equipment (safety specs and gloves) to prevent injury from glass shards. Clean the benchtop and the microscope stage at the end to make sure no glass is left. Since these steps cannot be carried out in a hood, make sure that the tools used are cleaned to minimize risk of infection.

• 6.
  Prepare the mounting implements.

  • a.
    Light a Bunsen flame with a well-defined blue cone.
  • b.
    Hold the glass capillary securely at both ends and gently roll it between the fingers, while holding its center horizontally over the tip of the blue cone of the flame.
  • c.
    As the center of the capillary becomes malleable, remove it from the flame and evenly and quickly pull from both ends. Cooling of the capillary will prevent further pulling.
  • d.
    Isolate the central pulled region by gently snapping the ends off and collect the pulled capillaries in a sample vial or 15 mL tube.

    Note: Practice is needed to determine the appropriate delay before, and the speed during, pulling. The width of the pulled capillaries can be determined by changing these parameters, making finer ones to mount earlier stage embryos.

  • e.
    Using a diamond knife make marks on a glass slide that are roughly the width of the imaging chamber.
  • f.
    Using the marks as a guide, hold the capillary firmly against the glass and use the diamond knife to break it into fragments that are roughly the same length.
• 7.
  Assemble the imaging chamber.

  CRITICAL: The chambers can be assembled on the day of the experiment. Throughout care must be taken to not scratch or dirty the bottom of the imaging chamber. This can be achieved by always handling it on a clean Petri dish lid.

  • a.
    Take an 8–chambered slide and fill the four end wells with 400 μl of pre-equilibrated culture medium each.
  • b.
    Fill a 5 mL syringe with vacuum grease and attach a 200 μl tip at its end as a nozzle. Compress the grease to make sure there are no air pockets.
  • c.
    Using a dissection microscope, extrude the grease to make two barriers at the top and bottom of the central (unfilled) wells (see Figure 1A) taking care not to smear grease on the center of the well.

    Note: The four central wells can be prepared in the same way, so that the best well can be later chosen for mounting and imaging. Depending on the clearance between the objective and the stage of the dissection microscope, tilting the slide can help apply the grease with more control.

  • d.
    Using fine tweezers, pick up a pre-cut capillary (step 3e–f) and fill it with culture medium from one of the filled wells. This can be done by immersing just the bottom of the capillary into the liquid and allowing the liquid to seep in through capillary action, while allowing air to be displaced from the opposite end.

    CRITICAL: Refraction and light scattering can interfere with imaging if there are air bubbles trapped within the glass capillaries, or if excess grease is smeared on the imaging surface in a region where the light path will cross. Air bubbles can also cause the capillaries to move during long term imaging. Air bubbles can be avoided by filling the glass capillaries by vertically dipping just one end of the capillary into the medium, so that air escapes from one end as the capillary slowly fills through capillary action.

  • e.
    Transfer two filled capillaries into a prepared central well (steps 3 c–d) embedding their ends within the grease and making sure the capillary lies flat against the glass bottom.
  • f.
    By gently pushing the capillaries towards or away from each other, adjust the distance between the capillaries to roughly the width of the embryos, so that they can sit snugly between them (see Fig B and B′).

    Note: When embedding larger embryos (7.0–7.5 dpc) the capillaries can be positioned and set slightly above the glass, so that the ectoplacental cones can be tucked underneath the capillaries (see Fig C and C′)

    Note: Making sure the capillaries are positioned as straight as possible (parallel to the side walls of the chamber) is important, so that the maximum possible volume of the embryo can be imaged using the least number of optical sections, and hence within the least amount of time.

  • g.
    Fill the wells containing the capillaries with 400 μl of culture medium each.
  • h.
    Return the sample carrier back to a humidified incubator at 37°C and 5% CO₂, for at least an hour or until needed. During this time the grease will set and help secure the glass capillaries in position.

Figure 1.

Mounting embryos for lattice light-sheet imaging

(A) Setting up of the 8-chambered slide for mounting embryos. The central wells are used for mounting the embryos for imaging while the holding wells are used for pre-culture.

(B) Image of 5.5 dpc embryos mounted between glass capillaries for imaging along the anterior-posterior axis.

(C) Image of 7.5 dpc embryos mounted beneath a glass capillary positioned with different orientations of the anterior-posterior axis, for imaging along the proximal-distal axis. B′ and C′ are magnifications of the boxed region from B and C respectively. The numbers in B′ indicate the order in which the embryos could be imaged to ensure minimal stage movement between tiles and time points. B″ and C″ are illustrations showing the position of a single embryo from B′ and C′ relative to the coverglass and the glass capillary. Scale bars represent 100 μm. ex- abembryonic half, em- embryonic half, epc- ectoplacental cone, pac- proamniotic cavity, ac- amniotic cavity, ec- exocoelemic cavity, am- amnion, al- allantois.

Setting up the microscope for imaging

Timing: 5 min and 30 min

• 8.
  Pre-set the microscope for imaging.

  • a.
    Turn on CO₂ supply to the microscope so that humidified mixed air can be supplied to the imaging chamber.
  • b.
    Turn on all other components of the Lattice light-sheet microscope and the computer, according to facility instructions/preference.
  • c.
    Open the software controlling the microscope (ZEN Blue for the ZEISS Lattice Lightsheet 7; version 3.10 or later).
  • d.
    Perform any automatic hardware initialization and stage/focus calibration necessary when prompted to do so on the software while there is no sample carrier in place.

    CRITICAL: Make sure there are no obstruction in the way while stage calibration is underway, otherwise it could cause significant damage to the microscope.

  • e.
    Make sure that the correct sample carrier holder is securely positioned on the tray, and that the lid is placed correctly.
  • f.
    On the software, select “Chamber 8x – 58 x 26 mm” as the Sample Carrier template.
  • g.
    Using the inbuilt software controlling the temperature unit and air mixer, set the lid, tray and base temperature to 37°C, and the atmosphere to 5% CO₂ and 65% humidity.
  • h.
    Fill the container at front of the microscope with distilled water, to make sure enough water in present for the motorized autoimmersion dispenser to last the total length of the experiment.
  • i.
    Leave to equilibrate for at least 30 minutes.

    CRITICAL: If the microscope is not equilibrated properly, components could expand due to changing temperature while imaging, leading to drift and poor-quality data.

