In Vivo Two-photon Imaging of Megakaryocytes and Proplatelets in the Mouse Skull Bone Marrow

Summary

We describe here the method for imaging megakaryocytes and proplatelets in the marrow of the skull bone of living mice using two-photon microscopy.

Abstract

Platelets are produced by megakaryocytes, specialized cells located in the bone marrow. The possibility to image megakaryocytes in real time and their native environment was described more than 10 years ago and sheds new light on the process of platelet formation. Megakaryocytes extend elongated protrusions, called proplatelets, through the endothelial lining of sinusoid vessels. This paper presents a protocol to simultaneously image in real time fluorescently labeled megakaryocytes in the skull bone marrow and sinusoid vessels. This technique relies on a minor surgery that keeps the skull intact to limit inflammatory reactions. The mouse head is immobilized with a ring glued to the skull to prevent movements from breathing.

Using two-photon microscopy, megakaryocytes can be visualized for up to a few hours, enabling the observation of cell protrusions and proplatelets in the process of elongation inside sinusoid vessels. This allows the quantification of several parameters related to the morphology of the protrusions (width, length, presence of constriction areas) and their elongation behavior (velocity, regularity, or presence of pauses or retraction phases). This technique also allows simultaneous recording of circulating platelets in sinusoid vessels to determine platelet velocity and blood flow direction. This method is particularly useful to study the role of genes of interest in platelet formation using genetically modified mice and is also amenable to pharmacological testing (study the mechanisms, evaluating drugs in the treatment of platelet production disorders). It has become an invaluable tool, especially to complement in vitro studies as it is now known that in vivo and in vitro proplatelet formation rely on different mechanisms. It has been shown, for example, that in vitro microtubules are required for proplatelet elongation per se. However, in vivo, they rather serve as a scaffold, elongation being mainly promoted by blood flow forces.

Introduction

Platelets are produced by megakaryocytes-specialized cells located in the bone marrow. The precise way megakaryocytes release platelets in the circulation has long remained unclear owing to the technical challenge in observing real-time events through the bone. Two-photon microscopy has helped overcome this challenge and led to major advances in understanding the platelet formation process. The first in vivo megakaryocyte observations were made by von Andrian and colleagues in 2007, with the visualization of fluorescent megakaryocytes through the skull¹. This was possible because the bone layer in the frontoparietal skull of young adult mice has a thickness of a few tens of microns and is sufficiently transparent to allow visualization of fluorescent cells in the underlying bone marrow².

Ensuing studies applied this procedure to evaluate proplatelet formation under various conditions and to decipher the underlying mechanisms^3,4,5,6. These studies provided definitive evidence that megakaryocytes dynamically extend protrusions, called proplatelets, through the endothelial barrier of the sinusoid vessels (Figure 1). These proplatelets are then released as long fragments that represent several hundred platelets in volume. The platelets will be formed after the remodeling of the proplatelets in the microcirculation of downstream organs, notably in the lungs⁷. To date, however, the precise process and molecular mechanisms remain subject to debate. For instance, the proposed role of the cytoskeletal proteins in the elongation of proplatelets differs between in vitro and in vivo conditions³, and differences in proplatelet formation have been demonstrated under inflammatory conditions⁶. Complicating things further, a recent study disputed the proplatelet-driven concept and proposed that in vivo, platelets are essentially formed through a membrane-budding mechanism at the megakaryocyte level⁸.

This paper presents a protocol for the observation of megakaryocytes and proplatelets in the bone marrow from the skull bone in living mice, using a minimally invasive procedure. Similar approaches have been previously described to visualize other marrow cells, notably hematopoietic stem and progenitor cells⁹. The focus here is on the observation of megakaryocytes and platelets to detail some parameters that can be measured, notably proplatelet morphologies and platelet velocity. This protocol presents how to insert a catheter into the jugular vein to inject fluorescent tracers and drugs and observe through the skull bone. The calvarial bone is exposed using minor surgery so that a ring is glued to the bone. This ring serves to immobilize the head and prevent movements due to breathing and form a cup filled with saline as the immersion medium of the lens. This technique is well suited to i) observe in real time the sinusoid geometry and megakaryocytes interacting with the vessel wall; ii) follow megakaryocytes in the process of proplatelet formation, elongation, and release; and iii) measure platelet movements to monitor the complex sinusoid blood flow. Data obtained using this protocol have been recently published³.

Protocol

All animal experiments were performed in accordance with European standards 2010/63/EU and the CREMEAS Committee on the Ethics of Animal Experiments of the University of Strasbourg (Comité Régional d'Ethique en Matière d'Expérimentation Animale Strasbourg, Animal Facility agreement N°: G67-482-10, project agreement N°: 2018061211274514).

1. Preparation of mice and insertion of a catheter in the jugular vein

NOTE: Here, male or female, 5-7-week-old mTmG reporter mice were used (B6.129(Cg)-Gt(ROSA)26Sor^{tm4(ACTB-tdTomato,-EGFP)Luo10}) crossed with Pf4-cre mice¹¹, allowing intense green fluorescence labeling in megakaryocytes and platelets¹². Before beginning the experiment, preheat the heating chamber of the microscope for a few hours.

