2.1.

Animals and Type 2 Diabetes Induction

All experimental procedures were approved by the Ethical Committee for Animal Experiments of South China Normal University. Institute of cancer research (ICR) mice, which are considered the optimal animals for inducing T2D, were selected as the experimental animals. Eight-week-old male ICR mice weighing approximately 30 g were purchased from Guangdong Medical Laboratory Animal Center. Each mouse was housed in a cage with an alternating $12:12\mathrm{h}$ of light/dark cycle at an ambient temperature of 22°C to 25°C. After adaptive breeding for 5 days, mice were fed with a high-fat diet for 4 weeks. Their diet consisted of 45% fat, 35% carbohydrate, and 20% protein, as well as essential elements to induce T2D with insulin resistance²⁶ Mice were fasted for 12 h with free access to water and then injected intraperitoneally with streptozocin (STZ, $60\mathrm{mg}/\mathrm{kg}$; Sigma, United States), a cytotoxic agent for pancreatic $\beta$-cells, for 3 days to induce T2D.²⁷ At 3 days after injection, the fasting body weight (FBW; after 12 h) was measured and fasting blood glucose (FBG; after 12 h) in the tail venous blood was assessed by using a glucose meter (Accu–Chek Active; Roche Diagnostics Brazil Ltd, Brazil). From this point, FBG and FBW were measured every week, and non-FBG was also measured every morning until the end of the experiment. Animals with FBG values of $\geqq 11.1\mathrm{mmol}/\mathrm{L}$ and that exhibited typical clinical polydipsia, polyuria, and weight loss were considered T2D mice.^28,29

2.2.

Preparation of Collagen Scaffolds

Fresh mouse skin was obtained from the dorsum and abdomen of ICR mice after anesthesia, and hair was removed. The subdermal fat tissue was excised, and the skin samples were cut into rectangular pieces with dimensions $10\times 10\times 3\mathrm{mm}$. The pieces were immersed in 0.25% dispase solution (Aoboxing, Beijing, China) at 4°C for 48 h to remove the epidermis and then extensively washed three times in distilled water. Subsequently, the pieces were incubated in 0.3% Triton X-100 solution with continuous stirring at room temperature for 48 h and then in 0.25% trypsin solution at 4°C for 2 h to remove cells in the dermal structure. The resulting dermal matrix was thoroughly rinsed three times in sterile phosphate buffer solution (PBS) at 4°C for 15 min each. Finally, the packed collagen scaffolds were stored in sterile distilled water at 4°C after being sterilized in 70% ethanol for 2 h.^30–33

2.3.

Immunofluorescence Analysis

Anti-AGE antibody was measured to quantify the glycation levels in the glycated collagen from T2D mice. For immunofluorescence detection, $5\text{-}\mu \mathrm{m}$ paraffin sections of collagen scaffolds were placed in citrate buffer (pH 6.0) inside a steamer for 10 min, blocked by incubation with peroxide block and normal goat serum (Beyotime, Shanghai, China), and incubated with primary anti-AGE antibody ($1:400$, ab23722; Abcam, Hong Kong) at 4°C overnight. The sections were washed in PBS, and Texas Red®-labeled secondary goat anti-rabbit Ig G-TR ($1:100$, sc-2780; Santa Cruz Biotechnology, United States) was applied at room temperature for 2 h. All slices were stained with 4',6-diamidino-2-phenylindole (Sigma) for 15 min to display the nuclei. Slice images were obtained using a confocal laser-scanning microscope (710 NLO; Carl Zeiss, Jena, Germany).

2.4.

Confocal Raman Microspectroscopy of Collagen Scaffolds

The specimens of collagen scaffold were cut into $50\text{-}\mu \mathrm{m}$ thick sections using a Leica CM 1950 cryostat (Leica, Germany). The Raman spectra were directly recorded on collagen scaffolds without further preparation by Renishaw (Renishaw Inc., New Mills) via Raman microspectroscopy system equipped with a 300-mW near-infrared diode laser at a wavelength of 785 nm semiconductor laser for excitation. The laser beam of approximately 15 mW was focused on the sample surface by a $50\times$ objective lens of a Leica DM2500 microscope (Leica, Germany). Experiments were performed using a microscope in a backscattering geometry, where the scattering light was collected inside the cone, defined by the objective. The spectra were recorded with a resolution of $1{\mathrm{cm}}^{-1}$ in the range of 700 to $1700{\mathrm{cm}}^{-1}$ and collected with 5 s exposure time and three accumulations. All the data were collected under the same conditions. The raw spectra acquired from the collagen scaffolds in the range of 700 to $1700{\mathrm{cm}}^{-1}$ represented a combination of prominent autofluorescence, weak Raman scattering signals, and noise. Therefore, baseline correction was carried out by using a modified third-order polynomial fitting followed by smoothing.