Key resources table

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Chemicals, peptides, and recombinant proteins
M2 medium	Sigma-Aldrich	M7167
CMRL-1066 medium	PAN-Biotech, UK	P04-84600
KnockOut serum replacement	Gibco	10828-010
L-glutamine	Sigma-Aldrich	G7513
Dow Corning high vacuum grease	SLS, UK	Z273554-1EA
Pen/Strep 10,000 U/mL (optional)	Gibco	15140-122
Cell-Tak (optional)	Corning	354240
SiR-Act (optional)	Spirochrome	SC001
SiR-DNA (optional)	Spirochrome	SC007
Isopropanol	Merck	24137
Ethanol	Merck	32221
Experimental models: Organisms/strains
Inducible reporter transgenic mouse lines (e.g., mTmG10); females	Jackson Laboratory	Strain #: 007676
Ca2+ reporter transgenic mouse lines (e.g., GCAMP6f9); males	Jackson Laboratory	Strain #: 024105
Tamoxifen-inducible Cre-recombinase mouse lines (e.g., ROSA26-CreERT217); males	Jackson Laboratory	Strain #: 008463
Cell-type-specific reporter mouse lines (e.g., Hhex-GFP18); males	Rodriguez et al.18	N/A
Wild-type mice (e.g., CD1); females	Charles River	Strain code: 022
Software and algorithms
ZEN Blue edition	Zeiss	Version 3.10 or later
IncuControl Software	Ibidi	Version 3.0.1
Other
35 mm dishes	Corning	353001
Glass capillaries (1.5 mm O.D. × 1.17 mm I.D.)	Harvard Apparatus	30-0062
Diamond-tip pen/knife	Sigma-Aldrich	933384
Tungsten needles (or tungsten wire for making needles)	Fine Science Tools	10130-05
Needle holder × 2	Fine Science Tools	26016-12
Dumont #55 forceps × 2	Fine Science Tools	11255-20
Extra fine Bonn scissors	Fine Science Tools	14085-08
Serrated forceps	Fine Science Tools	11000-12
Glass slide	VWR	e.g., 631-1550
Bunsen burner (DIN type cartridge burner)	Usbeck	e.g., Type: 1420
5 mL syringe	Terumo	SS-05S
8-well μ-Slide #1.5 (160-190 μm thickness)	Ibidi	IB-80807
Stereoscopic dissection microscope with bottom lighting	WILD Heerbrugg	e.g., M3z
Lattice Lightsheet 7	Zeiss	LLSM7
Ibidi gas incubation system with humidifier	Ibidi	e.g., 11923
Ibidi heating system	Ibidi	e.g., 12110
Cell culture incubator	BINDER	e.g., CB 56
CO2 supply	N/A	N/A

Materials and equipment

Note: The transgenic mouse lines listed in the key resources table above were the ones used to obtain the sample data shown included with this protocol under expected outcomes. However specific mouse strains that would be informative for individual studies should be carefully chosen.

Embryo culture medium

Reagent	Final concentration	Amount
CMRL	49.5% (v/v)	1.98 mL
KnockOut Serum	49.5% (v/v)	1.98 mL
200mM L-glutamine	1% (v/v)	40 μL
Total	N/A	4 mL

This culture medium should be made fresh each time under sterile conditions and left to equilibrate at 37°C and 5% CO₂ for at least 1 hour before use.

Step-by-step method details

Embryo isolation and pre-culture

Timing: 40 min (dissection; or quicker with more practice) and 30 min (pre-culture)

This step provides guidelines on how post-implantation embryos can be isolated from the uterine horns of pregnant female mice. The steps are outlined for dissecting 5.5 dpc embryos (see Methods video S3 and Srinivas 2010⁷ for visual demonstrations of the dissection steps). The same steps can be followed for later stages (6.5 dpc to 7.5 dpc), but the embryo will be more visible within the decidua and the dissection should be easier. See Elmore et al., 2022,¹⁹ to familiarize yourself with the positioning of the embryo within the decidua at different stages of development.

• 1.
  Isolation of the uterine horns.

  • a.
    Inside a laminar floor hood to minimize the spread of allergens, euthanize pregnant females carrying embryos of the right stage (determined by the date the vaginal plug was observed; see above) using an approved method and confirm death.
  • b.
    Placing the euthanized female supine on absorbent paper, spray the abdomen with 70% ethanol to wet the fur and to reduce the risk of contamination.
  • c.
    Open up the abdominal cavity of the mouse to expose the uterine horns.

    • i.
      Pinch the skin overlying the abdomen near the midline with a pair of serrated forceps.
    • ii.
      Make a lateral incision (roughly 1 cm across) using surgical scissors.
    • iii.
      With the thumbs and index fingers, pinch and pull apart the skin immediately above and below the incision towards the head and tail respectively, exposing the peritoneum.
    • iv.
      Cut open the peritoneum laterally to expose the abdominal cavity.
    • v.
      Using the forceps or forefingers, dislodge the intestines to one side to bring into view the dorsally located reproductive organs (ovaries, oviduct and uterine horns).
  • d.
    Using forceps and scissors, hold the uterine horns and isolate it from the abdominal cavity by making a cut rostrally below each oviduct and caudally at the cervix.
  • e.
    Remove any fat and maternal blood vessels associated with the isolated uterine horns.

    • i.
      Transfer the isolated uterine horns to a paper towel and stretch it out so that the uterine adipose tissue and vasculature sticks to the paper towel.
    • ii.
      Carefully trim away the fat and blood vessels using a pair of extra fine dissection scissors.

      Note: This step will help to keep the dissection medium clear and free of fat droplets, making dissection easier.

  • f.
    Transfer the cleaned uterine horns to a 35 mm dish containing warmed up (22°C to 30°C) M2 medium.

    CRITICAL: Extreme care must be taken not to nick the intestines or stomach while opening the abdominal cavity, or the uterine walls while removing the uterine horns.

• 2.
  Isolating deciduae from the uterine horns.

  • a.
    Under a dissecting microscope with low magnification and illumination from above, the sites of implantation will be clearly visible as bulges along the uterine horn from 5.5 dpc.
  • b.
    Using fine forceps, clamp the uterine horns between two bulges, so that the mesometrial side faces away from you.
  • c.
    Using one tip of a second pair of sharp fine forceps, pierce through the muscular uterine lining close to and parallel to the holding forceps, then tear the uterine tissue sideways and away from the holding forceps, opening a window to expose the decidua within.
  • d.
    While still firmly holding the uterine horn, gently peel back the muscular layers and slide the forceps beneath the decidua to free it from the uterine tissue. Starting at one end work along the entire length of both uterine horns to free all the decidua.
  • e.
    Wash the isolated decidua once with a change of M2 medium to remove any blood or fat.
  • f.
    Using forceps, transfer the decidua into the lid of 35 mm dish filled with just enough M2 medium to cover half the surface.
• 3.
  Isolation of the embryos from the deciduae.

  • a.
    Using only transmitted illumination from below, increase the magnification so that an individual decidua fills half the field of view.

    Note: A 5.5 dpc decidua will roughly be the shape of a prolate spheroid (or horizontally elongated cardioid), with one end being somewhat wider (by 6.5 dpc, it becomes more egg-shaped). The pointier end has the embryo and the opposite, more rounded end, is either open or sealed with a midline groove.

