Anti-angiogenic SPARC peptides inhibit progression of neuroblastoma tumors

Our previous studies demonstrated that peptide FS-E, representing the N-terminal EGF-like module of the SPARC follistatin domain, potently inhibits angiogenesis [26]. This module of SPARC is a β-hairpin, highly twisted by disulfide bonds that link cysteine 1 to cysteine 3 and cysteine 2 to cysteine 4. The crystal structure of peptide FS-E shows that the two central cysteines are closely located (Figure 1). By linking cysteine 4 with cysteine 3 in lieu of cysteine 2, separate N- and C-terminal loops of the peptide can be produced without disturbing the native structure. Using this strategy, we synthesized the N- and C-terminal loops of peptide FS-E as two separate peptides, FSEN and FSEC, as detailed in Figure 1. Both peptides were folded into their native conformation by linking the cysteines that were placed at both ends. In the FSEN peptide, alanine was substituted for the unpaired cysteine. Corresponding scrambled control peptides, scFSEN and scFSEC that were designed to contain the same amino acids as peptides FSEN and FSEC in a random order, were synthesized without special modifications.

Migration assays were performed with human umbilical vein endothelial cells (HUVEC) and each of the synthetic SPARC peptides to characterize their anti-angiogenic properties in vitro and to compare their potencies with the original peptide FS-E. As shown in Figure 2, the C-terminal loop peptide FSEC inhibited bFGF-stimulated endothelial cell migration with an EC₅₀ of ~1 pM, which was even lower than EC₅₀ of ~10 pM of the original FS-E peptide. Higher concentrations of peptide FSEC did not inhibit endothelial cell migration, which is also a property of peptide FS-E [26] and the full-length SPARC [10]. Biphasic inhibition of endothelial cell migration is a characteristic feature of other natural inhibitors of angiogenesis, including thrombospondin-1 [27] and endostatin [28]. The N-terminal loop peptide, FSEN, showed weaker inhibition of bFGF-stimulated endothelial cell migration, with an EC₅₀ of ~2 nM. Neither peptide, tested at concentrations from 0.1 pM to 100 μM, affected proliferation of HUVEC cells stimulated with bFGF (data not shown).

The anti-angiogenic properties of the peptides were further characterized in vivo using the Matrigel plug assay. Matrigel plugs containing bFGF and either SPARC peptides FSEN or FSEC were less hemorrhagic than the positive control Matrigel plugs containing bFGF with or without scrambled SPARC peptides (Figure 3A). To quantify the anti-angiogenic effects of the peptides, we stained the endothelial cells in paraffin sections of the Matrigel plugs with green fluorescence using anti-CD31 antibody and visualized the perivascular pericytes with red fluorescence using anti-α-smooth muscle actin (SMA) antibody (Figure 3B). The area occupied by each type of cells was quantified at low magnification in duplicate fields in each sample using ImagePro software. As shown in Figure 3C, the quantity of CD31-positive endothelial cells was statistically significantly decreased in the SPARC peptides-treated Matrigel plugs compared to the positive controls with bFGF alone (p < 0.005). Although the area, occupied by SMA-positive pericytes in the Matrigel plugs containing the SPARC peptides was significantly smaller compared to the positive controls (p < 0.005), there was no significant differences in the ratio of pericytes to endothelial cells (p > 0.8), indicating identical pericyte coverage. In the Matrigel plugs treated with scrambled peptides, no differences in endothelial cell (p > 0.7) or pericyte area (p > 0.3) were detected compared to the positive controls with bFGF alone.

To examine the anti-tumor activity of the SPARC peptides, mice with subcutaneous neuroblastoma xenografts were treated with either FSEN, FSEC, scrambled peptide scFSEN, or PBS 5 days/week for 2 weeks. Compared to control group, statistically significant inhibition of tumor growth was seen in the animals treated with the FSEC peptide, but not FSEN (Figure 4). In the FSEC-treated animals, the average tumor weight at the end of the experiment was 26% of the weight of the control tumors (0.38 ± 0.42 g vs 1.48 ± 1.24 g, respectively; p = 0.01). In the group of animals treated with the peptide FSEN, the average weight of tumors was 88% of the controls, which was not statistically significant (1.30 ± 1.77 g, p = 0.83). The average weight of tumors resected from animals treated with the scrambled peptide was not different from the PBS controls (1.51 ± 1 g, p = 0.97).

