C57BL/6J mice,12–20 wk old, (both sexes) were subjected to non-reperfused myocardial infarction (MI) as described previously.(13-17) Briefly, mice were anesthetized with intraperitoneal injections of ketamine hydrochloride (50 mg/kg) and sodium pentobarbital (50 mg/kg). Mice were orally intubated and ventilated; the ventilator’s room-air port was supplemented with oxygen. A 7–0 silk suture was passed under the left coronary artery and tied. The chest and skin were closed in layers. Mice were extubated upon recovery of spontaneous breathing. Analgesia (ketoprofen, 5 mg/kg) was provided prior to recovery and by 24 and 48 h post-surgery. The surgeon was blinded to mouse genotype. Seven days after MI, all mice were subjected to an echocardiogram to confirm sufficient depression of cardiac function (LVEF<60%). Mice were followed up through 4 wk.

Transthoracic echocardiography of the left ventricle was performed as previously described(13, 15, 16, 18, 19) with adjustments for the VisualSonics Vevo 3100 system. The sonographer was blinded to the treatment group.

The parasternal long axis (PLAX) view was used to obtain M-mode images for the measurement of the thickness of the left ventricle anterior wall (LVAW) and the left ventricular posterior wall (LVPW), and the left ventricular interior diameter (LVID). Fractional shortening (FS) was calculated by using the formula [(LVID;d – LVID;s)/ LVID;d] X 100. Four measurements were taken and averaged for each parameter. The parasternal long axis (PLAX) view, in B-Mode, was used to obtain two-dimensional images for the measurement of LV end-diastolic volume (LVEDV), LV end-systolic volume (LVESV), stroke volume (SV), ejection fraction (EF), cardiac output (CO), and heart rate, with six measurements taken and averaged. An apical four chamber view in Pulsed Wave (PW) Doppler was used to obtain two-dimensional images for the measurement of the isovolumic relaxation time (IVRT), with nine measurements taken and averaged.

Total RNA was isolated from cardiac fibroblasts as previously described for cardiac mesenchymal cells.(13) RNA quality was evaluated using NanoDrop ONEC (Thermo Scientific) and at least 2 μg of RNA was sent to the University of Louisville Sequencing Technology Center (Hyal2 and Has2 over-expression) and Admera Health (Has2 deletion) for poly-A RNA sequencing. At the University of Louisville Sequencing Technology Center, the TruSeq Stranded mRNA Library Prep Kit (Illumina Cat# 20020594) and TruSeq RNA CD Index Plate (Illumina Cat# 20019792) were used to generate a sequencing library. Size, purity, and semi-quantitation were performed on an Agilent Bioanalyzer 2100 system using the Agilent DNA HS Kit. Single-end sequencing was performed on an Illumina NextSeq 2000, using a P3 100 cycle cartridge with a P3 flow cell. Libraries were sequenced with a single 101 cycle read length with two index reads of 8 bp each. At Admera Health, isolated RNA sample quality was assessed by RNA Tapestation (Agilent Technologies Inc., California, USA) and quantified by Qubit 2.0 RNA HS assay (ThermoFisher, Massachusetts, USA). For library construction, the NEBNext^® Ultra^™ II Directional RNA Library Prep Kit for Illumina^® was utilized (New England BioLabs Inc., Massachusetts, USA). Final libraries quantity was assessed by Qubit 2.0, and quality was evaluated by TapeStation HSD1000 ScreenTape (Agilent Technologies Inc., California, USA). Equimolar pooling of libraries was performed based on QC values, and paired-end sequencing was performed on an Illumina^® NovaSeq X Plus (Illumina, California, USA). FASTQ files were generated utilizing BaseSpace DRAGEN analysis Version 1.2.1. (20-24)

Quality control (QC) of the raw sequence data was performed using FastQC (version 0.10.0) for each sequencing sample(25). The sequences were aligned to the mm10 mouse reference genome using STAR version 2.6.(26). Differential expression of ENSEMBL protein-coding transcripts was performed using DESeq2 (3), and raw counts were obtained from the STAR-aligned bam format files using HTSeq version 0.10.0.(27). The raw counts were normalized using the Relative Log Expression (RLE) method and then filtered to exclude genes with fewer than ten counts across the samples. DESeq2 guidelines were used to identify differentially expressed genes (sex was added as a covariate in analysis pipeline), and all P values were adjusted for testing multiple genes (Benjamini–Hochberg procedure; p≤0.05). Functional enrichment analysis was performed using the clusterProfiler R package to identify enriched Gene Ontology biological processes and KEGG pathways for each set of differentially expressed genes (DEGs)(28).

