Origin of transparency in scattering biomimetic collagen materials

Abstract

Living tissues, heterogeneous at the microscale, usually scatter light. Strong scattering is responsible for the whiteness of bones, teeth, and brain and is known to limit severely the performances of biomedical optical imaging. Transparency is also found within collagen-based extracellular tissues such as decalcified ivory, fish scales, or cornea. However, its physical origin is still poorly understood. Here, we unveil the presence of a gap of transparency in scattering fibrillar collagen matrices within a narrow range of concentration in the phase diagram. This precholesteric phase presents a three-dimensional (3D) orientational order biomimetic of that in natural tissues. By quantitatively studying the relation between the 3D fibrillar network and the optical and mechanical properties of the macroscopic matrices, we show that transparency results from structural partial order inhibiting light scattering, while preserving mechanical stability, stiffness, and nonlinearity. The striking similarities between synthetic and natural materials provide insights for better understanding the occurring transparency.

Keywords: collagen; mechanical properties; photonic materials; self-assembly; transparency.

Fig. 1.

Optical relationships between collagen fibrillar gradient and LC phase diagram. Images of a fibrillar collagen matrices (A and B). (A) The matrix is characterized by a gradient in concentration (20 to 50 to 100 mg/mL range) and exhibits a gap of transparency. (B) The matrix (red asterisk) which is delimited by the red dotted circle is fully transparent in the Petri dish; the final concentration of collagen is ∼45 mg/mL. Two other areas are observed: 1 is vacant and 2 is opaque after shearing. (C) Classical collagen phase diagram (scheme and polarized optical micrograph) from nematic (N) to cholesteric (C) observed between crossed polarizers (0 to 90°). Segments are molecules in the section plane; dots are molecules normal to this plane; nails are molecules in oblique position with the head representing the molecular extremity pointing toward the observer. The collagen phase diagram is implemented with a new precholesteric phase (P*) for which concentration is found around 45 mg/mL. For the P*, the birefringent texture appears as cubic platelets (D). This faceted pattern is still observed after a 45° polarizer rotation; the organization of collagen molecules for P* is presented schematically in the red framed image at higher magnification (SI Appendix, Fig. S2 A and B). The precholesteric phase (P) occurs by increasing the collagen concentration in the solution.

Fig. 2.

Ultrastructure investigations of optically transparent collagen-based materials by electron microscopies. (A) The collagen matrix (*) and the cornea (°) stored in Cornea Cold as classical storage medium. Both samples remain transparent when stored in the medium. (B) SEM and (C) TEM images of the biological cornea (°) and the synthetic transparent collagen matrix (*). (Scale bars, 1 μm [B] and 500 nm [C].) Insets in C are observations at higher magnification. (Scale bars, 50 nm.) The fibrils lay parallel (white line) or perpendicular (white dot) to the section cut. (D) Thick 3D fibrillar collagen matrix (1 cm in diameter) with a narrow gradient of concentration prepared by using the injection/reverse dialysis process to reach a final concentration around 45 mg/mL over the bulk. An optical gradient of transparency is observed going from opalescent (higher concentration) to transparent (lower concentration) from the curve side but is entirely transparent when observed from the flat end (Inset). (E and F) TEM images of the opalescent collagen matrix showing orthogonal arrangements of the fibrils reminding that of the corneal stroma.

Fig. 3.

Transparency of a freshly extracted pig cornea and of collagen matrices investigated by transmittance (A) and full-field OCT (B–E). (A) Transmittance spectra of collagen matrices for visible and near-infrared wavelengths are provided for concentrations ranging from 40 to 48 mg/mL. (B) Pig cornea over the black background of the sample holder (Top), and bar graph of the measured values of the scattering mean free path ℓ (Bottom) in four 1.3- × 1.3-mm² zones in the stroma (zones 1 through 3) and in the sclera (zone 4). Larger values of ℓ correspond to an increase in transparency. (C) In-depth OCT cross-section image in zone 3 (Right) from which ℓ is extracted by fitting the average OCT signal versus the depth z (Left, blue dots) with an exponential decay exp(−z/ℓ) (Left, red curve; values in the gray zone are discarded in the fitting procedure). (D) Collagen matrix (characterized by a gradient in concentration [20 to 50 to 100 mg/mL range; see Fig 1A]) over the black background of the sample holder (Top), and measured values of ℓ (Bottom) in seven 100- × 100-µm² zones, evenly spaced along a horizontal line, parallel to the collagen concentration gradient in the matrix. The general shape of the bar graph and the maximum value of ℓ measured in the third zone are consistent with the visual observation of maximum transparency in the center of the matrix. (E) Collagen matrix (with a narrow gradient of concentration around 45 mg/mL over the 1-cm bulk; see Fig. 2D) over the white background of the sample holder (Left), and in-depth OCT cross-section image (Right) along the collagen concentration gradient (black dashed line in E, Right). The change in the structural appearance of the matrix is clearly visible (from left to right in the OCT cross-section), with an increasing textured aspect.

Fig. 4.

Mechanical behavior (tension mode) of transparent collagen matrices and of freshly extracted pig cornea. (A) The three fibrillar collagen gels prepared by injection/reverse dialysis approach at concentrations in the range ∼43 to ∼45 mg/mL exhibit a “J-shaped” stress–strain response, as observed for biological tissues. As a guideline, the mechanical response of the pig cornea is given. No preconditioning was applied, that is, curves correspond to the first loading. Note that the maximal stress value before rupture, usually defined as the ultimate tensile strength (UTS), is only indicative since gel (as cornea) failure systematically occurred in the clamps, suggesting that clamps involve local premature damage. (B) Modulus increasing with extension, elastic moduli are given in the toe and linear region, respectively. Gel rigidities display moduli of same order of magnitude.