  • b.
    Hold the decidua at its pointy end with a pair of forceps and using a second pair of forceps make a horizontal incision along the long axis, up to the inside end of the midline groove.
  • c.
    Pivot the decidua so that it sits vertically, and the proximal tip of the embryo is visible (it will appear red at later stages).
  • d.
    While holding the decidua securely in this position, go in with closed forceps adjacent to the embryo and, by allowing the forceps to open, deepen the tear by flaring apart the decidual tissue.
  • e.
    Using closed forceps, scoop out or nudge the entire embryo (including the extraembryonic ectoplacental cone) freeing it form the decidua (at 5.5 dpc the embryo will be very loosely associated with the decidual tissue compared to subsequent stages).
  • f.
    Sweep the embryo to one side of the dish by gently moving the medium around it.
  • g.
    Repeat steps 3a–d for all remaining deciduae.
• 4.
  Removing the Reichert’s membrane.

  • a.
    Increase the magnification so that individual embryos and the tissue layers can be seen clearly.

    Note: The embryo will be enclosed within the Reichert’s membrane, composed of cells that appear more opaque than the other cell-types.

  • b.
    Using a fine tungsten needle mounted onto a long needle holder, pin the embryo down to the surface of the dish, by inserting the needle into the space between the Reichert’s membrane and the visceral endoderm, just below the ectoplacental cone or at the embryonic-extraembryonic boundary.
  • c.
    Using a second needle, remove the Reichert’s membrane from the rest of embryo, unshrouding the distal tip.
  • d.
    Cut off the membrane from its point of attachment to the ectoplacental cone using the tungsten needles.

    Note: Tungsten needles can be sharpened using electrolysis, to have a tip diameter as small as 1–5 μm. Since it is a hard and stable material, it has little spring action. With practice, it can be used for precision dissection steps such as removing a single cell layer without damaging outer layer. Precise cuts can be made with the tungsten needles, by crossing the tips of the needles against each other with the tissue to be cut positioned in between them.

    CRITICAL: Successfully removing the Reichert’s membrane from 5.5 dpc embryos will improve with practice. Practicing on 6.5 dpc embryos, that are larger, will be helpful. It is important to not nick the visceral endoderm while removing Reichert’s membrane. Additionally, keeping the ectoplacental cone intact helps embryos survive better in long-term culture.

  • e.
    Wash the embryos by transferring them, using a P20 pipette filter tip, through at least two drops/pools of M2 medium, and then two drops of the pre-equilibrated culture medium.

    Note: To minimize plastic use, the lid of the 35 mm dish used for dissections can be used to hold these drops.

  • f.
    Transfer the embryos to one of the end wells of the already prepared 8-chambered slide (from before you begin step 4), containing pre-equilibrated culture medium.
  • g.
    Return the slide to the incubator until needed. Up to four different litters can be collected into four separate wells.

    Note: Pre-culturing the embryos for ∼30 minutes will help embryos recover from dissection and will allow sub-optimal embryos that were damaged during dissection to be identified.

    Optional: If using live dyes (e.g., SiR-Act or SiR-DNA) or treatments such as pharmacological inhibition, use the pre-culture period for labeling and treatments of the embryos as needed.

Embryo mounting for live imaging

Timing: 10 min

This step explains how the embryos can be mounted in the correct position, to allow fast imaging of the maximum possible volume at each stage.

• 5.
  Mount the embryos.

  • a.
    Remove the 8-chambered slide containing the dissected embryos from the incubator. If necessary, screen the embryos under fluorescent illumination to identify those expressing the fluorescent reporter/s of interest.
  • b.
    Under a dissection microscope, with transmitted illumination from above, examine the four chambers prepared for imaging.
  • c.
    Using clean tweezers dislodge any bubbles that might have formed over the grease barriers or in the channels between the capillaries.
  • d.
    Using a P20 pipette and a filter tip, transfer the healthiest looking, undamaged embryos to one of the central wells and let them settle down in the center, to one side of the capillaries.

    Note: When selecting the best embryos, look for those that have entire Reichert’s membrane successfully removed, with an intact ectoplacental cone and with no nicks or perforations of the visceral endoderm and epiblast. If embryos were damaged during the dissection, this will become evident during the pre-culture period as cells will be extruded out from these regions, evident as interruptions in the continuous visceral endoderm around the embryo.

  • e.
    Using a tungsten needle, gently move individual embryos to position them in the channel between the glass capillaries (Fig B, B′ and B″) or underneath the capillaries (Fig C, C′ and C″), leaving roughly the length of an embryo in between, as space for them to grow.

    Note: It is better to have multiple glass capillaries with parallel channels between them, than to pack too many embryos along a single vertical channel. This will become important when imaging large number of embryos (e.g., up to fifteen 5.5 dpc embryos can be arranged as three columns of 5 each; Figures 1B and 1B′) as minimal stage movements will then be necessary to maneuver between positions before returning to the first position to start the next time point. If embryos need to be imaged perpendicular to the distal-proximal axis, the embryo can be mounted by tucking the ectoplacental cone underneath the capillary (Figures 1C, 1C′, and 1C″). However, the capillaries cannot be placed in the perpendicular orientation as it would cause refraction of the light sheet to occur in every optical section.

    Note: In situations where a genotyping step is required to identify the embryos of interest (e.g., those homozygous for a specific knockout allele from a cross between heterozygotes), it is recommended to image all the embryos in the litter. Embryos can then be recovered at the end of imaging for genotyping.

  • f.
    Once the embryos are mounted, taking care not to disturb them, move the imaging chamber onto the microscope stage and clip it into position.

Setting up imaging parameters and imaging

Timing: 15 min (for setup)

These steps will explain how to set up a time-lapse imaging experiment on the Lattice light-sheet microscope (Figure 2A), once the embryos have been correctly mounted. It will be necessary to switch the “continuous mode” off to change certain parameters, and every instance this is needed is not necessarily indicated below. However, the software will prompt if a change cannot be made while ‘continuous display’ is running. While some steps such as focusing the light-sheet can be done automatically through the software, it is recommended to learn how to do it manually.

• 6.
  Align the light-sheet.

  • a.
    Apply immersion fluid. This can be done manually using a Pasteur pipette, or automatically using the software.

    CRITICAL: If using “Autoimmersion”, make sure that there are no air bubbles or residues within the tubing, as they can get trapped between the sample and the meniscus lens, causing shadowing.

  • b.
    Using the joystick that controls the stage, move to a region within the well, away from the boundary between wells and where no sample or glass capillary is in the field of view.
  • c.
    Select the light-sheet to use and select the maximum field of view (2048 x 2048) under acquisition controls.

    Note: We have determined that for post-implantation embryos, it is best to use the Sinc3 100 x 1800 light-sheet (length or z_max [μm] x thickness [nm]) as it gives the widest field of optimal imaging with adequate spatial resolution to reveal subcellular dynamics of cells within the embryo. If imaging specific smaller, superficial regions within the embryo, other light-sheets can be used to obtain higher spatial resolution, but they are not recommended for imaging larger volumes (see Figure 3, comparison of results from different light-sheets).