1. Transport the mouse to the imaging room and proceed to anesthesia.
   1. Anesthetize the mouse by intraperitoneal (i.p.) injection of a solution of ketamine (ketamine (100 µg/g) and xylazine (10 µg/g) (5 µL/g).
      NOTE: Repeated administration of ketamine/xylazine is used here and works well but requires the interruption of recording for regular reinjection. Alternatively, if an anesthesia machine is available, the induction of anesthesia may be performed with ketamine/xylazine and subsequently maintained under inhalation of gases (mixture of isoflurane, air, and oxygen).
   2. Replace the animal in its cage until it reaches deep anesthesia. Place the mouse on a heated plate warmed at 37 °C for all subsequent manipulations.
      NOTE: While the animal reaches deep anesthesia, switch on the microscope and all equipment (see section 4.1) and set up the different parameters needed (see section 4.2) to start the experiment once the surgery is done. Here, the procedure lasts only 3 h and is terminal. The mouse is euthanized after the experiment without waking up. Hence, ethanol disinfection for skin and instruments is sufficient. However, depending on the institutional guidelines, or if the procedure would last longer, more complete disinfection methods should be used such as three alternating scrubs with betadine/povidone and alcohol.
2. Install a catheter in the jugular vein.
   NOTE: It is important to avoid microbial contamination during the surgery as far as possible. Take care to carefully disinfect the skin with 70% ethanol (EtOH) before proceeding with surgery. Always use sterilized scissors and tweezers, and regularly rinse them in 70% EtOH. Work in a clean place, and wear a face mask. A 22 G "over-the needle" catheter, consisting of a needle inside a plastic tube, is used for intravenous injection (see the Table of Materials).
   1. Place the mouse on its back, and fix the anterior legs with surgical tape to stretch the throat (Figure 2A). Disinfect the throat with 70% ethanol.
      NOTE: For convenience, the throat may be shaved before surgery. If the procedure lasts longer, careful shaving must be performed for proper aseptic technique.
   2. Make a 0.7-1 cm incision over the right or left jugular vein using sterile scissors. Stretch the adjacent connective tissue to expose the external jugular vein and the top of the pectoral muscle (Figure 2A).
   3. Fill the catheter with warm, sterile physiological saline. Insert the catheter into the vein after penetrating through the pectoral muscle (Figure 2B).
      NOTE: Take care to avoid air bubbles inside the catheter. It is important to insert the catheter through the muscle to prevent hemorrhage as the muscle will form a compression point. Even mice with hemostasis defects, such as Itgb3^-/- mice, do not bleed upon catheter insertion with this approach.
   4. Gently remove the needle, and if required, carefully move the catheter deeper into the vein. Connect the syringe and stabilize the catheter by adding a drop of surgical glue (Table of Materials).
      NOTE: It is possible to insert a catheter inside the tail vein, although it requires some practice becuase the vein has a smaller caliber. This is especially difficult when using mice with dark hair, whose tail vein is hardly visible. Due to its larger caliber, positioning the catheter inside the jugular vein may allow larger volumes or repeated treatments to be injected more reliably. Please note that the syringe is filled at first time with warm saline, and do not forget to inject the dead volume before injecting the drug solution.
   5. Carefully return the mouse to a prone position  (Figure 2C). Apply gel (Table of Materials) on the eyes to prevent dryness.

2. Surgery and installation of the cranial ring

NOTE: The full support for mouse installation comprises 4 pieces, a block and a plate, the ring to hold the mouse head, and a screw to fix the ring to the block (Figure 3A). All elements of the support have been obtained from i.materialise.com by 3D printing. The plate is made of acrylonitrile butadiene styrene (ABS) polymer, the block and screw of stainless steel, and the ring of high-grade stainless steel (high-detailed stainless steel) (See Supplemental Figure S1 for dimensions and Supplemental Materials for the 3D printing files). The block is fixed permanently to the support, either screwed or glued. Once the ring is fixed on the mouse skull, it should be screwed to the block holder (Figure 3A).

1. Moisten the hair on the scalp with a paper towel wet with 70% ethanol for disinfection. Remove all the loose hair with a wet paper towel.
   NOTE: For convenience, the scalp may be shaved before surgery.
2. Incise the scalp by forming a T at the midline up to 1 cm and between the ears to expose the calvarium using sterile fine scissors and tweezers.
   NOTE: The ring will be positioned to overlap the frontal and parietal bones (Figure 3B).
3. Use sterile tweezers to expose the skull. Carefully remove the periosteum with scissors and tweezers. Use a sterile cotton swab soaked with physiological saline to remove all the periosteum and any debris or hair, which could alter imaging.
   NOTE: To limit inflammatory reactions, take great care not to damage the bone.
4. Remove any blood traces by rinsing with saline, and rapidly dry the bone with a dry cotton swab. Apply the glue gel to the ring, position it on the exposed bone, and maintain it for a few seconds so that the ring is firmly attached to the skull.
   NOTE: No glue must enter the observation field as it may lead to undesirable autofluorescence.
5. Lightly moisten the skull with a cotton swab immersed in physiological saline.
6. Prepare the silicon dental paste by carefully mixing 1:1 blue and yellow components. Apply dental paste all around the ring for sealing to prevent leakage during imaging. Carefully remove any dental paste or glue that may have entered the ring.
7. Fill the ring with saline and check for leakage.
   NOTE: The saline serves as the immersion medium for the microscope objective. Bleeding may be more critical when using mice with hemostasis-related defects. It is essential to prevent blood from entering the ring as it may blur further imaging.
8. Place the mouse on the support, and screw the ring on the block holder (Figure 3C). Place folded compresses under the animal's head to raise the head and prevent detachment of the ring from the skull.
   NOTE: The head is directly fixed so that it prevents image movements due to breathing.
9. Position the support and mouse assembly in the heated microscope chamber (Figure 3D).
   ​NOTE: Be careful not to dislodge the inserted catheter at any time.

