Variations in mechanical stiffness alter microvascular sprouting and stability in a PEG hydrogel model of idiopathic pulmonary fibrosis

Objective: Microvascular remodeling is governed by biomechanical and biochemical cues


| INTRODUC TI ON
Microvascular network remodeling is necessary to heal tissues after injury, and critical first steps are the carefully regulated processes of capillary sprouting and pruning. Capillary sprouting and pruning are governed by interactions between endothelial cells (ECs) and pericytes, the two main cell types that comprise capillaries, and their interactions are regulated by both biomechanical and biochemical cues. [1][2][3] However, in many fibrotic diseases, such as idiopathic pulmonary fibrosis (IPF), heterogeneous microenvironments of mechanical stiffnesses and angiogenic growth factor gradients [4][5][6] locally disrupt EC and pericyte interactions, impacting capillary sprouting and pruning in ways that are not yet well understood. In IPF, this dysregulation manifests as fibrotic lesions, or foci, that are heterogeneously located throughout the lung. These fibrotic foci induce lung function decline and ultimately result in an average life expectancy of approximately 3 years from diagnosis. 7,8 Given the dynamic nature of capillary sprouting and pruning, and the reliance of ECs and pericytes on different environmental cues, it is not surprising that reports of microvascular remodeling in IPF are conflicting. Some studies show evidence of increased capillary sprouting within the hypoxic and stiff fibrotic foci, while others report microvascular pruning within fibrotic foci and excess vessel formation in the areas immediately surrounding fibrotic foci where mechanical stiffness is still relatively low. 9,10 As evident from these conflicting reports, unraveling the dynamics of lung microvascular network remodeling in IPF requires time-course imaging of ECs and pericytes to capture their aberrant interactions. Understanding how biochemical and biomechanical characteristics of the microenvironment impact EC-pericyte interactions necessitates a model system where these two variables can be independently and reproducibly modulated. As it is currently not possible to interrogate these complex and transient interactions in the lung in vivo, we developed a novel in vitro lung organoid model for dynamic imaging of microvascular network remodeling that has tunable biochemical and biomechanical microenvironments. We then used our in vitro model to study how characteristic biomechanical and biochemical cues in IPF impact the physical coupling and uncoupling of lung ECs and pericytes during capillary sprouting and pruning.
In stable microvessels, ECs and pericytes are considered coupled together when they maintain physical contact via several cell surface proteins including β-catenin and N-cadherin. [11][12][13] Microvascular cells can respond to complex angiogenic or anti-angiogenic signals by separating, or uncoupling, from one another to allow for angiogenesis or regression to occur and subsequently attaching, or recoupling, to stabilize the newly remodeled capillary. 11,14 Growth factors, such as vascular endothelial growth factor (VEGF), platelet derived growth factor (PDGF), and fibroblast growth factor (FGF), have long been implicated in orchestrating EC and pericyte uncoupling and recoupling. 11,[13][14][15] Further, high levels of VEGF in the environment stimulate EC sprouting leading to uncoupling of pericytes to allow for the formation of a new vessel. 14 Once a new vessel has formed, PDGF is released by ECs to recruit pericytes to the abluminal surface of capillaries to recouple, and, as a result, stabilize the new vessel. 14,16 Additionally, EC-pericyte uncoupling also allows for vessel pruning which occurs naturally to maintain vascular patterning and adequate blood flow to the region. 