Vascular architecture was initially evaluated on H&E stained sections. In the control tumors, the blood vessels were structurally abnormal with dilated, tortuous appearance and prominent microvascular proliferation (MVP) (Figure 4C). In xenografts treated with scrambled peptide, a similar architecture was noted with disorganized layers of endothelial and perivascular cells. Multiple areas of hemorrhage were present in the control and scrambled peptide-treated tumors. In the SPARC peptides-treated tumors, the blood vessel architecture was more normal, and there was no evidence of hemorrhage or MVP.

Immunofluorescent analysis of the tumors demonstrated a significant reduction in the quantity of endothelial cells in the SPARC peptide-treated xenografts compared to controls (Figure 5). The average area occupied by endothelial cells in FSEC-treated xenografts was 13 ± 6% of the blood vessel area in the control xenografts (5.9 ± 2.6 pixels ×10³ vs 44.9 ± 6.6 pixels ×10³, respectively; p < 0.001). Peptide FSEN also inhibited angiogenesis in the xenografts, although less potently. The average area occupied by endothelial cells in the FSEN-treated tumors was 33 ± 9% of the control (15.0 ± 4.0 pixels ×10³; p < 0.001). In contrast, no significant difference in the quantity of endothelial cells in the xenografts treated with scrambled peptide or control vehicle was detected (42.4 ± 13.1 pixels ×10³, p = 0.67).

The quantity of perivascular cells in each group corresponded to the level of angiogenesis. The average area occupied by pericytes in the tumors treated with peptides FSEC vs control tumors was 16 ± 10% (6.2 ± 4.0 pixels ×10³ vs 38.6 ± 10.4 pixels ×10³, respectively; p < 0.001) and 33 ± 16% (12.8 ± 6.2 pixels ×10³; p = 0.004) for peptide FSEN (Figure 5). No difference in the quantity of pericytes in tumors treated with the scrambled peptide and controls was detected (40.1 ± 13.1 pixels ×10³; p = 0.84). Although pericyte-to-endothelial cell ratio was not affected by the treatment, differences in the blood vessel architecture were apparent between the control and SPARC peptide-treated tumors. In the xenografts treated with vehicle or scrambled peptide, the blood vessels exhibited abnormal architecture with multiple layers of hypertrophic endothelial cells which often lacked co-localized pericytes. In contrast, blood vessels in the SPARC peptides-treated tumors had more normal architecture, mainly consisting of a single layer of spindle-shaped endothelial cells, enveloped by a layer of pericytes (Figure 6).

We used a novel approach to quantitatively assess blood vessel architecture by using 2-dimensional shape descriptors of Image J software that is similar to the method, previously described for 3-dimensional shapes [29]. Circularity is a measure of roundness of an object with value ranges from 0 to 1, where 1 is a perfect circle. As shown in Figure 6B, the blood vessels in the peptide-treated tumors are more circular compared to the PBS and scrambled peptide-treated controls (Figure 6B). Aspect ratio reflects the shape of an object as the ratio of its height to its width. We found that the aspect ratio was significantly closer to 1 in the treated versus the control tumors, indicating that these vessels are less elongated. Solidity, which approximates density of an object, was appreciably higher in the peptides-treated xenografts, compared to tumors treated with PBS or scrambled peptide, indicating that the endothelial cells are spaced more compactly in the blood vessels in the SPARC peptides-treated tumors. We integrated all three shape descriptors into a coefficient of abnormality to quantitatively assess roundness, elongation, and the border shape of the blood vessels using a scale from 0 to 1, with lower values reflective of more normal architecture. The scores for the blood vessels in the tumors treated with peptides FSEN and FSEC were significantly lower (0.53 ± 0.21 and 0.56 ± 0.20, respectively) compared to the vessels in tumors, treated with PBS or scrambled peptide (0.88 ± 0.12 and 0.93 ± 0.05, respectively).

SPARC peptides FSEN (CQNHHAKHGKVC) and FSEC (CELDENNTPMC) were synthesized at Alpha Diagnostics International (San Antonio, TX) using fmoc/tboc chemistry as cyclic cysteine-linked molecules, with disulfide bonds unambiguously formed during peptide synthesis. The purity of the peptides was assessed by high-performance liquid chromatography, and the molecular mass was checked by mass spectrometry. Control scrambled peptides scFSEN (KCGHKHQCAVHN) and scFSEC (MEPECNLNCTD), which contain the same amino acids as peptides FSEN and FSEC in a random order, were made without special modifications.

