A. Laser ablation creates holes in microtissues surrounded by non-viable tissue

In our previous studies, stromal microtissues were injured using a microdissection knife mounted to an xyz-micromanipulator.19–2119. M. Selman Sakar, J. Eyckmans, R. Pieters, D. Eberli, B. J. Nelson, and C. S. Chen, Nat. Commun. 7(1), 11036 (2016). https://doi.org/10.1038/ncomms1103620. J. B. Tefft, C. S. Chen, and J. Eyckmans, APL Bioeng. 5, 016102 (2021). https://doi.org/10.1063/5.002865121. S. L. Das, P. Bose, E. Lejeune, D. H. Reich, C. Chen, and J. Eyckmans, Tissue Eng., Part A 27, 1447 (2021). https://doi.org/10.1089/ten.tea.2020.0332 Due to the soft and compliant properties of microtissues, mechanical injury often resulted in slipping of the microtissue from the pillars before a gap was created. To overcome this limitation, we sought to deploy a Q-switched Nd:YAG nanosecond-pulsed laser to inflict wounds with laser ablation [Fig. 1(a)]. The Nd:YAG laser operates at 1064 nm, and our setup includes a removable KTP crystal capable of generating the second harmonic at 532 nm. Each pulse width was 5 ns. Given the different mode of injury, transection vs ablation, we set out to characterize the two wound types and performed a viability assay on microtissues injured with either a microdissection knife or laser pulse. Interestingly, we observed a key difference in cell viability at the wound periphery. In knife-injured tissues, the long edges of the wound were largely free of dead cells as if the dead tissue was almost entirely removed, leaving only a small patch of dead cells at the end of the incision [Fig. 1(c)]. In contrast, laser-injured microtissues showed a band of dead cells around the perimeter of the wound [Fig. 1(c)] indicating that tissue damage extended beyond the wound margins into the surrounding tissue.To establish a protocol for reproducibly inflicting full-thickness gaps with laser light, microtissues were injured with laser pulses of different wavelengths (532 or 1064 nm) and pulse energies ranging from 0.25 to 3 mJ, two parameters that affect the energy density of the focused laser pulse. Phase contrast images were captured following injury [Fig. 1(a)], and the initial gap area was quantified [Fig. 1(b)]. At both wavelengths, 532 and 1064 nm, there was a threshold energy below which no tissue gap was created, which was about 0.25 mJ at 532 nm and 1.5 mJ at 1064 nm. These laser pulse energies correspond to a fluence, which is the energy density at the focal point in the tissue, of 19 and 17.6 kJ/cm2 for 532 and 1064 nm, respectively (supplementary material, Table S1). Thus, the ablation threshold of collagenous microtissues was similar for both wavelengths. Because laser pulse energies were measured before each experiment, we reported laser pulse energies, and not estimated fluences, in the x-axis of the graphs. At the minimum energies, the gaps created were smaller (mean values of 17 600 ± 1540 μm2 at 532 nm and 14 300 ± 3000 μm2 at 1064 nm) than those created by knife (41 600 ± 3980 μm2). At intermediate energies, 0.4 and 0.55 mJ at 532 nm and 2 mJ at 1064 nm, the average gap size (34 200 ± 4270 μm2; 53 800 ± 8500 μm2; and 30 100 ± 6670 μm2, respectively) was on the same scale as knife injuries. Further increase in the pulse energy for both wavelengths yielded initial gap areas that were much larger than what the knife was capable of creating [Fig. 1(b)]. To assess damage to the tissue surrounding the gap area, knife and laser injured microtissues were stained with ethidium homodimer-1 to detect dead cells with damaged cell membranes, and the percentage of area that contained dead cells was quantified for different wavelengths and pulse energies [Figs. 1(c) and 1(d)]. The area of non-viable tissue in all but the lowest energy injuries was significantly larger than in the knife injuries, although there was a large range in size of the damaged zone [Fig. 1(d)]. Together, these data show that laser ablation can be used to reliably create gaps in microtissues with sizes comparable to knife inflicted wounds, but with more cell damage in the adjacent tissue.

