Preclinical Evaluation of ⁶⁸Ga-Labeled CSPG4-Targeting Peptides for PET/CT Imaging in Pancreatic and Gastric Carcinoma

Abstract

Chondroitin sulfate proteoglycan 4 (CSPG4) is highly expressed in various solid tumors and promotes tumor progression and migration. Targeting CSPG4 therapy, such as monoclonal antibodies 9.2.27 and CAR-M, has already been in clinical trials. Database analysis further correlated elevated levels of CSPG4 with reduced overall survival in both cancers. Achieving real-time visualization of aberrant CSPG4 expression in various solid tumors remains a critical issue that needs to be addressed. To address this, we pioneered the preparation of three ⁶⁸Ga-labeled PET probes targeting CSPG4, [⁶⁸Ga]­Ga-DOTA-LS10, [⁶⁸Ga]­Ga-DOTA-SH11, and [⁶⁸Ga]­Ga-DOTA-TH10. In pancreatic (ASPC1) and gastric (MKN45) carcinoma models, [⁶⁸Ga]­Ga-DOTA-TH10 showed prominent tumor uptake (ASPC1:1.63 ± 0.15%ID/g at 60 min), validated by Western blot and immunohistochemistry confirming CSPG4 overexpression. This study conducted the first PET imaging on CSPG4 expression in pancreatic cancer and gastric cancer models using ⁶⁸Ga-labeled peptides, aiming to provide guidance for clinical real-time monitoring of its expression.

Introduction

Early and accurate diagnosis, along with precision therapy, is essential for prolonging survival and improving the quality of life in cancer patients. Although fluorodeoxyglucose (FDG) and fibroblast activation protein inhibitors (FAPIs) have long been regarded as key tools in tumor imaging, their clinical applications are limited, particularly by false-positive results in inflammatory or fibrotic conditions. Therefore, developing more target-specific molecular probes and implementing effective radiotheranostic strategies have become urgent priorities in oncology research. Identifying novel biomarkers as precise diagnostic and therapeutic targets is a critical goal in advancing nuclear medicine with the potential to transform precision oncology.

Chondroitin sulfate proteoglycan 4 (CSPG4) is a type I transmembrane glycoprotein encoded by the CSPG4 gene on chromosome 15, consisting of a core protein and covalently attached chondroitin sulfate (CS) glycosaminoglycan (GAG) side chains. Although CSPG4 is absent or expressed at negligible levels in most normal tissues, it is strongly overexpressed in various human malignant tumor cells, the extracellular matrix (ECM) of the tumor microenvironment (TME), and tumor stem cells (TSCs). This aberrant expression pattern has been observed in several aggressive cancers, including pancreatic cancer, gastric cancer, triple-negative breast cancer (TNBC), and glioblastoma multiforme (GBM). Beyond its structural role, CSPG4 exerts multiple tumor-promoting functions, including stimulating cell proliferation, enhancing adhesion and invasiveness, promoting angiogenesis, inducing chemoresistance, enabling immune escape, and conferring radioresistance. We chose CSPG4 as a novel nuclear medical biotarget for three reasons: (1) it is widely expressed by various cancer cell subtypes but at negligible levels in most normal tissues, increasing its diagnostic multifunctionality across multiple cancer diseases; (2) it is abundantly expressed in multiple cancer stem cells, enabling early detection; (3) it is highly expressed in the tumor stroma microenvironment, increasing its diagnostic versatility. Collectively, these properties establish CSPG4 as a promising nuclear medicine biomarker with potential applications in early tumor diagnosis, prognostic assessment, and targeted radiotherapy development. It is noteworthy that monoclonal antibodies targeting CSPG4, such as 9.2.27 and 225.28, have also been modified into radioactive probes, and related trials have achieved promising results. However, due to the prolonged in vivo retention time of monoclonal antibodies, their clinical applications are often constrained.

Peptides combine the advantages of small molecules and biopharmaceuticals, integrating cell-penetrating properties with high target selectivity and the ability to engage large protein surfaces. They exhibit favorable pharmacokinetic profiles, including rapid tissue penetration, efficient systemic clearance, low immunogenicity, and straightforward radiolabeling feasibility. Consequently, radiolabeled receptor peptides have emerged as a promising approach for precise cancer diagnosis and radiotherapy. Notably, CSPG4-targeted peptides LS10 (LTLRWVGLMS) and TH10 (TAASGVRSMH) have demonstrated potent antiglioma and antilung metastasis activities in preclinical studies. ,

Here, we report the first synthesis and comprehensive evaluation of three ⁶⁸Ga-labeled CSPG4-targeted radiotracers: [⁶⁸Ga]­Ga-DOTA-LS10, [⁶⁸Ga]­Ga-DOTA-SH11, and [⁶⁸Ga]­Ga-DOTA-TH10. This study aimed to evaluate the potential of these agents for PET imaging of pancreatic carcinoma, gastric carcinoma, and hepatocellular carcinoma in subcutaneous xenograft mouse models. We systematically optimized the radiosynthetic protocols for these tracers and rigorously characterized their key physicochemical and biological properties, encompassing binding affinity to CSPG4, radiochemical purity, lipophilicity (Log P), in vitro and in vivo stability, pharmacokinetic profiles, microPET/CT imaging performance, expression of CSPG4, biodistribution characteristics, and biosafety assessments.