17 VEGF, PDGF, and FGF have all been shown to be differentially expressed and heterogeneously distributed in the IPF lung. The presence and concentration of these growth factors varies both temporally and spatially; they are dependent on the stage of disease and on the region in the lung from which the sample was acquired. Together, this suggests that spatiotemporal heterogeneity of growth factors likely contributes to a complex microenvironment for pruning and sprouting of lung capillaries in IPF. 9 Moreover, Nintedanib, one of the two FDA approved drugs used to treat IPF, targets the receptors for VEGF, PDGF, and FGF; however, the impact of this drug on EC-pericyte coupling and uncoupling is unknown-further motivating the study of vascular stability and remodeling in the context of IPF.
In addition to growth factor cues, stiffness of the extracellular matrix (ECM) in IPF also plays a role in EC-pericyte coupling. Studies of ECs suggest that high ECM stiffness can induce both apoptosis and capillary regression, but the presence of VEGF-A can increase VEGF-R2 internalization and EC proliferation. 18,19 Additionally, pericytes are also known to be sensitive to mechanical cues in their local environment, with elevated ECM stiffness recognized as a driver of myofibroblast differentiation, as evidenced by αSMA and collagen type 1 expression. Increased stiffness of the microenvironment has also been shown to cause physical uncoupling of pericytes from ECs, in both human and murine models of IPF. 20,21 There are several types of in vitro assays that have been used to study EC-pericyte interactions and vessel dynamics, such as cell co-culture models which capture de novo vessel formation, [22][23][24][25] intact vessel explant assays such as the rat mesentery explant model which captures microvascular remodeling dynamics, [26][27][28][29] and tissue explant angiogenesis assays which have the potential to capture both de novo sprouting and vessel remodeling. [30][31][32] Tissue explant angiogenesis assays have been previously used to study sprouting dynamics in various tissues including adipose, 30 choroid, 31 and skeletal muscle. 32 However, a limitation of these models is their use of Matrigel which has a complex biochemical formulation, such that it is not possible to decipher the role of any given protein signal in guiding vascularization. Additionally, Matrigel is not easily tunable to high (e.g., greater than 1 kPa) stiffnesses. Further, there are reports of high batch-to-batch variability in Matrigel, impacting experimental reproducibility. 33 The use of pro-angiogenic growth factors VEGF, 34 and the combination of PDGF and FGF 35 have been explored, with one recent study suggesting that the combination of PDGF and FGF may stimulate more dense and organized sprouting. 36 Overall, there is a need for an adaptable, in vitro model system that enables biomechanical and biochemical cues to be independently modulated in order to study microvascular sprouting dynamics and organization in fibrotic lung tissue.
Herein, we present a versatile lung explant angiogenesis assay using tailorable PEGDA hydrogels as an alternative sprouting matrix with variable stiffness (Figure 1). This assay is employed to examine the effects of crosstalk between growth factors and mechanical cues on capillary sprouting and vessel organization. We have chosen to model the mechanical cues that are representative of healthy lung tissue (2 kPa), fibrotic foci (10 kPa), and mature fibrosis (20 kPa), all of which can be found in the IPF lung, according to ranges that have been reported in the literature. 6 To further capture the key components of the biochemical microenvironment of the IPF lung, we incorporate VEGF or a combination of PDGF and FGF into the hydrogel matrix at concentrations previously used in similar PEGDA hydrogels. 34,35 Altogether, our experimental platform offers a controllable system for studying how angiogenic sprouting is influenced by the biochemical and mechanical environmental cues that are present in IPF and enables visualization of cell specific interactions between ECs and pericytes.