3. Follow-up of the anesthetized mice until the end of the experiment

NOTE: Anesthesia is re-induced every 35 min by alternating subcutaneous (s.c.) injections of ketamine (25 µg/g) in a volume of 5 µL/g body weight and a mixture of ketamine (50 µg/g) and xylazine (5 µg/g) (1.2 µL/g). As it is not possible to open the heated microscope chamber during acquisition, anesthetic monitoring is performed before and after re-administration of the anesthetic by toe pinching. Similarly, toe pinch is performed before beginning each new video recording.

1. If required, stop acquisition and re-induce anesthesia by s.c. injection.
   1. Being careful not to move the mouse head, pinch the skin of the back between the thumb and index finger of the left hand. Subcutaneously inject the anesthetic in the fold of the back skin using the right hand (or vice versa for left-handed people).
2. Check for the presence of saline in the ring, and if necessary, fill with fresh saline. Go on recording for the next 35 min, and proceed similarly for the recording duration.
3. At the end of the experiment, euthanize the mouse by cervical dislocation before it wakes up.
   ​NOTE: The mouse is kept anesthetized for a maximum duration of 3 h, in agreement with the local ethical and animal welfare committee.

4. Two-photon imaging

NOTE: See the Table of Materials for details about the microscope and related equipment. Images were recorded with a resonant scanner (12 or 8 kHz). The bidirectional mode was set up to increase speed acquisition as pixels are recorded in both directions; hence, any mismatch in the phase must be corrected with the control panel "phase correction." Finally, an adapted averaging was set up as a compromise between speed of acquisition and signal-to-noise ratio.