Results

Analysis of CSPG4 Expression and Prognostic Relevance Using Public Databases

Kaplan–Meier survival analysis revealed that high CSPG4 expression was significantly associated with worse overall survival in pancreatic cancer (HR = 1.85, 95% CI: 1.55–2.20, P = 2.8 × 10^{– 12}), gastric cancer (HR = 1.69, 95% CI: 1.40–2.04, P = 2.3 × 10^{– 8}), and liver cancer (HR = 1.67, 95% CI: 1.10–2.54, P = 0.015) (Figure a–c). Consistent with the transcriptomic data, IHC analysis demonstrated strong CSPG4 protein expression in pancreatic and gastric cancer tissues (Figure d,e), characterized by prominent membranous and cytoplasmic staining. In contrast, the level of CSPG4 expression was minimal in liver cancer samples, which served as a negative model in this study (Figure f). High-magnification views further revealed the subcellular localization of CSPG4 in positive tumors and confirmed the lack of signal in liver tissues. These findings highlight the tumor-specific expression of CSPG4 in pancreatic and gastric cancers and support its potential as a selective imaging or therapeutic target in these malignancies.

1.

CSPG4 is associated with poor prognosis and exhibits tumor-specific protein expression in gastric and pancreatic cancers. (a–c) Kaplan–Meier survival analysis of CSPG4 mRNA expression in (a) pancreatic cancer, (b) gastric cancer, and (c) liver cancer. Patients were stratified into high- and low-expression groups using the optimal cutoff. High expression of CSPG4 was significantly associated with shorter overall survival in all three types of cancer. (d–f) Representative immunohistochemical staining of CSPG4 in (d) pancreatic cancer, (e) gastric cancer, and (f) liver cancer samples using the antibody HPA002951. CSPG4 showed strong membranous and cytoplasmic localization in pancreatic and gastric tumors, while minimal staining was observed in liver cancer tissues. Lower panels show magnified views of the boxed regions to highlight subcellular localization. Scale bars = 200 μm (top panels), 20 μm (lower panels).

Radioactive Labeling of Three Probes and Molecular Docking

The precursors DOTA-LS10, DOTA-SH11, and DOTA-TH10 were purchased from Nanjing Peptide Biotech Ltd. The mass spectrometry data are attached in the Supporting Information (Figure S1).

In an acidic environment with a pH of approximately 4.2, DOTA-LS10, DOTA-SH11, and DOTA-TH10 were heated with HCl containing ⁶⁸Ga³⁺ at 100 °C for 10 min to obtain the labeled products of three probes (Figure a–c). The labeling rate was detected by radio-iTLC, and the results were close to 100% (Figure S2).

2.

Radioactive labeling and physicochemical property evaluation of three probes. (a) Radioactive labeling of [⁶⁸Ga]­Ga-DOTA-LS10; (b) radioactive labeling of [⁶⁸Ga]­Ga-DOTA-SH11; (c) radioactive labeling of [⁶⁸Ga]­Ga-DOTA-TH10.

To simulate the protein binding affinity between the probe and CSPG4, molecular docking was performed by MOE 2019 using the crystal structure of CSPG4 (RCSB ID: 7ML7). The schematic diagram of 2D binding between probe and protein is shown in Figures S3–S5. The docking scores for DOTA-LS10, DOTA-SH11, and DOTA-TH10 were −10.495, −11.473, and −11.945, respectively, which reflected the binding ability of the probe and CSPG4.

The Log P values of [⁶⁸Ga]­Ga-DOTA-LS10, [⁶⁸Ga]­Ga-DOTA-SH11, and [⁶⁸Ga]­Ga-DOTA-TH10 were −3.97 ± 0.32, −4.13 ± 0.23, and −4.20 ± 0.20, indicating that all three probes have strong hydrophilicity. The Log D_7.4 values of [⁶⁸Ga]­Ga-DOTA-LS10, [⁶⁸Ga]­Ga-DOTA-SH11, and [⁶⁸Ga]­Ga-DOTA-TH10 were −3.54 ± 0.31, −3.65 ± 0.31, and −3.61 ± 0.24.

Stability and Pharmacokinetics

The radioactive activity of [⁶⁸Ga]­Ga-DOTA-LS10, [⁶⁸Ga]­Ga-DOTA-SH11, and [⁶⁸Ga]­Ga-DOTA-TH10 in solution was approximately 6.5 GBq/μmol. Its stability was determined by radio-HPLC. Results of radio-HPLC showed that the retention time of the three probes was approximately 9 min (Figure a–c). The radiochemical purity of [⁶⁸Ga]­Ga-DOTA-TH10, [⁶⁸Ga]­Ga-DOTA-LS10 and [⁶⁸Ga]­Ga-DOTA-SH11 was over 95%. In vitro stability experiments confirmed that [⁶⁸Ga]­Ga-DOTA-LS10 and [⁶⁸Ga]­Ga-DOTA-TH10 could remain stable in PBS buffer for over 120 min, but the radiochemical purity of [⁶⁸Ga]­Ga-DOTA-SH11 showed a slight decrease. The radio-HPLC results of urine confirmed that [⁶⁸Ga]­Ga-DOTA-LS10 and [⁶⁸Ga]­Ga-DOTA-TH10 could remain stable in vivo for more than 120 min, but there was a deviation in the retention time of [⁶⁸Ga]­Ga-DOTA-SH11. These mean that [⁶⁸Ga]­Ga-DOTA-LS10 and [⁶⁸Ga]­Ga-DOTA-TH10 had better in vitro and in vivo stability than [⁶⁸Ga]­Ga-DOTA-SH11.