Heart tissue from infarct zone from 1 wk and 4 wk MI and from left ventricle from 1 wk and 4 wk sham hearts was powdered and weighed. 500 μL of Cell Lysis Buffer 2 (R&D Systems: 895347) was added to each sample and tissue was lysed for 30 min at room temperature with gentle agitation. Samples were centrifuged to remove debris. Samples were diluted to a 1:500 dilution. Quantikine ELISA: Hyaluronan Immunoassay (R&D Systems: DHYAL0) was used to assess HA concentration in each sample.

Adult mouse cardiac fibroblasts were isolated using collagenase II. The mice were anesthetized with 5% isoflurane and the hearts were quickly excised and washed with ice-cold phosphate buffered saline (PBS; pH: 7.4, Invitrogen 10010023). Excess fatty tissue was subsequently removed. The heart was then perfused with 5 mL of PBS, the aorta was removed, and the heart was finely chopped using a razor blade before being placed into a collagenase II solution (Worthington Biochemical, CLS2 LS004177) for 45 min. After digestion, the collagenase was neutralized using DMEM/F-12 containing 10% fetal bovine serum (VWR 97068-085), 1% penicillin-streptomycin (10,000 U/mL) (Invitrogen 15140122), 1% ITS-G (100×) (Invitrogen 41400-045), and 20 ng/mL human bFGF (Peprotech 100-18B). The collagenase-CFB Media solution was centrifuged, supernatant was aspirated, and the fibroblasts were resuspended in 5 mL of CFB media and filtered through a 70 μm cell strainer into a sterile 50 mL centrifuge tube. Fibroblasts were plated at a ratio of 1 heart per 75 cm² (10 mL/flask) and allowed to sediment for 2 h. Cardiac fibroblast media was changed following 2 h, 24 h, and subsequently every 2 d until reaching 70–80% confluency.

Cardiac fibroblasts were treated at passage 2. Cells (5 x 10⁵) were plated in 60 mm dishes (Corning 353002) and allowed to adhere to the bottom of the dish overnight. Cells were then serum starved for 24 h. Serum-containing media was added back to cells and 20 ng/mL of bFGF or TGFβ (Peprotech 100-21C) was added to reflect control conditions (bFGF) or activation (TGFβ). Conditioned media was collected from cells 24 h after treatment and hyaluronan was isolated from media.

Cardiac fibroblasts from Has2^fl/fl mice(29) were treated at P2. Cells (7 x 10⁵) were plated in 60 mm dishes and allowed to adhere to the bottom of the dish overnight. Cells were then serum starved for 24 h. To treat, fibroblasts were given Cre adenovirus (Vector Biolabs 1045N), a Has2 overexpression adenovirus (Vector Biolabs), or GFP adenovirus as the control. After 5 h, cells were given fresh media with 20 ng/mL of bFGF and TGFβ, respectively. After 48 h of stimulation, the conditioned cell media was collected, and HA was isolated from it. After isolation, the HA was sized using gel electrophoresis.

Cardiac fibroblasts from homozygous Hyal2^fl/fl mice were treated in the exact same manner as Has2^fl/fl fibroblasts. An adenovirus harboring Hyal2 (Vector Biolabs) was used to overexpress Hyal2.