1. Switch on the two-photon microscope, including the laser and resonant scanner (output laser power usually ~1.8-2.5 W depending on the laser).
2. Set the appropriate laser wavelength for the chosen fluorophores. Set the appropriate wavelength for the recovery of emitted lights.
   NOTE: Here, a wavelength of 913 nm and hybrid detectors (500-550 nm and 575-625 nm) were used to simultaneously excite and detect AlexaFluor-488 and Qtracker-655, respectively.
3. Using the intrajugular catheter, inject the fluorescent tracer to label the vasculature (Qtracker-655; Table of Materials). Inject 50 µL/mouse of a solution at 0.2 µM, renewed once every hour for a maximum of 150 µL/experiment.
   NOTE: Other tracers can be used such as high molecular weight dextran¹. In that case, consider that blood viscosity can be altered by dextran and modify some hemorheological parameters.
4. Place the microscope stage with the support and mouse under the objective of the microscope. Check that the ring is always filled with saline and immerse the objective. Use epifluorescence to locate the bone marrow vessels (red) and the presence of megakaryocytes (green) aligned along sinusoid vessels.
   NOTE: The skull bone has a thickness lower than 100 µm². The marrow is located beneath the bone and is easily recognized by the fluorescent green megakaryocytes and the dense red anastomosed sinusoid vessels labeled by Qtracker-655.
5. Long acquisitions (megakaryocytes, proplatelets)
   1. Search for the region of interest. Apply image acquisition parameters as mentioned above.
   2. For long acquisitions (e.g., visualization of proplatelet formation or elongation), acquire z-stack images with optimized intervals (usually between 0.85 to 1.34 µm) using an 8 kHz resonant scanner.
      NOTE: A 12 kHz resonant scanner can increase acquisition speed.
   3. Acquire 384 x 384 pixel images to visualize the entire megakaryocyte and vessel. Use line averaging, as desired, for rapid acquisition with good resolution. Determine the time interval of image stack acquisition, usually 10 s or 30 s for proplatelet visualization.
      NOTE: Here, a line averaging of 8 was used to obtain a higher signal-to-noise ratio, which allowed the acquisition of 1 image/10 s with an 8 kHz resonant scanner. Ensure that line averaging during live acquisition is allowed.
6. Short and rapid acquisition (platelets)
   NOTE: For shorter acquisition of rapid events, such as platelet velocity, it is essential to maximize the frame acquisition rate. Selecting the proper acquisition type will optimized the frame acquisition rate (i.e. a space-time acquisition type “xyt”). 
   1. Minimize (e.g., 256 x 90 pixels) and adjust image size to the vessel by rotating the imaging field, if required.
   2. Use the highest available scanning speed.
   3. Use bidirectional scanning.
      NOTE: For bidirectional scanning, ensure that the phase is well-adjusted.
   4. Minimize line averaging to find the optimal compromise between image definition and rapidity of acquisition.
      NOTE: For platelet velocity measurements, a pixelized image may be sufficient if it helps increase the rapidity of acquisition.
   5. Minimize z-stack to increase acquisition rapidity. Acquiring only 1 z-stack is sufficient to quantify platelet velocity. Acquire for 10-20 s to measure platelet velocity.
      NOTE: A maximum of 133 frames/s can be obtained using these conditions with a 12 kHz resonant scanner. The parameters must be adapted to the conditions. In some vessels, platelet velocity is nevertheless too high to quantify platelet velocity with these settings reliably.
7. Proplatelet width and length measurement
   NOTE: To load LIF-files obtained on Leica software with ImageJ, first download and install the LOCI plug-in on ImageJ to use the Bio-Formats Importer.
   1. Proplatelet width
      1. Open the movie with ImageJ. To perform width measurement, use only the movie of the proplatelets without the vessel. To separate the different channels, click on Image|Color and select Split Channels.
      2. For a z-stack movie, make an average z-projection for each time point by clicking on Image|Stacks and select Z Project. For Projection type, enter average Intensity.
      3. Treat the image to remove the noise. Subtract the background by clicking on Process|Subtract background and apply Rolling ball radius: 50.0 pixels. Smooth the image by clicking on Process and Smooth.
      4. Trace a line close to the base of the proplatelet (near the megakaryocyte, avoiding any other object).
         NOTE: Do not forget to set the scale of the image to obtain the value in the desired unit by clicking on Analyze and Set Scale. By default, the unit of the image is in pixels.
      5. Plot the intensity profile, fit a gaussian curve, and calculate the Full Width at Half Maximum (FWHM). For that, write the following script in a macro by clicking on Plugins|New and select Macro and run it by clicking on Macro|Run Macro: y=getProfile(); x=Array.getSequence(y.length); Fit.doFit("Gaussian",x,y); Fit.plot(); sigma=Fit.p(3); FWHM = (2 * sqrt( 2 * log(2) ) ) * sigma; print(FWHM).
         NOTE: Ensure that there is only a single peak in the line scan, corresponding to the proplatelet intensity.
      6. Perform the actions described in step 4.7.5 on 1 slice or repeated over time to get the mean width of the proplatelet
   2. Proplatelet length
      1. Trace a line along the proplatelet, from the base close to the megakaryocyte to the top. Use a segmented line to follow the proplatelet shape. Repeat it for each time point.
      2. Remember to save each line using the ROI Manager: click on Analyze | Tools and select ROI Manager. Click on Add in the ROI Manager window to add each line to the manager as is it selected. To save all the lines in a folder, click in the ROI Manager on Deselect | More and click on Save. From the ROI Manager, extract the length value of each line (Measure) from the results table.
         NOTE: Do not forget to set the scale of the image to obtain the value in the desired units by clicking on Analyze and Set Scale. By default, the unit of the image is in pixels.
8. Platelet velocity measurement
   NOTE: Platelet velocity is calculated from the video recording of fluorescent platelets flowing inside the vessel over 10-20 s. A first-image treatment is performed to obtain cyclically repeated line-scan data. Motion leads to streaks whose angle depends on the velocity, allowing its calculation. Here, velocity was calculated with a method that uses the Radon transform, which is more robust than the classical Singular Value Decomposition (SVD) method¹⁵.
   1. Image treatment with ImageJ
      1. Open the movie with ImageJ. Apply a Gaussian Blur filter by clicking on Process | Filters, select Gaussian Blur, and apply Sigma (Radius): 2.00. Choose a small region to measure the flow, and trace 3 adjacent lines following the flow direction.
         NOTE: Do not forget to save each line by using the ROI Manager by clicking on Analyze | Tools | ROI Manager. Click on Add in the ROI Manager window to add each line and save them; click first on Deselect | More and then click on Save.
      2. Using ImageJ line-scan software, build a kymograph for each line by clicking on Image | Stacks; select Reslice and apply the following parameters: Output spacing (microns): 1.000 / Slice count: 1 / Avoid interpolation. Save the resulting space-time images.
         NOTE: The resulting image shows streaks corresponding to the linear motion of the platelets in the blood flow over time. The angle of a streak is a function of the velocity.
   2. Velocity measurement using Matlab or GNU Octave software
      NOTE: The velocity of the platelets in the blood flow is estimated from the space-time images using a computational analysis method based on the Radon transform described by Drew and collaborators¹⁵. The description of this mathematical method is beyond the scope of this review, and the reader is referred to the publication by Drew et al. for details regarding the calculation¹⁵. Both the Matlab code and more information are available on their website https://www.drew-lab.org/code.
      1. Open the code using Matlab or GNU Octave software.
         NOTE: The code may require adaptation according to the desired study.
      2. Extract the value from the 3 space-time images, corresponding to the adjacent lines drawn in the same vessel portion. Calculate the mean platelet flow velocity value for this portion of the vessel.
         NOTE: Do not forget to convert the result in the correct unit of measure (e.g., µm/s).

