Novel positron emission tomography imaging targeting cell surface glycans for pancreatic cancer: ¹⁸F‐labeled rBC2LCN lectin

Abstract

Advancement in early detection modalities will greatly improve the overall prognoses of pancreatic ductal adenocarcinoma (PDAC). For this purpose, we report a novel class of tumor‐specific probes for positron emission tomography (PET) based on targeting cell surface glycans. The PDAC‐targeting ability of rBC2LCN lectin, combined with fluorine‐18 (¹⁸F) ([¹⁸F]FB‐rBC2LCN), resulted in reproducible, high‐contrast PET imaging of tumors in a PDAC xenograft mouse model. [¹⁸F]N‐succinimidyl‐4‐fluorobenzoate ([¹⁸F]SFB) was conjugated to rBC2LCN, and [¹⁸F]FB‐rBC2LCN was successfully prepared with a radiochemical purity >95%. Cell binding and uptake revealed that [¹⁸F]FB‐rBC2LCN binds to H‐type‐3‐positive Capan‐1 pancreatic cancer cells. As early as 60 min after [¹⁸F]FB‐rBC2LCN (0.34 ± 0.15 MBq) injection into the tail vein of nude mice subcutaneously bearing Capan‐1 tumors, tumor uptake was high (6.6 ± 1.8 %ID/g), and the uptake increased over time (8.8 ± 1.9 %ID/g and 11 ± 3.2 %ID/g at 150 and 240 min after injection, respectively). Tumor‐to‐muscle ratios increased over time, up to 19 ± 1.8 at 360 min. High‐contrast PET imaging of tumors relative to background muscle was also achieved as early as 60 min after injection of [¹⁸F]FB‐rBC2LCN (0.66 ± 0.12 MBq) and continued to increase up to 240 min. Our ¹⁸F‐labeled rBC2LCN lectin warrants further clinical development to improve the accuracy and sensitivity of early‐stage pancreatic cancer detection.

In this study, we generated an engineered, ¹⁸F‐labeled rBC2LCN lectin, [¹⁸F]FB‐rBC2LCN, to target cell surface glycans differentially upregulated in pancreatic cancer. In our pancreatic cancer cell xenograft model, [¹⁸F]FB‐rBC2LCN reproducibly yielded high‐contrast positron emission tomography (PET) imaging of tumors. Our ¹⁸F‐labeled rBC2LCN lectin warrants further clinical development to improve the accuracy and sensitivity of early‐stage detection of pancreatic cancer.

Abbreviations

CT

computed tomography

EUS

endoscopic ultrasonography

FDG

18‐fluorodeoxyglucose

MRCP

magnetic resonance cholangiopancreatography

PDAC

pancreatic ductal adenocarcinoma

PET

positron emission tomography

ROI

region of interest

SFB

N‐succinimidyl‐4‐fluorobenzoate

SUV

standard uptake value

¹⁸F

fluorine‐18

1. INTRODUCTION

Pancreatic cancer, especially pancreatic ductal adenocarcinoma (PDAC), remains a lethal malignancy. Asymptomatic pathogenesis often leads to discovery at advanced disease stages (where resection is not useful) and current therapies are eventually overcome by resistance, resulting in a 5‐year survival rate of only around 10%. ¹ While patients diagnosed at early cancer staging have significantly higher survival rates (35%–45%), only 10%–15% of patients have localized disease at the time of diagnosis. ² Advancements in early detection modalities could thus greatly improve overall prognoses, but such screening technologies must rapidly and accurately diagnose PDAC.

