A refined method of tissue clearing was developed and applied to the adult mouse heart. This method was designed to clear dense, autofluorescent cardiac tissue, while maintaining labeled fibroblast fluorescence attributed to a genetic reporter strategy.
Cardiovascular disease is the most prevalent cause of mortality worldwide and is often marked by heightened cardiac fibrosis that can lead to increased ventricular stiffness with altered cardiac function. This increase in cardiac ventricular fibrosis is due to activation of resident fibroblasts, although how these cells operate within the 3-dimensional (3-D) heart, at baseline or after activation, is not well understood. To examine how fibroblasts contribute to heart disease and their dynamics in the 3-D heart, a refined CLARITY-based tissue clearing and imaging method was developed that shows fluorescently labeled cardiac fibroblasts within the entire mouse heart. Tissue resident fibroblasts were genetically labeled using Rosa26-loxP-eGFP florescent reporter mice crossed with the cardiac fibroblast expressing Tcf21-MerCreMer knock-in line. This technique was used to observe fibroblast localization dynamics throughout the entire adult left ventricle in healthy mice and in fibrotic mouse models of heart disease. Interestingly, in one injury model, unique patterns of cardiac fibroblasts were observed in the injured mouse heart that followed bands of wrapped fibers in the contractile direction. In ischemic injury models, fibroblast death occurred, followed by repopulation from the infarct border zone. Collectively, this refined cardiac tissue clarifying technique and digitized imaging system allows for 3-D visualization of cardiac fibroblasts in the heart without the limitations of antibody penetration failure or previous issues surrounding lost fluorescence due to tissue processing.
Although cardiomyocytes comprise the greatest volume fraction in the heart, cardiac fibroblasts are more plentiful and are critically involved in regulating the baseline structural and reparative features of this organ. Cardiac fibroblasts are highly mobile, mechanically responsive, and phenotypically ranging depending on the extent of their activation. Cardiac fibroblasts are necessary to maintain normal levels of extracellular matrix (ECM), and too little or too much ECM production by these cells can lead to disease1,2,3. Given their importance in disease, cardiac fibroblasts have become an increasingly important topic of investigation towards identifying novel treatment strategies, especially in attempting to limit excessive fibrosis4,5,6,7. Upon injury, fibroblasts activate and differentiate into a more synthetic cell type known as a myofibroblast, which can be proliferative and secrete abundant ECM, as well as have contractile activity that helps remodel the ventricles.
While cardiac fibroblasts have been extensively evaluated for their properties in 2-D cultures6,8,9,10, much less is understood of their properties and dynamics in the 3-D living heart, either at baseline or with disease stimulation. Here, a refined method has been described to tissue clear the adult mouse heart while maintaining the fluorescence of fibroblasts labeled with a Rosa26-loxP-eGFP x Tcf21-MerCreMer genetic reporter system. Within the heart, Tcf21 is a relatively specific marker of quiescent fibroblasts4. After tamoxifen is given to activate the inducible MerCreMer protein, essentially all quiescent fibroblasts will permanently express enhanced green fluorescent protein (eGFP) from the Rosa26 locus, which allows for their tracking in vivo.
Numerous well-established tissue clearing protocols exist, some of which have been applied to the heart11,12,13,14,15,16,17. However, many of the reagents used in different tissue clearing protocols have been found to quench endogenous fluorescence signals18. Additionally, the adult heart is difficult to clear due to abundant heme group-containing proteins that generate autofluorescence19. Therefore, the goal of this protocol was to preserve fibroblast marker fluorescence with the simultaneous inhibition of heme autofluorescence in the injured adult heart for optimal 3-D visualization in vivo12,13,14,16,17,20.
Previous studies attempting to examine the cardiac fibroblast in vivo employed perfused antibodies to label these cells, although such studies were limited by antibody penetration and cardiac vascular structure14,16,17,20. Although Salamon et al. have shown tissue clearing with maintenance of topical neuronal fluorescence in the neonatal heart, and Nehrhoff et al. have shown maintenance of fluorescence marking myeloid cells, maintenance of endogenous fluorescence through the entire ventricular wall has not yet been demonstrated, including the visualization of adult cardiac fibroblasts at baseline or following injury13,20. This tissue clearing protocol refines a mixture of previous protocols based on the CLARITY method (clear lipid-exchanged acrylamide-hybridized rigid Imaging/immunostaining/in situ-hybridization-compatible tissue hydrogel) and PEGASOS (polyethylene glycol (PEG)-associated solvent system). This refined protocol permitted a more robust examination of cardiac fibroblasts in the mouse heart at baseline and of how they respond to different types of injury. The protocol is straightforward and reproducible and will help characterize the behavior of cardiac fibroblasts in vivo.