Total cardiac RNA was extracted and used to make cDNA. The relative levels of mRNA transcripts were quantified by real-time PCR using Power SYBR Green (Thermo Fisher Scientific) on a real-time PCR system (QuantStudio5). Most primers were made using NCBI Primer Blast except HPRT primers (PPM03559E-200, QIAGEN). The data were normalized to mouse HPRT mRNA threshold cycle (CT) values by using the ΔΔCT comparative method. All primer sequences are listed in the (Supplemental Table 1).

Total cardiac protein was isolated, and protein concentration was determined by the Bradford assay with a protein assay dye (Bio-Rad, Cat. No. 5000006) using different concentrations (0, 1, 2, 4, and 8 mg/mL) of bovine serum albumin as standards. We measured protein concentrations using a Thermo Multiskan Spectrum spectrophotometer.

Following experimental treatments, conditioned medium was collected, centrifuged and supernatant mixed with Laemmli sample buffer (125 mM Tris-HCl, 10% SDS, 50% Glycerol, 0.05% Bromophenol Blue, pH 6.8) and incubated at 95°C for 5 min. Each sample was then loaded on a 7.5% acrylamide/bis SDS-PAGE gel for electrophoresis at 120 V. SDS-PAGE-resolved proteins were then transferred overnight onto a PVDF membrane (Cytiva, 10600021) at 4oC at 20 V for 16 h. Subsequently, membranes were blocked for 1 h at room temperature with Tris-buffered saline containing Tween-20 (TBS-T) and 5% bovine serum albumin (Sigma-Aldrich, A7906). Membranes were probed with primary antibodies overnight at 4°C. The following day, membranes were washed three times using TBS-T and then incubated with secondary antibody for 1.5 h. Membranes were then washed three times with TBS-T before exposure to Pierce^™ ECL Plus Western Blotting Substrate (Thermo Scientific, 32132) and imaging on a ChemiDoc imager (BioRad). Protein was normalized to total cell protein and parallel blots were normalized to an anchor protein. Antibodies used include primary anti-COL1A1 (1:20,000; Thermo Fisher, PA5-29569) and secondary anti-rabbit IgG HRP-linked (1:2,500; Cell Signaling, 7074). Because of potential promiscuity of collagen antibodies, we referred to immunopositivity with the COL1A1 antibody simply as ‘collagen’.

Following final echocardiography, hearts were excised and arrested in diastole with 2% KCl. Hearts were then sectioned into 1-mm short-axis sections. A midventricular section was fixed with 10% formalin; the samples were subsequently embedded, cut, and mounted. MATLAB version 9.7.01216025 (R2019b) was used to compile data files exported from the Keyence analysis software to generate a summary file for data analysis, and the microscopist was blinded to group assignment.

Transverse heart sections from sham, 1-week, and 4-week post MI mice were fixed in neutral-buffered 10% formalin and embedded in paraffin wax. The hearts were sliced into 4 μm sections, deparaffinized in an oven, dehydrated in xylene, and rehydrated with ethanol and water. All stained tissue slides were imaged with a Keyence BZX-810 microscope and analysis was performed with the Keyence software couple with customized Matlab macros.

To visualize fibrosis accumulation, the heart sections were stained with Picro-Sirius Red made from Direct Red 80 and saturated Picric Acid (Sigma-Aldrich 365548-5G, P6744-1GA). The right ventricle was removed from the image and the total fibrosis percentage was quantified in the remote zone, infarct zone, and left ventricle.

To examine HA accumulation, the tissue was incubated in hyaluronic acid binding protein (Sigma-Aldrich 385911-50μG) with 1× PBS and stained with Alexa Fluor 568 (ThermoFisher Scientific S11226). The nuclei were stained with DAPI. HA percentage was quantified using the same methods above.

Cardiomyocyte hypertrophy and capillary density were assessed. The borders of the cardiomyocytes were stained with wheat germ agglutinin (WGA CF^™ 488A conjugate, Biotium) and the capillaries with Isolectin B4 DyLight 649 (Vector Labs DL-1208-.5) in 10 mM HEPES solution with calcium chloride dihydrate. DAPI was used to visualize the nuclei. Only myocytes with centrally located nuclei and a circularity of 50–100 were used for quantification. Capillary density was calculated by dividing the number of capillaries by the total tissue area in the image (mm²).