Current modalities for detection of pancreatic cancer rely mainly on endoscopic ultrasonography (EUS), magnetic resonance cholangiopancreatography (MRCP), and computed tomography (CT). ³ , ⁴ Although routinely employed, the discriminatory power of these methods relies on morphological changes associated with cancer which, logically, increase with disease progression, rendering them less useful for early detection. ⁵ More recently, molecular imaging techniques (e.g., positron emission tomography [PET]) that provide functional and molecular data to supplement earlier decision‐making are gaining wider adoption. ⁶ , ⁷ In fact, 18‐fluorodeoxyglucose (FDG) PET/CT has often been reported to play an important role in the initial management of pancreatic cancer. ⁸ , ⁹ However, the utility of FDG for early‐stage pancreatic cancer detection is not definitive because its reliance on altered glucose metabolism is not exclusive to cancer cells and accumulation may also occur in other diseased cells with altered metabolism (e.g. pancreatitis), giving rise to false positives. ¹⁰ , ¹¹ , ¹² , ¹³ , ¹⁴ To overcome these limitations, PET radiotracers conjugated to malignancy‐specific targeting agents, such as antibodies targeting cancer‐associated/specific antigens, have been developed to improve specificity toward cancer cells. ¹⁵ , ¹⁶ , ¹⁷ , ¹⁸ These engineered antibodies, while highly specific, depend on the surface presence of their unique epitope. Moreover, foreign antibodies may cause immunogenic responses or unwanted bioaccumulation due to high molecular weight. ¹⁹ Nevertheless, this concept of targeting cancer‐specific cell surface molecules opens up possibilities for developing novel imaging probes. ²⁰

One underexplored cancer cell surface‐specific target is the glycocalyx. ²¹ Cancer cells with altered cellular metabolism have been reported to display altered glycosylation patterns in glycoconjugates, such as increased sialylation and fucosylation, truncation of O‐glycans, and branching of N‐ and O‐linked glycans in cell surface‐resident glycans. ²² , ²³ , ²⁴ We have demonstrated specific upregulation of fucosylated glycans (H‐type‐3 motifs) in pancreatic cancers, ²⁵ and a recombinant lectin called rBC2LCN ²⁶ can specifically bind to these moieties. ²⁷ Importantly, rBC2LCN does not induce hemagglutination (induced when exogenous lectins are introduced into the blood), ²⁸ allowing it to be harmlessly administered intravenously. ²⁷ Exploiting these unique properties, we previously employed rBC2LCN to deliver cytocidal drugs to PDAC cells. ²⁷ , ²⁹ By combining rBC2LCN specificity toward PDAC cells in concert with conjugated payloads, such as exotoxins ²⁷ and/or theragnostic dyes, ²⁹ we now seek to explore the potential application of an rBC2LCN–fluorine‐18 (¹⁸F) conjugate as a PET probe. The PET imaging technique has advantages in sensitivity and high tissue penetration depth, which could overcome the limitations of shallow fluorescent dye tissue penetration in molecular imaging. ³⁰ Through combining the advantages of rBC2LCN specificity and ¹⁸F tissue penetration, we developed a novel lectin‐based PET probe, ¹⁸F‐labeled rBC2LCN lectin ([¹⁸F]FB‐rBC2LCN), for detection of pancreatic cancer and evaluated its biodistribution in a xenograft mouse model.

2. MATERIALS AND METHODS

2.1. Reagents

The rBC2LCN lectin was purchased from FUJIFILM Wako Pure Chemical Corporation (Cat# 025‐18063). O‐18 enriched water (98%) was purchased from Taiyo Nippon Sanso. Butyl‐4‐tetrabutylammoniumbenzoate (TBAB) was purchased from ABX Gmbh (Cat# 4839.0005). All other reagents were of reagent grade.

2.2. Lectin histochemistry

To verify the binding of rBC2LCN lectin to human clinical pancreatic cancer tissues, lectin staining was performed. Antigen retrieval from 3‐μm slide sections of formalin‐fixed and paraffin‐embedded (FFPE) tissues was performed by autoclaving before endogenous peroxidase activity was blocked with 3% H₂O₂ in methanol. Next, horseradish peroxidase‐labeled rBC2 was applied and visualized by applying the chromogen diaminobenzidine. Images were obtained using a BZX710 microscope (Keyence Corporation). Tissue staining was judged as 0, +1, +2, +3 with regard to rBC2 reactivity as previously described, ²⁷ and the correlation with different cancer stages was assessed.

