18F-fluorobenzoate-labeled cystine knot peptides for PET imaging of integrin αvβ6

Abstract

Integrin αvβ6 is a cell surface receptor minimally expressed by healthy tissue but elevated in lung, colon, skin, ovarian, cervical, and pancreatic cancers. A molecular PET agent for integrin αvβ6 could provide significant clinical utility by facilitating both cancer staging and treatment monitoring to more rapidly identify an effective therapeutic approach.

Methods: Here, we evaluated 2 cystine knot peptides, R01 and S02, previously engineered with a 3-6 nM affinity for integrin αvβ6, for (18)F radiolabeling and PET imaging of BxPC3 pancreatic adenocarcinoma xenografts in mice. Cystine knot peptides were labeled with N-succinimidyl-4-(18)F-fluorobenzoate and evaluated for binding affinity and serum stability. Peptides conjugated with (18)F-fluorobenzoate (2-3 MBq) were injected via the tail vein into nude mice xenografted with BxPC3 (integrin αvβ6-positive) or 293 (integrin αvβ6-negative) tumors. Small-animal PET scans were acquired at 0.5, 1, and 2 h after injection. Ex vivo γ-counting of dissected tissues was performed at 0.5 and 2 h.

Results: (18)F-fluorobenzoate peptides were produced in 93% ((18)F-fluorobenzoate-R01) and 99% ((18)F-fluorobenzoate-S02) purity. (18)F-fluorobenzoate-R01 and (18)F-fluorobenzoate-S02 had affinities of 1.1 ± 0.2 and 0.7 ± 0.4 nM, respectively, and were 87% and 94%, respectively, stable in human serum at 37°C for 2 h. (18)F-fluorobenzoate-R01 and (18)F-fluorobenzoate-S02 exhibited 2.3 ± 0.6 and 1.3 ± 0.4 percentage injected dose per gram (%ID/g), respectively, in BxPC3 xenografted tumors at 0.5 h (n = 4-5). Target specificity was confirmed by low tumor uptake in integrin αvβ6-negative 293 tumors (1.4 ± 0.6 and 0.5 ± 0.2 %ID/g, respectively, for (18)F-fluorobenzoate-R01 and (18)F-fluorobenzoate-S02; both P < 0.05; n = 3-4) and low muscle uptake (3.1 ± 1.0 and 2.7 ± 0.4 tumor to muscle for (18)F-fluorobenzoate-R01 and (18)F-fluorobenzoate-S02, respectively). Small-animal PET data were corroborated by ex vivo γ-counting of dissected tissues, which demonstrated low uptake in nontarget tissues with only modest kidney uptake (9.2 ± 3.3 and 1.9 ± 1.2 %ID/g, respectively, at 2 h for (18)F-fluorobenzoate-R01 and (18)F-fluorobenzoate-S02; n = 8). Uptake in healthy pancreas was low (0.3% ± 0.1% for (18)F-fluorobenzoate-R01 and 0.03% ± 0.01% for (18)F-fluorobenzoate-S02; n = 8).

Conclusion: These cystine knot peptide tracers, in particular (18)F-fluorobenzoate-R01, show translational promise for molecular imaging of integrin αvβ6 overexpression in pancreatic and other cancers.

Keywords: cystine knot; integrin αvβ6; positron emission tomography.

FIGURE 1

Cystine knot peptides. R₀1 and S₀2 are cystine knot peptides that contain 3 disulfide bonds, an active binding loop (black), and a sole primary amine at N terminus. N-terminal amine was coupled with ¹⁸F-SFB in 0.1 M sodium phosphate, pH 7.5, at 50°C for 45 min. Peptide sequences are presented with conserved residues highlighted.

FIGURE 2

Serum stability. ¹⁸F-fluorobenzoate-R₀1 (A) and ¹⁸F-fluorobenzoate-S₀2 (B) were incubated in 50% human serum at 37°C for 2 h. Samples were separated by RP-HPLC on C18 column with gradient of 5%–85% acetonitrile in water (both with 0.1% trifluoroacetic acid) from 5 to 35 min and analyzed with a γ-ray detector.

FIGURE 3

Small-animal PET imaging. BxPC3 pancreatic adenocarcinoma cells (integrin α_vβ₆–positive) or 293 (integrin α_vβ₆–negative) were xenografted into nude mice. ¹⁸F-fluorobenzoate-R₀1 (2–3 MBq) (A) or ¹⁸F-fluorobenzoate-S₀2 (2–3 MBq) (B) were injected via tail vein. Five-minute static scans were acquired at 0.5, 1, and 2 h after injection. Decay-corrected coronal and transverse slices are presented. Tumor (T) and kidneys (K) are marked on images. Images are all normalized to quantitative color scale shown on right of each figure set.

FIGURE 4

Small-animal PET quantification of data in Figure 3. Signals in tumor and muscle were quantified with AsiPro VM. Value and error bars represent mean and SD for ¹⁸F-fluorobenzoate-R₀1 (A and B; n = 5 for BxPC3, n = 3 for 293) and ¹⁸F-fluorobenzoate-S₀2 (C and D; n = 4 for BxPC3, n = 4 for 293). *P < 0.05.

FIGURE 5

Dynamic PET. BxPC3 pancreatic adenocarcinoma cells (integrin α_vβ₆–positive) or 293 (integrin α_vβ₆–negative) were xenografted into nude mice. ¹⁸F-fluorobenzoate-R₀1 (2–3 MBq) (A) or ¹⁸F-fluorobenzoate-S₀2 (2–3 MBq) (B) were injected via tail vein. Ten-minute dynamic scan was acquired via PET. Mean signals in ellipsoid regions of interest were quantified with AMIDE.

FIGURE 6

Biodistribution. BxPC3 pancreatic adenocarcinoma cells (integrin α_vβ₆–positive) or 293 (integrin α_vβ₆–negative) were xenografted into nude mice. ¹⁸F-fluorobenzoate-R₀1 (2–3 MBq) (A; 0.5 h: n = 3 for BxPC3, n = 3 for 293, n = 6 for nontumor values; 2 h: n = 5 for BxPC3, n = 3 for 293, n = 8 for nontumor values) or ¹⁸F-fluorobenzoate-S₀2 (2–3 MBq) (B; 0.5 h: n = 3 for BxPC3, n = 3 for 293, n = 6 for nontumor values; 2 h: n = 4 for BxPC3,n = 4 for 293, n = 8 for nontumor values) were injected via tail vein. Mice were euthanized at 0.5 or 2 h after injection. Tissues were collected, and decay-corrected activity relative to injected dose was determined per gram of tissue (%ID/g). Value and error bars represent mean and SD. Values are presented for Bl = blood;Bo = bone; Br = brain; Bx = BxPC3 tumor; H = heart; I = intestine; K = kidney; Li = liver; Lu = lungs; M = muscle; P = pancreas; Sk = skin; Sp = spleen; St = stomach.