Representative Results

Using this protocol, the fluorescent tracer, Qtracker-655, was intravenously administered to image anastomosed marrow sinusoid vessels in the skull bone marrow and the flow direction as depicted by the arrows (Figure 4A, left). Using mTmG mice, eGFP-fluorescent platelets were recorded over 20 s in each vessel branch, and their velocity was measured using ImageJ and GNU Octave software (Figure 4A, right). Note the heterogeneity in flow velocity and direction. Sinusoid vessels present complex flows due to the anastomoses, with the presence of flow-reflow and even stasis, as shown in the bifurcation recorded in Video 1. The same bifurcation is shown in Figure 4B (left), with the red and blue arrows highlighting opposite flows. Platelet velocity in this bifurcation has been measured and reported in the graph (Figure 4B, right). The red tracing corresponds to the velocity in the left vessel branch and the blue tracing to that in the right vessel branch. This shows irregularity over time in each vessel branch, with phases of acceleration, stasis, and deceleration. It is important to note that in some vessels, blood flow was too rapid for reliable analysis under the present acquisition conditions.

Proplatelet elongation recording allows the visualization of their various morphologies over time (Figure 5). Some proplatelets elongate with irregular morphologies (Figure 5Ai and Video 2). Others, usually thinner ones, may extend over long distances necessitating the movement of the acquisition window to follow their extension (Figure 5Aii and Supplemental Video 1). Others are shorter and thicker and usually elongate slowly (Figure 5Aiii and Supplemental Video 2). When megakaryocytes extend proplatelets in areas of complex flows, such as the bifurcation shown in Figure 4B, proplatelets are tossed from one vessel branch to the other according to flow direction (Figure 5B and Video 3). The importance of hemodynamic forces for proplatelet elongation in the flow direction was evidenced when the mouse unexpectedly underwent cardiac arrest. Stopping blood flow relaxed the proplatelets extended by the megakaryocyte (Figure 5C and Video 4).

When the quality/resolution of acquisition is appropriate, morphological parameters can be measured, such as the length of proplatelets until detachment, their width at the base of the proplatelet or any other defined location, or the size of the buds (Figure 6A). A mean maximal proplatelet length of ~185 µm (range 41.5-518.5 µm) and a mean proplatelet width of 5.2 µm (range 2.8-8 µm) was calculated (Figure 6A). Plotting proplatelet length as a function of time allows the visualization of the behavior of proplatelets during phases of elongation, stasis, or even retraction (Figure 6B). Measurement of proplatelet elongation velocity is an important parameter, which directly reflects the elongation, stasis, or retraction behavior of proplatelets³.

Proplatelet elongation velocity can be measured if the proplatelet remains attached to its mother cell; once detached, it will be immediately carried away by the flow and disappear from the visualization window. In wild-type (WT) mice, the mean proplatelet elongation velocity was ~10 µm/min (Figure 6C), in agreement with previous publications¹. Finally, it is possible to inject drugs through the catheter inserted in the jugular vein. For instance, it has been observed that intravenous administration of vincristine, a drug that depolymerizes microtubules, led to the retraction of proplatelets from WT mice but had no effect on proplatelets from myosin IIA-deficient mice(Myh9^-/-) (Video 5 and Video 6)³.

Figure 1: Schematic representation of proplatelet formation in the bone marrow. After differentiation from hematopoietic stem cells, large megakaryocytes align along sinusoid vessels and extend cytoplasmic projections, called proplatelets, through the endothelial barrier. Proplatelets elongate and detach under the influence of hemodynamic forces to further remodel into platelets in the downstream microcirculation. Please click here to view a larger version of this figure.

Figure 2: Insertion of the catheter. (A) An incision is made to expose the jugular vein and the top of the pectoral muscle. (B) The catheter filled with saline is inserted in the jugular vein by passing through the muscle, creating a compression point. (C) The mouse is carefully turned so that it lies ventral surface downward with the catheter well-positioned. Please click here to view a larger version of this figure.

Figure 3: Experimental setup used for two-photon imaging through the skull bone. (A) The 3D printed stainless steel block is fixed on the ABS polymer support (see Supplemental Figure S1 and Supplemental Materials for the 3D designs). The block and the skull ring are designed so that the ring can be screwed to the block. Note that, for ease of use, the screw head has been embedded in Epon. (B) Schematic showing the positioning of the ring on the exposed skull bone. (C) Photographs illustrating the mouse with the ring glued to the skull, the head on a tissue pad, and (D) the positioning under the microscope objective. Abbreviations: 3D = three-dimensional; ABS = acrylonitrile butadiene styrene. Please click here to view a larger version of this figure.

Figure 4: Measure of platelet velocity in skull bone marrow anastomosed sinusoid vessels. (A) Left, two-photon imaging of sinusoid vessels labeled by Q-Dot injection. z-projection showing the anastomosis of the sinusoid vessels in the skull bone marrow. Arrows indicate the direction of flow, illustrating the complexity of the flows in sinusoids. Right, mean flow velocity estimated by the measurement of the platelet velocity in each vessel branch. The numbers on the x-axis correspond to the arrow numbers in the image on the left. (B) Left, sinusoid bifurcation delineated by dotted lines, showing platelets flowing in opposite directions (red and blue arrows) (corresponding to Video 1). Two-photon single plane image. Right, a graph showing the flow velocity in each portion of the vessel as a function of time (red line, left side; blue line, right side), showing phases of stasis, acceleration, and deceleration. Scale bars = 20 µm. Please click here to view a larger version of this figure.