2.3. Synthesis of [ ¹⁸F]FB‐rBC2LCN

First, an aqueous solution of [¹⁸F]fluoride was produced by proton irradiation of O‐18‐enriched water using an Eclipse HR Cyclotron (Siemens) at the Advanced Imaging Center Tsukuba (Tsukuba, Japan). [¹⁸F]N‐succinimidyl‐4‐fluorobenzoate ([¹⁸F]SFB), which was prepared according to a previous report ³¹ in DMSO (100 μL), was added to rBC2LCN (16 kDa) lectin (1 mg) in 0.1 M phosphate buffer solution to a final concentration of 63 μM (pH 8.5, 1 mL). This mixture was reacted for 1 h at room temperature. The obtained product ([¹⁸F]FB‐rBC2LCN) was purified by PD‐10 gel filtration column (Cytiva, Cat# 17085101) and analyzed by HPLC (LC‐20A; Shimadzu) using a TSKgel G2000SW column (7.5 × 300 mm; Tosoh, Cat# 0005788) and 0.1 M phosphate buffer (pH 7.0) as eluent. Radioactivity was monitored using an NaI (Tl) scintillation detector (3 inch; OHYO KOKEN KOGYO) and analyzed by SIC μ7 software (GL Sciences Inc.). [¹⁸F]FB‐rBC2LCN was identified by comparing its retention time to that of rBC2LCN recorded using a UV detector (220 nm).

2.4. Cell lines and culture conditions

Capan‐1, an H‐type‐3‐positive human pancreatic cancer cell line, was obtained from American Type Culture Collection (ATCC; Cat# HTB‐79). SUIT‐2, an H‐type‐3‐negative human pancreatic cancer cell line, was obtained from the Japanese Collection of Research Bioresources (JCRB) Cell Bank (Cat# JCRB1094, CVCL_3172). Cells were routinely tested for Mycoplasma and authenticated by the JCRB Cell Bank using short tandem repeat analysis. Capan‐1 cells were cultured in Iscove's Modified Dulbecco's Medium (FUJIFILM Wako Pure Chemical Corporation, Cat# 098‐06465) supplemented with 20% fetal bovine serum (FBS; Thermo Fisher Scientific, Cat# 10270‐106) and 1% penicillin–streptomycin (FUJIFILM Wako Pure Chemical Corporation, Cat# 168‐23191). SUIT‐2 cells were cultured in Dulbecco's modified Eagle's medium (FUJIFILM Wako Pure Chemical Corporation, Cat# 044‐29765) supplemented with 10% FBS and containing 1% penicillin‐streptomycin. Cells were maintained at 37°C in a humidified incubator in 5% CO₂.

2.5. In vitro cell‐binding assays

The in vitro binding characteristics of [¹⁸F]FB‐rBC2LCN were assessed using saturation and displacement cell‐binding assays. Capan‐1 cells (2 × 10⁵) were seeded into 35 mm dishes and incubated for 48 h. For the saturation binding assay, the cells were incubated (37°C, 1 h) with increasing concentrations (100 pM–2 μM) of [¹⁸F]FB‐rBC2LCN, while for the displacement assay, the cells were incubated (37°C, 1 h) with 4.8 nM of [¹⁸F]FB‐rBC2LCN and increasing concentrations (10 pM–50 μM) of nonlabeled rBC2LCN lectin. The incubation medium was then removed, and the cells were washed with cold PBS and trypsinized before the radioactivity of the cells was measured with a γ‐counter (AccuFlex γ7000; ALOKA). The total specific binding was plotted, and nonlinear regression analysis was performed using GraphPad Prism 9.

2.6. Experimental animals

Female BALB/c nude mice (aged 6 weeks, weighing approximately 20 g) were purchased from CLEA Japan. Mice were housed and maintained in specific‐pathogen‐free conditions according to animal use guidelines of the University of Tsukuba (Ibaraki, Japan). During the procedure, mice were anesthetized with inhaled isoflurane. Tumor volumes were determined using the following formula: (width)² × (length)/2.

2.6.1. Subcutaneous cell xenograft model

Capan‐1 cells (2 × 10⁶) were injected subcutaneously into the right dorsa of nude mice. For xenografts in the dual‐dorsal tumor model, Capan‐1 and SUIT‐2 cells (2 × 10⁶) were injected into the right and left dorsa, respectively. Tumors were studied after they reached volumes of approximately 100 mm³.