| Lung harvest and explant preparation
Fourteen male TdTomato Mice (The Jackson Laboratory) were humanely euthanized via carbon dioxide asphyxiation between 14 and 16 weeks of age. The thorax was opened, and cardiac perfusion was performed with 10 mL of PBS. Next, the lungs and heart were removed from the chest cavity and separated. Lungs were placed in cold PBS to preserve them until slicing. Each lobe was sliced into 1 mm long segments using a tissue slicer (Stoelting); then, those segments were sliced again along the other axis to yield 1 mm × 1 mm lung explants. The explants were stored on ice in a petri dish of cold PBS until implantation. Each set of lungs yielded on average 25 lung explants.

| Synthesis of PEG-RGDS, PEG-PQ-PEG, and PEG-growth factor conjugates
The cell adhesive peptide sequence RGDS, the MMP-2 and MMP-

| Formation of hydrogels of varying stiffness
Previous studies by Gill et al. 37 characterized 5% (w/v) PEG-PQ as having a modulus of approximately 20 kPa, and this was used as the formulation for the stiffer hydrogels used in this study. Recently, Chapla et al. 38 described a method of modulating PEG hydrogel stiffness using the addition of small vinyl-containing monomers into hydrogel precursor solutions. This method was utilized to synthesize softer gels through the addition of alloc-protected lysine at a 2:1 ratio of alloc:acrylate to form hydrogels having a modulus of 10 kPa and a ratio of 3.5:1 alloc:acrylate to form gels with a modulus of 2 kPa. Hydrogels were formed in 6-mm silicon ring molds adhered to 24-well glass bottom plates that were pretreated with ethanol with 2% (v/v) 3-(trimethoxysilyl)propyl methacrylate for 48 h to methacrylate the glass surface. Thirty-five μL of PEG precursor solution containing acetophenone was pipetted into the center of each mold, and the plate exposed to UV light. The resulting gels were rinsed in PBS overnight and then incubated in cell culture media.

| Matrigel preparation
Aliquots of Matrigel Basement Membrane Matrix, phenol red-free and LDEV-free (Corning), were thawed overnight on ice at 4°C.
Pipette tips, silicone ring molds like those described above, and a 24well glass bottomed plate (Cellvis) were also maintained at 4°C overnight. On Day 0, silicone ring molds were placed in each well of the 24-well plate to ensure the Matrigel matrix would match the dimensions of the PEGDA hydrogels. Forty μL of Matrigel was pipetted into each mold using the pre-chilled tips to prevent premature gelation.

| Explant culture
On Day 0, one lung explant was placed in the center of each hydrogel in a 24-well plate with Endothelial Growth Media 2, Microvascular Formula (EGM2-MV; Lonza). Half of the media in each well was refreshed every other day. There were six experimental groups that were a combination of two growth factor treatments and three stiffnesses:

| Brightfield and immunofluorescent imaging
On Day 4 and Day 7, brightfield images were taken with a Leica THUNDER Imager using the 10× objective (Leica). Four of six wells per condition were imaged, and the same four wells were imaged on both Day 4 and Day 7. Day 4 images were grouped based on cell infiltration and vessel formation. Day 7 images were separated into four categories: type 1, type 2, type 3, and type 4, based on organization as shown in Figure 2. The percentage of samples containing vessels was defined as the number of type 1 and type 2 images in a well added together and divided by the total number of images in that well times one hundred. Images in categories 1 and 2 were analyzed using ImageJ 39 to quantify total number of vessels, longest vessel, and shortest vessel as measures of angiogenesis. Each plate was considered to be an independent sample, N, and for each F I G U R E 2 Sample images of vessel formation categories. experimental group (e.g., PDGF 2 kPa), all images were averaged for each sample N. Average longest sprout represents the average of all of the longest vessels as quantified by ImageJ analysis across an independent sample. Shortest average sprout was calculated similarly except using the shortest sprout identified by ImageJ analysis in each image.

| Time course imaging
Explants were incubated in 5% carbon dioxide at 37°C for the first 24 h of culture to allow for their adhesion to the gels. Twenty-four hours later, the 24-well plate was transferred to a stage top incubator (Okolab) where culture conditions could be held constant at 5% carbon dioxide and 37°C. Every well was imaged once an hour using a Leica THUNDER Imager at the 10× objective (Leica).
Images were then compiled into one video per well in the Leica Las X software.

| Immunofluorescent staining
On Day 7, samples were fixed in 4% paraformaldehyde for 20 min then permeabilized with 0.5% Triton X-100 in PBS. Gels were blocked overnight with 5% serum from the secondary host species (goat) (Sigma-Aldrich) and 3% Bovine Serum Albumin (Jackson Immuno-Research). Primary antibody (1:200) for CD31 (Santa Cruz) was applied for 48 h at 4°C followed by a thorough washing with PBS before a goat anti-mouse Alexa Fluor 647 secondary antibody (1:200) (Invitrogen) was applied for 24 h. After washing with PBS, cell nuclei were stained with DAPI (1:1000) (Invitrogen), and the plate was imaged the next day on a Leica THUNDER Imager using the 10× and 20× objectives (Leica). Z stacks were taken for each region of interest, and a maximum projection of the z stack, created using the Las X software (Leica), was exported for image analysis.