3.

Stability evaluation of the probes and pharmacokinetics. The radio-HPLC analyses of [⁶⁸Ga]­Ga-DOTA-LS10, [⁶⁸Ga]­Ga-DOTA-SH11, and [⁶⁸Ga]­Ga-DOTA-TH10 in vivo and in vitro are shown in parts (a–c). The radio-HPLC analysis was performed using a gradient method, using mobile phase A: water containing 0.1% trifluoroacetic acid (TFA) and mobile phase B: acetonitrile containing 0.1% TFA. The percentage concentration of phase A was reduced from 100% to 95% from 0 to 3 min, reduced again from 95% to 5% from 3 to 10 min, and maintained at this concentration for 3 min. The percentage concentration of phase A was then increased to 100% from 13 to 14 min and maintained at this concentration for 1 min for equilibration prior to the next analysis. (d) Time-activity curve of [⁶⁸Ga]­Ga-DOTA-LS10 (n = 5). (e) Time-activity curve of [⁶⁸Ga]­Ga-DOTA-SH11 (n = 5). (f) Time-activity curve of [⁶⁸Ga]­Ga-DOTA-TH10 (n = 5).

Pharmacokinetic studies showed that the circulating half-lives of [⁶⁸Ga]­Ga-DOTA-LS10, [⁶⁸Ga]­Ga-DOTA-SH11, and [⁶⁸Ga]­Ga-DOTA-TH10 were 21.64 ± 1.74 min, 23.54 ± 3.09 min, and 20.75 ± 7.77 min (Figure d–f).

A table summarizing key parameters of three probes is presented in Table S1.

Screening of Three Probes via MicroPET/CT Imaging

The tumor-targeting efficacy of the three probes was comparatively assessed via microPET/CT imaging in dual xenograft models: MKN45 and ASPC1. The images showed that [⁶⁸Ga]­Ga-DOTA-SH11 (Figure b) and [⁶⁸Ga]­Ga-DOTA-TH10 (Figure c) exhibited persistent focal radioactivity accumulation at tumor sites across all time points (30/60/90 min). However, the imaging capability of [⁶⁸Ga]­Ga-DOTA-LS10 (Figure a) was poor. Based on the stability assessment results mentioned above, [⁶⁸Ga]­Ga-DOTA-TH10 was selected for subsequent experiments.

4.

MicroPET/CT imaging and Western blot. MicroPET/CT images of [⁶⁸Ga]­Ga-DOTA-LS10 (a), [⁶⁸Ga]­Ga-DOTA-SH11 (b), and [⁶⁸Ga]­Ga-DOTA-TH10 (c) on MKN45 and ASPC1 tumor-bearing mice. (d) MicroPET/CT images of ASPC1, ASPC1 block, and HepG2 at 60 min. (e) ROI analysis of different organs from images (n = 3, ASPC1 vs ASPC1 block, ** P = 0.001881; ASPC1 vs HepG2, *** P = 0.000369). (f) Tumor-muscle ratio in different experimental groups (n = 3). (g) MicroPET/CT images of the head-to-head comparison between [⁶⁸Ga]­Ga-DOTA-TH10 and [¹⁸F]­FDG. (h) Tumor-muscle ratio of [⁶⁸Ga]­Ga-DOTA-TH10 and [¹⁸F]­FDG at 60 min (n = 3). (i) Evaluation of CSPG4 expression by Western blot.

To further evaluate the imaging effect of [⁶⁸Ga]­Ga-DOTA-TH10, head-to-head blockade imaging was performed on ASPC1 tumor-bearing mice and HepG2 tumor-bearing mice, which with relatively low expression of CSPG4 were selected for imaging (Figure d). The tumor uptake of [⁶⁸Ga]­Ga-DOTA-TH10 was 2.07 ± 0.25%ID/g at 30 min, 1.63 ± 0.15%ID/g at 60 min, and 1.43 ± 0.12%ID/g at 90 min (Figure e). In the blocking group, the ASPC1 tumor uptake of [⁶⁸Ga]­Ga-DOTA-TH10 was 0.88 ± 0.09%ID/g at 60 min, while HepG2 tumor uptake was 0.60 ± 0.05%ID/g, both significantly lower than the ASPC1 experimental group (ASPC1 vs ASPC1 block, P = 0.001881; ASPC1 vs HepG2, P = 0.000369). 60 min was considered the appropriate imaging time point, during which the tumor-muscle ratio (T/M) of [⁶⁸Ga]­Ga-DOTA-TH10 was highest (Figure f). T/M was 2.90 ± 0.36 at 30 min, 3.62 ± 0.52 at 60 min, while it was 2.59 ± 0.47 at 60 min in the blocking group, and 2.00 ± 0.05 at 60 min in the HepG2 group.