2.7. Biodistribution study

Subcutaneous Capan‐1 cell xenograft mice were injected with 0.34 ± 0.15 MBq (68 ± 15 μg) of [¹⁸F]FB‐rBC2LCN into the tail vein. At 60, 150, 240, and 360 min after injection, mice were sacrificed and their organs were dissected. Then, major organs (including brain, heart, lung, liver, spleen, kidneys, pancreas, stomach, small intestine, large intestine, muscle, and bone), as well as blood and tumors, were collected and weighed. Using a γ‐counter, the remaining radioactivity in the tissues was measured against the injected dose (AccuFlex γ7000; ALOKA).

2.8. PET study

Subcutaneous cell xenograft mice were injected with 0.66 ± 0.12 MBq (98 ± 38 μg) of [¹⁸F]FB‐rBC2LCN into the tail vein. PET scans were performed using the Genisys4 small bench‐top preclinical PET scanner (Genisys4; Sofie Biosciences). ³² About 2 min after the injection, whole‐body scans (60, 150, 240, and 360 min; 10 min for each scan duration) and dynamic data acquisition (1 frame, 10 min, up to 6 frames) were initiated and recorded with a 150–650 keV energy window. No attenuation or scatter correction was applied when reconstructing the images using the two‐dimensional ordered‐subsets expectation maximum (2D‐OSEM) algorithm. For each scan, region(s) of interest (ROI) were placed on the tumor, normal tissue, and kidneys. The mean pixel values inside the multiple ROI were used to calculate the radioactivity accumulation within the tumor or organs. The data were calculated as a standard uptake value (SUV; [injected dose/body weight] × [tissue weight/tissue radioactivity]). The cross‐calibration factor was separately obtained from the acquisition of a phantom image containing a known concentration of ¹⁸F radioactivity. Biodistribution results (%ID/g) were converted to SUV for their comparison to PET data. SPSS (version 28.0.1.1; IBM Corp.) was used for the comparison.

3. RESULTS

3.1. Reactivity of rBC2LCN lectin for clinical PDAC samples

To evaluate the reactivity of rBC2LCN lectin toward different stages of PDAC, 21 human clinical PDAC specimens were analyzed by rBC2LCN lectin histochemical staining. We classified each specimen into localized (confined to primary site), regional (spread to regional lymph nodes), or distant (cancer has metastasized) stages, all based on previously published guidance ² that directly correlates cancer staging and survival rates. All PDAC samples were stained either strongly (2+) or very strongly (3+), regardless of stage and including localized cases corresponding to early stages (Figure S1), thus providing evidence that PET probes using rBC2LCN lectin should enable earlier detection of pancreatic cancer.

3.2. Radiosynthesis

Radiosynthesis of [¹⁸F]FB‐rBC2LCN is shown in Figure 1A. [¹⁸F]SFB, which was prepared according to a previous report, ³¹ was added to rBC2LCN lectin. The obtained product ([¹⁸F]FB‐rBC2LCN) was purified by PD‐10 gel filtration column. The decay‐corrected radiochemical yield of [¹⁸F]FB‐rBC2LCN was 7 ± 2% in 110 ± 8 min (n = 16) based on [¹⁸F]SFB. Temperature‐dependent improvement of radiochemical yield was observed (Figure 1B). Size‐exclusion chromatography separation profile demonstrated that [¹⁸F]FB‐rBC2LCN can be effectively separated from the small unconjugated [¹⁸F]SFB. A representative profile is shown in Figure 1C, and fractions from 11 to 20 were used for the experiment. The radiochemical purity of [¹⁸F]FB‐rBC2LCN was over 95% by analytical HPLC analysis (Figure 1D), and apparent molar activity was 95 ± 54 MBq/μmol at the end of synthesis.

FIGURE 1.

(A) Reaction scheme. (B) Temperature‐dependent improvement of radiochemical yield as revealed by Pearson's regression analysis. (C) Purification of [¹⁸F]FB‐rBC2LCN with gel filtration column. (D) HPLC profile of [¹⁸F]FB‐rBC2LCN. TSKgel G2000SW column (7.5 × 300 mm, Tosoh) and eluted with 0.1 M phosphate buffer (pH 7.0) at 0.7 mL/min. UV detection was conducted on 220 nm, and radioactivity was monitored by NaI(Tl) scintillation detector.

