*3.5. Laser Scanning Microscopy (LSM)*

LSM recordings of a collagen–elastin hydrogel with primary collagen antibody staining support the above claims of lateral collagen–elastin polymerization (Figure 5). The left image (Figure 5a) represents an exemplary primary collagen antibody-stained sample. In total, images from seven random positions were recorded which all displayed the discussed features.

**Figure 4.** (**a**) Polar plot of the 2D anisotropy analysis of each 10 random positions in one 2 mg/mL collagen network and one collagen–elastin (33.3 w% elastin) network. The coordinates refer to the determined angle (polar angle) and the standard deviation (radial length) while the circles refer to the median values of the distributions of the standard deviations, i.e., elastin increases the standard deviation of the network and, thereby, the 2D anisotropy. (**b**) Comparison between the standard deviation of the angle distribution of the samples already displayed in Figure 4. Significance was tested with the Mann–Whitney U test. This standard deviation is referred to as "2D anisotropy".

**Figure 5.** Fluorescence images of a 33.3 w% elastin–collagen gel. (**a**) collagen type I antibody and (**b**) Col-F. The dots in each image are most likely clustered elastin–collagen monomers.

The right image (Figure 5b) shows all the network features (collagen + elastin). A comparison with the primary collagen antibody-stained image reveals identical features in both images while any observable differences must be attributed to the brightness thresholding of the image software. The left image contains the well-known features of the R + G collagen mixture, namely, the nodes and fibers, while the right image does not convey any additional structural features; therefore, a lateral alignment of the elastin and collagen monomers appears likely. This is further plausible, given the architecture of the elastic fiber under physiological conditions. Elastin is synthesized in the endoplasmic reticulum and then transported outside of a cell by binding to an elastin-binding protein. Upon binding of this protein to the galactosugars of micro fibrils outside of a cell, elastin is released from the elastin-binding protein and interacts then with the microfibrils. Elastin is then incorporated in a complex way into the microfibrils resulting finally in a fiber which contains elastin on the inside and a microfibrillar shell outside [15]. Basically, elastin needs a scaffold to be deposited on and several proteins of the fibrillin class as well as MAGP-1 where it is shown to interact with elastin [34]. Furthermore, although the elastin–fibrillin interaction is very complex, it has been shown that elastin binds to a glycine and proline-rich region in fibrillin-2 [35]. Consequently, some homology to collagen is given, which lends credibility to the elastin–collagen interaction as seen in the above described LSM recordings; however, other proteins such as fibulin-5 are also central to elastic fiber formation [36]. The list of important proteins continues and their non-existence in our system may be a likely explanation for the lack of formation of distinct elastic fibers. This consideration hardens our claim of a lateral, or at least, homogenous incorporation of elastin into collagen fibers, as elastin monomers simply do not experience guidance and as such are subjected to following collagen fibrillogenesis.

The image in Figure 5 contains the clustered collagen and elastin which was already discussed in the section, "plate reader", where the high elastin concentration sample displayed an absorption peak prior to the main maximum. These clusters are said to contribute to clouding as the early binding of elastin to collagen might form these clusters which, by chance, are not polymerized into the final network. As we can see the clusters also in the "collagen only" channel (Figure 5a), they must have at least contained some collagen, but as the clusters also appeared after the elastin addition, they must have also contained elastin.

An interesting accordance is seen when the 3D pore size of the networks is compared (Figure 6). The addition of elastin lowered the median pore size by only ~4%. Additionally, the interquartile range was smaller after the elastin addition (0.46 μm for the collagen and 0.32 μm for the collagen–elastin). A likely explanation for this effect is an increase in the fiber diameter because of a lateral fiber alignment between the elastin and collagen chains. The resulting thicker fibers would automatically lead to a decreased pore size when the network architecture remains unchanged, which was shown earlier. Indeed, Figure 6 reveals an increase in the fiber thickness by ~10% which is, again, similar to the percentage changes for the pore size. In fact, other imaginable polymerization types, i.e., a branched fiber alignment, should significantly reduce the median fiber thickness. Consider also the network illustration shown in Figure 7. If the addition of elastin to the network would connect random points along the turquois fibers, the pore size would be halved or at least significantly reduced. This effect must lead to a significant decrease in the median pore size which, presently however, was not observed.

**Figure 6.** (**a**) Pore diameter of a 2 mg/mL collagen gel and a 33.3 w% elastin collagen gel. (**b**) Fiber diameter of the same gel. Ten positions for each condition were used and 100 planes were summed each prior to analysis. Significance was tested with the Mann–Whitney U test.

**Figure 7.** Drawn intersection of collagen fibers (blue lines), which enclose a pore (circle); however, a hypothetical elastin fiber (red line) will divide the pore in two much smaller pores. As we did not see a significant decrease in the pore size, the only other mechanism must be lateral polymerization.