Figure 5: Various proplatelet morphologies observed in the skull marrow by two-photon microscopy. (A) Three representative in vivo proplatelets: z-projection images from time-lapse experiments (shown in Video 2, Supplemental Video 1, and Supplemental Video 2) showing the various morphologies of proplatelets (arrows) extending within bone marrow sinusoids. Proplatelets and megakaryocytes are in green; sinusoid vessels are in red. (B) Time-lapse images of a proplatelet in the same bifurcation as in Figure 4B, oscillating according to the direction of flow (also shown in Video 3). (C) z-projection time-lapse images showing relaxation of 3 proplatelets after cessation of blood flow due to cardiac arrest of the mouse (shown in Video 4), previously aligned in the same flow line (indicated by the white, yellow, and blue arrows). Scale bars = 20 µm. Please click here to view a larger version of this figure.

Figure 6: Proplatelet morphological analyses. (A) Upper, z-projection image depicting the site where the proplatelet width was measured along with maximal length; below are graphs representing the maximal length of 25 individual proplatelets and the mean proplatelet width of the same proplatelets (same gray code) throughout elongation. i) The proplatelet width was measured by drawing a perpendicular line at a fixed position close to the base of the proplatelet, and the profile intensity was recorded at this position every minute. A Gaussian curve was fitted to this profile, and the FWHM was considered to represent the width of the proplatelet. The mean width was then calculated by averaging the widths of the proplatelet measured at each time point at the fixed position. Bars are mean ± sem of the width measured every minute during acquisition; ii) the maximal proplatelet length was also determined, from the base of the proplatelet to its tip. (B) Individual tracings showing the elongation behavior of WT proplatelets, some of them presenting pause and retraction phases. To better highlight the pauses and retractions upon the process of proplatelet growth, the length measured every 10 s was represented as a percentage of the maximal length of each proplatelet. (C) Scatter plot representing the proplatelet elongation speed. The net mean proplatelet elongation speed (including pauses and retractions) was calculated as the change in proplatelet length at 1 min intervals, averaged over the whole time sequence. Individual values and mean ± sem. Abbreviations: FWHM = Full Width at Half Maximum; sem = standard error of the mean; WT = wild-type.

Video 1: Reverse flow within sinusoid vessels. Circulating platelets (green) within the sinusoid vessels (red, Qtracker) (single confocal z-plane). An anastomosed sinusoid vessel is shown presenting bifurcations where the flow is unstable and displays phases of stasis, acceleration, and deceleration. Acquired with a Leica SP5 microscope equipped with a 12 kHz resonant scanner and a 25x water objective with a numerical aperture of 0.95 (Leica), one plane, 256 x 90 pixels, bidirectional acquisition, line averaging 2, 133 frame/s. Please click here to download this Video.

Video 2: Elongating proplatelets display various morphologies. Representative time-lapse video recording (z-projection) in a WT mouse with megakaryocytes extending proplatelets with various morphologies (green) within bone marrow sinusoids (red, Qtracker). Acquired with a Leica SP5 microscope equipped with a 12 kHz resonant scanner and a 25x water objective with a numerical aperture of 0.95 (Leica), z-stack acquisition at 10 s intervals, z-depth of 4 µm, x-y resolution 384 x 384 pixels, line averaging 8. Whenever necessary, the profiles were aligned using the ImageJ template matching plug-in to obtain an average z-projection. Depending on the image quality, the background was subtracted, the image was smoothed, and the brightness and contrast adjusted. Please click here to download this Video.

Video 3: Proplatelets are tossed by changes in the flow direction. Time-lapse video showing a wild-type megakaryocyte in the sinusoid bifurcation shown in Video 1, in the process of extending a proplatelet that oscillates in one vessel branch depending on the direction of flow (z-projection). Acquired with a Leica SP5 microscope equipped with a 12 kHz resonant scanner and a 25x water objective with a numerical aperture of 0.95 (Leica), z-stack acquisition at 10 s intervals, z-depth of 4 µm, x-y resolution 384 x 384 pixels, line averaging 8. The profiles were aligned using the ImageJ template matching plug-in, and an average z-projection was obtained. The background was subtracted, the image was smoothed, and the brightness and contrast adjusted. Please click here to download this Video.

Video 4: Impact of flow on proplatelet alignment and tension. Time-lapse video recording of a proplatelet before and after cardiac arrest. The video shows three proplatelets (green, labeled with AF-488 anti-GPIX antibody derivative) in the same flow line before cardiac arrest, thus poorly individualized at the resolution of the two-photon microscopy. Following cardiac arrest, the three proplatelets separate and become relaxed in the absence of blood flow. Video acquired with a Leica SP8 confocal microscope equipped with an 8 kHz resonant scanner, image with a 25x water objective with a numerical aperture of 0.95 (Leica). The profiles were aligned using the ImageJ template matching plug-in, and an average z-projection was obtained. The background was subtracted, the image was smoothed, and the brightness and contrast adjusted. Please click here to download this Video.