3.3. Binding specificity in vitro

Cell binding was determined by measuring the radioactivity in the H‐type‐3‐positive cell line Capan‐1 during incubation with increasing concentrations of [¹⁸F]FB‐rBC2LCN (Figure 2A). [¹⁸F]FB‐rBC2LCN bound to Capan‐1 cells was competitively displaced by unlabeled rBC2LCN, indicating that specificity of the conjugate is unaltered (Figure 2B). The IC₅₀ of this competitive analysis was 240.4 nM.

FIGURE 2.

(A) A representative saturation assay result using Capan‐1 cells. (B) Inhibition of [¹⁸F]FB‐rBC2LCN binding to Capan‐1 cells by unlabeled rBC2LCN (n = 3).

3.4. Biodistribution studies

We then analyzed the biodistribution of [¹⁸F]FB‐rBC2LCN as summarized in Table 1 and Figure 3. Capan‐1 tumor uptake was high as early as 60 min after injection of [¹⁸F]FB‐rBC2LCN (6.6 ± 1.8 %ID/g), and the uptake increased over time (8.8 ± 1.9 %ID/g and 11 ± 3.2 %ID/g at 150 and 240 min after injection, respectively). Tumor‐to‐muscle ratio increased over time, up to 19 ± 1.8 at 360 min after injection, indicating image contrast. The tumor uptake and tumor‐to‐background ratios were not excessive and were similar to previous reports using FDG. ³³ , ³⁴ Although kidneys (76 ± 22 %ID/g), liver (6.8 ± 1.7 %ID/g), and lungs (7.4 ± 2.6 %ID/g) showed initial accumulation up to 60 min after injection, the uptake of those organs decreased over time. The blood activity concentration of the tracer also peaked at 60 min after injection (13 ± 4.3 %ID/g) and decreased over time (8.0 ± 2.6, 4.0 ± 1.2 and 2.3 ± 0.28 at 150, 240, and 360 min after injection, respectively). Blood SUV curves revealed a monoexponential decay with a half‐life of 118 min (Figure S2). Bone uptake was very low (below 1.8 %ID/g throughout the study), indicating that defluorinating metabolism was negligible.

TABLE 1.

Biodistribution of [¹⁸F]FB‐rBC2LCN in mice bearing Capan‐1 xenografts.

Organ %ID/g	60 min	150 min	240 min	360 min
Blood	13 ± 4.3	8.0 ± 2.6	4.0 ± 1.2	2.3 ± 0.28
Heart	4.8 ± 1.1	3.2 ± 0.86	1.8 ± 0.57	1.1 ± 0.14
Lung	7.4 ± 2.6	5.2 ± 1.7	4.0 ± 1.2	3.5 ± 0.73
Liver	6.8 ± 1.7	6.3 ± 1.0	6.0 ± 1.2	6.1 ± 2.0
Kidney	76 ± 22	64 ± 30	45 ± 22	23 ± 9.3
Pancreas	3.1 ± 0.88	2.6 ± 0.53	2.2 ± 0.55	1.4 ± 0.13
Spleen	5.6 ± 1.7	4.3 ± 1.3	3.3 ± 1.0	2.6 ± 0.60
Stomach	4.1 ± 0.83	5.4 ± 1.9	5.5 ± 1.5	3.6 ± 0.26
S. intestine	5.6 ± 1.7	7.4 ± 2.6	8.1 ± 2.3	6.4 ± 1.9
L. intestine	2.3 ± 0.84	3.2 ± 0.81	3.5 ± 0.90	2.9 ± 0.28
Brain	0.37 ± 0.12	0.26 ± 0.090	0.16 ± 0.050	0.10 ± 0.010
Tumor	6.6 ± 1.8	8.8 ± 1.9	11 ± 3.2	10 ± 0.57
Muscle	0.90 ± 0.23	0.93 ± 0.26	0.70 ± 0.090	0.52 ± 0.060
Bone	1.8 ± 0.54	1.4 ± 0.41	1.2 ± 0.34	0.85 ± 0.15
Tumor/muscle	7.5 ± 2.1	10 ± 2.0	15 ± 3.5	19 ± 1.8

Note: Numbers represent mean uptake expressed in % ID/g ± SD. n = 6 (60 min); n = 9 (150 min); n = 7 (240 min); n = 3 (360 min).