Hyaluronan was isolated as described(13, 30), with slight modifications. For cultured fibroblast media, hyaluronan (HA) was isolated from conditioned media of treated fibroblasts. Proteins were digested using 1× proteinase K (Invitrogen 25530015) at 60°C for 4 h and first precipitation was done with 200 proof ethanol at −20°C (Day 1). Samples are then washed with 75% ethanol, resuspended in 100 mM ammonium acetate (Sigma Aldrich A1542) and nucleic acids were digested with Benzonase (EMD Millipore 70664-3 25 U/μl) at 37°C (Day 2). Second precipitation was done with 200 proof ethanol and samples were washed with 75% ethanol. Samples were air-dried, resuspended in 100 mM ammonium acetate (Sigma Aldrich A1542) and incubated at 4°C (Day 3). Samples were lyophilized and resuspended in 10 M formamide (Sigma Aldrich F9037) and incubated at 4°C (Day 4). A 1% agarose gel was made with 1× TAE was used to size HA. After casting, gel was pre-run in 1× TAE for 6 h at 80 V to remove impurities. Loading buffer (0.2% bromophenol blue (Sigma Aldrich B0126) in 10 M formamide) was added to each sample and HA^HMW (R&D Systems GLR002) positive control. Samples were run on gel in 1× TAE at 100 V for 2.5–3.0 hours and equilibrated in 30% ethanol for 1 hour before being stained with 1 mL of 400× stock stains-all solution (12.5 mg of stains-all (Sigma Aldrich E9379) to 5 mL of 200 proof ethanol) in 400 mL of 30% ethanol in the dark, overnight (Day 5). Gels were equilibrated in dd-water for 1 h in the dark and destained under ambient light on the bench for 15-20 min. Color images were taken on an EPSON flatbed scanner for presentation purposes. Image for quantitation was taken using a Bio-Rad ChemiDoc using the Stain-Free setting. Stain-Free images were analyzed using Fiji-Image J. Confirmation of Hyaluronan isolation are shown in (Supplemental Figure 1).

For heart tissue, mouse hearts were chopped into small pieces using a razor blade and suspended in 5 ml of 100 mM ammonium acetate (Sigma Aldrich A1542). Proteins were digested using 1× proteinase K (Invitrogen 25530015), shaken at 60°C for 4 h, and first precipitation was done with 200 proof ethanol. Isolation from this point was done as “Cultured fibroblast media” (vide supra).

Cardiac fibroblasts from wild-type mice were treated at P2. Cells (5 x 10⁵) were plated in 60 mm dishes and allowed to adhere to the bottom of the dish for 24 h and then serum starved overnight. To treat, fibroblasts were given various concentrations of glucose, mannitol, and stimulant for various time points. The media was changed and collected at various time points and HA was isolated.

Cardiac fibroblasts from Has2-floxed mice were treated at P2. Cells (6 x 10⁴) were plated in 6 well plates and allowed to adhere to the bottom of the dish for 24 h and then serum starved overnight. Fibroblasts were given Cre adenovirus (Vector Biolabs 1045N), a Has2 overexpression adenovirus (Vector Biolabs), or GFP adenovirus as the control. After 5 h, cells were given fresh media. 24 h and 48 h later, cells were lifted from the plate and counted using a flow cytometer. Media was collected from cells before lifting were imaged and cells were kept for RNA isolation after flow. bFGF treatment (naïve cells) was used as a control and TGFβ was used to stimulate cells into activation.

Fibroblasts were given HA^LMW and HA^HMW for 24 h and 48 h, then harvested, and counted using a flow cytometer. Cells were kept for RNA isolation after flow. bFGF was used a naïve control and TGFβ was used to stimulate cells into activation.