B. Ablated microtissues close following an initial opening phase

Gap closure in our model system is initiated by the fibroblasts adjacent to the wound edge.2222. E. Mailand, B. Li, J. Eyckmans, N. Bouklas, and M. S. Sakar, Biophys. J. 117, 975 (2019). https://doi.org/10.1016/j.bpj.2019.07.041 Given the increased levels of cell death in laser-ablated wounds vs knife wounds [Figs. 1(c) and 1(d)], we hypothesized that ablation wounds would show impaired closure. To investigate this hypothesis, we monitored gap closure for stromal microtissues injured by knife and by laser, both 532 and 1064 nm, using time-lapse microscopy. As previously reported, knife-injured tissues smoothed the wound margins and commenced closure within minutes, and closure finished in about 24 h1919. M. Selman Sakar, J. Eyckmans, R. Pieters, D. Eberli, B. J. Nelson, and C. S. Chen, Nat. Commun. 7(1), 11036 (2016). https://doi.org/10.1038/ncomms11036 [Fig. 2(a), Multimedia view]. The laser-ablated tissues displayed a prolonged smoothing phase that was characterized by opening of the gap that lasted between 4 and 6 h, which was then followed by closure similar to knife-injured tissues [Fig. 2(b)].To quantitatively assess tissue closure between the different conditions, we compared closure rates during two phases of the healing process. We defined the initial phase between 0 and 4 h because ∼80% of tissue gaps reached their maximum area by 4 h. We considered the closure phase to occur between 6 and 16 h post injury as most tissue gaps (>90%) reached the maximum area by 6 h, and the closure rate was linear for the ensuing 10 h [Fig. 2(b)]. The rate of closure during each phase was quantified as change in gap area per hour, where a negative value indicates the gap was opening rather than closing [Fig. 2(c)]. During the initial phase [Fig. 2(c) left], knife injured tissues displayed an average closure rate of 1380 ± 270 μm2/h. For 532 nm injuries, the closure rate during this phase was −890 ± 820 μm2/h (p μm2/hour (p Fig. 2(c) right]. Together, these data suggest that a prolonged opening phase, rather than an impaired closure phase, delayed gap closure in laser ablated wounds.

