*2.3. Repopulated Sca*ff*olds*

After characterization of scaffold properties, we continued to study the cellular response of primary human lung fibroblasts cultured up to 9 days on scaffolds derived from IPF patients and healthy individuals. For these experiments we seeded fibroblasts, derived from a healthy donor, on scaffolds derived from four patients for each group and cultured in duplicates for each patient (*n* = 4 per group). To start, we examined if cellular attachment and viability varied between the two types of scaffolds. Cellular viability, measured as metabolic activity, showed no difference between the two types of scaffolds after 1 day of culture, indicative of equivalent numbers of attached cells (Figure 3A), also visualized by confocal microscopy (Figure 3B).

This result was also confirmed by counting the number of cells left in the wells after cell seeding, where no difference between the groups was seen. No significant difference in cellular viability was detected in any of the scaffolds measured up to 9 days in culture (Figure 3A). SEM imaging showed differences in cell orientation between the groups, with cells densely packed on top of dense areas of the IPF scaffolds and heavily repopulated less dense structures, whereas cells cultured on healthy derived scaffolds followed and maintained open lung structures (Figure 3C). To visualize cellular attachments and organization in the scaffolds, repopulated scaffolds were antibody labeled for collagen type VI, a cell binding protein, in combination with the mesenchymal cell marker vimentin and the focal adhesion protein vinculin (Figure 3D). Results showed similar staining of collagen type VI in both types of scaffolds, again representing thin alveolar septa in the healthy scaffold compared to heavily remodeled parenchyma in the IPF scaffold. Interestingly, repopulating fibroblasts on IPF scaffolds appeared to show more intense vimentin staining compared to cells on healthy scaffolds, which is in line with other studies demonstrating a correlation between vimentin and substrates stiffness [23]. Furthermore, cells on IPF scaffolds were primarily situated in less dense tissue areas and appeared stretched and elongated, lining surface edges of pulmonary structures (Figure 3D). To

further examine cellular attachments, immunofluorescence (IF) staining for vinculin was performed, an integrin involved in intracellular signaling [24] which showed no difference in cellular distribution between healthy and IPF derived scaffolds.

**Figure 3.** Cell viability and cell attachment of repopulated IPF and healthy tissue scaffolds. (**A**) Cellular viability of primary lung fibroblast repopulated on tissue scaffolds (biological replicates *n* = 4), shown as mean ± SD. (**B**) Confocal live imaging after 1 day of culture, showing equal cell attachment. Cell staining (blue), autofluorescent scaffold (green). Arrows indicate cells. Scale bar = 500 μm. (**C**) SEM after 9 days of culture, visualizing cellular differences in orientation in the scaffolds and repopulation. Arrows indicate elongated cells. Scale bar = 100 μm. (**D**) Visualization of repopulated scaffolds, showing fibroblasts attaching to the surrounding tissue (here: collagen type VI) and cytoskeleton (vimentin) and patterns of focal adhesions (vinculin). Hematoxylin/eosin staining of corresponding scaffolds shown in the right panel. Arrows: different intensities of vimentin in fibroblasts repopulating healthy vs. IPF scaffolds. Scale bar = 20 μm.

#### *2.4. Proteomic Profiling of Matrisome Proteins in Repopulated Healthy and IPF Lung Sca*ff*olds*

Healthy primary human fibroblasts were cultured in SILAC-medium 5 days before cellular seeding and over the whole culture period on the scaffolds. In this experiment the cells take up heavy amino acids from the media and start to produce proteins with heavy amino acids that are distinguishable in the mass spectrometer from residual scaffold proteins that only contain light amino acids (Figure 1A). This enabled us to follow protein turnover over time from day 1, by differentiating between newly synthesized cell-derived proteins (heavy) and pre-existing matrisome proteins in the scaffold (light). As with the mass spectrometry (MS) data for the decellularized scaffolds, the data for the repopulated scaffolds was adjusted for differences in tissue density and a mean value was

calculated for each group at each time point. Density adjustment allowed us to study an equal tissue volume and thereby the same number of cells in the two types of sca ffolds. The matrisome protein di fferences between the groups over time were analyzed through a Spearman correlation test, which clearly demonstrated that newly synthesized proteins (Figure 4A, heavy) from cells cultured on IPF sca ffolds had a di fferent protein composition compared to healthy individuals. Sca ffolds in the IPF group correlated within its group over time as did the healthy individuals (Figure 4A, light). The temporal changes of overall matrisome compositions for each type of sca ffold were shown as heavy and light matrisome protein groups over time (Figure 4B). Interestingly, fibroblasts diverged in their production (heavy intensity/mm3) of matrisome proteins, detected as early as day 1 of culture on IPF derived sca ffolds. At day 1, the fibroblasts produced a significantly (*p* = 0.0069) higher level of proteoglycans compared to cells cultured on healthy sca ffolds (Figure 4B). Over time, we observed a tendency of increased collagens production in repopulated IPF sca ffolds, however the level of ECM glycoproteins remained unchanged. Examination of the preexisting sca ffold composition (light intensity/mm3) (Figure 4B) representing ongoing ECM remodeling, showed increased amounts of proteoglycans (*p* = 0.0231) such as perlecan and lumican in IPF sca ffolds at day 1 (Figure S3). However, at day 9, IPF sca ffolds showed significantly decreased amounts of ECM regulators (*p* = 0.029) (Figure 4B) e.g., TIMP-3 (Figure S3) and secreted factors (*p* = 0.089) (Figure 4B).