Cardiac fibroblasts from Has2-floxed mice were treated at P2. Cells (2 x 10⁵) were plated in 6 well plates and allowed to adhere to the bottom of the dish overnight. Cells were then serum starved for 24 h. Fibroblasts were given HA^LMW or HA^HMW for 24 h. Afterwards, cells were scratched for 6 h and 12 h time periods. Images were taken after timepoints with a Keyence imager. 0 h scratch images were taken immediately after 24 h HA treatment. Cells were kept for RNA isolation. BFGF was used a naïve control and TGFβ was used to stimulate cells into activation.

Fibroblasts were transduced with Cre- (Vector Biolabs 1045N), a Has2 overexpression adenovirus (Vector Biolabs), or GFP adenovirus (Vector Biolabs 1060) as the control. After 5 hours, cells were given fresh media for 24 h. After 24 h, cells were scratched, then followed for 0 h, 6 h, and 12 h. Images were taken after timepoints with a Keyence imager. 0 h scratch images were taken immediately after 24 h virus treatment. Media was collected from cells after scratches were imaged and cells were kept for RNA isolation. bFGF was used a naïve control and TGFβ was used to stimulate cells into activation.

Cardiac fibroblasts from Has2-floxed mice were treated at P2. Cells (7 x 10⁵) were plated in 60 mm dishes and allowed to adhere to the bottom of the dish overnight. Cells were then serum starved for 24 h. Fibroblasts were transduced with Cre- (Vector Biolabs 1045N), a Has2 overexpression adenovirus (Vector Biolabs), or GFP adenovirus (Vector Biolabs 1060) as the control. After 5 hours, cells were given fresh media for 24 h. After 24 h media was collected, and cells were left to produced ECM for 2 wk. Media was changed every 2 days. After 2 wk, dishes were decellularized using decellularizing solution (1× PBS containing 50 μM ammonium hydroxide [NH₄OH, 0.05% Triton X-100) to lyse cells and DNAse solution (20 U/mL DNase 1, 20 μL/mL DNase 1+MgCl₂ buffer, in sterile H₂O)] to degrade the DNA. The ECM left was removed from the dish and hydrolyzed using 12 N NaOH and boiling for 1 h at 120°C. Hydrolyzed ECM was neutralized using 12 N HCl and hydroxyproline concentration in ECMs were quantified using a commercial hydroxyproline assay (Abcam Ab222941). Confirmation of decellularization is shown in Supplemental Figure 2.

To confirm that HA accumulates in our mouse MI model (Figure 1F and 1G, Supplemental Table 2), we investigated hyaluronan accumulation in the heart acutely and chronically after myocardial infarction (MI). Our data indicate that HA accumulates in the left ventricle by 7 d post-MI (Figure 1A and 1B) and persists through 28 d post-MI (Figure 1C and 1D), and this was corroborated via HA ELISA (Figure 1E). This, coupled with our(13) and others’(31) prior observations indicate that HA accumulation is an early and durable occurrence following MI. This has important but largely unexplored implications for ventricular remodeling. From a spatial perspective, accumulation of HA was most notable in the infarct region and generally coincided with collagen, which was highly abundant at both 7 d (Figure 2A and 2B) and 28 d after MI (Figure 2C and 2D). HA did not accumulate significantly in the remote zone at either time point. We did not observe any sex differences in both HA accumulation and collagen in the infarct zone.

HA exhibits polydispersity (i.e., a range of sizes) (32) and is most often categorized into two main groups: high molecular weight (HMW; low MDa range) and low molecular weight (LMW; below 0.5 MDa). Because hyaluronan accumulates in the heart acutely and chronically post MI, we determined the relative size of the accumulating HA. These analyses indicate that HA^HMW accumulates in the infarct zone at 7 d and at 28 d (Figure 3A and 3B), with accumulation of HA^HMW being most notable at 28 d. Given the spatial coincidence of hyaluronan accumulation and scar formation, we next queried whether fibroblasts produce HA.