All experiments involving mice were approved by the Institutional Animal Care and Use Committee (IACUC) at Cincinnati Children’s Hospital Medical Center. The institution is also AAALAC (American Association for Accreditation of Laboratory Animal Care) certified. Mice were euthanized via cervical dislocation, and mice undergoing survival surgical procedures were given pain relief (see below). All methods used for pain management and euthanasia are based on recommendations of the Panel on Euthanasia of the American Veterinary Medical Association. All mice were housed in corn cob bedding units with water and food available at all times. Mice were housed 4 to a cage with the same sex. For surgery or uninjured tissue clearing, equal numbers of 6–8-week-old male and female mice were used.
NOTE: Sterile surgical conditions were maintained in all surgeries. The surgeon changed into clean scrubs and a sterile gown and then donned shoe covers and a hairnet. The surgeon then scrubbed their hands with chlorhexidine and donned sterile surgical gloves. The surgeon was assisted by a technician who sedated, shaved, and scrubbed the incision site 3 times each, alternating between 2% chlorohexidine gluconate and 70% isopropanol. The mice were then brought to the surgeon, and surgery was performed. Between animals, the instruments were sterilized in a bead sterilizer.
1. Cre recombination
2. Surgical models
3. Clearing adult mouse hearts using a modified active CLARITY protocol
4. Imaging cleared hearts using an upright single photon confocal microscope
NOTE: The imaging apparatus consists of the bottom half of a 10 cm glass Petri dish, a 3D printed bottom reservoir, a round glass coverslip, and a 3D printed top reservoir (Figure 1C–E). 3D printed materials were made in-house by the Cincinnati Children’s Hospital Clinical Engineering Department.
Cardiac fibroblasts are essential for baseline function of the heart as well as for response to cardiac injury. Previous attempts to understand the arrangement and morphology of these cells have been conducted largely in 2-D settings. However, a refined cardiac tissue clearing (Figure 2) and 3-D imaging technique has been published, which allows for the advanced, more detailed visualization of cardiac fibroblasts. With this imaging technique, fibroblasts were found to be densely packed and have a spindled morphology in uninjured hearts (Figure 3, Supplemental Videos S1–4).
After left ventricular tissue clearing had been accomplished in an uninjured heart, the protocol was applied to several injury models to examine how this clearing protocol would perform when studying injured heart tissue. Mice were subjected to I/R injury by temporary closure of the left coronary artery (LCA) for 1 h followed by reperfusion lasting 3, 7, 14, or 28 days. These experiments showed that there was a loss of cardiac fibroblasts in the ischemic region right after I/R injury, but that by day 7 and day 14, fibroblasts migrated or proliferated to repopulate this area of the injured heart. By day 28, the cardiac fibroblast population surrounding the injured areas of the heart were at their greatest density (Figure 4, Supplemental Videos S5–8).
MI injury is a surgical model resulting from permanent LCA ligation (no reperfusion). Hearts that had undergone MI surgery were excised at 1.5 and 3 days following surgery for fixation, clearing, and analysis. There were very few fibroblasts remaining in the injured left ventricle at 1.5 days following surgery (Figure 5A, Supplemental Video S9). However, by day 3, fibroblasts expanded and were present in most of the left ventricle except for one small region, presumably having migrated in and/or proliferated from a population in the border zone (Figure 5B, Supplemental Video S10). Again, analysis software was used to better visualize fibroblast localization, and areas of loss of cardiac fibroblasts were outlined in orange (Figure 5). This analysis better showed the initial loss of cells at day 1.5 following MI and how fibroblasts either proliferated or migrated into that damaged area by day 3 to ostensibly repair the area and form a scar.