FIGURE 3.

Biodistribution of [¹⁸F]FB‐rBC2LCN in athymic nude mice bearing Capan‐1 tumors at several time points following injection. n = 6 (60 min); n = 9 (150 min); n = 7 (240 min); n = 3 (360 min).

3.5. PET imaging

Representative whole‐body, decay‐corrected coronal images obtained from mice bearing Capan‐1 tumors at different time points after tracer injection are shown in Figure 4A. High‐contrast PET imaging of tumors relative to background muscle was achieved as early as 60 min after injection, and continued to increase over the 240‐min time point measured (Figure 4B), which was consistent with the biodistribution result. Strong kidney accumulation is likely indicative of blood clearance predominantly through renal excretion (Figure 4C). PET data strongly correlated with those from the biodistribution study (r = 0.97, p < 0.01 by Pearson's regression analysis). To evaluate target specificity, a PET imaging study for dual‐dorsal tumor mice bearing Capan‐1 (H‐type‐3‐positive, left flank) and SUIT‐2 (H‐type‐3‐negative, right flank) xenografts was also performed. [¹⁸F]FB‐rBC2LCN showed higher accumulation in Capan‐1 tumors over the 240‐min timepoint series (Figure 4D,E). After the 240‐min imaging timepoint, mice were sacrificed and the radioactivity of each tumor was measured. After adjusting for tumor weight, Capan‐1 tumors showed significantly higher radioactivity (Figure 4F; p < 0.01), indicating the specificity of [¹⁸F]FB‐rBC2LCN in vivo.

FIGURE 4.

(A) Positron emission tomography (PET) images obtained at 60, 150, and 240 min after injection of [¹⁸F]FB‐rBC2LCN in athymic nude mice bearing Capan‐1 xenografts. Red arrows indicate Capan‐1 tumors. Yellow arrows indicate kidneys. (B) Tumor‐to‐background ratio calculated from the standard uptake value (SUV), and (C) SUV for Capan‐1 tumors and kidneys at different time points after tracer injection. n = 4 (dynamic data up to 60, 150, and 240 min); n = 3 (360 min). (D) PET images obtained at each time point after injection of [¹⁸F]FB‐rBC2LCN in the dual‐dorsal tumor mice bearing Capan‐1 and SUIT‐2 xenografts. Coronal images are separately displayed for each tumor to show maximum tumor size. Red and yellow arrows indicate Capan‐1 and SUIT‐2 tumors, respectively. (E) SUV for Capan‐1 and SUIT‐2 tumors at different time points after tracer injection (n = 3). (F) Radioactivity of each tumor at 240 min after injection of [¹⁸F]FB‐rBC2LCN with tissue dissection method. n = 3; **p < 0.01, by Student's t‐test.

4. DISCUSSION

In this study, we generated an engineered, ¹⁸F‐labeled rBC2LCN lectin, [¹⁸F]FB‐rBC2LCN, to target cell surface glycans differentially upregulated in pancreatic cancer. In our pancreatic cancer cell xenograft model, [¹⁸F]FB‐rBC2LCN reproducibly yielded high‐contrast PET imaging of tumors (Figure 5).

FIGURE 5.

Graphic summary of positron emission tomography (PET) imaging targeting cell surface glycans for pancreatic cancer by using ¹⁸F‐labeled rBC2LCN lectin.