Fibroblasts can be activated by a variety of cues. The cytokine, TGFβ, is one of the commonest inducers of fibroblast activation in vitro. We used this well described activator in cardiac fibroblasts to assess the impact of fibroblast activation on HA levels. We show that after 24 h of TGFβ, there was a significant increase in HA accumulation compared to naïve (bFGF) fibroblasts (Figure 4A and 4B). We also show that the accumulated HA was HA^HMW. This is consistent with the observations depicted in Figure 3. Given that activated fibroblasts produce HA, we next determined whether the likeliest enzyme, HAS2 (encoded by Has2), was responsible for TGFβ-induced accumulation of HA.

There are three known hyaluronan synthases. Based on the literature and our prior observations of increased expression, it was likely that HAS2 was a source. When we evaluate RNA sequencing data from activated fibroblasts, we notice that the Has2 gene is the only one of the three hyaluronan synthases significantly increased compared to naïve fibroblasts (Figure 4C). Further supporting this idea, HAS2 is considered the most important isoform because it is finely regulated(33, 34) and its deletion is embryonic lethal.(35) For loss-of-function studies (Figure 5A and 5B), Has2 deletion significantly decreased HA^HMW accumulation in the media; this observation was true in naïve fibroblasts and TGFβ-activated fibroblasts. For gain-of-function experiments (Figure 5A and 5B), overexpression of Has2 (in the absence of TGFβ stimulation) increased HA^HMW accumulation in the media compared to naïve control, which effectively recapitulated the effect of TGFβ stimulation in terms of HA^HMW accumulation. These proof-of-concept studies definitively show that Has2 is the primary source of HA^HMW production by fibroblasts. To determine the extent to which deletion or overexpression of Has2 impacts the expression of relevant elements of relevant regulatory elements of the matrix, we also performed RNA-Seq.

To understand the substrate-level regulation of HA production, naïve and TGFβ-activated fibroblasts were treated with varying concentrations of glucose. In naïve fibroblasts, low glucose showed no significant change compared to basal glucose with high glucose showing the same (Figure 6A and 6B). Of the TGFβ-induced activated fibroblasts both low and high glucose showed no significant differences compared to basal glucose (Figure 6C and 6D). Interestingly, zero glucose shows significant decrease in HA^HMW accumulation in TGFβ-induced activated fibroblasts compared to normal glucose treatments. This effect was significant only after 72 h post treatment (Figure 6J), with the 48 h post treatment trending towards a decrease (Figure 6G). In naïve fibroblasts, the zero glucose treatment appears to be slightly fainter after both 48 h (Figure 6E) and 72 h (Figure 6H). Overall, these data suggest that extracellular glucose concentrations affect the amount of HA^HMW accumulated by activated cardiac fibroblasts.

To determine the extent to which changing HA production influences gene expression in fibroblasts, we used loss- and gain-of-function approaches as described above, isolated RNA, and performed bulk RNA-Seq.

In samples in which we forced expression of Has2, there were limited significant changes (only 13 genes downregulated) in gene expression (other than an increase in Has2). Of these 13 genes, four genes (Gbp3, Herc6, Ddx60, and Iigp1) were also downregulated in the context of TGFβ treatment. Whether these four downregulated genes are part of a HAS2-dependent suppression of expression, which also occurs in TGF-induced fibroblast activation, could be speculated. There is not, however, a clear pathway or role of these genes in fibroblast function. Although TGFβ changed the expression of thousands of genes, deletion of Has2 had limited impact on the response to TGFβ. When we deleted Has2, we observed upregulation of only five genes in naïve fibroblasts and zero gene changes after deletion of Has2 in activated fibroblasts (Supplemental Figure 3).

Initially, we were surprised to note that the total counts of Has2 mRNA did not significantly decrease after Cre adenoviral transduction; however, differential exon usage RNA sequencing (DEXseq) analysis revealed that the sequencing counts for the non-deleted region of Has2 were disproportionately higher than for the deleted region of Has2 (Supplemental Fig 4). To further confirm deletion of Has2, we performed qRT-PCR for the recombined exon, and generated UCSC Genome Browser sample expression tracks positioned at the Has2 gene and Integrative Genome Brower Sashimi plots depicting exon 2 skipping in Has2 deleted samples (Supplemental Fig 4). This is an important technical note for deletion studies, i.e., that the RNA-Seq data may not fully reflect artificial interventions into living systems.

