FT-IR imaging of native and tissue-engineered bone and cartilage

Abstract

Fourier transform infrared (FT-IR) imaging and microspectroscopy have been extensively applied to the analyses of tissues in health and disease. Spatially resolved mid-IR data has provided insights into molecular changes that occur in diseases of connective or collagen-based tissues, including, osteoporosis, osteogenesis imperfecta, osteopetrosis and pathologic calcifications. These techniques have also been used to probe chemical changes associated with load, disuse, and micro-damage in bone, and with degradation and repair in cartilage. This review summarizes the applications of FT-IR microscopy and imaging for analyses of bone and cartilage in healthy and diseased tissues, and illustrates the application of these techniques for the characterization of tissue-engineered bone and cartilage.

Figure 1

Typical infrared spectra of bone (a) and cartilage (b). a) Infrared spectra of healthy adult human cortical bone showing the absorbance bands for vibrations of phosphate, carbonate, and protein amide bonds (AmI, AmII, AmIII). Only the frequencies (wavenumbers, cm⁻¹) used in mid-infrared imaging are shown. b) Infrared spectrum of bovine cartilage showing absorbance bands for vibrations of collagen amide bonds (AmI, AmII, AmIII), amide side chain groups, and proteoglycan sugar ring absorbance (PG).

Figure 2

Infrared images and pixel histograms for osteonal bone in an adult human female femoral neck: (left) mineral/ matrix ratio (M/M), (center) crystallinity (XST), and (right) collagen maturity (XLR). The same area of bone is shown in each figure; note that the crystallinity parameter occupies a lesser area than the collagen maturity parameter, as there is unmineralized matrix (osteoid) in the sample. The mean ± SD for the pixels in this image are 4.1± 0.45 (M/M), 1.30± 0.08 (XST), 2.1 ± 0.6 (XLR).. These pixel histograms can be combined to provide statistics for multiple samples.

Figure 3

Mineral properties in osteoporotic bone show different patterns when the maximum value in the center of the trabeculae is compared to the lowest value (generally at the edge). This figure, summarizing data described in Boskey et al [50], shows the mean (±SD) % change in iliac crest biopsies of both female and male patients with high-(HTOP) and low-(LTOP)- turnover osteoporosis as contrasted with age-matched normal controls for mineral/matrix ratio (top left), carbonate/phosphate ratio (top right), crystallinity (bottom left), and acid phosphate content (bottom right).

Figure 4

Qualitative images of matrix (a) and mineral(b) distribution on surface roughened titanium implants with bmp-2 (20 ng/ml) treated fetal rat calvarial osteoblasts. The cells were cultured on titanium implant materials (PT, pretreated; SLA-course grit blasted and acid etched, and TPS – Ti plasma sprayed). Only SLA and TPS showed close association of mineral and matrix with the greatest yield on the TPS plates. Reprinted with permission from B.D. Boyan et al [61].

Figure 5

FT-IR images of collagen content (amide I area) (i), PG content (PG sugar ring C-O absorbance (985–1140 cm⁻¹) normalized to amide I area) (ii), collagen integrity (1338 cm⁻¹ normalized to amide II area ratio) (iii), and collagen orientation (amide I/amide II ratio in polarized images) (iv) from the articular surface (superficial zone, SZ) down to the tidemark (lower boundary of deep zone, DZ), acquired from the medial femoral condyle of an OA (right panel) and non-surgical control (left panel) rabbit at two weeks post-surgery. The color scale indicates the pixel values for collagen and PG content, and collagen integrity index, where red and dark blue represent highest and lowest values, respectively. For the polarized data, the boundaries on the color bar indicate the three collagen fibril orientation categories with respect to the articular surface, corresponding to amide I / amide II peak area ratios ≥ 2.7 (parallel fibrils), between 2.7 and 1.7 (random fibrils), and ≤ 1.7 (perpendicular fibrils), respectively. A marked reduction in PG content is noted in the OA compared to Control cartilage.

Figure 6

Images from cross-sections of engineered mammalian cartilage grown in a hollow fiber bioreactor for 2, 4 and 8 weeks. Data include immunohistochemical type II collagen distribution, histological PG distribution (Alcian Blue stain), and FT-IR-determined collagen (amide I area) and proteoglycan ((PG sugar ring C-O absorbance (985–1140 cm⁻¹) normalized to amide I area) distribution. Qualitatively, increased collagen and a change in the distribution of PG can be observed at 4 and 8 weeks compared to 2 weeks of growth.

Figure 7

FT-IR images of rabbit osteochondral defect at 6 weeks post-repair with TP 508 protein. The defect repair cartilage has lower collagen content (amide I area) compared to the adjacent articular cartilage, but also has regions with PG content ((PG sugar ring C-O absorbance (985–1140 cm⁻¹)) of a similar magnitude to that in the native cartilage. There is some orientation present in the repair cartilage, as evidenced by the layer of fibrils parallel to the surface. For the orientation data, the boundaries on the color bar indicate the three collagen fibril orientation categories with respect to the articular surface, corresponding to amide I / amide II peak area ratios ≥ 2.7 (parallel fibrils), between 2.7 and 1.7 (random fibrils), and ≤ 1.7 (perpendicular fibrils), respectively.

Figure 8

(i). FT-IR and SEM images of PVA/PLGA hydrogels comprised of 10, 25, or 50% PLGA. Infrared absorbance peaks unique to PLGA and PVA are at 1756 cm⁻¹ and 1092 cm⁻¹, respectively, and the ratio of the peaks was used to monitor the quantity and distribution of the PLGA within the hydrogel. The hydrogel morphology changed with increaseing PLGA content. (ii). Extracellular matrix (ECM) produced by chondrocytic infiltration into the hydrogel was evident on Alcian Blue stained hydrogel sections, and was monitored by FT-IR imaging of the protein Amide I absorbance.