**Figure 4.** *Cont*.

**Figure 4.** Proteomic characterization of matrisome proteins in tissue scaffolds repopulated with SILAC labelled fibroblasts. (**A**) Spearman correlations of matrisome proteins in repopulated healthy and IPF scaffolds at day 1, 3, and 9 after repopulation. Scaffold group mean MS-intensities for each time point presented (biological replicates *n* = 4, technical replicates *n* = 2). (**B**) Statistics for matrisome groups for repopulated healthy and IPF scaffolds over time calculated from summed matrisome groups. Student's *t*-test with Benjamin–Hochberg corrected *p*-values for significance between patient groups of the same time point with *p*-values \* *p* < 0.05, \*\* *p* < 0.01. Light # *p* = 0.089 (Secreted Factors). (**C**) Heatmap of matrisome proteins over time. Unsupervised hierarchical clustering of Z-scored values (ward.D2). Scaffolds are presented as patient means (biological replicate *n* = 2) and repopulated scaffolds (light and heavy) are presented as patient group mean intensities for each time point (*n* = 4) with technical replicates for each group (*n* = 2). Light and heavy intensities were selected and visualized according to previous scaffold clustering.

To exclude cell number variabilities in between the two types of scaffolds after adjusting the values, we examined heavy and light labeled histones in each group over time, showing no significant difference between IPF and healthy repopulated scaffolds at any time point (data not shown). These results indicate that the identified diversity in protein synthesis was dependent on the original scaffold properties and not by variety in cellular content. To further describe this finding, we selected the matrisome proteins found to be significantly different and descriptive for the respective decellularized scaffold group, and analyzed these further showing temporal differences in repopulated scaffolds. We compared both newly synthesized matrisome proteins (Figure 4C, heavy) as well as changes in the original scaffold composition over time (Figure 4C, light) based on the significantly different matrisome proteins in the scaffolds from the starting point (Figure 4C, scaffold). The top cluster in the heat map presents the matrisome proteins that are more abundant in healthy scaffolds in comparison to IPF scaffolds. Within its own group, newly synthesized proteins from IPF and healthy scaffolds showed similar protein expressions over time (top cluster). The overall pattern of newly synthesized matrisome

proteins appeared to overlap with the original scaffold composition representative for each group. These results indicate that the characteristics of the original scaffolds can influence cellular activity, stimulating scaffold specific protein production in primary fibroblasts, thus mimicking the composition found in the original scaffold.

Furthermore, to connect to our previous findings in the decellularized scaffolds, we examined how BM protein production was affected over time in each type of scaffold (Figure 5A). Most of the significantly different expressed BM proteins showed a reduced expression in IPF repopulated scaffolds (top cluster, Figure 5A) as compared to healthy scaffolds. Changes in protein intensity were analyzed over time and significantly differently expressed BM proteins (11 out of 20 BM proteins) were presented as mean heavy intensity for each group (Figure 5B). Interestingly, healthy scaffolds showed an increased production of nidogen-1 and laminins (subunit α3, β3, and α5) over time, whereas in the IPF scaffolds the synthesis was low or undetected. Similar responses were seen with collagen IV production of α3 and α4 chains. For repopulated IPF scaffolds we found an increased production of the following structural BM proteins; basement membrane-specific heparan sulfate proteoglycan core protein (perlecan), collagen type VI chains α1, α2, and α3. The diverse expression and downregulation of several heavy labeled BM proteins in repopulated IPF scaffolds further supports the manifestation of a disorganized BM as previously visualized in IPF derived scaffolds. Scaffold changes over time were shown with light labeled protein intensity, showing several BM proteins with similar temporal patterns as heavy intensities (Figure S4).

Based on the quantitative data in Figure 5B and Figure S4, we show with antibody labeling, the spatial expression pattern of collagen type VI in repopulated scaffolds, showing an intensified expression level at day 9 in healthy scaffolds as compared to day 1 (Figure 5C). In IPF scaffolds, collagen VI appeared stable over time showing no distinct visual difference in antibody labeling. The overall expression of collagen VI appeared visually increased in IPF scaffold vs. healthy, which could be explained by the higher density of the tissue.