C. Characterization of fiber damage and movement at gap periphery

To investigate the gap opening in laser inflicted wounds, we carefully analyzed the time lapse phase contrast images during the opening phase and observed a darker band of damaged tissue at the periphery of the gap [Fig. 2(a)] that corresponded to the area of dead cells [Figs. 1(c) and S2). In phase contrast microscopy, changes in apparent brightness are due to phase shifts in light that occur when light passes through the sample. Therefore, visually darker tissue indicates that a different phase shift occurred in these parts of the tissue, potentially due to altered topography or density (increased or decreased density relative to the bulk tissue could both cause this visual distinction). This band of damaged tissue was present in most ablated tissues; very few knife injured tissues contained a small patch of damaged tissue, but not a circumferential band (Fig. S2). During the initial opening phase, this band grew smaller [Fig. 2(a)] as the gap area increased [Fig. 2(a) white dotted lines, Fig. 2(b)]. We did not observe significant differences between the two laser wavelengths in the clearance and repair process. Although not statistically significant, the laser seems to generate more peripheral tissue damage at 1064 nm. For these reasons, and to simplify experiments, the laser was used in 1064 nm mode from this point onward in the study.From the time-lapse images in Fig. 2, we postulated that the ECM at the gap periphery was compromised and, therefore, cleared by the remaining fibroblasts in the microtissue. In order to characterize the mechanical damage to surrounding ECM caused by each type of injury, we sought to visualize the microstructure of the fibrous matrix. We used high-resolution, reflection microscopy to capture z-stack images of the matrix topography at the wound edges. From these images, we made two main observations. First, the depth of the tissues in z was significantly smaller for knife (110 ± 14 μm) than for laser (152 ± 18 μm) injuries (p Fig. 3(a)], suggesting that the mechanical injury compacts the microtissue, likely due to the pressure from the microdissection knife. To quantify fiber density, the 25 μm region closest to the gap edge was divided into subregions and the density of each calculated. The histogram showing fiber density within these subregions reveals that laser injuries contain more empty space than do knife injuries [Fig. 3(b), left, p 2323. P. Bose, J. Eyckmans, T. D. Nguyen, C. S. Chen, and D. H. Reich, ACS Biomater. Sci. Eng. 5, 3843 (2019). https://doi.org/10.1021/acsbiomaterials.8b01183 based on the Fibril Tool method2424. A. Boudaoud, A. Burian, D. Borowska-Wykrȩt, M. Uyttewaal, R. Wrzalik, D. Kwiatkowska, and O. Hamant, Nat. Protoc. 9, 457 (2014). https://doi.org/10.1038/nprot.2014.024 to determine the distribution of fiber orientation angles within 25 μm of the gap edge. This analysis revealed that the majority of fibers in knife-injured tissues aligned along the long axis of the tissue, parallel to the direction of movement of the knife. The peak alignment angle shifts to about −20°, and a wider spread of orientation angles was observed for ablation injuries compared to mechanical injuries [Fig. 3(b), right, p We wondered next how the fibrous matrix structure changes as the fibroblasts remodel the microtissue at the gap edge. Using reflection microscopy, we performed live microscopy and captured images every 10 min [Fig. 3(c), Multimedia view]. In mechanical injuries, the gap edge smooths and begins closure within 4 h. For ablation injuries, we observed the tissue from the damaged gap periphery moving into the bulk tissue over a period of 4–10 h. In order to maximize the amount of damaged tissue that we could observe, wounds in this experiment only were made roughly three times larger than those in Fig. 2. Therefore, they take roughly three times longer to complete the clearance phase and reach their maximum wound size. In some places, fiber breakages occurred as the damaged matrix was remodeled [Figs. 3(c′) and 3(c‴), white arrows]. At the conclusion of this tissue clearance phase, we observed compacted tissue at the wound edges [Fig. 3(c), yellow arrows]. In order to confirm that the fibrous tissue remodeling was due to cell activity, rather than a passive feature of the collagenous hydrogel, we treated the microtissues with a potent inhibitor of actin polymerization, latrunculin A, which caused the cells in the microtissue to round up and prevented all remodeling of the wound edges (Fig. S3).Given the fiber breakages and compaction of the damaged matrix, we postulated that actomyosin contractile forces were required for tissue clearance [Fig. 3(d)]. The data in Fig. 3(d) are reported as relative opening rate for each treatment compared to the vehicle control (DMSO). Laser ablation was used for all conditions; because we sought to observe changes in damaged tissue clearance, and knife injuries did not display this clearance phase, the knife condition was not included in these experiments. To decrease cellular contractility, the myosin-2 inhibitor blebbistatin (100 μM), myosin light chain kinase (MLCK) inhibitor peptide 18 (P18, 10 μM) or Y27632 (10 μM), and a Rho-associated kinase (ROCK) inhibitor were added prior to microtissue ablation. Interestingly, blebbistatin and P18 did not significantly affect the tissue opening rate, while Y27632 significantly decreased the mean tissue opening rate to 50% ± 1% of the vehicle control, DMSO (p 2525. A. Yoneda, H. A. B. Multhaupt, and J. R. Couchman, J. Cell Biol. 170, 443 (2005). https://doi.org/10.1083/jcb.200412043 To investigate if fibroblasts cleared the wound by engulfing damaged matrix, we added dynasore (60 μM) to inhibit dynamin activity, which is necessary for phagocytosis26,2726. E. S. Gold, D. M. Underhill, N. S. Morrissette, J. Guo, M. A. McNiven, and A. Aderem, J. Exp. Med. 190, 1849 (1999). https://doi.org/10.1084/jem.190.12.184927. G. J. K. Praefcke and H. T. McMahon, Nat. Rev. Mol. Cell Biol. 5, 133 (2004). https://doi.org/10.1038/nrm1313 and clathrin-mediated endocytosis.2727. G. J. K. Praefcke and H. T. McMahon, Nat. Rev. Mol. Cell Biol. 5, 133 (2004). https://doi.org/10.1038/nrm1313 Indeed, addition of dynasore lowered the mean opening rate to 57% ± 9% of control (p 

D. Transmission electron microscopy reveals intracellular fibrillar structures, suggesting phagocytosis of ECM after ablation

