**2. Results**

#### *2.1. Morphology and Biomechanical Properties of Native Tissue and Sca*ff*olds Derived from Healthy and IPF Lung Tissue*

The schematic layout of the study and sca ffold preparation is illustrated in Figure 1A. Macroscopic characterization of healthy and IPF lung tissue slices, using scanning electron microscopy (SEM), displayed an evident di fference in tissue morphology, with a dense meshwork of matrix and compact lung architecture in IPF tissue (Figure 1B), which correspond to the end stage of long-term IPF. After decellularization, the sca ffolds maintained original tissue integrity and characteristics in both IPF and healthy decellularized lung tissue, examined by SEM and in histology (Figure 1B–E). Overview images of sca ffolds with SEM illustrated the heterogeneity in the IPF patient material with more or less dense

tissue. The decellularized scaffolds from IPF and healthy individuals showed no signs of visible cells in the tissue, as seen with hematoxylin/eosin staining (Figure 1C). Furthermore, the cellular content was examined in decellularized IPF derived scaffolds, measuring dsDNA. In support of our previous study on healthy lung tissue by Rosmark et al. [13], dsDNA content was efficiently removed following decellularization showing only 1.5% residual dsDNA per mg tissue in IPF derived scaffolds (data not shown). In the stress–strain measurements, the native lung tissues from IPF patients showed significantly higher tensile stiffness in comparison to healthy individuals (*p* = 0.0003), as well as higher ultimate force (*p* = 0.0097) (*n* = 4) (Figure 1D). One duplicate of native lung tissue from one patient examined for stiffness was excluded and regarded as an outlier with a value (115.39) exceeding more than three standard deviations from the mean. These properties remained in the decellularized IPF scaffolds. The healthy scaffolds, on the other hand, showed a higher stiffness (*p* = 0.0485) and ultimate force (*p* = 0.0146) compared to the native tissue, although with a larger variability. Within the scaffold groups, differences in stiffness (*p* = 0.06676) and ultimate force (*p* = 0.0594) were maintained compared to difference in between native tissue groups. We did not observe any differences in stress-relaxation behavior for native lung tissue and the decellularized scaffolds for neither the healthy nor the IPF samples (Figure S1A). Force to failure curves revealed a clear shift towards higher tensile strength, with increased force to tissue displacement in IPF tissue (Figure S1B,C). Despite high patient variability, tissue density (mg/mm3) was significantly higher (*p* = 0.0022) in IPF scaffolds in comparison to healthy scaffolds (Figure 1E).

**Figure 1.** *Cont*.

**Figure 1.** Characterization of native lung tissue and scaffolds (**A**) Schematic of experimental layout. Dissection and decellularization of 350 μm human lung tissue slices (1). Mounting of repopulated scaffolds pre-cultured in SILAC medium (2). Schematics of culture conditions and sample extractions (3). Mass spectrometry (MS) analysis on light (green bars) and heavy (purple bars) protein intensities (m/Z, protein mass/protein charge) illustrating the mass shift of 6 Da (Arg) or 8 Da (Lys) between pre-existing (scaffold extracellular matrix (ECM)) and newly produced matrisome proteins (4). Intensity/μg was adjusted for tissue density resulting in intensity/mm<sup>3</sup> (5). (**B**) Representative scanning electron microscopy (SEM) images with the same magnification (scale bar = 100 μm) of native tissue (left) and decellularized tissue (scaffold) (middle) and scaffolds at an overview (right, scale bar = 1 mm) for illustration of sample variability (right). (**C**) Hematoxylin and eosin staining of native lung tissue and corresponding scaffold after decellularization of the tissue (scale bar = 100 μm). (**D**) Stiffness and ultimate force measurements of biological replicates (*n* = 3, with two technical replicates except for native healthy tissue) from native healthy and idiopathic pulmonary fibrosis (IPF) lung tissue and corresponding scaffolds (*n* = 4, with two technical replicates) derived from healthy and IPF tissue. (**E**) Density measurements of healthy and IPF scaffolds (*n* = 2, with three technical replicates). Unpaired *t*-test for significance between patient groups with *p*-values \* *p* < 0.05, \*\* *p* < 0.01, \*\*\* *p* < 0.001. Stiffness # *p* = 0.068, Ultimate force # *p* = 0.059.

#### *2.2. Proteomic Profiling of Lung Sca*ff*olds*

In the next step, we used quantitative mass spectrometry to determine the ECM composition using a matrisome classification system [14,19,20] to investigate if the molecular composition of the sca ffolds could be explained by the di fferences in matrisome properties between healthy and IPF sca ffolds. Each group, healthy and IPF, was analyzed in triplicates from each donor, with two donors per group (Figure S2). The analysis showed protein groups containing comparable numbers of identified matrisome proteins in both healthy and IPF derived sca ffolds, indicative of an equivalent protein extraction from each type of sca ffold (Figure 2A). However, the number of identified non-ECM proteins (other) were higher in IPF sca ffolds (530 proteins) in comparison to healthy derived sca ffolds (417 proteins), a di fference that could be explained by slightly increased cellular remnants in the compact decellularized IPF tissue. Nonetheless, the low content of dsDNA in IPF sca ffolds verified the matrices as decellularized tissue with > 98% DNA removal [21].