In addition to ischemic injury, the reaction of cardiac fibroblasts to high blood pressure is not well understood. To discover the response of these cells to high blood pressure, angiotensin II and phenylephrine were administered. Angiotensin II and phenylephrine (Ang/PE) are drugs that cause persistent high blood pressure and cardiac fibroblast activation with areas of interstitial fibrosis, confirmed through application of this tissue clearing protocol to ang/PE treated hearts (Figure 6A)5,28. In contrast to ischemic surgery, infusion of Ang/PE over several weeks does not result in loss of cardiac tissue and wall thinning. Instead, a different result was observed in fibroblast behavior following this injury.
As the heart pumps, the myocardium twists, following a right-handed helix pattern29,30,31. In the Ang/PE model of injury, the fibroblasts aligned along the axis of this right-handed helix contraction pattern using the refined tissue clearing protocol (Figure 6B). The hypothesis to explain this behavior is that cardiac fibroblasts were sensing the direction of ventricular wall strain and aligning within the myofibers to provide the greatest support within the ECM as the fibrotic response acutely unfolded during agonist infusion (Figure 7). Another interesting finding was that the fibroblasts appeared small and rounded, as opposed to the spindle shape seen in models of I/R and MI injury (Figure 6, Supplemental Video S11).
Figure 1: Mice and materials used for tissue clearing hearts.
(A) Schematic of breeding strategy for Tcf21-MerCreMer (mcm) x Rosa26eGFP mice used for tissue clearing. (B) Timeline of tamoxifen treatment and surgery performed in mice. Uninjured mice sacrificed on day 14. (C) 3-D printed well to hold heart tissue sealed to a glass Petri dish with vacuum grease. (D) Addition of a round coverslip vacuum grease sealed to the top of the 3-D printed well set on top of the bottom well. (E) Tissue cleared heart in the bottom well of the tissue holding apparatus, filled with Refractive Index Matching Solution (RIMS). Coverslip (as in D) and top 3-D printed well piece laid over the bottom 3-D printed reservoir component, containing cleared heart tissue. Glycerol is added to the top reservoir for glycerol immersion microscopy. Scale bars = 1 cm. Reservoir blueprints can be found in Supplemental Figure 1. Abbreviation: eGFP = enhanced green fluorescent protein. Please click here to view a larger version of this figure.
Figure 2: Visual appearance of cardiac tissue clearing stages.
(A) From left to right: uncleared left ventricle, electrophoresed left ventricle, electrophoresed left ventricle treated with crosslinker, electrophoresed left ventricle treated with crosslinker and incubated in RIMS. (B) Uncleared and cleared whole mouse heart. (C) Left: uninjured, uncleared left ventricle. Right: uninjured, cleared left ventricle. (D) Left: Uncleared MI injured left ventricle. Right: Cleared MI injured left ventricle. Scale bars = 0.5 cm. Abbreviation: MI = myocardial infarction. Please click here to view a larger version of this figure.
Figure 3: Efficacy of novel clearing method for uninjured and sham-operated cleared hearts.
Cleared (A) uninjured (scale bar = 400 μm) and (B) sham-operated hearts (scale bar = 500 μm) with Tcf21mcm x Rosa26eGFP fibroblasts (green) and pseudocolored areas of increased fluorescence (purple) showing fibroblast localization. Accompanying videos showing fibroblast videos can be found in Supplemental Videos S1–2. Still images show background clearing and maintenance of fibroblast-endogenous fluorescence (green). Please click here to view a larger version of this figure.
Figure 4: Attenuation of loss of fibroblasts in ischemia/reperfusion-injured cleared hearts over time.
Cleared cardiac tissue from the indicated I/R time points with Tcf21MerCreMer (mcm) x Rosa26eGFP fibroblasts (green), pseudocolored areas of increased fluorescence (purple) showing fibroblast localization, and orange outlining areas devoid of fibroblasts. Top row: still images of left ventricles from I/R-injured cleared hearts. Bottom row: still images of left ventricles from I/R-injured cleared hearts with Imaris spots function used to see fibroblast patterns more easily in the whole left ventricle. Videos of I/R-injured cleared hearts showing not only gross patterning of fibroblasts, but also clear images of individual fibroblasts and their morphologies can be found in Supplemental Videos S5–8. Scale bars = 500 μm. Abbreviation: I/R = ischemia/reperfusion. Please click here to view a larger version of this figure.