2.5.

Statistical Analysis

All measurements were performed in triplicate sections from each sample. Data preprocessing consisted of three steps: removal of spectra containing cosmic rays, background subtraction, and normalization. Normal distribution of the fluorescence intensity and their homogeneity of variance were analyzed by the Shapiro–Wilk test and Levene’s test, respectively. Differences in fluorescence intensity at different timepoints were compared using one-way independent ANOVA and posthoc procedures for multiple comparisons. A Pearson correlation was used to measure the strength of the relationships between Raman intensity and fluorescence intensity of AGEs. All statistics were performed using the Statistical Package for Social Science (SPSS 20, SPSS Inc., Chicago, Illinois). All tests were two-tailed, and $P<0.05$ was considered statistically significant.

3.1.

Physiological Characteristics of STZ-Treated Type 2 Diabetes Mice

To examine whether T2D mice were successfully induced by STZ injection, we observed the behavior and measured the FBW and FBG for 3 days and every week after STZ injections of the control and diabetic mice. The T2D mice exhibited typical clinical symptoms, such as cachexia, polydipsia, and polyuria. Fig. 1(a) shows the average FBG levels of the control and diabetic mice from 1 to 12 weeks after SZT injections. The T2D mice had higher FBG values than the control mice ($P<0.05$). The FBG levels of the diabetic mice significantly increased during 1 week after STZ injections and nearly reached $25\mathrm{mmol}/\mathrm{L}$ during the whole experiment. All the diabetic levels were above $11.1\mathrm{mmol}/\mathrm{L}$, which is the standardized criterion for the FBG levels of the diabetic mouse models. However, the FBG levels of the control mice were stable at nearly $5\mathrm{mmol}/\mathrm{L}$. The average FBW levels of the control and diabetic mice from 1 to 12 weeks are shown in Fig. 1(b). The FBW levels of the diabetic mice significantly decreased during 1 week after STZ injections and reached approximately 32 g in the following experiment compared with those of the control mice ($P<0.05$). However, the FBW levels of the control mice were stable at nearly 42 g in the following experiment.

Fig. 1

(a) Average fasting blood glucose (FBG) levels from 0 to 12 weeks ($n=5/\text{group}$). At 0 day, the mice were injected with streptozocin (STZ). (b) Average fasting body weight (FBW) levels from 0 to 12 weeks ($n=5/\text{group}$). At 0 day, the mice were injected with STZ. Compared with the control group at the same time, *$P<0.05$.

Numerous differences were observed in the mice after 9 weeks of diabetes, especially the morphological variations of the skin. Figures 2(a) and 2(b) show that the diabetic mice were smaller and weaker with withered hair ($n=20/\text{group}$). The diabetic mice were about 10.5 cm long, while the control ones were only about 9 cm long. Figures 2(c) and 2(d) show that no significant pathological changes were observed in the skin of both the control and diabetic mice, but the diabetic skin was thinner than the normal skin with reduced elasticity. Figures 2(I) and 2(II) show that the number of the diabetic epidermal cells decreased. Moreover, the cell layer was less clear and lacked a multilayer arrangement. The skin collagen became atrophic, swollen, and degenerated, which were considered the major changes of histology caused by AGEs. The subcutaneous fat also exhibited progressive atrophy or even disappearance and was difficult to distinguish from dermis collagen.