A head-to-head comparison between [⁶⁸Ga]­Ga-DOTA-TH10 and [¹⁸F]­FDG was performed on ASPC1 tumor-bearing mice; the microPET/CT images are shown in Figure g. Although the T/M of [¹⁸F]­FDG was 7.60 ± 0.27 (Figure h), higher than that of TH10 (4.83 ± 0.88), its imaging effect was more significantly affected by healthy organs.

Western blot experiments confirmed that CSPG4 was highly expressed in ASPC1 and MKN45, while that of HepG2 was low. This was consistent with tumor uptake in microPET/CT images, which confirmed the targeting ability of [⁶⁸Ga]­Ga-DOTA-TH10. Full images of the Western blot are shown in Figure S6.

Immunohistochemistry, Biodistribution, and Biosafety

Histological sections of ASPC1, MKN45, and HepG2 tumors subjected to immunohistochemical staining for CSPG4 are presented in Figure a–c. Compared with HepG2, the CSPG4 expression in ASPC1 and MKN45 was more pronounced. Consistent with the tumor uptake in microPET/CT imaging. Confirmed the high targeting ability of [⁶⁸Ga]­Ga-DOTA-TH10.

5.

Immunohistochemistry, biodistribution, and biosafety studies. (a–c) Immunohistochemistry of ASPC1, MKN45, and HepG2. (d) Biodistribution of ASPC1 tumor-bearing mice (n = 3, 60 min vs 60 min block, * P = 0.019379). (e) Blood biochemistry of normal and [⁶⁸Ga]­Ga-DOTA-TH10 injected mice. (f) RBC, WBC, and PLT of normal and [⁶⁸Ga]­Ga-DOTA-TH10 injected mice. (g) Representative images for H&E staining of heart, liver, spleen, lungs, and kidneys in normal and [⁶⁸Ga]­Ga-DOTA-TH10 injected mice.

The results of biodistribution are shown in Figures d and S7. The results showed that the ASPC1 tumor uptake was 1.74 ± 0.12 at 30 min, 1.38 ± 0.16 at 60 min, and 0.85 ± 0.10 at 90 min, with a significant decrease of 0.9 ± 0.15 (34.8% reduction) in the 60 min blocking group compared to the nonblocking group (60 min vs 60 min block, P = 0.019). Except for the kidneys, the uptake of [⁶⁸Ga]­Ga-DOTA-TH10 by healthy organs was generally lower than that of tumors. Detailed data are shown in Tables S2 and S3. Results of the dosimetry program for nuclear medicine diagnostics are shown in Table S4.

Hematological and histopathological assessments demonstrated favorable biosafety profiles of [⁶⁸Ga]­Ga-DOTA-TH10. Serum biochemistry indicators, including hepatic/renal function markers and metabolic panels, showed comparable levels across both groups, indicating the absence of systemic toxicity (Figure e). Concurrently, routine complete blood count (CBC) analysis revealed no statistically significant alterations in erythrocyte (RBC), leukocyte (WBC), or thrombocyte (PLT) parameters between the [⁶⁸Ga]­Ga-DOTA-TH10 cohort and control animals (Figure f). Histological evaluation via hematoxylin and eosin (H&E) staining further corroborated these findings: cardiac, hepatic, splenic, pulmonary, and renal architectures in radiotracer-administered mice exhibited intact cellular morphology and tissue integrity, with no evidence of inflammation, necrosis, or degenerative changes relative to controls (Figure g). Collectively, these data validate the biocompatibility of [⁶⁸Ga]­Ga-DOTA-TH10 for in vivo applications.

Discussion

The identification of novel theranostic biomarkers for pancreatic and gastric carcinomas has become a critical research priority driven by rapid advancements in radiopharmaceuticals and molecular-targeted therapies. CSPG4, a type I transmembrane protein, is expressed at significantly higher levels in these malignancies than in adjacent normal tissues, establishing it as a promising candidate for theranostic applications across multiple carcinoma types. This study provides the first validation that CSPG4 can serve as a novel and viable molecular target for nuclear-based molecular imaging in pancreatic and gastric cancers.

The CSPG4-targeted radiotracer [⁶⁸Ga]­Ga-DOTA-TH10 exhibited favorable binding affinity, robust in vivo stability, optimal pharmacokinetic profiles, favorable biosafety, and superior micro-PET/CT imaging performance. While [¹⁸F]­FDG remains the clinical gold standard for imaging glucose metabolism, its inability to dynamically reflect biomarker expression profiles limits its utility in optimizing precision molecular-targeted therapies. Beyond its role in cancer stem cell (CSC) maintenance, CSPG4 plays a critical role in regulating tumor cell proliferation, survival, and microenvironmental remodeling. Developing novel radiopharmaceutical probes that target CSPG4 represents a paradigm-shifting strategy for precision cancer diagnosis and targeted therapy.

Notably, our data show that [⁶⁸Ga]­Ga-DOTA-TH10 exhibits robust PET imaging efficacy in pancreatic (ASPC1) and gastric (MKN45) carcinomas, particularly during early tumor development. Furthermore, the superior specificity of these radiolabeled peptides for CSPG4 over normal tissues minimizes off-target signals, thereby reducing false-positive rates observed with conventional agents such as FDG and FAPI. This selective binding also enhances diagnostic accuracy, effectively distinguishing tumors from inflammatory or fibrotic lesionsa well-documented limitation of current radiotracers.