Figure 5: Attenuation of loss of cardiac fibroblasts following injury in myocardial infarction-injured cleared hearts over time.
Cleared cardiac tissue from MI hearts with Tcf21MerCreMer (mcm) x Rosa26eGFP fibroblasts (green), pseudocolored areas of increased fluorescence (purple) showing fibroblast localization, and orange outlining areas devoid of fibroblasts. Top row: Still images of tissue cleared left ventricles from MI-injured hearts (green). Bottom row: still images of tissue cleared left ventricles from MI-injured hearts (green) with Imaris spots function (purple) overlaid to show gross fibroblast distribution. Videos of tissue cleared left ventricles from MI-injured hearts show gross fibroblast arrangement in the heart as well as the positioning and morphology of individual fibroblasts in this 3D in vivo model (Supplemental videos S7–8). Scale bars: 1.5 day = 1000 μm, 3 day = 700 μm. Abbreviation: MI = myocardial infarction. Please click here to view a larger version of this figure.
Figure 6: Cleared tissue from Angiotensin/Phenylephrine-treated hearts.
(A) Schematic showing how Angiotensin/Phenylephrine pumps are used to administer drugs over a two-week period following tamoxifen activation of Tcf21MCM x eGFP. (B) Still image, still image + Imaris spots, and video showing fibroblast organization and morphology in Ang/PE-treated hearts (Supplemental Video S11) Scale bars = 500 μm. Abbreviations: Ang/PE = angiotensin II/phenylephrine; eGFP = enhanced green fluorescent protein. Please click here to view a larger version of this figure.
Figure 7: Representative images of observed fibroblast pattern in Angiotensin/Phenylephrine-treated hearts.
(A, C, and D): Images of right-handed helix twisting pattern of fibroblasts from the perspective of the outside of the ventricle. (B) Image of fibroblast patterning from the perspective of the inside of the ventricle. Arrows highlight linear groups of fibroblasts that make up the twisting pattern. Scale bars = 500 μm. Please click here to view a larger version of this figure.
Supplemental Figure 1: 2D renderings of imaging reservoir apparatus. (A) 2D rendering of bottom half of imaging reservoir. This reservoir is sealed to the Petri dish with vacuum grease. The heart is placed in the center opening and submerged in RIMS. (B) 2D rendering of top half of imaging reservoir. Glass coverslip is adhered to flat bottom of this piece with vacuum grease. This is gently placed on top of bottom reservoir. Top reservoir can then be filled with glycerol for imaging. Units in mm. Abbreviations: 2D = two-dimensional; RIMS: Refractive Index Matching Solution. Please click here to download this figure.
Video S1: Uninjured tissue cleared heart. Cardiac fibroblasts (green) in an uninjured left ventricle were small, and many were rounded with no more than two cellular projections. Please click here to download this video.
Video S2: Sham tissue cleared heart. Cardiac fibroblasts (green) in the left ventricle of a sham-operated mouse heart were small and mostly round with few projections, similar to what was seen in uninjured hearts. Please click here to download this video.
Video S3: Detailed fibroblast view through the ventricular wall of an uninjured heart. This video begins on the interior ventricular wall and zooms toward the exterior of the heart. Cardiac fibroblasts (green) were shown in detail and were observed to have rounded or slightly elongated cell bodies with no more than two cellular projections. Please click here to download this video.
Video S4: Detailed view of section of left ventricular wall of uninjured heart. This ~300-μm-wide section of the cleared uninjured mouse heart shows the 3-D arrangement of cardiac fibroblasts (green) and their morphologies in detail. Please click here to download this video.
Video S5: Cleared left ventricle of 3 day I/R. Detailed view of the left ventricle showed that following I/R injury, there is an area of cardiac fibroblast loss. It was also apparent that fibroblasts (green) developed a much more elongated shape in this injured condition in comparison to the more rounded morphologies seen in uninjured and sham hearts. Abbreviation: I/R = ischemia/reperfusion. Please click here to download this video.