Video 5: Impact of vincristine drug injection on the proplatelet behavior of WT mice. Time-lapse video recording of a proplatelet before and after administration of vincristine (1 mg/kg) (z-projection). The elongating proplatelet (green) within the sinusoid vessels (red) starts to retract after intravenous administration of vincristine. Video was acquired with a Leica SP8 confocal microscope equipped with an 8 kHz resonant scanner, image with a 25x water objective with a numerical aperture of 0.95 (Leica). The profile was aligned using the ImageJ template matching plug-in, and an average z-projection was obtained. The background was subtracted, the image was smoothed, and the brightness and contrast adjusted. Please click here to download this Video.

Video 6: Impact of vincristine drug injection on the proplatelet behavior of myosin-deficient mice. Time-lapse video recording of a proplatelet before and after administration of vincristine (1 mg/kg) (z-projection). The elongating proplatelet (green) within the sinusoid vessels (red) continues to elongate in the Myh9^-/- mice. Video was acquired with a Leica SP8 confocal microscope equipped with an 8 kHz resonant scanner, image with a 25x water objective with a numerical aperture of 0.95 (Leica). The profile was aligned using the ImageJ template matching plug-in, and an average z-projection was obtained. The background was subtracted, the image was smoothed, and the brightness and contrast adjusted. Please click here to download this Video.

Supplemental Figure S1: Design for the 3D printing of the skull ring and support. This design was used to 3D print (A) the skull ring in high-quality stainless steel, (B) the screw, (C) the block in stainless steel, and (D) the support in ABS polymer. Abbreviations: 3D = three-dimensional; ABS = ABS = acrylonitrile butadiene styrene. Please click here to download this Figure.

Supplemental videos:

Supplemental Video 1: Representative time-lapse video recording (z-projection) in a wild-type mouse with megakaryocytes extending proplatelet (green) within bone marrow sinusoids (red, Qtracker). Acquired with a Leica SP5 microscope equipped with a 12 kHz resonant scanner and a 25x water objective with a numerical aperture of 0.95 (Leica), z-stack acquisition at 10 s intervals, z-depth of 1.34 µm, x-y resolution 384 x 384 pixels, line averaging 8. Whenever necessary, the profile was aligned using the ImageJ template matching plug-in to obtain an average z-projection. Depending on the image quality, the background was subtracted, the image was smoothed, and the brightness and contrast adjusted. Please click here to download this Video.

Supplemental Video 2: Representative time-lapse video recording (z-projection) in a wild-type mouse with megakaryocytes extending proplatelet (green) within bone marrow sinusoids (red, Qtracker). Acquired with a Leica SP5 microscope equipped with a 12 kHz resonant scanner and a 25x water objective with a numerical aperture of 0.95 (Leica), z-stack acquisition at 10 s intervals, z-depth of 1.34 µm, x-y resolution 384 x 384 pixels, line averaging 8. Whenever necessary, the profile was aligned using the ImageJ template matching plug-in to obtain an average z-projection. Depending on the image quality, the background was subtracted, the image was smoothed, and the brightness and contrast adjusted. Please click here to download this Video.

Supplemental Materials:

Coding File 1: 3D printing-Block. 3D object file for the printing of the block. Please click here to download this Coding File.

Coding File 2: 3D printing-Skull ring. 3D object file for the printing of the skull ring. Please click here to download this Coding File.

Coding File 3: 3D printing-Support. 3D object file for the printing of the support. Please click here to download this Coding File.

Coding File 4: 3D printing-Screw ring. 3D object file for the printing of the screw. Please click here to download this Coding File.

Discussion

The mechanisms of platelet formation are highly dependent on the bone marrow environment. Hence, intravital microscopy has become an important tool in the field to visualize the process in real-time. Mice with fluorescent megakaryocytes can be obtained by crossing mice expressing the Cre recombinase in megakaryocytes with any floxed reporter mice containing a conditional fluorescent gene expression cassette. Here, mTmG reporter mice were used (B6.129(Cg)-Gt(ROSA)26Sor^{tm4(ACTB-tdTomato,-EGFP)Luo10}) crossed with Pf4-cre mice¹¹, allowing intense green fluorescence labeling in megakaryocytes and platelets¹². It is important to note that some leakage has been reported with Pf4-cre mice in leukocytes, including monocytes/macrophages and fibroblasts^12,13.

However, megakaryocytes are easily recognized by their large size and nucleus. Experimenters new to the field could compare observations in [mTmG; Pf4-cre] mice with those obtained after megakaryocyte-specific labeling. This is possible through intravenous administration of an Alexa-fluor-labeled antibody directed against the platelet-specific protein, GPIX (CD42a)^3,14. This latter approach is also beneficial for imaging fluorescent megakaryocytes from any genetically engineered mouse without performing tedious and time-consuming mouse crossings with fluorescent reporter mice. Here, 5-7-week-old mice were used as younger mice present with a thinner bone that is more translucent and facilitates imaging compared to older mice. However, mice that are too young may present difficulties in the installation of the cranial ring without leakage.

In this study, intravital imaging was performed using a Leica microscope equipped with a 25x water objective with a numerical aperture of 0.95 (Table of Materials). The objective should be conical so that it can enter the cup formed by the cranial ring with an optimal working distance. Images were recorded with a resonant scanner (12 or 8 kHz). The overall settings will depend on the type of scientific question to be answered and should be optimized depending on the magnification, resolution, and acquisition speed needed. For instance, with a 12 kHz resonant scanner, a 25x objective would be better suited to measure platelet velocity, whereas an 8 kHz scanner and a lower magnification could be sufficient to record proplatelet elongation.