  • d.
    To allow reflection of the light-sheet from the coverglass at the bottom of the 8-chambered slide, tick on the 640 nm laser and set the light path with SBS640 in the dichroic position and “no emission filter” in the emission filter drop down.
  • e.
    Select camera LLS1 only. The laser power is automatically limited by the software.
  • f.
    Set the exposure time at 10 ms.
  • g.
    With the dithering turned off under acquisition controls, and while running on ‘continuous’ mode, move in Z, to visualize the first then the second reflection from the bottom and top surface of the coverglass of the imaging chamber respectively (see Figure 2B).

    Note: As the stage is moved through Z (with the value of the Z-position decreasing from 0 into negative values) a band of white vertical stripes will come into view as the light-sheet hits and is first reflected from the bottom surface of the coverglass of the sample carrier. This is at the immersion water/glass boundary and is called the first reflection. Further descent in Z will bring into view a tighter band of stripes, which is the second reflection of the light-sheet from the top surface of the coverglass at the glass/culture medium boundary (seeFigure 2B). As the embryo sits directly on top of the coverglass, it is important to align the light-sheet horizontally at this position, so that movements of the stage in X and/or Y will not cause the light-sheet to shift in Z through the embryo.

    Note: Dithering creates a uniform light-sheet by quickly deflecting the sheet across the field of view, faster than the camera frame time, using a galvo. The lattice structure is visible for focusing only when dithering is off.

  • h.
    Align the horizontal crosshairs with the second reflection.

    Note: Record the Z position of the second reflection and use this as a guide when moving through the sample in Z to find regions of interest. The further you are from this value (more negative), the deeper you will be within the sample. Different stage inserts will have to be used for different sample carriers and the Z position of the second reflection will vary depending on it. For the 8-chambered slides used in this protocol, the insert used is part no. and the second reflection is visible at a Z position of ∼375 μm.

  • i.
    Using the cross hair as a guide, change “Tilt Y-Axis” under aberration control if necessary to ensure the reflection is horizontal across the field of view on the screen. (Figure 2B).
  • j.
    Move a set distance in X using the joystick, note the displacement of the reflection from the crosshairs, and adjust “Tilt X-Axis” under aberration control until the reflection does not drift up or down on the screen for this movement in X.
• 7.
  Locate the sample and focus the light-sheet.

  • a.
    Depending on the fluorophores being imaged, set the beamsplitter, emission filters and lasers under imaging setup as needed.

    Note: For example, to visualize Lifeact-eGFP – a fluorescent peptide that decorates filamentous actin in live eukaryotic cells – we initially use 488 nm laser with a SBS L640 beamsplitter coupled with a Band Pass 495-550/570-620 emission filter, at 5% laser power and 20 ms exposure time; equivalent to 74 μW of power at the objective).

  • b.
    With the white-light LED in the lid switched on and set to 5%–7%, run on “continuous” mode.
  • c.
    While visualizing on the screen, move the joystick to move through the imaging chamber towards where the glass rods are (Figure 2C) and locate with a bright-field view, the first embryo (top or bottom of the well) in between them. Having the maximum field of view (2048 x 2048) makes navigating across the well and locating the embryos much easier.
  • d.
    Switch off the transmitted light.
  • e.
    Making sure “Dithering” is still switched off, adjust the display setting as needed to visualize the fluorescent signal from the embryo without the bright field.
  • f.
    Move in X, Y and Z to locate cells within the embryo closest to the coverglass (or any other region of interest).
  • g.
    By altering the “Focus Sheet” and “Focus Waist” in the “Lightsheet/Abberation Contol” drop-down menu, adjust the light-sheet so that the lattice structure is clearly visible in the channel being imaged (Figures 2A′ and 2D). If necessary, change the “Aberration Control” setting to make the lattice structure clearer.

    Note: In the ZEISS Lattice Lightsheet 7 the core optics module is made up of a 13.3x/0.4 excitation objective and 44.83x/1.0 detection objective positioned at 30° and 60° angles to the bottom of the sample carrier respectively, housed underneath the meniscus lens (Figure 2A).

    Note: Ensure dithering is turned off so that the lattice structure can be seen in the fluorescent signal from your sample. The aim of this step is to align the optimal portion of the light-sheet exiting the illumination objective, with the focal point of the collection objective (seeFigures 2A and 2A′). “Focus Sheet” moves the sheet along the optical axis of the collection objective by deflecting the sheet exit from the illumination objective with a mirror. “Focus Waist” moves a tube lens before the illumination objective to align the focal point of the sheet perpendicular to the collection objective. As the sheet becomes aligned with the objective, the stripe pattern from undithered optical lattice becomes more apparent and sharper. “Aberration Control” is similar to collar correction on other detection and illumination objectives. This is most important for the sphericity of dot-like structures if present in the sample.

  • h.
    When two or more channels are being imaged on separate cameras, make sure the pixel shift is adjusted, so that all channels are aligned correctly.

    Note: This would ideally be done with a specimen or test sample where two colors are present on the same structure such as large beads or plasma membrane (Figure 2E). The pixel shift is specific for the beam splitter used. For applications where sub-pixel corrections are required, reference images should be taken and post-processing used. Higher laser powers can be used for the pixel alignment step to get a brighter image without the risk of damage to the sample, as the laser powers at the objective are much lower on the LLSM compared to conventional confocal microscopes. For example, at 30% and 5 ms exposure, the 488 nm, 561 nm and 640 nm lasers generate 146 μW, 119 μW and 99 μW of power respectively, at the objective of the Zeiss Lattice Lightsheet 7, compared to 4.5 mW, 2.3 mW, 0.68 mW for the Zeiss LSM 880 confocal microscope.

• 8.
  Set the imaging volume.

  • a.
    Under “Acquisition” controls, switch to “optimal” field of view for the light-sheet being used (Figure 2C).

    Note: This automatically crops down the collected field-of-view to the portion of the camera chip covered by the optimal light-sheet, which is close to the length described by the sheet setting e.g., for the Sinc3 100 x 1800, this is 759 px or 110 μm.

  • b.
    Switch on “Dithering”, which will smooth the lattice structure to create a uniform light-sheet (see Figure 2D).
  • c.
    Set the volume to be imaged using the “First/Last” setting.

    • i.
      Move in X and Z so that the surface where the sample first contacts the light-sheet is positioned towards the bottom of the field of view.
    • ii.
      Move through X only, until the sample disappears from the top of the field of view, or until you see the lightsheet blurring due to reflection from the glass rods.
    • iii.
      Move in Y leaving enough room at the distal tip of the embryo for it to grow into and still stay in view.

      Note: These steps will make sure that the surface of the embryo (apical surface of visceral endoderm cells) is captured. However, if desired, the sample can be moved in Z to set the light-sheet deeper within the sample to acquire a cross-sectional midsagittal volume through the epiblast for example (seeFigures 2F, 4A, and 4B).

  • d.
    Set the “x-Interval” to 0.4 μm (see Figures 5A and 5B).