| Immunofluorescent image analysis
All immunofluorescent images used for analysis were obtained with a 20× objective on the Leica THUNDER Imager (Leica). Images were first processed in FIJI 40 to remove any part of the explant's bulk mass that was captured in the image as the signal from the bulk explant interfered with the ability to quantify vessel sprouts emerging from it. Images were then imported into AngioTool 41 and analyzed for several metrics, including total number of junctions, total vessel length, average vessel length, and total number of sprout end points.

| MMP-9 analysis using ELISA
At each media change (Days 2, 4, and 6), the 250 μL of spent media from PDGF and FGF-containing plates was saved and frozen at −20°C. For each row in a 24 well plate, the six wells were split evenly into three pooled samples (i.e., spent media from wells A1 and A2 were combined, and wells A3 and A4 were combined). An MMP-9 ELISA (R&D Systems) was run according to manufacturer's instructions and imaged on a plate reader at both 450 and 540 nm to quantify absorbance. The background signal was then subtracted out of the absorbance reading, and the data were analyzed. To remain consistent with imaging measurements, the data for each experimental group per plate were averaged to yield an N of 3.

| Statistical analysis
Each 24 well plate was considered an independent sample, N, and all images acquired for that plate were averaged for each experimental sub-group. For comparisons between Matrigel and PEGDA gel groups, statistical analysis was conducted using a two-way ANOVA with Tukey's post hoc test. One-way ANOVA with Tukey's post hoc test was used to compare groups of different stiffnesses within one gel treatment (VEGF or PDGF + FGF). Student's t-tests were used for statistical analyses between groups with the same stiffness with statistical significance asserted at p-values <.05. All data are presented as average +/− standard deviation.

| Endothelial cell sprouting confirmed with immunofluorescent imaging
To confirm that the sprout-like structures emerging from the explant surface were EC lined vessels, on Day 7 samples were immunostained for CD31 and imaged using both brightfield ( Figure 3A) and fluorescent imaging ( Figure 3B). Overlap of CD31 staining with vessel-like structures in the brightfield images confirmed EC identity. The presence of pericytes was identified by the expression of TdTomato, as well as by the classic pericyte morphology of cell processes enwrapping an EC. In the Matrigel, sprout formation between hours 77 and 87 is observed (blue arrow) but this sprout has shortened or regressed by hour 93 (blue arrow). Network growth and competitive branch formation (red arrow) is also observed. In the 2 kPa VEGFtreated gel, bridging between one multicellular sprout and another is observed (yellow arrow).

| Cell infiltration and early vessel formation 4 days after implantation
Using a stage top incubator, samples were imaged on Day 4 using brightfield imaging to assess the infiltration of cells from the explant into the gel and early vessel formation. There were six experimental groups that were a combination of two growth factor treatments and three stiffnesses: 2 kPa VEGF-treated, 10 kPa VEGFtreated, 20 kPa VEGF-treated, 2 kPa PDGF + FGF-treated, 10 kPa PDGF + FGF-treated, and 20 kPa PDGF + FGF-treated. Matrigel was used as a control. At least four wells per condition were imaged for analysis. Wells where the explant did not fully adhere to the gel were excluded. Two separate 24-well plate experiments were imaged for an N of 2. Each condition had at least eight samples (n) that were averaged to calculate the averages seen in Table 1. At least 88% of samples in all study groups contained cells that were either on the surface of the explant or cells that had migrated into the gel, indicating that the cultured lung tissue was viable and the gel was permissive for cell attachment or migration (Table 1)

| Softer gels support more vessel formation
For each 24 well plate, brightfield images acquired on Day 7 were grouped into the four categories previously described in Figure 2.  Figure S1).