The most critical step of the protocol is the correct attachment of the ring to the skull bone so that no leakage occurs, and so that it does not detach during recording. For that, it is essential that the skull bone be dry just before applying the ring. Once the ring is glued, rapidly re-humidify the bone. Furthermore, due to the force exerted by the weight of the head on the ring, the ring is subjected to a tension that may cause it to detach partially during the recording. This may cause leakage and thus, dryness of the bone, leading to both poor-quality images and undesirable inflammatory reactions. This problem can be easily avoided by adding a folded compress under the nose of the mouse to support the head (Figure 3C). Another critical point is to minimize bleeding and the presence of blood inside the ring as it may lead to blurred images. This must be kept in mind when studying mice that present defective hemostasis and are prone to bleeding. Some tracers, such as dextran, may present leakage in the bone marrow cavity if injected in a concentrated form, which worsens vessel imaging, though not the green fluorescence of megakaryocytes. Qtracker-655 does not leak much but is cleared from the circulation within 30 min so that new injections of the tracer are required to visualize vessels over longer periods.

One limitation of this method described here is the necessity to stop acquisition every 35 min to re-inject anesthetics. This limitation can be overcome if an anesthesia machine is available. In that case, anesthesia can be induced similarly by i.p. injection of a mixture of ketamine and xylazine, which is the easiest way to perform the insertion of the catheter and the minor surgery. Once positioned below the microscope objective, anesthesia is then maintained by inhalation of gases (mixture of oxygen, anesthetics such as isoflurane and ambient air). In this way, observations can be made over several hours without interruption. However, for ethical reasons, observations were limited to 3 h in this study. Another limitation in using ketamine/xylazine mixture could be its potential deleterious impact on platelet functions¹⁶ that could require alternative anesthesia schemes.

Overall, the principal merit of this well-established protocol is its noninvasive aspect with only minimal surgery. Other protocols have been developed to perform time-lapse intravital imaging in long bones. These rely on either invasive surgery that requires muscle tissue removal and bone abrasion^17,18, implanted endoscopic probes close to the femur head^19,20, or installing a window chamber in the femur²¹. All these procedures result in major trauma and varying degrees of inflammation. Inflammation is not only detrimental to the mouse, but it has also been reported to modify the process of platelet formation⁶ so that uncontrolled inflammatory conditions might lead to discrepancies and misinterpretation of the data. That is why this procedure was chosen to avoid inflammatory reactions. However, depending on the scientific question, experimenters may prefer to have a deeper observation field and higher resolution (less scattering), which is possible by using a skull-thinning-based approach at the expense of a low degree of inflammatory reaction.

Because of its noninvasive nature, the major limitation of this method is the maximal depth that can be reached. Due to the density of the bone and its scattering properties, it is possible to image regions within a depth of only a few hundred micrometers, preventing observations of the whole calvarial marrow. The development of three-photon microscopy, which relies on even higher infrared excitation wavelengths, seems to be a promising approach with superior deep-tissue resolution. It has already been successfully used to image deep into the brain even through the bone^22,23, opening up the exciting possibility to image bone marrow within long bones without having to thin the bone or implant a surgical window.

In summary, this method is increasingly used to study the behavior of megakaryocytes in the bone marrow and visualize the extension of proplatelets.In addition, by allowing the visualization of the outer calvarial compact bone as well as part of the underlying marrow, this method has already been used in many applications outside the platelet field. It has been used for the study of both bone and marrow cell dynamics in their environment, including osteoblasts and osteoclasts, leukocyte trafficking and marrow exit, endothelial cells and microvasculature architecture, blood flow dynamics in marrow microvessels, or cell homing²⁴.

Materials

Name	Company	Catalog Number	Comments
GNU Octave software	GNU Project		https://www.gnu.org/software/octave/
Histoacryl 5 x 0, 5 mL	Braun	1050052	injectable solution of surgical glue
HyD hybrid detectors Leica Microsystems 4tunes	Leica Microsystems
ImageJ	GNU project		Minimum version required
Imalgene/Ketamine 1000 fl/10 mL	Boehring	03661103003199	eye protection
Leica SP8 MP DIVE microscope equipped with a 25x water objective, numerical aperture of 0.95	Leica Microsystems		simultaneous excitation of AlexaFluor-488 and Qtracker-655
Matlab	MathWorks		https://www.mathworks.com/
Ocrygel 10 g	Laboratoires T.V.M.	03700454505621	Silicon dental paste blue and yellow
Picodent twinsin speed	Rotec	13001002
Qtracker 655 vascular label	Invitrogen	Q21021MP	injectable solution
Resonant scanner, 8 or 12 kHz
Rompun Xylazine 2% fl/25 mL	Bayer	04007221032311
Superglue gel			to glue the ring to the bone
Surflo IV catheter - Blue 22 G	Terumo	SR-OX2225C1
Ti:Saph pulsing laser (Coherent) (femtosecond)	Coherent