    Note: When imaging multiple embryos, 8c can be skipped as this can be done along with “Tile Selection” under “Multidimensional Acquisitions”.

    Note: Light scattering from the glass capillary can deteriorate the quality of the image in sections where the light-sheet passes through the glass capillary before entering the embryo (seeFigure 2F and troubleshooting below).

• 9.
  Set up to image multiple embryos.

  • a.
    Select “Tiles”, which will open a new tab under “Multidimensional Acquisitions”.
  • b.
    Switch on the transmitted light and run on continuous mode, with the “Advanced Tiles Viewer” displayed.
  • c.
    While moving in Y and X, roughly locate each embryo to be imaged by drawing an approximate tile region with the rectangular contour tool. Register each region by adding it to the list of “Tile Regions”.
  • d.
    Switch off the transmitted light and run on continuous mode, adjusting the display to visualize the fluorescent signal from the embryo.
  • e.
    Return to each registered tile region and adjust the width (X dimension, i.e., first and last slice) and height (Y dimension) of the rectangle to image the desired volume.
  • f.
    Adjust Z as needed and register the new Z position with “Set Current Z”.

    Note: If the set Y is greater than the maximum length of the light-sheet, each tile region will be imaged multiple times to acquire the full area. Different Z positions can be set for each position, but it may not be possible to select different XY Tile dimensions for each position. In this case, the largest Tile dimensions will be applied to all positions.

  • g.
    Repeat steps 9e–f for each position.

    Note: when setting the order in which the embryos/tiles are imaged, minimizing large movements will reduce imaging time, allowing shorter intervals between time points and reducing the risk of immersion overflowing (seeFigure 1B′).

• 10.
  Set the time lapse parameters and start the experiment.

  • a.
    Under “Experimental Information”, note down the “Theoretical Duration” it would take to image all the Tiled positions selected in step 9.
  • b.
    Select “Time Series” under “Acquisitions”, which will open a new tab under “Multidimensional Acquisitions”.
  • c.
    Set the total “Duration” for the experiment and “Interval” between time points.

    Note: When setting the interval between time points, consider the “Theoretical Duration: noted in 10a (sum of the exposure time for all optical sections for all times) and the dead time for the mechanical movement from tile to tile and back to the start position and the time taken to compress and/or save the data if using those options. If it is set too close to the “Theoretical Duration”, the actual Interval will be greater than set Interval. In this case, the actual interval will have to be later calculated at the end of the experiment by dividing the total duration by the number of time points acquired.

  • d.
    Select “Auto Save” for the experiment and define the file path of the folder the data is to be saved to and a Name.
  • e.
    Switch on the “Autoimmersion” and set the time, asking the experiment to paused to apply autoimmersion. For the atmosphere and temperature setting used in this protocol, autoimmersion every 20 minutes is adequate to prevent the immersion fluid from drying out or overflowing.
  • f.
    Press “START” Experiment.

Note: Long time lapses can create hundreds of gigabytes of data making autosaving essential to preserve RAM capacity and prevent computer lagging, freezing or acquisition failure. It is recommended to save to a fast SSD or to make sure the data can be saved fully to another drive before the next acquisition time point. This will prevent data accumulating in the RAM.

Figure 2.

Setting up the initial imaging parameters on the lattice light-sheet

(A) The optical set up of the ZEISS Lattice light-sheet 7 microscope, showing the angle of the excitation objective and detection objective relative to the coverglass bottom (or xy-plane of the sample). A′ shows the direction of the movement of the light-sheet while focusing the sheet and the waist.

(B) Images of the first and second reflection from the coverglass, showing a focused lattice structure.

(C) Full 2048 x 2048 view of a 5.5 dpc embryo mounted between glass capillaries visualized with transmitted light switched on, used to locate the samples and register their initial positions. The boxed area represents the optimal region for imaging using the Sinc3 100 x 1800 light-sheet.

(D) Comparison of an unfocused and focused light-sheet with and without dithering turned on.

(E) Comparison of a region of an embryo expressing two fluorophores (green and magenta) visualized with a focused light-sheet with dithering off, showing the difference between misaligned and aligned channels.

(F) Illustration showing the light-sheet as it passes through the embryo, indicating the optimally imaged region (green dotted lines and shaded region) and portions of the light-sheet (red solid lines) that would be scattered as it passes through the glass capillaries used for mounting. The images underneath show what to expect from the first optical section (right), an optical section from the middle of the x-stack (middle) and one where the light-sheet is scattered as it passes through the glass capillary before contacting the embryo (left). Scale bars represent 20 μm.

Figure 3.

Comparison of volumetric data acquired using different light-sheets

(A) The optimal 2D sections (Z^∗ by Y) that can be acquired for a 5.5 dpc embryo, with six different light-sheets. Y was fixed at 2048 pixels (297 μm).

(B) Comparison of the thickness of the volume (in Z; between orange arrowheads) acquired with three different light-sheets imaging a 6.5 dpc embryo expressing the Lifeact-GFP transgene, with the same number of optical sections (1694 x-sections each) and x-interval (0.2 μm). 8.12 μm, 16.38 μm and 54.95 μm for the 15 x 650, 30 x 1000, and 100 x 1800 light-sheets respectively. Scale bar in A and the minor scale graduation marks in B represent 10 μm.

Figure 4.

Comparison of volumetric data that can be acquired for different stages of embryos

(A) Illustrations and 3D rendering of data showing the maximum volume that can be acquired using the Sinc3 100 x 1800 light-sheet. 5.5, 6.5 and 7.5 dpc embryos expressing the Lifeact-GFP transgene are represented being imaged along the anterior-posterior axis at 0.4 μm x-interval. Y and Z were kept at the maximum possible dimension for the light-sheet being used, while the X-range was set to acquire the maximum possible volume covering the anterior-posterior axis of the embryo (385, 668 and 1048 x-sections for 5.5, 6.5 and 7.5 dpc embryos respectively). The limits of the x-range are the diameter of the glass rods on either side of the mounted embryo, and how close they are to the embryo itself.

(B) 3D maximum intensity renderings of a 6.5 dpc embryo expressing the Lifeact-GFP transgene, where the optimal light-sheet is shifted 40 μm in Z for each subsequent volume (with a 10 μm overlap between volumes), showing the deterioration of spatial resolution due to light scattering in volumes acquired from deeper within the embryo. The second reflection of the coverglass is taken as Z=0. The scale grid units represent 50 x 50 x 50 μm.

Figure 5.

Deskewing volumetric data acquired on the lattice light-sheet

(A) Illustration showing the axes of an embryo before and after deskewing. While imaging, each optical section is a Z^∗ by Y rectangle, which moves through the embryo along X. Z^∗ is at 30° to X. The x-interval can be set as desired, and the Y and Z dimensions shown for each optical section are the maximum dimensions that can be acquired using the Sinc3 100 x 1800 light-sheet. After deskewing (bottom), three mutually orthogonal axes, X, Y, Z are derived, where each optical section is a X, Y rectangle, and a z-interval of 145 nm between each optical section is automatically generated.