Video S6: Cleared left ventricle of 7 day I/R. Detailed view of the left ventricle shows that by 7 days following I/R injury, the area lacking in fibroblasts was smaller than that seen by day 3 following I/R injury. There were more fibroblasts (green) present, and interestingly, new cell morphologies were present. Specifically, some fibroblasts had rounded cell bodies with multiple protrusions, potentially indicating a new or developed role of fibroblasts in this environment. Abbreviation: I/R = ischemia/reperfusion. Please click here to download this video.
Video S7: Cleared left ventricle of 14 day I/R. Detailed view of left ventricle showed that there was still an area central to the ventricular wall that had very few fibroblasts but fibroblasts (green) on the periphery of this void maintained the highly elongated morphology seen in day 7 post-I/R samples. Interestingly, the few fibroblasts that were present in the injured region had a morphology like that in uninjured hearts—small and rounded. Abbreviation: I/R = ischemia/reperfusion. Please click here to download this video.
Video S8: Cleared left ventricle of 28 day I/R. Detailed view of cardiac fibroblasts (green) on day 28 following I/R injury showed that there was a small area in the injured region that lacked fibroblasts. It was also observed that there was a region of high fibroblast density surrounding this region, and that the morphologies in this dense area were highly elongated. Abbreviation: I/R = ischemia/reperfusion. Please click here to download this video.
Video S9: Cleared left ventricle of 1.5 day MI. There were very few cardiac fibroblasts (green) remaining in the injured left ventricle at 1.5 days after MI, cardiac fibroblast death throughout this region of the ventricle. Abbreviation: MI = myocardial infarction. Please click here to download this video.
Video S10: Cleared left ventricle of 3 day MI. There was a large area devoid of cardiac fibroblasts 3 days after MI, but some cardiac fibroblasts (green) re-appeared in the ventricle, unlike the results found following 1.5 days of MI. Also, cardiac fibroblast morphology profiles were different than those seen in I/R injured hearts. Here fibroblasts were elongated but there were others that had rounded cell bodies (larger than those seen in uninjured hearts) and a subpopulation of these had many cell projections. Abbreviations: MI = myocardial infarction; I/R = ischemia/reperfusion. Please click here to download this video.
Video S11: Cleared left ventricle of Ang/PE-treated mice. There was no apparent loss of cardiac fibroblasts (green) following Ang/PE treatment. However, cardiac fibroblasts were mostly small and rounded or small and elongated and seemed to align with the contractile patterns of the heart. Abbreviation: Ang/PE = angiotensin II/phenylephrine. Please click here to download this video.
This article presents a refined method for tissue clearing that allows for visualization of cardiac fibroblasts in vivo, both at baseline and following injury, to characterize and better understand fibroblasts in the mouse heart. This enhanced protocol addresses limitations in existing tissue clearing protocols that have attempted to identify specific cell types in the adult or neonatal heart12,13,14,16,17,20. In the initial attempts to clear the mouse heart, the passive CLARITY technique was used, wherein the heart was left in a clearing buffer on a nutator for approximately 1 week to allow for passive distribution of the buffer throughout the tissue15,18. This process only produced clearing of approximately 80 μm of depth into the tissue (data not shown), which is consistent with previous observations whereby several weeks were needed to clear a 1 day-old neonatal mouse heart13. Active CLARITY using an active electrophoresis system allowed for deeper clearing through the entire ventricular wall, approximately of 700 μm–1 mm deep, over a shorter time frame12,18.
However, using this previously published active CLARITY protocol led to a loss of fibroblast fluorescence with high levels of tissue autofluorescence, which had previously been noted to occur in active CLARITY by Kolesova et al.12. To allow for maintenance of fibroblast fluorescence, the appropriate fixation protocol was found to be of utmost importance. Too short of a fixation process caused loss of fluorescence during electrophoresis. Too long of a fixation process caused loss of fluorescence itself. Therefore, overnight fixation at 4 °C in 4% PFA was found to be optimal. To ameliorate the background fluorescence issue, a decolorization soak (a principle borrowed from the CUBIC method of clearing) was employed to reduce heme binding within myoglobin, and therefore reduce autofluorescence caused by this chromophore18. Decolorization treatment resulted in more robust clearing of background fluorescence, with maintenance of reporter fluorescence so that the labeled fibroblasts were more pronounced (Figure 2A).