These results are consistent with those of previous preclinical studies demonstrating the therapeutic potential of CSPG4-targeted peptides. For example, LS10 and TH10 peptides have shown antiglioma and antilung metastasis activity, reinforcing CSPG4 as a therapeutic target for peptide-based radiotherapy. The incorporation of gallium-68, a radionuclide commonly used in PET imaging, adds a dimension of real-time tumor visualization, enabling clinicians to track tumor progression and evaluate therapeutic responses noninvasively.

Despite these promising findings, several limitations must be acknowledged to guide future development. Although the absolute tumor uptake of [⁶⁸Ga]­Ga-DOTA-TH10 was anticipated (1.63 ± 0.15%ID/g at 60 min, with FAPI-02 exhibiting tumor SUV < 1), it remains insufficient for radionuclide therapy. To address this, we plan to optimize probe design through strategies such as albumin binding moiety incorporation, covalent conjugation with cell-penetrating peptides, or the introduction of hydrophobic groups to enhance tumor retention and uptake. Second, while this study focused on three primary tumor models (pancreatic, gastric, and hepatocellular carcinoma), the diagnostic efficacy of [⁶⁸Ga]­Ga-DOTA-TH10 should be validated in additional models, including glioma, lung cancer, and breast cancer. Finally, we plan to expand radionuclide options to include ¹⁸F (for increasing production) and ¹⁷⁷Lu (for theranostic applications), allowing for the comprehensive evaluation of both diagnostic performance and therapeutic efficacy.

Collectively, these findings underscore that [⁶⁸Ga]­Ga-DOTA-TH10 is a promising nuclear medicine radiotracer with substantial translational potential for clinical application.

Conclusions

The study provides strong evidence that ⁶⁸Ga-labeled CSPG4-targeting peptides, especially [⁶⁸Ga]­Ga-DOTA-TH10, have significant potential as diagnostic and therapeutic agents for cancers that overexpress CSPG4. By improving the specificity and sensitivity of tumor detection, these probes could serve as valuable tools for early diagnosis, prognosis prediction, and personalized treatment strategies. Further clinical studies are required to validate the effectiveness of these probes in human cancer patients and to explore their potential in combination with radiotherapy or immunotherapy, aiming for a more comprehensive cancer treatment approach.

Experimental Section

Device and Reagents

The ⁶⁸Ge/⁶⁸Ga radionuclide generator (Eckert & Ziegler, Berlin, Germany) served as the gallium-68 source. Radiochemical purity was evaluated via dual analytical platforms. High-performance liquid chromatography (HPLC): Agilent 1200 system (Agilent Technologies, Santa Clara, USA) equipped with a Raytest GABI flow-through γ-detector. Radioactive thin-layer chromatography (TLC): Mini-GITA radio-TLC scanner (Raytest GmbH, Straubenhardt, Germany). In vivo imaging was conducted on an Inveon multimodality PET/CT scanner (Siemens Medical Solutions, Malvern, USA). Radioactivity quantification utilized a γ-counter (SN-6110, Shanghai Rihuan Photoelectric Instrument Co.). Cell lines, sourced from the Shanghai Institute of Biochemistry and Cell Biology (Chinese Academy of Sciences), were cultured as monolayers in Dulbecco’s Modified Eagle Medium (DMEM; Gibco, Carlsbad, USA) supplemented with 10% (v/v) fetal bovine serum (FBS; Gibco) and 1% penicillin-streptomycin (Invitrogen). Incubation proceeded at 37 °C under a humidified atmosphere of 5% CO₂/95% air.

Tumor-Bearing Mice

The animal experimental design was approved by the Ethics Committee of the Fudan University Shanghai Cancer Center (Ethics Approval Number: FUSCC-IACUC-S2022-0611).

ASPC1, MKN45, and HepG2 tumor-bearing mice selected T/B/NK triple immunodeficient female mice. All of the mice were purchased from Gempharmatech (Jiangsu, China).

Synthesis of Probes

The preparation of DOTA-LS10, DOTA-SH11, and DOTA-TH10 employs solid-phase synthesis on Rink Amide MBHA Resin, where sequential coupling of amino acid building blocks was automated via machine-assisted protocols to achieve high-purity peptide conjugation. The amino acid sequence of DOTA-LS10, DOTA-SH11, and DOTA-TH10 were [Leu­(DOTA)-Thr-Leu-Arg-Trp-Val-Gly-Leu-Met-Ser], [Thr-Arg-His-Val-Asn-Ile-Leu-Leu-Phe-Met-Lys­(DOTA)], and [Thr­(DOTA)-Ala-Ala-Ser-Gly-Val-Arg-Ser-Met-His]. All precursors used in this study had a purity of >95% and were obtained from Nanjing Peptide Biotech Ltd. (Nanjing, China).