| Increased gel stiffness associated with less adhesion, migration, and vessel sprouting
Within the VEGF-treated group, there was a significant difference between the percentage of samples containing sprouts in the 10 and 20 kPa groups when compared to the 2 kPa group ( Figure 6A).
The higher the stiffness, the lower the percentage of samples that contained vessels, although the percentages were not significantly different between the 10 and 20 kPa groups. The average number of sprouts ( Figure 6B) and average longest sprout length were decreased ( Figure 6C) in both the 10 and 20 kPa groups when compared to the 2 kPa group. Of note, the shortest vessels in the 20 kPa group were significantly shorter than those in the 2 kPa group ( Figure 6D).
While a similar trend was seen in the percentage of samples containing vessels in the PDGF + FGF-treated gels, with the percentage in the 2 kPa gel group being significantly higher than the 10 or 20 kPa groups ( Figure 7A), the relationship between stiffness and vessel properties did not follow the same trends as in the VEGF-treated gels. The average number of sprouts was lower in stiffer gels, although this difference was not significant ( Figure 7B). The 20 kPa group had a similar average longest sprout length to the 2 kPa group, while the average longest sprout of the 10 kPa group was shorter by at least 100 μm ( Figure 7C). Additionally, the 2 kPa group had significantly shorter average shortest sprout compared to the 20 kPa group, which was the opposite trend of that observed in the VEGFtreated gels ( Figure 7D).

| Early MMP-9 secretion promotes sprouting in PDGF + FGF-treated gels
Media was saved at each media change timepoint to observe changes in MMP-9 levels over time in PDGF + FGF-treated gels of different stiffnesses (Figure 9). In the 2 kPa gels, MMP-9 levels were significantly higher on Day 2 than Days 4 and 6. In the 10 and 20 kPa gels, there was a decreasing trend in MMP-9 levels over time but there was no significant difference between the timepoints.

| Immunofluorescent analysis confirms increased stiffness affects the amount of vessel formation
A subset of PEDGA gels that were treated with PDGF + FGF was also used to perform immunofluorescent analysis of CD31-positive EC sprouts. The 2 kPa gels had significantly more samples with sprouts compared to the 10 and 20 kPa groups ( Figure 10A). There was no significant difference between the percentage of samples with sprouts between 10 and 20 kPa. The images were then run through an optimized editing workflow to remove the image artifact caused by the bulk explant mass being contained within the imaging field of view in FIJI 40 ( Figure 10B). These images were then input into AngioTool,41 and the most relevant metrics were plotted. Average vessel length represented the average length of all vessels in the sample ( Figure 10C). Average total vessel length is the average length of the total summed length of all vessels in an image ( Figure 10D). Two additional outputs related to vessel morphology, including average total junctions ( Figure 10E) and average total endpoints ( Figure 10F), were also calculated. None of these metrics yielded significant differences between the groups.