In terms of manipulation of Hyal2, overexpression of Hyal2 in naïve fibroblasts changed the expression of no genes except Hyal2. In the context of TGF-activation, deletion of Hyal2 had little impact on the fibroblast transcriptional profile. Overall, 66 genes changed after deletion of Hyal2 in naïve cardiac fibroblasts. None of those genes were related to hyaluronan metabolism, hyaluronan receptors, or ECM proteins and when evaluating the Gene Ontology visualizations, most of the biological processes with the most changed genes involved were related to inflammation. Understanding HA’s role in inflammation is important in understanding HA’s role in cardiac wound healing (Supplemental Figure 5).

Given the history of extracellular hyaluronidase function ascribed to HYAL2 (encoded by Hyal2), we determined the extent to which Hyal2 may contribute to HA^HMW degradation. To address a potential role of fibroblast-mediated degradation of HA, we used a similar gain- and loss-of-function approach for Hyal2. The rationale for focusing on Hyal2 was driven by several lines of evidence. First, several studies show that Hyal2 has hyaluronidase activity. Second, Hyal2 is expressed in fibroblasts. Third, Hyal2 is reported to have hyaluronidase activity for HA^HMW. Fourth, HYAL2 is expressed at the surface of cells. Fifth, HYAL1 has only been shown to be expressed inside of cells (where it cannot access secreted HA). Sixth, the other expressed HYAL, Hyal3, has been shown to have activity only in sperm(36). Given this robust premise, we focused on Hyal2. Deletion of Hyal2 did not cause an increase in HA accumulation (Figure 7), suggesting little constitutive Hyal2-dependent hyaluronidase activity. Likewise, overexpression of Hyal2 did not decrease HA levels in the media, which calls into question its capacity to function as a universal hyaluronidase—though additional experiments would be needed to address this separate question. Nevertheless, these collective data indicate that fibroblast-borne Hyal2 does not contribute significantly to degradation of HA.

To better understand how HA may play a role during wound healing, we assessed whether changing HA accumulation caused changes in fibroblast proliferation. We first deleted and overexpressed Has2 in naïve fibroblasts but observed no significant changes in proliferation compared to control. We then deleted and overexpressed Has2 in TGFβ-induced activated fibroblasts and saw similar results (i.e., no group differences) (Supplemental Figure 6A). Next, we determined whether different-sized exogenous HA influences fibroblast proliferation. We treated naïve fibroblasts with HA^LMW or HA^HMW and found that changing the size of the HA did not influence fibroblast proliferation compared to control. We then treated TGFβ-induced activated fibroblasts with HA^LMW or HA^HMW and obtained similar results (Supplemental Figure 6B).

To further examine the effect of HA on fibroblast function, we assessed the impact of HA on fibroblast migration. Neither deletion nor overexpression of Has2 in naïve fibroblasts impacted migration compared to GFP. Even in the context of TGFβ-induced activation, neither deletion nor overexpression of Has2 differed from control (GFP). To determine whether different sizes of HA impact fibroblast migration, we treated naïve fibroblasts with HA^LMW or HA^HMW and observed no impact fibroblast migration compared to control. When treating TGFβ-induced activated fibroblasts in the same manner, no significant changes in migration were found (Supplemental Figure 6C).

Lastly, it is known that, after activation, fibroblasts increase collagen secretion. To determine if HA plays a role in the increased collagen secretion, we deleted Has2 in TGFβ-induced activated fibroblasts, but we found no difference in soluble collagen levels in media from treated fibroblasts. We also assessed Has2 deletion and overexpression in naïve fibroblasts and did not identify changes in soluble collagen secretion compared to naïve control. To test whether there were changes in collagen deposited into the ECM, we deleted and overexpressed Has2 in naïve fibroblasts and quantified hydroxyproline concentration from decellularized and hydrolyzed ECMs. There were no significant changes in hydroxyproline either after Has2 deletion or Has2 overexpression compared to the control (Figure 8).