Development of a fibrin-targeted theranostic for gastric cancer

Abstract

Patients with advanced gastric cancer (GCa) have limited treatment options, and alternative treatment approaches are necessary to improve their clinical outcomes. Because fibrin is abundant in gastric tumors but not in healthy tissues, we hypothesized that fibrin could be used as a high-concentration depot for a high-energy beta-emitting cytotoxic radiopharmaceutical delivered to tumor cells. We showed that fibrin is present in 64 to 75% of primary gastric tumors and 50 to 100% of metastatic gastric adenocarcinoma cores. First-in-human ⁶⁴Cu-FBP8 fibrin-targeted positron emission tomography (PET) imaging in seven patients with gastric or gastroesophageal junction cancer showed high probe uptake in all target lesions with tumor-to-background (muscle) uptake ratios of 9.9 ± 6.6 in primary (n = 7) and 11.2 ± 6.6 in metastatic (n = 45) tumors. Using two mouse models of human GCa, one fibrin-high (SNU-16) and one fibrin-low (NCI-N87), we showed that PET imaging with a related fibrin-specific peptide, CM500, labeled with copper-64 (⁶⁴Cu-CM500) specifically bound to and precisely quantified tumor fibrin in both models. We then labeled the fibrin-specific peptide CM600 with yttrium-90 and showed that ⁹⁰Y-CM600 effectively decreased tumor growth in these mouse models. Mice carrying fibrin-high SNU-16 tumors experienced tumor growth inhibition and prolonged survival in response to either a single high dosage or fractionated lower dosage of ⁹⁰Y-CM600, whereas mice carrying fibrin-low NCI-N87 tumors experienced prolonged survival in response to a fractionated lower dosage of ⁹⁰Y-CM600. These results lay the foundation for a fibrin-targeted theranostic that may expand options for patients with advanced GCa.

Fig. 1.. Fibrin deposition is abundant in human primary and metastatic gastric tumors.

(A) Representative microscopic images of normal stomach and gastric adenocarcinoma seen on the biopsy cores from human TMAs, stained for fibrin using IHC staining (top) and H&E (bottom). Scale bar, 2000 μm (B) Quantification of fibrin deposition in primary gastric tumors and matched healthy or noncancerous adjacent stomach tissues. (C) Fibrin deposition among primary and metastatic gastric tumor types and adjacent stomach. (D and E) Comparison of the fibrin content among different grades of primary adenocarcinoma cores (P < 0.0001) (D) and between the available two grades of metastatic adenocarcinoma cores (P < 0.0001) (E). (F and G) Comparison of fibrin content between the male and female patients’ tumor cores with primary (P > 0.05) (F) and metastatic adenocarcinomas (P > 0.05) (G). Chi-square test was performed for comparative analyses. N/A, not available.

Fig. 2.. First-in-human ⁶⁴Cu-FBP8 fibrin–targeted PET imaging of gastric cancer.

Representative axial ⁶⁴Cu-FBP8 PET scans acquired at 2.5 to 3 hours after injection fused with nonfat saturated T1-weighted MRI and axial T1 MRI scans are shown (left two columns), alongside ¹⁸F-FDG PET/CT images (center). H&E and fibrin IHC staining (right two columns) of the available tumors are presented as well. (A) Patient #1 is a 71-year-old male with newly diagnosed poorly differentiated adenocarcinoma of the distal stomach. Tumor is shown with a white arrow, and adjacent noncancerous stomach is shown with a blue arrow. Follow-up standard ¹⁸F-FDG PET/CT was performed after receiving one cycle of immunotherapy. Tumor tissue core biopsy was obtained 3 weeks before fibrin scan. K, kidney; Li, liver; P, pancreas; A, aorta; white arrow, tumor site; blue arrow, noncancerous stomach. (B) Patient #2 is a 55-year-old female with newly diagnosed proximal gastric cancer (GCa) shown with a white arrow. ¹⁸F-FDG PET/CT was acquired 2 weeks before the fibrin scan. A tumor core biopsy was obtained 1 day after fibrin scan. (C) Patient #3 is a 65-year-old female with metastatic gastroesophageal junction cancer (GEJ Ca). The primary tumor (white arrow) and metastatic lesions in the right chest wall (yellow arrow) and inside the gluteal muscles (red arrow) are shown. ¹⁸F-FDG PET/CT was obtained 2 weeks before fibrin scan. Correlative H&E and fibrin staining of the primary tumor and chest wall metastatic lesion are from the diagnostic cores obtained 3 months before fibrin scan. (D) Quantitative analysis of ⁶⁴Cu-FBP8 uptake in primary and metastatic gastric tumors in all participants and incidental findings (uptake along the fibrin sheaths and intravascular clot) are reported as target SUV_max relative to muscle SUV_mean (T/M). Data are shown as means ± SD, and each data point demonstrates one target lesion. Scale bar, (histology image) 200 μm.

Fig. 3.. ⁶⁴Cu-FBP8 fibrin PET probe demonstrates persistent uptake in primary and metastatic gastroesophageal tumors 1 day after administration.

(A) Coronal PET maximum intensity projections (MIPs) at 2.5 and 21 hours after ⁶⁴Cu-FBP8 injection are shown for patient #7, a 66-year-old male with metastatic poorly differentiated adenocarcinoma of the distal esophagus and GEJ. (B) Representative axial ⁶⁴Cu-FBP8 fibrin PET images fused with fat-saturated T1-weighted MRI and T1-MRI scans are shown for patient #5. Images are focused on the subcarinal node (top row, white arrow), multiple metastatic lung nodules (top row, yellow arrowheads), left axillary and subpectoral nodes (middle row, white arrows), and left supraclavicular lymph nodes (bottom row, white arrows). (C) Representative H&E (left) and fibrin IHC stainings (right) of one of the left supraclavicular nodes are shown. Scale bar, 200 μm; hr., hour; p.i., post-injection.