Fig. 2

(a) Appearance of control mouse. (b) Appearance of diabetic mouse. (c) Image of the body part of control mouse without hair. (I): skin hematoxylin-eosin (H and E) image of the control mouse; (d) image of the body part of diabetic mouse without hair. (II): Skin H and E image of the diabetic mouse. Epidermal layer (EL), dermis layer (DL), connective tissue (CT), muscular layer (ML), and folliculus pili (FP) inserted in the dermis are shown from the epidermis to subcutaneous. $\text{Scale}\text{bar}=50\mu \mathrm{m}$.

3.2.

Characteristics of Collagen Scaffolds

We prepared the collagen scaffolds by degrading the fresh skin tissue derived from the mice. The collagen scaffolds were porcelain white and transparent. They were soft and resorbable for tissue engineering [Fig. 3(a)]. Collagen fibers in the skin are disassembled into collagen fibrils by decellularizing the dermal matrix. These scaffolds were composed of collagen fibrils without any cells, as shown in the scanning electron microscopy (SEM) images of the control and diabetic mice. The collagen fibrils ($\text{diameter}<200\mathrm{nm}$) of the control mouse were evenly distributed and extended in all directions [Fig. 3(b)]. However, the structure of the collagen fibrils in the diabetic mice was destroyed. Parts of the diabetic collagen fibrils were arranged in clusters. Also, the horizontal grain and the cross-linking structure of the collagen fibrils in the diabetic mice were more obvious [Fig. 3(c)].

Fig. 3

(a) Image of collagen scaffolds. (b) Scanning electron microscopy (SEM) of collagen scaffold of the control mouse. (c) SEM of collagen scaffold of the diabetic mouse. Electron microscope magnifications of both (b) and (c) are $50000\times$, and the extra high tension is 5 kV. $\text{Scale}\text{bar}=200\mathrm{nm}$.

3.3.

Immunofluorescence Assay of Glycated Collagen from Type 2 Diabetes Mice

In a complementary approach, the accumulation of collagen AGEs was detected by immunofluorescence assay with $5\text{-}\mu \mathrm{m}$ paraffin sections of control and diabetic mice at different timepoints. Figures 4(a) to 4(d) display the results of red fluorescence assay of 0, 4, 8, and 12 weeks diabetic glycated collagen scaffolds, respectively, which were emitted by Texas Red®-labeled secondary antibody. Figure 4(e) shows an obvious increase in the red fluorescence of glycated collagen scaffolds as a response of glycation levels. This result indicates that the content of AGEs increased as the diabetic time progressed. The mean fluorescence intensity was calculated by summing the total fluorescence intensity of all measurements and then dividing by the total fluorescent area of measurements.

Fig. 4

Immunofluorescence images of collagen advanced glycation end products (AGEs) from type 2 diabetic (T2D) mice at different timepoints. (a) Collagen AGEs of 0 week; (b) 4 weeks; (c) 8 weeks; and (d) 12 weeks. (e) Mean fluorescence intensity analysis was performed in diabetic collagen AGEs at different timepoints ($n=3/\text{group}$). Values are the mean of three independent experiments (***$P<0.001$, compared with the 0 week glycated collagen scaffolds from T2D mice, $\text{mean}\pm \mathrm{SEM}$).

3.4.

Confocal Raman Microspectroscopic Analysis of Collagen Scaffolds

The samples were analyzed by confocal Raman microspectroscopy to determine the differences of molecular groups between the control and diabetic collagen scaffolds. Each sample had three slices, and each slice had three spots; that is, nine spots were detected by confocal Raman microspectroscopy for one sample. The mean Raman spectra of the spots on each specimen are shown in Fig. 5, exhibiting the characteristic spectral features with associated band assignments (Table 1).

Fig. 5

Raman spectral analysis of glycated collagen scaffolds from T2D mice at different timepoints. Comparisons between Raman spectra of native nonglycated and glycated collagen scaffolds of T2D mice at different timepoints (4, 8, and 12 weeks). Spectra represent means of the three independent measurements and are normalized in the range of 700 to $1700{\mathrm{cm}}^{-1}$. Excitation wavelength, 785 nm; laser power, 15 mw; and laser spot diameter $\sim 1\mu \mathrm{m}$.

Table 1

Raman-selected bands (cm−1) and assignments in the range of 700 to 1700  cm−1 region of glycated collagen scaffolds from type 2 diabetic (T2D) mice at different timepoints.