| DISCUSS ION
We have presented a novel experimental assay for forming lung ex- investigation of the role of stiffness and growth factor treatments in the formation of these organoids.
Our system was able to support angiogenesis in a comparable manner to Matrigel in soft 2 kPa gels while also allowing for tunability to study the effects of stiffness on sprout formation. Matrigel was used as a control gel, as it is a standard hydrogel used in many tissue explant angiogenesis assays. [30][31][32] We aimed to create a tunable hydrogel assay that could support as robust sprout formation as Matrigel does but is also tailorable to the user's specific study design needs. The 2 kPa gels treated with VEGF or PDGF + FGF both performed similarly to the Matrigel control with over 75% of samples in all three groups containing sprouts with an average of over seven sprouts per sample in all groups ( Figure 5). These results were not significantly different than the Matrigel, indicating that the 2 kPa PEGDA gels supported angiogenesis in a comparable manner to Matrigel. This finding is also in agreement with previous co-culture studies in the field suggesting that softer gels increase EC organization into microvessels as quantified by measurements such as total network length. [42][43][44][45] As gel stiffness increased to 10 kPa, both the VEGF-treated and PDGF + FGF-treated groups had significantly less samples with sprouts and significantly fewer sprouts per sample demonstrating the effect of stiffness on hindering angiogenesis. In the VEGF-treated gels, these outcomes were further decreased by the 20 kPa stiffness with an average of 12% of samples containing sprouts. Notably, this decrease was not seen in the PDGF + FGFtreated 20 kPa gel which had a slightly higher average percentage of samples containing sprouts than the 10 kPa PDGF + FGF-treated gels (33% vs. 29%, respectively). Since the 2 kPa groups performed similarly to Matrigel but were tailored to our specific study, we could then move forward with direct comparisons between PEG hydrogel groups to investigate how stiffness and growth factor treatment independently and in combination affected vessel formation.
The soft 2 kPa groups both had significantly more samples containing vessels than their respective 10 and 20 kPa groups indicating that softer gels better supported sprout formation. The role of suggesting that the vessels in the 2 kPa group are more mature. For example, there could be less regression of early sprouts in the 2 kPa F I G U R E 9 MMP-9 ELISA analysis of cell culture media collected at Days 2, 4, and 6 from PDGF + FGF-containing gels with stiffnesses of (A) 2 kPa, (B) 10 kPa, and (C) 20 kPa. Statistical analysis was done using a one-way ANOVA with Tukey's post hoc test to quantify differences between groups. N = 3 per condition, *p < .05.

