Collagen mimetic peptides as novel therapeutics for vascular disease in the central nervous system

Abstract

Background: Loss of vascular integrity is a common comorbidity of neurodegenerative diseases of the central nervous system (CNS). Compromised blood flow to the brain and excessive vascular remodeling is evident in chronic systemic cardiovascular diseases such as atherosclerosis, driving neurodegeneration and subsequent cognitive decline. Vascular remodeling occurs in response to changes in the microenvironment, with the extracellular matrix (ECM) as a major component. Collagens within the ECM and vascular basement membrane are integral to endothelial cell (EC) function and maintenance of the blood-brain barrier. Disruption of the ECM and breakdown of collagen with disease may lead to vascular dysfunction and neurodegeneration.

Methods: We induced hyperglycemia in ApoE-deficient (ApoE-/-) mice by intraperitoneal injection of streptozocin (STZ; 50 mg/Kg) for 5 days and accelerated diabetic atherosclerotic disease through a high fat diet (HFD). Over a 12 weeks period, mice received weekly intravenous treatment of collagen mimetic peptide (CMP) or vehicle (phosphate buffered saline) to assess efficacy in promoting vascular integrity in central brain structures.

Results: Following the STZ/HFD regimen, diabetic atherosclerotic ApoE-/- mice treated with CMP exhibited increased vascular integrity compared to vehicle in the cortex and in the CA1 and dentate gyrus regions of the hippocampus, as assed by higher levels of the endothelial cell adhesion glycoprotein CD31 and intravascular collagen IV, increased vascular area, and diminished leakage. Interestingly, in hippocampus, astrocytes were closer in proximity to vessels despite being less numerous in the CMP group.

Conclusion: Collagen integrity is important for maintaining cerebrovascular architecture in disease. Application of CMP which intercalates with and repairs damaged collagen may have therapeutic use in neurodegenerative diseases by preserving vasculature structure and promoting blood-brain barrier integrity. These findings underscore the need to further explore the role of collagen repair as a novel therapeutic for diseases of the brain involving vascular degradation.

Keywords: atherosclerosis; cerebrovascular disease; collagen; collagen mimetic peptides; diabetes; extracellular matrix; neurodegeneration; vascular dysfunction.

FIGURE 1

Blood glucose and body weight were not affected by treatment regimen in mouse model (ApoE−/− + STZ). (A) ApoE knockout animals (ApoE−/−) were used for the atherosclerosis-diabetes-high fat model. Baseline weights and blood glucose measurements were taken 5 days before intraperitoneal streptozocin (STZ) (50 mg/Kg) injection. On day 7 post-final STZ injection blood glucose and weights were measured. Once hyperglycemia was confirmed, animals were switched to a high-fat diet (HFD; 45% fat). At initiation of HFD mice were randomly assigned to receive once weekly intravenous collagen mimetic peptide (CMP) (CMP13A, 0.1 mg/Kg) or vehicle control (1x PBS) and blood glucose and weight measurements for the remainder of the study. At 12 weeks post-STZ, animals were euthanized, tissue processed, and endpoints measured. (B) Mean body weight for ApoE−/− males and females increased over time following initiation of HFD at week 1 (dashed line) for both vehicle and CMP cohorts with considerable overlap in standard error of the mean (Veh: stippled gray; CMP: stippled green). (C) Compared to wildtype (WT, n = 2), at week 0 (baseline) ApoE−/− males and females were significantly lower in weight (#p < 0.0001). Following the experimental period, mean weight increased from baseline after HFD in both vehicle and CMP groups for males (*p < 0.0001) and females (*p ≤ 0.005). The weights of mice did not differ between vehicle and CMP groups for either males or females for the duration of the study (p ≥ 0.11). (D) Mean blood glucose levels for ApoE−/− males and females also increased over time with overlap in standard error of the mean (Veh: stippled gray; CMP: stippled green). (E) Blood glucose levels at baseline (day 0) for ApoE−/− males and females did not differ from WT (p > 0.05). Following the experimental period, glucose for males and females increased in both PBS and CMP groups compared to baseline (*p < 0.001). For either sex, there was no significant difference between vehicle and CMP groups (p ≥ 0.36). Cohorts: for vehicle, seven males, six females; for CMP, six males, eight females. For all comparisons, data presented as mean ± SEM. Statistical analyses = Two-way ANOVA, Tukey’s multiple comparisons test.

FIGURE 2

Collagen IV vascular labeling in the brain is increased after collagen mimetic peptide (CMP) treatment. (A) Representative montage of confocal images of collagen IV immunolabeling in the brain of vehicle-treated and CMP-treated ApoE−/− mice with wildtype (WT) shown for comparison. Scale bar = 500 μm. (B) Representative high-magnification confocal images of collagen IV immunolabeling in the cortex and CA1 and dentate gyrus (DG) areas of the hippocampus in vehicle-, CMP-treated, and WT mouse. (C) Quantification of collagen IV intensity (normalized to vehicle). CMP treatment significantly increased collagen IV intensity in all areas compared to vehicle (*p = 0.04, ****p < 0.0001), which was slightly less than WT (ns, p > 0.05). Two-way ANOVA Tukey’s multiple comparisons tests. n = 9 sections for CMP, seven for vehicle, and four for WT. Data presented as mean ± SEM. Scale = 30 μM.