Finally, to allow for full optic transparency, the tissue was equilibrated in Refractive Index Matching Solution (RIMS) (Figure 2B–D). By matching the refractive index of the tissue with its surroundings, this increased the optic transparency of the tissue, allowing for deeper imaging. High speed resonant scanning was then used to image tissue as it is faster than galvanometric scanning. Because mouse hearts are slightly different sizes, individual XY imaging parameters and Z intensity corrections were set. With these parameters, it was possible to image through the uninjured heart to visualize fluorescent fibroblasts in approximately 4 h (Figure 3). Image quality was improved in post-imaging processing by using the unmixing and denoising analysis software to reduce background and clarify the fluorescence present in the image. Additionally, secondary analysis software was used to highlight the localization of fluorescent fibroblasts. This post-imaging analysis was used to clearly annotate cardiac fibroblasts by eliminating background voxel-by-voxel (Figure 3, Figure 4, Figure 5, Figure 6).
This optimized CLARITY protocol has been applied to optically clear both injured and uninjured hearts. This allows for a better understanding of the reaction of cardiac fibroblasts to injury. These injuries included a time course of MI and I/R, as well as Ang/PE dosing. As ischemic injury weakens cardiac tissue, it is critical that greater care is taken during electrophoresis to maintain tissue integrity. Indeed for both I/R and MI injury, a shorter period of electrophoresis (≤1.5 h) is required32. Previous studies have not considered the effects of injury on the tissue clearing process. The newly optimized protocol presented here accommodates for injury, allowing for clearing without further destruction of the tissue.
The authors have nothing to disclose.
The authors would like to acknowledge the CCHMC Confocal Imaging Core for their assistance and guidance in development of this model, as well as Matt Batie from Clinical Engineering for the design of all 3D printed parts. Demetria Fischesser was supported by a training grant from the National Institutes of Health, (NHLBI, T32 HL125204) and Jeffery D. Molkentin was supported by the Howard Hughes Medical Institute.
4-0 braided silk | Ethicon | K871H | |
8-0 prolene | Ethicon | 8730H | |
40% Acrylamide Solution | Bio-Rad | 1610140 | |
Angiotensin II | Sigma | A9525-50G | |
Artificial Tear Ointment | Covetrus | 048272 | |
DABCO (1,4-diazabicyclo[2.2. 2]octane) | Millipore Sigma | D27802-25G | |
GLUture topical tissue adhesive | World Precision Instruments | 503763 | |
Heparin | Sigma | H0777 | |
Imaris Start Analysis Software | Oxford Instruments | N/A | |
Micro-osmotic pumps | Alzet | Model 1002 | |
Nikon Elements Analysis Software | Nikon | N/A | |
Nikon A1R HD upright microscope | Nikon | N/A | |
Normal autoclaved chow | Labdiet | 5010 | |
Nycodenz, 5- (N-2, 3-dihydroxypropylacetamido)-2, 4, 6-tri-iodo-N, N'-bis (2, 3 dihydroxypropyl) isophthalamide |
CosmoBio | AXS-1002424 | |
Paraformaldehyde | Electron Microscopy Sciences | 15710 | |
Phenylephrine Hydrochloride | Sigma | P6126-10G | |
Photoinitiator | Wako Chemicals | VA-044 | |
Rosa26-nLacZ [FVB.Cg-Gt(ROSA)26Sortm1 (CAG-lacZ,-EGFP)Glh/J] | Jackson Laboratories | Jax Stock No:012429 | |
Sodium Azide | Sigma Aldrich | S2002-5G | |
Sodium Chloride solution | Hospira, Inc. | NDC 0409-4888-10 | |
Tamoxifen | Sigma Aldrich | T5648 | |
Tamoxifen food | Envigo | TD.130860 | |
Tween-20 | Thermo Fisher Scientific | BP337-500 | |
Quadrol, N,N,N′,N′-Tetrakis(2-Hydroxypropyl)ethylenediamine, decolorizing agent | Millipore Sigma | 122262-1L | |
X-Clarity electrophoretic clearing chamber | Logos Biosystems | C30001 | |
X-Clarity electrophoretic clearing solution | Logos Biosystems | C13001 | |
X-Clarity electrophoresis tissue basket | Logos Biosystems | C12001 | |
X-Clarity electrophoresis tissue basket holder | Logos Biosystems | C12002 |