At further examination of heavy labeled proteins, we identified that the synthesis of tenascin and periostin was significantly altered in repopulated IPF scaffolds, matrix components that have been associated with the progression of IPF [25,26]. Tenascin and periostin were also found to be elevated in the original scaffold composition of IPF (Figure 6A). Fibroblasts on IPF scaffolds produced significantly higher amounts of tenascin (*p* = 0.044 at day 3, *p* = 0.027 at day 9) and periostin (*p* = 0.039 at day 1) (Figure 6A), protein expression patterns that have been implicated in fibrosis [25,27,28]. IHC staining of repopulated scaffolds visualized periostin distribution (Figure 6B). Periostin was found in certain areas of the thin alveolar septa in healthy scaffolds. IPF scaffolds, on the other hand, had a stronger staining in less remodeled areas and very low periostin signal in the heavily remodeled and fibrotic tissue areas. No apparent intracellular periostin could be detected by IHC.

**Figure 5.** *Cont*.

**Figure 5.** Synthesis of basement membrane proteins in repopulated scaffolds. (**A**) Heatmap of basement membrane (BM) proteins over time. Repopulated scaffolds as group mean heavy intensities for each time point (biological replicates *n* = 4, technical replicates *n* = 2). With Student's *t*-test, significantly differentially expressed proteins are marked with \*. Perlecan = Basement membrane-specific heparan sulfate proteoglycan core protein. Production of significantly different basement membrane proteins in repopulated IPF and healthy scaffolds. Protein intensity shown as mean heavy intensity (**B**) over time for each group with SD. Student's *t*-test with Benjamin–Hochberg corrected *p*-values for significance between patient groups of the same time point with *p*-values \* *p* < 0.05. Blue = Healthy, Red = IPF. (**C**) Antibody labeling of repopulated scaffolds, at day 1 and day 9 of culture, showing collagen type VI (α1) (red) with DAPI staining (blue). Images of collagen VI staining, illustrate newly synthesized protein and original scaffold composition, with arrows exemplifying positive staining. Scale bar = 50 μm.

Synthesis of the proteoglycan decorin was upregulated (*p* = 0.043) in fibroblasts cultured on IPF scaffolds on day 1 of repopulation (Figure 6C). Decorin labeling showed clear intracellular and periocellular staining with a general enhanced overall tissue expression in IPF derived scaffolds (Figure 6D). Interestingly, these early changes in proteoglycan production were also detected for biglycan (*p* = 0.040) and versican (*p* = 0.027), showing significantly increased levels in repopulated IPF derived scaffolds (Figure 6C). The increased production of proteoglycans was further supported with antibody labeling (Figure 6D). IPF scaffolds showed higher levels of decorin compared to healthy scaffolds, and intracellular staining of decorin in fibroblasts were found on both scaffold types. Biglycan staining showed a strong intrinsic accumulation in IPF scaffolds, whereas healthy scaffolds only had sporadic staining apart from vessels. Intracellular biglycan could be found in fibroblasts cultured on both scaffold types. As for versican, staining was largely absent in healthy scaffolds, but a distinct intrinsic accumulation in IPF scaffolds could be seen as well as prominent cellular signal in fibroblasts cultured on IPF scaffolds. No intracellular staining of versican could be seen in cells cultured on healthy scaffolds.

In summary, these results demonstrate that the material properties of the ECM affected fibroblast activity, thus supporting a profibrotic phenotype when cultured in a diseased milieu.

**Figure 6.** *Cont*.

**Figure 6.** (**A**) Graphs of heavy and light intensity/mm<sup>3</sup> of tenascin and periostin of repopulated scaffolds (biological replicates *n* = 4, technical replicates *n* = 2) with calculated grand mean. Student's *t*-test for significance between patient groups means for each time point (\* *p* < 0.05). (**B**) Periostin in repopulated scaffolds (day 1). Antibody labeling indicating periostin (brown) in certain areas in alveolar septa in the healthy scaffolds, and to a stronger degree in less remodeled tissue areas in the IPF scaffolds. Largely absent periostin staining in heavily remodeled IPF tissue areas (\*). Arrows highlight differences in intracellular staining. Scale bar overview 500 μm, details 10 μm. (**C**) Graphs of original (light) and newly synthesized (heavy) proteoglycans decorin, biglycan, and versican of repopulated scaffolds (biological replicates *n* = 4, technical replicates *n* = 2) are presented with calculated grand mean of light and heavy intensity at day 1. Student's *t*-test for significance between patient groups (\* *p* < 0.05, \*\* *p* < 0.01). (**D**) Antibody staining illustrating proteoglycan increase in IPF vs. healthy scaffolds (day 1) of newly synthesized proteoglycans including original scaffold composition. Decorin (brown): IPF scaffold showing intrinsically more decorin, with cellular signal in fibroblasts on both scaffold types. Biglycan (red): rare intrinsic biglycan in healthy scaffolds apart from vessels, but strong accumulation in IPF scaffolds. Cellular signal of biglycan in fibroblasts on both scaffold types were seen. Versican (red): absent intrinsic versican in healthy scaffolds, but distinct accumulation in IPF scaffolds. Prominent cellular signal in fibroblasts on IPF scaffolds, but not on healthy scaffolds. \* extracellular deposition in repopulated scaffolds. Arrows highlight differences in intracellular staining. Corresponding hematoxylin/eosin staining of the repopulated scaffolds shown in bottom row. Scale bar 50 μm.