(B) The 3D volumes generated at acquisition and after deskewing. On the right, a single optical section from the middle of the x-range is shown with the axes indicated by color.

(C) 3D maximum intensity renderings of deskewed data from a 6.5 dpc embryo expressing the Lifeact-GFP transgene, acquired with different x-intervals showing the deterioration of spatial resolution with larger x-intervals.

(D) Comparison of 3D maximum intensity renderings of a 5.5 dpc embryo expressing the Lifeact-GFP transgene, acquired with an x-interval of 0.4 μm, and deskewed with and without deconvolution. C and D show maximum intensity renderings generated directly using the ZEN software. Scale bars represent 20 μm.

For easy review of the data before processing, XY maximum intensity projections (MIPs) can be automatically generated upon saving by selecting the checkbox in Tools > Options >Acquisition>Online MIP. This MIP is a quickly generated 8 bit approximation which should not be used for analysis. To create a MIP for analysis, deskewing should be done first (see step 11).

Image processing

Timing: 3–6 h

This last step describes the minimal post-processing required to prepare data acquired on the lattice light sheet microscope using the ZEN software, for any downstream analysis on other platforms.

Note: Since the light-sheet penetrates the sample through the excitation objective at an angle of 30° and the detection objective is positioned perpendicular to it underneath the meniscus (Figure 2A), the resulting volume acquired will be a parallelepiped in shape (Figure 5A, top). However, this is initially displayed as a cuboid, making projections of the image skewed (Figure 5B, top). The pixel resolution of the data will also not be isometric with two dimensions being 145 nm per pixel, and the third being equivalent to the x-interval set in step 8d. To generate a volume where x, y and z are mutually orthogonal and of uniform axial resolution, deskewing of the data is necessary. Deskewing uses interpolation and black pixel padding to achieve a parallelepiped volume or structure matrix with orthogonal axes (Figure 5A, bottom), within which the embryo is projected correctly and can be interpreted for computation (Figure 5B, bottom). Often, this process is combined with ‘coverglass transformation’ which rotates the data, again using interpolation, such that the XYZ co-ordinates correspond to the stage geometry and not the objective geometry.

• 11.
  Deskew the files.

  • a.
    At the end of the experiment, view the MIP files for each tile and select those with good quality data (healthy embryo throughout the culture period, minimal drifting of the sample).
  • b.
    Open the .czi file to be deskewed using the Zen software.
  • c.
    Under the “Processing” tab, select the file and define a subset of time points or if possible, region of interest, to minimize the size of the output file.
  • d.
    Apply the inbuilt “Lattice Lightsheet” deskewing parameters with linear interpolation and using “Coverglass Transformation” as the processing method.
  • e.
    Define the file path for the “Output Folder”.

    Note: The x-interval used ultimately determines the sharpness of the deskewed 3D rendering (Figure 5C) but for thicker light-sheets such as Sinc3 100x1800 used to image embryos, x-intervals smaller than 0.4 μm will not be necessary.

    Note: The deskewed image is of an isometric axial resolution of 145 x 145 x 145 nm/voxel and always contains a constant number of z-optical sections (344 sections for the Sinc3 100x1800 light-sheet) irrespective of the number of x-sections used to acquire the data. This means that compared to the original file, the size of the deskewed file will increase significantly for smaller embryos (where the initial number of x-sections to cover the embryo might be fewer than 380) and decrease for large embryos (where more than 380 x-sections were likely acquired).

    Note: While deskewing, deconvolution can also be applied to the data (Figure 5D). This will however increase the processing time. Therefore, deconvolution is not recommended as standard and is better done for selected images in downstream processing pipelines.

  • f.
    Save the deskewed images with a recognizable suffix, as a .czi file or .tiff stacks. These files can be used for downstream analysis with other compatible software.

    Note: Alternatively, Batch Processing can be applied to run the same Lattice light-sheet deskewing parameters on multiple files and save them to a designated Output Folder. While “Batch Processing”, distinct spatial and temporal cropping parameters cannot be applied to individual images. The speed at which the processing can be done will also depend of the specs of the computing set up.

Expected outcomes

The LLSM was designed for high spatial resolution imaging across time of live cells, offering reduced phototoxicity and photobleaching compared to conventional imaging modalities.¹³ In combination with optimal in vitro culture conditions, this protocol enables researchers to image post-implantation embryogenesis over several hours to characterize the cellular and subcellular dynamics driving morphogenetic events. About half the volume of each of as many as 15, 5.5 dpc embryos can be imaged simultaneous for 8 hours at 3-5-minute intervals. This is extremely beneficial especially when working with mutant lines, to get enough data for meaningful downstream statistical analysis. By using reporters that are specific to cell populations, their behavioral dynamics (migration, cell division events) during the culture period can be visualized and quantified. For example, using the Hhex-eGFP allele to label anterior visceral endoderm cells on a background where all cells express membrane-tdTomato (mTomato), we can study the behavior of this cell population as they actively migrate from the distal tip of the embryo towards the boundary between the embryonic and extraembryonic halves (Figures 6A and 6B). The high spatial resolution allows us to visualize the basal projections produced by the cells along the basement membrane and underneath adjacent cells. The high temporal resolution on the other hand shows their changing directionality and help us determine the ‘decisiveness’ of these projections.

Figure 6.

Examples of representative data acquired using lattice light-sheet microscopy

(A) Maximum intensity projection of a 6.25 dpc embryo expressing ubiquitous membrane-td-Tomato (mTomato) with the anterior visceral endoderm cells labeled with a Hhex-GFP reporter.

(B) Maximum intensity projection of selected frames from an embryo expressing Hhex-GFP, showing the migration of the anterior visceral endoderm cells from the distal tip to the prospective anterior over 8 hours, imaged at 5-minute intervals.

(C) Maximum intensity projection of a gastrulating 7.5 dpc embryo ubiquitously expressing mTomato and with mosaic labeling of cells with membrane-GFP (mGFP) using a tamoxifen-inducible ROSA-CreERT2 (also see Methods video S1).

(D) Selected (non-continuous) frames from the boxed region in C, showing the change in position and morphology of cells during a 6-hour period, imaged at 3-minute intervals. The change in morphology of a mesodermal cell (arrowhead) is clearly visible as it migrates within the three-dimensional context of the embryo. D′ shows an enlarged view of a single mesodermal cell from the gastrulating embryo within the boxed region in the last time point in D, highlighting the subcellular resolution that can be achieved using the Sinc3 100 x 1800 light-sheet, with white arrows pointing to nm-scale cellular projections.

(E) Maximum intensity projection of selected time frames from a 6.5 dpc embryo ubiquitously expressing the Ca²⁺ reporter GCaMP6f, imaged at 30 s interval for 6 hours, showing the propagation of a Ca²⁺ wave originating in a cell of the visceral endoderm (arrowhead) and through cells of the epiblast.