As cancerous malignancies often express differential (tumor‐associated) and sometimes atypical (tumor‐specific) cell surface proteins compared with noncancerous tissues, ³⁵ the most conducive approach for developing molecular imaging probes unique to cancer is to target cancer cell‐specific/‐associated surface antigens. For this purpose, antibody conjugates are excellent tools and can be readily adapted from those developed to treat other cancers and diseases. ³⁶ While several studies investigated the utilization of such radiolabeled antibodies as imaging agents, ¹⁵ , ¹⁶ , ¹⁷ , ¹⁸ there are currently no single antibodies that can target the majority of PDAC. We previously reported that rBC2LCN has strong reactivity toward the majority of human PDAC, albeit with varying degrees of avidity. ²⁷ In fact, rBC2LCN is likely to have a broader coverage than carcinoembryonic antigen 5 (CEA), one of the widely accepted gold‐standard PDAC markers, as we previously showed using a proteomics approach that CEA is one of the fucosylated glycan targets of rBC2LCN but not the only one. ³⁷ Here, we demonstrate that rBC2LCN conjugated to a PET contrasting agent ([¹⁸F]SFB) can be used to accurately detect pancreatic cancer cell xenografts, extending the usefulness of rBC2LCN lectin into the realm of noninvasive detection of pancreatic cancer. While widely used PDAC biomarkers like CEA and CA19‐9 demonstrate increased predictive value at the more advanced stage of PDAC, ³⁸ we show here that rBC2LCN is equally capable of detecting all stages of clinical PDAC samples from localized to advanced metastatic stages. Moreover, we previously reported that rBC2LCN tends to be more strongly reactive toward well‐ and moderately differentiated PDAC samples than poorly differentiated samples. ²⁷ As early‐stage PDAC cases have been reported to be more differentiated than late stages, ³⁹ our probe is likely to be more effective at detecting earlier stages of PDAC.

For visualization, the high sensitivity (picomolar detection) and superior tissue penetration of PET imaging ⁴⁰ complement the avidity and specificity of rBC2LCN. Although several positron‐emitting isotopes have been evaluated as potential radiosynthons for imaging pancreatic malignancies, ^{41 18}F may be an ideal positron emitter for labeling rBC2LCN lectin with regard to half‐life (109.7 min), as we previously demonstrated maximum tumor uptake and contrast within 2–4 h of near‐infrared dye‐labeled rBC2LCN lectin administration. ²⁹ Furthermore, ¹⁸F has the additional benefits of wide availability and high positron yield (almost 100%), making it more appropriate for routine clinical use.

To synthesize our reagent, we selected [¹⁸F]SFB because of its popularity as a prosthetic agent for ¹⁸F‐labeling of peptides and proteins in the present literature, ⁴² despite reports on other agents. ⁴³ To achieve higher yields and function of our final product, pH was discovered to be the critical variable. Similarly, Kuchar et al. ⁴⁴ reported that benzoylation preferentially proceeds on the ε‐amino group of lysine residues at elevated pH in contrast to the acylation on terminal amino groups at pH 7. Reports also indicate rapid decomposition of [¹⁸F]SFB at pH values near 9 that accelerate with heating. ⁴⁴ For these reasons, we conducted [¹⁸F]SFB labeling of rBC2LCN lectin in pH 8.5 phosphate buffer. We also observed [¹⁸F]SFB remaining in the reaction mixture, especially when the radiochemical yield was lower, indicating competition between the acylation reaction and the decomposition. Temperature‐dependent improvement of radiochemical yields was observed between 22 and 28°C (Figure 1B); however, further heating did not provide positive results. In this series of experiments, our obtained [¹⁸F]FB‐rBC2LCN product was purified by gel‐filtration column. While obtained labeled proteins are co‐eluted with their precursor proteins in this method, ⁴⁵ other methods for purifying radiolabeled proteins, such as using click reactions with non‐native biorthogonal proteins, ⁴⁶ would have added another layer of complexity and uncertainty. Finally, since labeling in the present study was carried out only on a very small laboratory scale, the apparent specific activity was as low as 95 ± 54 MBq/μmol. This value will be dramatically improved once scaled up using an automated synthesizer in a shielded system that permits higher amounts of radiolabel.