F I G U R E 1 0 (A) The percentage of samples containing sprouts in each PDGF + FGF-treated gel condition according to IF analysis. (B)
Overview of the workflow and challenges of immunofluorescent image analysis using AngioTool, which quantified: (C) average vessel length, (D) average total vessel length, (E) average number of junctions, and (F) average total number of end points. Statistical analysis was done using a one-way ANOVA with Tukey's post hoc test to quantify differences between groups. N = 4 per condition, *p < .05, **p < .01.
VEGF-treated group allowing more sprouts to mature compared to the 20 kPa group, although this will need to be explored further.
The role of stiffness in the PDGF + FGF-treated groups was more convoluted. While the 2 kPa group still had a significantly higher per- Previous studies have shown that hydrogels containing PDGF + FGF are able to induce a more robust angiogenic response compared to hydrogels containing VEGF, 36 and that immobilizing PDGF + FGF into the hydrogel matrix leads to prolonged angiogenic signaling. 35 To further evaluate the efficacy of our modified angiogenesis assay, we assayed a subset of hydrogels containing immobi- In the PDGF + FGF-containing gels, the presence of MMP-9 was confirmed at Days 2, 4, and 6 in gels of all three stiffnesses levels ( Figure 9). The only significant change in MMP-9 levels was observed in the 2 kPa group where the amount of MMP-9 on Day 2 was significantly higher than Days 4 or 6, suggesting that MMP-9 may play an important role in early phase of sprouting. In the 10 and 20 kPa gels, no significant differences were observed, although there was a decreasing trend in MMP-9 levels from Day 2 to Day 6.
Finally, we performed immunofluorescent image analysis to confirm that the sprouts we observed in the PDGF + FGF-treated gels were in fact CD31+ ECs and TdTomato+ pericytes. We observed the same trend in samples containing vessels as we did in our brightfield analysis with the 2 kPa PDGF+FGF-treated gels having significantly more samples with sprouts than the 10 or 20 kPa groups ( Figure 10A).
We then explored the use of a publicly available immunofluorescent image analysis toolkit, AngioTool, 41  should also be taken to not edit any vessels out of the images, as well.
For each N average vessel length ( Figure 10C), average total vessel length ( Figure 10D), average number of junctions ( Figure 10E), and average vessel endpoints ( Figure 10F) were calculated. While these outputs differed from our brightfield analysis outputs, we observed that there were no significant differences in these metrics, as we did for number of vessels and longest vessel in the brightfield analysis.
This supports our hypothesis that the combined treatment of PDGF and FGF allows the vessels to overcome any stiffness-related barrier to organization. It is worth noting that there is more variability in the immunofluorescence data, and we believe this variability can be attributed to a potential loss of bulk tissue mass or vessels during the  9 While our present work features gels of homogeneous stiffnesses, we can adapt our hydrogel to better recapitulate the IPF microenvironment by also incorporating regional variations in stiffness 46,47 or start with a soft gel that progressively stiffens [48][49][50] to study the effect of varied stiffening on vessel formation.
Our results align with previously published findings. The sprouts in our assay look qualitatively similar to sprouts that have been observed in Matrigel. 30 Additionally, while the extracellular matrix in IPF is denser than that in healthy lung 51 and some studies have suggested that the density of stiff matrix has been identified as a barrier to angiogenesis. 52 Alternatively other studies of matrix stiffness independent of matrix density suggest that angiogenesis can increase with stiffness induced by increased ECM crosslinking, but that the vessels formed are typically disorganized and leaky. 53  In the pulmonology field, lung explants and precision cut lung slices have been popular models for studying pharmacology, [54][55][56] infection, 57-59 and physiology. 60,61 In the context of IPF, one study by Alsafadi et al. 62 uses a profibrotic cocktail, including tumor growth factor beta (TGFβ) and PDGF, to imitate the early fibrotic lung environment in IPF in their precision cut lung explant assay. A potential extension of our assay would be to combine our gel model with this profibrotic cocktail to analyze the effects on sprouting and MMP secretion, since TGFβ can stimulate MMP-2 63 and MMP-9 secretion and TGFβ can be activated by MMP-9. 64,65 Another recent study used precision cut lung slices from bleomycin and saline treated mice to screen potential drug candidates for IPF. 66 Bailey et al. 67 proposed using biomaterials in the study of precision cut lung explants to promote extended ex vivo culture of lung explants. Importantly, our study is the first to combine the use of biomaterials simulating a profibrotic microenvironment with pro-angiogenic conditions to study their effects on microvascular network dynamics at different stages of fibrosis.
Future studies should explore the role of pericyte migration and EC-pericyte coupling in overcoming stiffness associated antiangiogenic trends. Another area of interest would be to track the forces that these cells exert on their environment as they sprout, as has been done previously by Juliar et al. 68 Additionally, this organoid assay can be used to study how treatment with tyrosine kinase inhibitors, such as those used to treat IPF which target VEGFR, PDGFR and FGFR, affect vessel organoid remodeling in different environmental settings. Our assay is also versatile and can be adapted to perform dose response studies and/or to test the effects of growth factor gradients on vessel formation. Another interesting extension of this work would be to compare angiogenic sprouting from explants obtained from lungs harvested from mice that were treated with bleomycin (a widely used murine model of IPF), to angiogenic sprouting from lung explants obtained from healthy mice.
In conclusion, we have developed a novel, tunable hydrogel assay for the creation of lung explant vessel organoids that can isolate the effects of different biomechanical and biochemical signals.
This assay can be used to study the effects of growth factor availability and stiffness on vessel formation over time and can be easily adapted to the design goals of the user.

| Perspective
1. Developed a novel tunable hydrogel lung explant angiogenesis assay that can take some uncertainties out of angiogenesis assays by isolating the effects of biomechanical and biochemical signals.

2.
Our system was able to support angiogenesis in a comparable manner to Matrigel in soft 2 kPa gels while also allowing for tunability to study the effects of stiffness on sprout formation.
3. This highly versatile assay can be adapted in future studies to investigate the effects of other growth factors, heterogeneous stiffnesses, and progressive stiffening on sprout formation.

ACK N OWLED G M ENTS
The authors would like to thank Yixuan Yuan and Samuel Agro for their assistance with polymer and hydrogel synthesis.

CO N FLI C T O F I NTER E S T S TATEM ENT
The authors have declared no conflict of interest.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study are available from the corresponding author upon reasonable request.