Analysis of CSPG4 Expression and Prognostic Relevance Using Public Databases

To comprehensively evaluate the expression profile and prognostic significance of CSPG4 across multiple solid tumors, we utilized two publicly available databases: Kaplan–Meier Plotter (https://kmplot.com/analysis) and The Human Protein Atlas (HPA, https://www.proteinatlas.org). Kaplan–Meier Plotter was used to assess the association between CSPG4 mRNA expression and overall survival (OS) in patients with pancreatic ductal adenocarcinoma, gastric cancer, and hepatocellular carcinoma. Patients were stratified into high- and low-expression groups using the optimal cutoff (autoselected), and survival differences were evaluated by the log-rank test. Hazard ratios (HRs) and 95% confidence intervals (CIs) were automatically computed. To determine whether the transcriptional differences were reflected at the protein level, we examined immunohistochemistry (IHC) data for CSPG4 of the same tumor types using the HPA002951 antibody in the HPA database. Representative tumor samples were evaluated based on staining intensity, subcellular localization, and visual accessibility to assess the potential of CSPG4 as a molecular imaging or therapeutic target.

Labeling of the Radiopharmaceuticals

Gallium-68 (⁶⁸Ga³⁺) was eluted from a ⁶⁸Ge/⁶⁸Ga radionuclide generator (Eckert & Ziegler, Berlin, Germany) using 5.0 mL hydrochloric acid solution (0.1 M, pH 1.0), yielding an activity of 555 MBq. About 1 mL eluent was taken and buffered to pH 4.2 by adding 160 μL sodium acetate solution (1.0 M, pH 9.3). About 20 μL of DOTA-TH10 ligand solution (1 mg/mL in ultrapure water) was combined with the buffered ⁶⁸Ga solution. The mixture underwent thermal incubation at 100 °C for 10 min to facilitate chelation between Ga³⁺ and DOTA macrocycles, forming the [⁶⁸Ga]­Ga-DOTA-TH10 complex. The labeling methods for DOTA-SH11 and DOTA-LS10 were the same as DOTA-TH10. Radio-iTLC and radio-HPLC analyses were performed to determine the labeling rate of probes.

The radio-HPLC analysis was performed using a gradient method, using mobile phase A: water containing 0.1% trifluoroacetic acid (TFA) and mobile phase B: acetonitrile containing 0.1% TFA. The percentage concentration of phase A was reduced from 100% to 95% from 0 to 3 min, reduced again from 95% to 5% from 3 to 10 min, and maintained at this concentration for 3 min. The percentage concentration of phase A was then increased to 100% from 13 to 14 min and maintained at this concentration for 1 min for equilibration prior to the next analysis. The stability of [⁶⁸Ga]­Ga-DOTA-SH11 and [⁶⁸Ga]­Ga-DOTA-LS10 in vitro and in vivo uses the same method as that for [⁶⁸Ga]­Ga-DOTA-TH10.

Molecular Docking

Molecular docking simulations were executed in Molecular Operating Environment (MOE) 2019 to quantify the binding affinity between the macrocyclic chelator DOTA-TH10 and CSPG4. The crystal structure of CSPG4 (PDB ID: 7ML7, 3.17 Å) was sourced from the RCSB Protein Data Bank. Structural preprocessing involved the removal of all crystallographic water molecules to eliminate potential interference with the ligand–receptor interactions. Hydrogen atoms were subsequently added, which optimizes protonation states at physiological pH (7.4) based on electrostatic potential calculations. Both DOTA-TH10 and CSPG4 then underwent energy minimization. The site of activity was created by the MOE site finder. Ligand DOTA-TH10 was docked into the active sites by using MOE-Dock. The placement of ligands into the receptor binding site was achieved using the Triangle Matcher algorithm, which employs geometric hashing and pharmacophore-based feature matching to generate initial ligand conformations that optimally align with key receptor interactions. Refinement was conducted under the Rigid Receptor strategy, where the receptor structure was kept static while the ligand’s atomic coordinates were optimized via energy minimization to refine binding poses and reduce steric clashes. During docking, 10 poses are formed, and then 5 more promising ones are selected for subsequent screening.

The docking methods for DOTA-SH11 and DOTA-LS10 were the same as those for DOTA-TH10.

Determination of the Partition Coefficient

The partition coefficient of [⁶⁸Ga]­Ga-DOTA-TH10 was quantified using the shake-flask method. Briefly, 74 kBq of the radiolabeled peptide was introduced into a biphasic solvent system (2 mL total volume; 1:1 v/v n-octanol/ultrapure water, n = 6 replicates). The mixture was vortexed for 5 min to ensure an equilibrium distribution between the two phases. The aqueous and organic layers were isolated by centrifugation at about 12,000 rpm for 10 min. The radioactivity of each sampling layer (100 μL) was measured by using a γ-counter. The partition coefficient was computed as Log P = lg ((counts in the octanol layer)/(counts in the water layer)). The method for determining the partition coefficient of [⁶⁸Ga]­Ga-DOTA-SH11 and [⁶⁸Ga]­Ga-DOTA-SL10 was the same as that of [⁶⁸Ga]­Ga-DOTA-TH10.

The determination method for Log D _7.4 is analogous to that of Log P, with the key difference being the substitution of PBS (phosphate-buffered saline) for ultrapure water in the assay.