The peak of $1005{\mathrm{cm}}^{-1}$ (assigned to Phe groups) remained highly constant upon glycation. The ratio intensities (${\mathrm{I}}_{\text{peak of interest}}/{\mathrm{I}}_{1005}$) were calculated for the peak of interest. Amide I (1635 and $1672{\mathrm{cm}}^{-1}$) and amide III (1244 and $1274{\mathrm{cm}}^{-1}$) bands did not exhibit position changes, confirming the conservation of the triple-helix structure of collagen during glycation. However, as shown in Fig. 6, the intensity of the Raman bands increased upon glycation at these positions, suggesting conformational changes, which may be due to the increase of cross-linking degree of collagen caused by nonenzymatic glycation. The correlation analysis suggested that the spectral maps corresponding to these bands had a high positive correlation with the expression of anti-AGE antibody obtained by immunofluorescence imaging of the same collagen scaffolds as shown in Figs. 6(I) to 6(IV). Furthermore, other important changes in peak intensities were revealed upon glycation, especially for residues of prolines (Pro) and hydroxyprolines (HyPro), which were the second and third aminoacids in the collagen triple-helix composition, respectively. The observed changes are represented in Fig. 7, showing a steady increase of the 921, 1033, and $1346{\mathrm{cm}}^{-1}$ (Pro) peaks overtime. The bands of Pro residues in the spectral maps exhibited a high positive correlation with the expression of anti-AGE antibody obtained by immunofluorescence imaging of the same collagen scaffolds as shown in the correlation analyses in Figs. 7(I) to 7(III). However, the 878 and $1180{\mathrm{cm}}^{-1}$ (HyPro) peaks had no obvious variety regulation, which suggested that collagen glycation had an excellent priority selectivity of Pro rather than HyPro. These data suggested changes in the amino acid exposure and protein backbone of collagen because of high glucose concentration in T2D mice.

Fig. 6

Analysis of Raman signals of amides I and III residues from glycated collagen scaffolds of 0-, 4-, 8-, and 12-week diabetes. The $1005{\mathrm{cm}}^{-1}$ peak corresponded to the Raman signal of Phe group. The 1635 and $1672{\mathrm{cm}}^{-1}$ peaks corresponded to the Raman signal of amide I groups, and the 1244 and $1274{\mathrm{cm}}^{-1}$ peaks corresponded to the Raman signal of amide III. The ratio of the intensities of each of these peaks (${\mathrm{I}}_{1635}/{\mathrm{I}}_{1005}$, ${\mathrm{I}}_{1672}/{\mathrm{I}}_{1005}$, ${\mathrm{I}}_{1244}/{\mathrm{I}}_{1005}$, and ${\mathrm{I}}_{1274}/{\mathrm{I}}_{1005}$) was calculated. (I), (II), (III), and (IV): Correlation analyses between the relative Raman intensity of 1635, 1672, 1244, and $1274{\mathrm{cm}}^{-1}$ bands with the mean fluorescence intensity of AGEs.

Fig. 7

Analysis of Raman signals from Pro and HyPro residues from glycated collagen scaffolds of 0-, 4-, 8-, and 12-week diabetes. The $1005{\mathrm{cm}}^{-1}$ peak corresponded to the Raman signal of the Phe group. The 921, 1033, and $1346{\mathrm{cm}}^{-1}$ peaks corresponded to the main Raman signal of the Pro group, and the 878 and $1180{\mathrm{cm}}^{-1}$ peaks corresponded to the main Raman signal of the HyPro group. The ratio of the intensities of each of these peaks (${\mathrm{I}}_{1921}/{\mathrm{I}}_{1005}$, ${\mathrm{I}}_{1033}/{\mathrm{I}}_{1005}$, ${\mathrm{I}}_{1346}/{\mathrm{I}}_{1005}$, ${\mathrm{I}}_{878}/{\mathrm{I}}_{1005}$, and ${\mathrm{I}}_{1180}/{\mathrm{I}}_{1005}$) was calculated. (I), (II), (III), (IV) and (V): Correlation analyses between the relative Raman intensity of 921, 1033, 1346, 878, and $1180{\mathrm{cm}}^{-1}$ bands with the mean fluorescence intensity of AGEs.