FIGURE 3

CD31 vessel area and intensity in the brain is increased after collagen mimetic peptide (CMP) treatment. Representative confocal images of CD31 immunolabeling in the cortex, CA1 and dentate gyrus (DG) in vehicle, CMP, and wildtype (WT). Scale = 100 μM.

FIGURE 4

Quantification of CD31 area and intensity in cortex, CA1, and dentate gyrus (DG). (A) Mean area of CD31+ vessel segments decreased in vehicle animals compared to wildtype (WT) in cortex and CA1 (**p = 0.002; ****p < 0.0001), but not DG (ns, p = 0.06). With collagen mimetic peptide (CMP) treatment, CD31 area significantly increased in cortex, CA1, and DG regions relative to vehicle (**p = 0.03; ****p < 0.0001) and relative to WT in the cortex (*p = 0.03) while reaching WT levels in CA1 and DG. (B) Mean CD31 intensity in vessel segments (normalized to vehicle) was not changed between vehicle-treated and WT animals (p > 0.05). However, CMP increased CD31 intensity beyond vehicle-treated or WT in all brain regions (*p = 0.01; **p = 0.007; ****p < 0.0001). n ≥ 1,021 vessel segments for CMP, ≥ 1,208 for vehicle, and ≥ 642 for WT across nine, seven, and fiur sections, respectively. Two-way ANOVA Tukey’s multiple comparisons test for statistical analyses. Data presented as mean ± SEM.

FIGURE 5

Astrocyte-vessel interactions in the hippocampus. (A–C) Representative confocal images of glial fibrillary acidic protein (GFAP)-positive astrocytes with immunolabeling against a vascular marker, either CD31 as in (A) and (B), or Isolectin-B4 (IB4) as in (C), in CA1 region of hippocampus of (A) vehicle-, (B) collagen mimetic peptide (CMP)-treated, and (C) wildtype (WT) mouse. Scale = 30 μM. (D) The minor difference in GFAP intensity between CMP- and vehicle-treated and WT groups was not significant (ns; p > 0.05). (E) The number of astrocytes per equally-sized image field in the CA1 was less in CMP-treated mice (*p = 0.01) but not significantly different to WT (p > 0.05). (F) The shortest distance from the centroid of a given astrocyte to the nearest blood vessel was less in CMP-treated mice (n = 1,156 astrocytes measured) compared to vehicle-treated (n = 1,035 astrocytes) and WT mice (n = 518 astrocytes; ****p < 0.0001). (G–I) Representative confocal images of GFAP-labeled astrocytes with immunolabeling against a vascular marker, either CD31 as in (G) and (H), or Isolectin-B4 (IB4) as in (I), in DG region of hippocampus of (A) vehicle-, (B) CMP-treated, and (C) WT mouse. Scale = 30 μM. (J) GFAP intensity was again slightly higher with CMP treatment but not significant (p > 0.05). (K) The number of astrocytes per image in the DG region was similar in all groups (p > 0.05). (L) The shortest distance from the centroid of a given astrocyte to the nearest blood vessel not significantly different across all groups (p > 0.05; vehicle n = 1,527, CMP n = 1,360, and WT n = 300 astrocytes measured. All data are presented as mean ± SEM. Two-way ANOVA Tukey’s multiple comparisons test for statistical analyses.

FIGURE 6

Collagen mimetic peptide (CMP) crosses the blood-brain barrier and reduces leakage. (A) Intravenous injection of a fluorescently labeled CMP (green) led to punctate binding in the cortex, CA1, and dentate gyrus (DG). Immunolabeling against glial fibrillary acidic protein (GFAP) (red) is shown with vasculature using IB4 (purple) to reveal local pockets of astrocyte processes and vessels, respectively, for comparison. Scale = 100 μM. (B,C) Representative confocal images of cortical sections immunolabeled against biotin following intravenous (tail-vein) injection of (B) vehicle- and (C) CMP-treated mice. Perivascular leakage of biotin was most evident in vehicle tissue (arrows, zoomed area on right) compared to CMP. Scale = 50 μM. (D) Histograms of number of biotin-fluorescing pixels (following background removal) in cortical sections within a criterion distance (15 pixels, ∼7 μm) of the nearest blood vessel. Vehicle tissue demonstrated greater high-intensity biotin signal compared to CMP; this is reflected in significant difference between the mean pixel intensity (E), **, p = 0.0044; Welch’s t-test, for unequal variances. Data presented as mean ± SEM; n = 9 images for CMP; four for vehicle).

FIGURE 7

Collagen mimetic peptide (CMP) preserves vessel area with increased CD31. Total fluorescent intensity of CD31 immunolabel (in arbitrary units, AU) summed over the total area of CD31+ vessel segments in cortical (A), CA1 (B), and dentate gyrus (DG) (C) regions of the brain from vehicle- and CMP-treated ApoE−/− mice; Pearson correlation coefficients for the best-fitting linear regression are given for each set of data. For vehicle, CD31 intensity decreased in cortex and CA1 even as the area of CD31+ segments increased, indicating diminishing CD31 prior to vessel loss (p ≤ 0.038), while for CMP, the opposite occurred, reaching significance for CA1 and DG (p ≤ 0.013) but not in cortex (p = 0.061) due to outlying samples (indicated by *). For the cortex and CA1 region, the slopes of the regression lines for vehicle and CMP were significantly different (p-values indicated; Analysis of Covariance; n = 9–12 sections for CMP, 6–8 for vehicle).