Stability In Vitro and In Vivo

To assess the stability in vitro, about 3 MBq of [⁶⁸Ga]­Ga-DOTA-TH10 was incubated in 200 μL phosphate-buffered saline (PBS, pH 7.4) at 37 °C for 120 min. , Samples were analyzed by radio-HPLC. In addition, to evaluate the stability in vivo, Female Balb/c mice received intravenous tail vein injections of 5.5 MBq of [⁶⁸Ga]­Ga-DOTA-TH10 in 200 μL of sterile saline. Urine samples were collected at 120 min postinjection. After dilution with PBS (1:3 v/v), samples were subjected to radio-HPLC conditions identical to those in in vivo studies.

Pharmacokinetics

Pharmacokinetic profiling of [⁶⁸Ga]­Ga-DOTA-TH10 was conducted in healthy female Balb/c mice (n = 5) via serial tail vein blood sampling at 1, 3, 5, 10, 15, 30, 60, 90, and 120 min postintravenous injection of 5.5 MBq radiotracer. Blood aliquots (50 μL) were collected using capillary tubes, transferred to preweighed microtubes, and measured for wet mass. Radioactivity was quantified with a γ-counter, with decay correction to injection time. Then, the percentage of the injected dose per gram of tissue was calculated. Pharmacokinetic analyses of [⁶⁸Ga]­Ga-DOTA-SH11 and [⁶⁸Ga]­Ga-DOTA-LS10 use the same method as that for [⁶⁸Ga]­Ga-DOTA-TH10.

The curve graph of pharmacokinetic results was drawn using GraphPad prism 8 software, and the values were calculated using DAS2.0 software.

MicroPET/CT Imaging

MKN45, ASPC1, and HepG2 tumor models were collected by Gempharmatech (Jiangsu, China). MKN45 and ASPC1 tumor-bearing mice (n = 3 per group) received intravenous injections of 5.5 MBq [⁶⁸Ga]­Ga-DOTA-TH10 in 100 μL saline. At 15 min postinjection, anesthesia was induced with 4% isoflurane/97% oxygen and maintained at 2% during imaging. Sequential microPET/CT scans were performed at 30, 60, and 90 min using a 20 min static acquisition per time point. MKN45 and ASPC1 tumor-bearing mice’s microPET/CT images of [⁶⁸Ga]­Ga-DOTA-SH11 and [⁶⁸Ga]­Ga-DOTA-LS10 were obtained by the same method as [⁶⁸Ga]­Ga-DOTA-TH10. The probe [⁶⁸Ga]­Ga-DOTA-TH10, demonstrating optimal imaging performance, was selected for subsequent blockade imaging. For competitive inhibition, tumor-bearing mice were coinjected with 5.5 MBq [⁶⁸Ga]­Ga-DOTA-TH10 and 2 mg of unmodified TH10 peptide. MicroPET/CT was acquired at 60 min postinjection using identical scanning parameters. Head-to-head comparison with [¹⁸F]­FDG: ASPC1 tumor-bearing mice (n = 3) underwent dual-tracer imaging: first with [⁶⁸Ga]­Ga-DOTA-TH10 (5.5 MBq), followed by [¹⁸F]­FDG (6 MBq) after a 20 h washout period. Prior to [¹⁸F]­FDG injection, mice were fasted for 8 h to minimize myocardial glucose uptake. Both scans were acquired at 60 min postinjection, with PET data corrected for decay, scatter, and attenuation.

The CT scans were performed first (Voltage: 80 kV, Current: 500 μA, Exposure 300 ms, Projections: 120, effective pixel size:112.85 μm). Then, 10 min PET scans were performed, and an OSEM3D/SP-MAP algorithm (2 OSEM iterations, 18 MAP iterations, Target resolution 1.5 mm, and matrix size: 128 × 128) without filtering and smoothing was used to reconstruct the PET images. The PET images were converted into SUV units by normalizing the activity concentration to the dosage of the injected drug and the mouse body weight. The animal-PET and animal-CT images were generated separately and then fused using an Inveon Research Workplace (Siemens Preclinical Solution, Knoxville, TN). The region of interest (ROI) was manually drawn covering the whole tumor on the fused images for further analysis. The highest uptake point of the entire tumor was included in the ROI, and no necrosis area was allowed.

Western Blot

ASPC1, MKN45, and HepG2 cell suspensions were pre-added to a 6-well plate (10⁶ cells per well) and incubated overnight for cell adhesion. The culture medium was aspirated, and each well was washed thrice with PBS. The cell protein concentration was measured using RIPA. Electrophoresis and membrane transfer were initiated after measuring protein concentration using the BCA assay. The nitrocellulose membrane was placed in NG2 antibody (1:2000) and shaken overnight at 4 °C. Afterward, the cells were incubated with IgG (1:5000) for 1 h. Photographs were taken after incubation with the antibodies. Antibodies were from huabio: ET1703–16 and epizyme: LF101.

Immunohistochemistry Staining

Tumor tissues obtained from the biodistribution study were formalin-fixed, paraffin-embedded, and sectioned at a 4 μm thickness. Sections were deparaffinized in xylene, rehydrated through graded ethanol, and subjected to heat-mediated antigen retrieval in Tris-EDTA buffer (pH 9.0) for 20 min. Endogenous peroxidase activity was quenched with 3% hydrogen peroxide for 20 min at room temperature. After being rinsed with PBS, sections were blocked with 1% bovine serum albumin (BSA) for 30 min at room temperature and then incubated with anti-CSPG4 antibody (ET1703-16, 1:200 dilution) overnight at 4 °C. After washing, slides were incubated with an HRP-conjugated secondary antibody for 1 h at room temperature. Immunoreactivity was visualized using DAB as the chromogen, followed by hematoxylin counterstaining. Finally, sections were dehydrated, cleared in xylene, and mounted with a DPX mounting medium.