Fig. 4.. ⁶⁴Cu-CM500 PET probe detects and accurately quantifies fibrin in gastric tumors.

(A) Representative fused PET/MRI and PET of male nude mice with fibrin-high SNU-16 and fibrin-low NCI-N87 subcutaneously implanted gastric tumors (white circle) using fibrin-binding ⁶⁴Cu-CM500 and fibrin nonbinding ⁶⁴Cu-CM120. (B) Quantitative analyses of ⁶⁴Cu-CM500 (n = 16 per tumor type) and ⁶⁴Cu-CM120 tumor uptake by PET (n = 4 per tumor type) using a two-tailed unpaired t test. (C) Biodistribution analyses in a subset of mice at 75 min after probe injection (n = 4 per group, unpaired t test, two-tailed). (D) Representative H&E (top) and fibrin IHC staining (bottom) of extracted SNU-16 (left) and NCI-N87 (right) tumor tissues. Image scale bar, 200 μm. (E) Comparison of FPA between tumor types (n = 8 per group, unpaired t test). (F and G) correlation of tumor fibrin content with tumor PET uptake for ⁶⁴Cu-CM500 (F) and ⁶⁴Cu-CM120 (G). Each data point represents one mouse. Triangles represent tissues for mice with NCI-N87, and circles represent tissues from mice with SNU-16 tumors. The bars include means ± SD. %ID/cc, percentage of injected dose per cc.

Fig. 5.. ⁹⁰Y-CM600 binds to tumor fibrin and reduces tumor growth in mouse models of GCa.

(A) Schematic illustration of biodistribution experimental design in male nude mice. (B and C) Biodistribution of fibrin-specific ⁹⁰Y-CM600 and control fibrin nonbinding ⁹⁰Y-CM620 at 1 hour (B) and 24 hours (C) after injection (n = 3 to 8 per group per time point). Each data point represents one mouse, and data for each group are shown as means ± SD. (D) Schematic illustration of therapeutic study design. (E and F) Tumor volume changes (relative to baseline) in response to single administration of ⁹⁰Y-CM600 (37 or 18.5 MBq iv), control nonbinding ⁹⁰Y-CM620 (37 MBq iv), fibrin-specific precursor (CM600), or vehicle in the fibrin-high SNU-16 (n = 4 to 6 per group) (E) and fibrin-low NCI-N87 (n = 8 or 9 per group) groups (F) over 20 days after treatment initiation. Tumor volume changes are shown as means ± SEM and compared using a one-way ANOVA with post hoc Tukey test. (G) Representative histological staining of the extracted tumor tissues 4 days after vehicle or 37 MBq ⁹⁰Y-CM600 treatment with H&E images in the first row and fibrin IHC staining in the second row. Scale bar, 2000 μm for the whole tumor area and 100 μm for the magnified image. The IHC images for cleaved caspase 3 are shown in the third row (scale bar, 100 μm). Immunofluorescence staining images of γ-H2AX are in the fourth row. Blue color, DAPI; red color, γ-H2AX; scale bar, 50 μm. (H to J) Quantitative comparison of fibrin content (H), cleaved caspase 3 (I), and γ-H2AX (J) between the vehicle- and ⁹⁰Y-CM600–treated tumors (n = 4 per group, unpaired t test, two-tailed). Triangles represent tissues from mice with NCI-N87 tumors, and circles represent tissues from mice with SNU-16 tumors. ns, not significant.

Fig. 6.. Toxicity assessments of the fibrin-binding ⁹⁰Y-CM600 radiopharmaceutical.

Healthy nontumor-bearing male BALB/c mice (n = 3 per treatment group) were injected with ⁹⁰Y-CM600 at the indicated dosages or vehicle and followed for 21 days. (A) Hematological and blood chemistry values at days 1, 7, 14, and 21 after injection with the indicated dosages of ⁹⁰Y-CM600 or with phosphate-buffered saline as vehicle control. Each data point demonstrates one laboratory value for each mouse. Truncated plots demonstrate the maximum and minimum values, and the middle line indicates the mean. Dashed lines demonstrate the normal laboratory values. (B) Mouse weight changes relative to baseline over 21 days after injection of ⁹⁰Y-CM600 or vehicle. (C) Representative H&E-stained histology micrographs of the kidney, liver, and lungs 21 days after intravenous administration of vehicle control (left) or ⁹⁰Y-CM600 at different dosages. image scale bar, 100 μm; WBC, white blood cell; AST, aspartate aminotransferase.

Fig. 7.. Fibrin-specific ⁹⁰Y-CM600 improves survival in mouse models of GCa.

(A to C) Tumor growth curves as groups (A) and for each mouse (B) and Kaplan-Meyer survival curves (C) of fibrin-high SNU-16 tumor–bearing male nude mice treated with one injection of ⁹⁰Y-CM600 (37 MBq iv X1, red), fractionated ⁹⁰Y-CM600 (18.5 MBq iv X3, injected at days 0, 8 and 16; blue) or phosphate-buffered saline (vehicle, black) as control (n = 6 to 10 per group). (D to F) Tumor growth curves as groups (D) and for each mouse (E) and survival comparison (F) of fibrin-low NCI-N87 tumor–bearing male mice treated with fractionated ⁹⁰Y-CM600 (18.5 MBq X3, blue, n = 4) or vehicle (black, n = 5). Individual and average (means ± SEM) tumor volumes are shown. Log-rank test was performed for survival comparison. *, days of treatment.