(F) Selected time frames from an 8.0 dpc embryo ubiquitously expressing the GCaMP6f Ca²⁺ reporter transgene, imaged for 10 seconds at 5 ms interval, showing the propagation of a Ca²⁺ wave across the cardiac crescent within the anterior portion of the embryo. (also see Methods video S2).

(G) Comparison of labeling achieved with a genetic reporter (Lifeact-GFP) and two different live dyes (SiR-Act-652 and SiR-DNA-652), showing differences in the depth and uniformness of labeling with different approaches. The Lifeact-GFP and SiR-Act-652 are shown in the same embryo. The 3D volume renderings are maximum intensity projections. Scale bars represent 20 μm.

Alternatively, using tamoxifen-inducible Cre-mediated labeling, in combination with floxed reporter alleles, can help ‘titer’ the proportion of cells labeled. Such approaches offer the advantage of providing high contrast for labeled cells amongst unlabeled neighbors. This facilitates visualizing all surfaces of individual labeled cells which could otherwise be obscured if surrounding cells are also labeled. For example, an inducible ROSA26-CreERT2 (Gt(ROSA)26^{tm1(cre/ERT2)Tyj}/J)- allele can be used with the conditional double fluorescent mT/mG (Gt(ROSA)26-Sor^{tm4(ACTB-tdTomato; eGFP)Luo}/J) reporter allele, to sparsely label cells throughout the peri-gastrulation embryo. In combination with Lattice light-sheet microscopy, this allows us to visualize the migration of individual mesodermal cells from the site of gastrulation in the posterior of the embryo to the anterior, along characteristic migratory paths (Figure 6C; Methods video S1). During the culture period, division events involving the labeled cell can be documented, to compare the migratory behavior of daughter cells, to see if and how their behaviors (and fates) diverge. The cellular dynamics of individual mesodermal cells can be visualized within the context of the embryo (Figure 6D) with high enough subcellular spatial resolution to discern individual projections (Figure 6D′).

The low phototoxicity of the LLSM permits higher laser powers to be used with shorter exposure times (as low as 1.67 ms) without causing significant alteration to the physiological state of the embryo. This is not possible with other comparable imaging modalities (e.g., confocal microscopy where a large volume of the specimen is illuminated). This allows faster events to be visualized with volumetric imaging. For example, using a ubiquitously-expressed the GCaMP6f (Gt(ROSA)26Sor^{tm95.1(CAG-GCaMP6f)Hze}/J) calcium (Ca²⁺) reporter line, we can record the dynamics with which Ca²⁺ transients propagate through cells of the epiblast in the 6.5 dpc embryo (Figure 6E), or how Ca²⁺ transients travel from cell-to-cell along the cardiac crescent of the developing heart of the 8.25 dpc embryo (Figure 6F; Methods video S2). The mounting of embryos for imaging within 8-chambered slides, as described in this protocol, further provides good accessibility to the embryos. Embryos can be imaged to record baseline reporter dynamics, reflective of physiological processes or active signaling pathways, and then reimaged after perturbation of the pathways involved. The ease of accessibility also makes remounting of the embryos feasible, where full volumes of the embryo cannot be acquired from one fixed position.

The LLSM technology also makes it practical to image multiple channels over long periods of time, without causing noticeable phototoxicity. While breeding mice with multiple reporter alleles can be cumbersome and expensive, genetic reporters can be used in combination with live dyes to this purpose (Figure 6G). One limitation of live dyes is their ability to penetrate and effectively label deeper tissues (see Figure 6G comparing SiR-Act and SiR-DNA to the genetic reporter LifeAct-GFP). However, they can be of great utility in labeling superficial tissues such as the visceral endoderm of the post-implantation embryo.

Limitations

All live imaging modalities need to balance the trade-offs between i) spatial and temporal resolution, ii) total volume imaged and iii) photodamage to the sample. Amongst these, the main limitation of Lattice light-sheet microscopy, with the approach described in this protocol, is the small volume imaged at optimal resolution and speed. In comparison to other light-sheet microscopes, it is not possible to acquire images from multiple camera angles or rotate the sample. While a near-full volume of 5.5 dpc embryos can be acquired, only a fraction of the volume of 6.5+ dpc embryos is possible (Figure 4A). This makes the orientation the embryos are mounted in extremely important, to make sure that the region of interest for imaging lies closest to the bottom coverglass. While stacking multiple volumes in Z can help image deeper within the embryo, the post-processing needed to align the tiles correctly can sometimes pose problems. Image quality also deteriorates significantly the deeper within the embryo a volume stack is acquired due to scattering of light by the tissue (Figure 4B). While the Lattice light sheet gives a much higher spatial resolution compared to other light sheet approaches, it does not achieve super resolution quality to observe some subcellular dynamics. The resolution of the final image can be improved by using thinner light sheets, but this will be at the cost of the total volume of the embryo that can be imaged (Figures 3A and 3B).

Finally, when imaging for long durations, the embryo will grow by cell proliferation and increase in volume, as cavities within the embryo expand. This means that not all the cells at the start of an experiment will stay within the field of view for the complete duration of the experiment, limiting the number of cells that can be tracked through the full duration. It is important to note that when using faster time intervals for higher temporal resolution or narrow light-sheets for better spatial resolution, the volume imaged becomes even more limited (Figures 7A and 7B).

Figure 7.

Spatial and temporal limits of imaging

(A) 3D surface renderings of a 5.5 dpc embryos expressing the Lifeact-GFP reporter transgene, imaged with the same number of optical sections (492 x-sections) and x-interval (0.4 μm) but with different exposures times per optical section. The minimum interval indicates the fastest temporal resolution that can be achieved with the different exposures. Resolution (and the number of gray levels) deteriorates with faster imaging, if keeping the laser power constant.

(B) The different volumes that can be acquired using a fixed exposure time of 30 ms for each optical section, but with different temporal resolutions (i.e., intervals between volumes). The minor scale marks represent 10 μm.

Troubleshooting

Problem 1

Unable to locate the coverglass reflections or the second reflection does not show the lattice structure expected (related to Steps 6g–h).

This can be due to dirt on the cover glass of the sample carrier or residues on the meniscus lens (see Figure 2A). Any obstructing material such as grease, shards of glass or the glass rods can cause refraction disrupting the lattice structure. Figure 2B shows the expected, in focus lattice structure without any obstructions.

Potential solution

• •Check the bottom of the sample carrier and clean it using lens tissue and isopropanol.
• •It could also be due to not enough immersion being applied or due to a bubble in the immersion itself. Check the immersion liquid interface and reapply immersion if needed.
• •Make sure the region being viewed to locate the reflections is not at the boundary between two wells or obstructed by grease, glass rods or a sample. Figure 2F (left image), shows an example of what refraction from the glass rods would look like.

Problem 2

Embryo drifting out of frame during the imaging period (related to Step 5e and 11a).