**Figure 2.** *Cont*.

**Figure 2.** *Cont*.

**Figure 2.** Proteomic and histological characterization of healthy and IPF derived tissue scaffolds. (**A**) Number of identified proteins in decellularized scaffolds derived from healthy individuals (biological replicates *n* = 2, technical replicates *n* = 3) and IPF patients (*n* = 2, *n* = 3). Protein groups assigned to matrisome affiliation. (**B**) Matrisome grouped summed raw intensities for proteins in healthy and IPF derived scaffolds, left panel. Matrisome grouped summed intensities after density adjustment in healthy and IPF scaffolds, right panel. Mean values for groups presented. (**C**) Statistics for summed matrisome groups. Student's *t*-test with Benjamini-Hochberg corrected *p*-values \* *p* < 0.05. (**D**) Unsupervised hierarchical clustering of significantly different matrisome proteins characteristic for the scaffold types using Z-scored values. Basement membrane proteins marked with \*. (**E**) Spearman correlation between scaffold groups. (**F**) Histological verification and spatial tissue distribution of selected matrisome (or basement membrane) proteins. Hematoxylin/Eosin staining showing distinct morphological differences between thin, alveolar septa (healthy scaffold) compared to thickened, fibrotic remodeled septa (IPF scaffold). (**G**) This is accentuated (arrows) by staining for collagen type IV (brown), indicating clear basement membrane staining lining the alveolar septa in the healthy scaffold compared to disorganized fragments in the IPF scaffold, with large areas devoid of collagen type IV signal or reduced intensity (\*). (**H**) Inversely, collagen type VI showed accumulation in these fibrotic structures in the IPF scaffold. Scale bar 50 μm.

The distributions of proteins in the two types of scaffolds were presented as summed intensities by matrisome groups (Figure 2B). To compensate for the discrepancy in tissue morphology between healthy and IPF derived scaffolds, the summed intensity (intensity/μg) of all proteins was adjusted for tissue density (mg/mm3) (Figure 1A:5). Data showed a distinct difference in intensity of matrisome groups between IPF and healthy derived scaffolds following density adjustment. The summed intensity of each matrisome group was increased in IPF scaffolds, in comparison to healthy derived scaffolds, reflecting the difference in matrisome composition. Further examination of matrisome group intensities showed a significantly higher amount of nearly all matrisome groups in IPF scaffolds compared to healthy scaffolds, seen as intensity/mm<sup>3</sup> (Figure 2C). To select matrisome proteins significant for respective scaffold group we used a threshold of fold change 2 and a Benjamini-Hochberg corrected

*p*-value below 0.05 between the healthy and diseased group. Visualization of significantly different matrisome proteins using unsupervised hierarchical clustering of Z-scored values, identified matrisome proteins characteristic for healthy and IPF derived scaffolds respectively (Figure 2D). The top cluster in the heatmap depicts matrisome proteins that were less abundant in IPF scaffolds, while the bottom cluster showed proteins more abundant in IPF scaffolds, as compared to healthy scaffolds (Figure 2D). Despite recognized lung tissue heterogeneity coupled to IPF patients [22] and limited number of patient samples, IPF derived scaffolds showed high correlation in matrisome composition, seen both in biological replicates and in between donors (rank = 0.81) (Figure 2E). Healthy derived scaffolds showed similar matrisome correlation (rank = 0.88) within its group. Further examination of the significantly expressed matrisome proteins showed that nine out of 20 of them were assigned as BM associated proteins (Figure 2D) [20]. Nidogen-2 and Collagen type VI (α 1,2,3 chain) were clustered together as more abundant in IPF scaffolds in comparison to healthy scaffolds, whereas laminin γ2, laminin β3, laminin α3 and collagen type IV (chains α3 and α4) were significantly decreased in IPF scaffolds. Hematoxylin/eosin staining showed distinct morphological differences with thin alveolar septa in the healthy scaffolds compared to thickened and remodeled septa in the IPF scaffolds (Figure 2F). This was also clearly illustrated by immunohistochemistry (IHC) for collagen type IV (Figure 2G) and type VI (Figure 2H). In the healthy lung scaffold, the BM of both sides of the alveolar septum displayed a thin, even double-line indicating the presence of collagen type IV. In the IPF lung scaffold however, this staining was uneven and often poorly defined, with areas of decreased staining intensity and both thickened and thinned BM structures. Most obvious was the increased alveolar septum thickness, representing large areas without collagen type IV signal. Inversely, collagen type VI showed accumulation in exactly these fibrotic structures in the IPF scaffold. Furthermore, the IPF scaffolds showed disorganized BM-fragments and possibly increased microvasculature. In summary, this illustrates the loss of normal lung organizations in IPF with altered BM membrane composition and structure.