To further corroborate the hypothesis that fibroblasts clear wounds by engulfing ECM, we performed transmission electron microscopy (TEM) on uninjured microtissues and at the wound edge of laser injured microtissues at 0, 4, and 20 h post injury (Fig. 4). Immediately following ablation, cells at the wound edge displayed fragmented or misshapen nuclei and showed cellular blebbing [Figs. 4(c) and 4(c′), red arrows]. Other cells showed the formation of numerous, apparently vacant, cytoplasmic vacuoles, a reduction in cell size, and chromatin condensation at the edge of the nucleus [Fig. 4(c′)], which is reminiscent of a necrotic or oncotic cell.28,2928. B. F. Trump, I. K. Berezesky, S. H. Chang, and P. C. Phelps, Toxicol. Pathol. 25, 82 (1997). https://doi.org/10.1177/01926233970250011629. J. Balvan, A. Krizova, J. Gumulec, M. Raudenska, Z. Sladek, M. Sedlackova, P. Babula, M. Sztalmachova, R. Kizek, R. Chmelik, and M. Masarik, PLoS One 10, e0121674 (2015). https://doi.org/10.1371/journal.pone.0121674 At 4 h after ablation, cells at the wound edge contained vacuoles with distinct fibrillar structures [Fig. 4(d), yellow arrows and Fig. 4(d′)]. These vacuoles were relatively large (>1 μm in diameter), and the fibrillar structures appeared to have a similar morphology to those of the surrounding damaged extracellular matrix. At 20 h after ablation, cells at the wound edge lacked vacuoles, but instead contained numerous organelles, including mitochondria [Fig. 4(e)]. This observation suggests increased metabolic activity, as would be expected of fibroblasts actively migrating and producing new ECM, two ATP intensive processes. Close inspection of the cell surfaces near the wound edge revealed that fibrillar structures appear to be emanating from the cell [Fig. 4(e′), blue arrowheads], which is in line with our previous studies showing that fibroblasts in microtissues close wounds by producing a provisional matrix that is rich in fibronectin and collagen type III.19–2119. M. Selman Sakar, J. Eyckmans, R. Pieters, D. Eberli, B. J. Nelson, and C. S. Chen, Nat. Commun. 7(1), 11036 (2016). https://doi.org/10.1038/ncomms1103620. J. B. Tefft, C. S. Chen, and J. Eyckmans, APL Bioeng. 5, 016102 (2021). https://doi.org/10.1063/5.002865121. S. L. Das, P. Bose, E. Lejeune, D. H. Reich, C. Chen, and J. Eyckmans, Tissue Eng., Part A 27, 1447 (2021). https://doi.org/10.1089/ten.tea.2020.0332 Remnants of necrotic cells can still be seen at 20 h after injury [Fig. 4(e), red arrowheads]. However, following closure, the number of necrotic cells was drastically reduced compared to one hour after injury (Fig. S4), suggesting that dead cell debris may also be cleared by fibroblasts. Although this observation is in line with a recent study showing that dermal fibroblasts phagocytose apoptotic cells,30,3130. M. Schwegler, A. M. Wirsing, A. J. Dollinger, B. Abendroth, F. Putz, R. Fietkau, and L. V Distel, Biol. Cell 107, 372 (2015). https://doi.org/10.1111/boc.20140009031. B. Romana-Souza, L. Chen, T. R. Leonardo, Z. Chen, and L. A. DiPietro, FASEB J. 35, 21443 (2021). https://doi.org/10.1096/fj.202002078R it remains to be seen if 3T3 fibroblasts phagocytose dead cell debris in collagenous microtissues.In order to determine whether ECM phagocytosis occurred at all times in the microtissues, or whether it was an injury-specific response, we quantified the number of vacuoles that contained distinct fibrillar structures [as in Fig. 4(d′)] for 10 fields of view at each timepoint. In uninjured tissues, 6 such vacuoles were observed, and in tissues 4 h post injury, 17 such vacuoles were observed. Only vacant vacuoles were seen immediately following injury, and thus, not included in this quantification. No vacuoles were observed at 20 h post injury. Due to differences in number of cells and cell sizes, the number of fiber-containing vacuoles was normalized to cellular area [Fig. 4(f)].