HUVEC cells (VEC Technologies, Rensselaer, NY) were maintained at 5% CO₂ in EGM media (Lonza, Walkersville, MD) supplemented with 5% FBS (Life Technologies, Carlsbad, CA). Neuroblastoma cell line SMS-KCNR was grown at 5% CO₂ in RPMI 1640 (Life Technologies) supplemented with 10% heat-inactivated FBS and 1% penicillin/streptomycin.

To characterize the anti-angiogenic properties of the peptides in vitro, migration assays were performed in a modified Boyden chamber with HUVEC cells in EBM media (Lonza) containing 0.01% BSA as described [42]. Serial dilutions of each of the synthetic SPARC peptides, starting at 100 μM were assayed with or without 10 ng/ml bFGF (National Cancer Institute Preclinical Repository, Frederick, MD). At least three independent experiments were performed for each peptide. To compare multiple experiments, the data were represented as the percentage of bFGF-induced stimulation using the difference between bFGF-induced migration and background migration in EBM alone as 100% control.

The effect of SPARC peptides on endothelial cell proliferation was measured in 96-well plates with HUVEC cells in EBM media supplemented with 0.5% FBS. Serial ×10 dilutions of SPARC peptides from 0.1 pM to 100 μM with and without 10 ng/ml bFGF were added to quadruplicate wells, cells were grown at 37°C in 5% CO₂ for 24 and 48 h, and proliferation was determined with the CellTiter cell proliferation assay (Promega, Madison, WI).

A neuroblastoma xenograft model was used to examine the anti-tumor activity of peptides FSEN and FSEC. Briefly, 4-6 week old athymic nude mice were injected subcutaneously with 0.2 ml PBS containing 1 × 10⁷ SMS-KCNR neuroblastoma cells. Once tumors were palpable (~70 mm³), animals were randomized into four treatment groups with at least 6 animals per group. Experimental animals received intraperitoneal injection of 10 mg/kg of the SPARC peptides FSEN or FSEC, or scrambled peptide scFSEN (0.3 mg of the peptide in 150 μl of PBS) 5 days a week for 2 weeks. Animals in the control group were injected with PBS without the peptide. Possible toxic effects of the treatment were assessed by monitoring the animals for signs of distress such as difficulty in ambulating, ulceration, cachexia, respiratory distress, loss of 15-20% body weight, or gain of 10% body weight. The size of the tumors was determined every 2-3 days by external measurements with a caliper. At the end of treatment, animals were euthanized using CO₂ followed by cervical dislocation. Tumors were removed, measured, weighed, and photographed. Tissue was fixed with 10% buffered formalin, embedded in paraffin and 4 μm-thick sections were prepared for histologic evaluation. All animal studies were approved by the IACUC at the University of Chicago.

Four-μm-thick sections were stained with H&E for histological evaluation. The entire tissue section was evaluated for vascular morphology. Vessels with thickened walls containing complete layers of hypertrophied endothelial cells plus additional layers of vascular mural cells were classified as positive for MVP as previously described [38]. Vessels with normal vessel architecture and no more than a single layer of flat, spindle shaped endothelial cells were characterized as MVP negative. For quantitative analysis of angiogenesis, 4-μm-thick sections were stained with anti-CD31 antibody (Santa Cruz Biotechnology Inc., Santa Cruz, CA) at a 1:50 dilution followed by FITC-labeled secondary antibody. To characterize the blood vessel architecture, pericytes were visualized with red fluorescence using anti-α-SMA antibody (Sigma-Aldrich, St. Louis, MO) at 1:100 dilution. The area occupied by each cell type was quantified at ×100 magnification in duplicate fields in each sample using ImagePro software. Two-dimensional shape descriptors of Image J software were used to quantitatively assess blood vessel architecture. Circularity was calculated as 4π × area/perimeter², modified aspect ratio was measured as object's width/length, and solidity was calculated as a ratio of an object area/convex hull (outermost outline). Coefficient of abnormality was calculated as 1 - (circularity × aspect ratio × solidity).