Biodistribution

Female ASPC1 and MKN45 tumor-bearing mice (n = 3 per time point) were intravenously injected with 5.5 MBq of [⁶⁸Ga]­Ga-DOTA-TH10. At 30, 60, and 90 min postinjection, blood samples were collected via eyeballs, followed by euthanasia via cervical dislocation. The heart, liver, spleen, lungs, kidneys, stomach, intestines, bladder, muscles, bones, brain, and tumors were surgically removed. Tissues were weighed using an analytical balance. Tissue radioactivity was measured by a γ-counter. Counts were decay-corrected to the injection time.

In this study, we utilized the IDAC-Dose 2.1, an internal dosimetry program for nuclear medicine diagnostics based on the International Commission on Radiological Protection’s adult reference voxel model. Based on the uptake of [⁶⁸Ga]­Ga-DOTA-TH10 by different tissues in mice at 5, 10, 15, 30, 60, and 120 min, we used software to estimate the uptake of the probe by different organs in clinical patients.

Biosafety

Biosafety was evaluated after injecting about 5.5 MBq of [⁶⁸Ga]­Ga-DOTA-TH10 into the tail veins of normal mice on the 15th day. Blood samples were collected for biochemical analyses. The heart, liver, spleen, lungs, and kidneys were stained with hematoxylin and eosin.

Statistical Analysis

The quantitative data are expressed as the mean ± the standard deviation. Statistical analysis was conducted using the Student’s t test or one- or two-way ANOVA with GraphPad Prism version 8.0. A P-value below 0.05 was deemed statistically significant. In the figures, statistical significance is denoted by asterisks: * P < 0.05, ** P < 0.01, *** P < 0.001, and **** P < 0.0001.

Glossary

Abbreviations

%ID/g

percentage injected dose per gram of tissue

¹⁸F

fluorine-18

⁶⁸Ga

gallium-68

ALP

alkaline phosphatase

ALT

alanine aminotransferase

AST

aspartate aminotransferase

BCA

bicinchoninic acid

BSA

bovine serum albumin

BUN

blood urea nitrogen

CBC

complete blood count

CIs

confidence intervals

CREA

creatinine

CS

chondroitin sulfate

CSC

cancer stem cell

CSPG4

chondroitin sulfate proteoglycan 4

DMEM

Dulbecco’s Modified Eagle Medium

DOTA

1,4,7,10-tetraazacyclododecane-1,4,7,10-tetraacetic acid

FAPI

fibroblast activation protein inhibitor

FBS

fetal bovine serum

FDG

fluorodeoxyglucose

GAG

glycosaminoglycan

GBM

glioblastoma multiforme

H&E

hematoxylin and eosin

HPLC

high-performance liquid chromatography

HRs

hazard ratios

IHC

immunohistochemistry

iTLC

instant thin-layer chromatography

Log D _7.4

distribution coefficient at pH 7.4

Log P

partition coefficient (octanol/water)

LS10

LTLRWVGLMS

MOE

molecular operating environment

OS

overall survival

PBS

phosphate-buffered saline

PET/CT

positron emission tomography/computed tomography

PLT

platelets

RBC

red blood cells

RIPA

radioimmunoprecipitation assay

ROI

region of interest

rpm

revolutions per minute

SUV

standardized uptake value

T/M

tumor-to-muscle ratio

TFA

trifluoroacetic acid

TH10

TAASGVRSMH

TME

tumor microenvironment

TNBC

triple-negative breast cancer

TSCs

tumor stem cells

WBC

white blood cells

The original data are available upon request.

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acs.jmedchem.5c02187.

• Molecular formula strings (CSV)

• Mass spectrometry report; radio-iTLC analyses; active site and 2D binding; full images of western blot; biodistribution; key parameters of three probes; detailed data on biodistribution; and analysis of radiation dose estimation (PDF)

Project design: Z.W., X.P., and J.L. Experiment construction and implementation: Z.W., X.P., Y.L., L.B., and F.Z. Data acquisition and processing: Z.W., X.P. Writing the manuscript: Z.W., X.P., and J.L. Project supervision and coordination: J.L. All authors have read and approved the final version of the manuscript. Z.W. and X.P. contributed equally to this work.

This work was supported by the Shanghai Pujiang Program (22PJ1401700), National Isotope Engineering Technology Research Center Funding (GJTWSGCZX-202306), the construction project of Shanghai Key Laboratory of Molecular Imaging (18DZ2260400), the Open Research Fund of the State Key Laboratory of Complex, Severe, and Rare Diseases (2025-O-PY-005), and the Open Research Fund of the State Key Laboratory of Oral Diseases (SKLOD2025OF03).

The animal experimental design was approved by the Ethics Committee of Fudan University Shanghai Cancer Center (Ethics Approval Number: FUSCC-IACUC-S2022-0611).

The authors declare no competing financial interest.