If the drift is significant, it can be observed while setting up the tiles as the embryo slowly drifts out of the field of view. If the drift is slower, it will only be apparent only when viewing the data once it has been acquired. Minor X and Y drifts can be corrected by registration during post-processing, however drifts in Z will lead to loss of data by reducing the volume that can be through the entire time course.

Potential solution

• •Minimize drift in X and Y by making the channel within which the embryos sit narrower, by placing the glass capillaries closer to each other.
• •To reduce the loss of data due to drift in Y, include space on the proximal and distal sides of the embryo, beyond the field of interest within the tile being imaged.
• •To minimize drift in Z, the embryo can also be further immobilized, by having small glass shards balanced horizontally across the glass capillaries rods and over the embryos. The glass shards need to be narrow enough to allow efficient nutrient exchange to the embryo from the culture medium. They should ideally be placed over a region that is not being imaged, such as the ectoplacental cone or the most proximal extraembryonic half.
• •Alternatively, drifting of the embryos during imaging can be minimized by sticking the embryos down onto the coverglass using a live-tissue competent adhesive such as Cell-Tak (see materials and equipment). Care should be taken to make sure the embryo makes contact with the adhesive surface in a region that is not being imaged, to avoid optical artifacts being introduced from the adhesion proteins themselves.
• •Z-drift can also be caused by components of the objective expanding during imaging, if the system has not been adequately equilibrated in advance. Pre-equilibrate for longer than an hour if necessary.

Problem 3

Few embryos survive for the total duration of the experiment (related to Steps 5d and 11a).

During the pre-culture, unhealthy embryos can be identified as those with nicks to the visceral endoderm, or damage to the ectoplacental cone. Unhealthy embryos also look more opaque compared to when they were first isolated. During the experiments, poor embryo survival is apparent from an unusual number of cells undergoing apoptosis, the buildup of cellular debris inside the proamniotic cavity. In addition, there is often increased blebbing and the visceral endoderm becomes cobblestoned in appearance.

Potential solution

• •Selecting embryos during the pre-culture period can ensure that the most healthy and undamaged embryos are used for the experiment (see steps 1–4). In case of experiments where all embryos in a litter need to be imaged (e.g., when mutants cannot be identified without genotyping), this might not be feasible. However, with practice, efficient dissection techniques will become more natural, allowing the isolation of entire litters without damaging any embryos.
• •Leaving the ectoplacental cone intact will also improve embryo survival during culture.
• •Other culture media, different sera and serum concentrations can be tested to further optimize the culture conditions for the embryo stage or strain being used.
• •Make sure the temperature, humidity and CO₂ levels remain stable throughout the culture period. Temperature and gas levels can be logged using the IncuControl Software (Ibidi; see key resources table) to see if this could be the reason for poor embryo survival. If that is the case, the gas supply, humidifying chamber and heating components of the set up will need to be checked.
• •If poor embryo survival is due to photodamage, then the laser powers and exposure times will need to be reduced, or the imaging intervals increased. While this could reduce the signal to noise ratio, the dynamic range of data acquired, or the resolution (temporal and/or spatial) downstream processing such as denoising, or deconvolution can help restore data.

Problem 4

Reflection from the glass capillaries leading to reduced good quality volume (related to Step 8c).

The glass rods are necessary to restrain the embryos and minimize drift during imaging. However, as demonstrated in Figure 2F, for optical sections where the light-sheet passes through the glass capillaries before making contact with the embryo, the structure of the light-sheet will be lost and the image acquired will be of poor resolution. The volume lost due to this is minimal for 5.5 dpc embryos but could become significant for larger embryos (Figures 1B″ and C″). However, the total volume that can be imaged at high quality depends on the diameter of glass rods used and their proximity to the embryo as the light-sheet will refract as it travels through it.

Potential solution

• •One solution is to place the glass capillaries further apart and hold the embryo in position using a glass shard horizontally balance across the glass rods.
• •Alternatively, the ectoplacental cone can be tucked underneath a third horizontally placed glass capillary held in position with vacuum grease. But this could cause the embryo to drift in XY.
• •If imaging deeper sections within the embryo, make sure to choose glass rods with larger diameters to minimize the loss of volume that can be imaged as high resolution as the light-sheet is shifted in Z.

Problem 5

Large datasets with high storage demand and take an excessive time to post-process (related to Steps 11a–f).

The datasets generated for time-lapse experiments using the lattice light-sheet are often large, in the range of 0.25–1 TB per embryo imaged at 5-minute intervals for 6–8 hours. The data is always acquired at 16 bits and the file size could further increase with deskewing.

Potential solution

• •The first step to help with curating the data is to make sure that a maximum intensity projection for each position/embryo is automatically generated throughout the experiment, if the time intervals allow. This will help to select the data files at the end of the experiment from embryos that survived and stayed within the frame of view for the full duration of the experiment.
• •The file sizes can also be significantly reduced by spatially cropping the data to cover the regions of interest and temporally limiting the dataset to events of interest.
• •The ZEN software allows for automatic saving as a Zstd file with loss-less compression. Care needs to be taken with the compatibility of downstream analyses with this compressed file type. Alternatively, data can be saved at acquisition in an uncompressed state and later compressed in ZEN or companion software before storage/archiving.
• •To make post-processing more manageable, the data can be saved as multiple temporal subsets (e.g., 12 x 30-minute fragments from a 6-hour time lapse experiment).

Resource availability

Lead contact

Further information and requests for resources and reagents should be directed to and will be fulfilled by the lead contact, Shankar Srinivas (shankar.srinivas@dpag.ox.ac.uk).

Technical contact

Further information and requests for any technical issue should be directed to and will be fulfilled by the technical contact, Shifaan Thowfeequ (shifaan.thowfeequ@dpag.ox.ac.uk).

Materials availability

No unique reagents or equipment was developed for this protocol. The availability of all resources and mouse strains used to generate the representative data are detailed within the protocol.

Data and code availability

All data relating to the protocol are including as figures. No unique code was generated for this protocol. All commercially available software used are detailed in the key resources table.

Acknowledgments

We thank the Oxford-ZEISS Centre of Excellence in Biomedical Imaging for assistance with the use of the lattice light-sheet microscope and the Pathology Services Building and Biomedical Services staff for excellent animal support. We are grateful to Dr Tristan Rodriguez for the Hex-GFP line, Dr Roland Wedlich-Söldner for the Lifeact-GFP line, and Dr Liqun Luo for the ROSA26^mTmG line. This work was funded by the BBSRC Pioneer Award BB/Y513106/1 and Wellcome Discovery Award 227926/Z/23/Z to S.S.

Author contributions

S.T. and S.S. developed and refined the embryo isolation, mounting, and culture protocols; S.T. and H.C. developed the LLSM imaging parameters. S.T. wrote the manuscript, and H.C., M.F., and S.S. helped in its review and revision.

Declaration of interests

The authors declare no competing interests.