Our in vitro study showed the specificity of [¹⁸F]FB‐rBC2LCN for targeting glycans with an H‐type‐3 motif. Although glycan‐lectin interactions are generally of lower binding affinity than antibody–antigen interactions (1 × 10⁻⁶ M vs. 0.1 × 10⁻⁹), ⁴⁷ our previous reports with a lectin‐based drug carrier confirmed that high avidity (total binding across all valid surface binding points) sufficiently compensates for lower binding affinity as evidenced by tumoricidal activity. ²⁷ , ²⁹ Thus, while maintaining high‐contrast tumor delineation in our in vivo work, [¹⁸F]FB‐rBC2LCN's lower binding affinity is compensated for by the larger number of binding sites, potentially increasing whole‐tumor imaging resolution. ⁴⁸ Despite the fact that glycans should be attractive targets due to their specificity, development of probes targeting glycans for in vivo imaging has been limited. ⁴⁹ Theoretically, an antibody could be generated against any glycan epitope; however, the synthesis of even small glycan structures can be unwieldy, complicating the generation of haptens for immunization. In contrast, lectins, the naturally occurring glycan‐binding proteins used in this study, are often dismissed due to their ability to induce hemagglutination, precluding their use in vivo. Nevertheless, we previously singled out rBC2LCN as a lectin that demonstrated specificity toward PDAC and does not induce hemagglutination ²⁷ as a novel, PDAC‐specific probe. Similar to small engineered antibody fragments, such as Fv fragments (25 kDa), the half‐life value of our probe using rBC2LCN (16 kDa) was short (118 min), which is consistent with previous reports for agents of similar size. ⁵⁰ , ⁵¹ The short half‐life and high avidity of our model system are therefore suitable for clinical imaging development as they offer signal without the drawback of undesirable systemic bioaccumulation. Although we expect our [¹⁸F]FB‐rBC2LCN probe to have high tissue penetration due to its small molecular weight (compared with antibodies), ⁵² this will need biorthogonal confirmation using patient‐derived xenografts or genetically engineered PDAC mouse models to properly assess the effect of desmoplastic stroma on [¹⁸F]FB‐rBC2LCN penetration ability.

This study has some limitations. [¹⁸F]FB‐rBC2LCN showed high uptake in kidneys, even while it exhibited high accumulation in targeted tumors and achieved high‐contrast PET imaging of tumors. Although high kidney deposition included nonspecific pathways from urinary excretion, persistent uptake indicated that [¹⁸F]FB‐rBC2LCN accumulated in the kidneys as bladder uptake lessened. These results were consistent with a report that renal tubular epithelial cells were positive for rBC2LCN staining. ⁵³ However, we previously reported that rBC2LCN lectin alone showed no toxicity to any organ, including kidneys and liver. ²⁷ Therefore, any potential rBC2LCN lectin harm would depend on the chosen payload and not the molecule itself. ²⁹ An immediate concern is the fact that the kidney is in physical proximity to the pancreas and thus may mask signal from the PDAC. This may be overcome through the application of PET/CT or PET/MRI that provides 3‐dimensional spatial information, which should enable high‐fidelity signal resolution from the kidney and the pancreas. ⁵⁴ Finally, we must acknowledge the limitation of a lack of efficiency comparisons between [¹⁸F]FB‐rBC2LCN and FDG, the gold standard for clinical PET imaging. While the tumor contrast was sufficiently high compared with previous reports using FDG in pancreatic cancer xenograft models, ⁵⁵ , ⁵⁶ , ⁵⁷ , ⁵⁸ , ⁵⁹ the sensitivity of [¹⁸F]FB‐rBC2LCN versus FDG should be directly evaluated in future animal studies to compare and contrast our modality with current gold standards in the detection of early‐stage pancreatic cancer.

We have previously demonstrated that our lectin‐based carrier system is also promising as a therapeutic strategy for pancreatic cancer. ²⁷ , ²⁹ Additional versatility against prostate, ⁶⁰ gastric, ⁶¹ and colorectal cancers ⁵³ reinforces its usefulness as a multispectrum therapy that exemplifies the potential of lectin development with regard to cancer screening and treatment. Future lectin engineering could simultaneously create similar dual‐use imaging and therapy molecules that exploit favorable reactivity to diverse organ and tissue malignancies, linking side chains and effector molecules to customize imaging/tumoricidal functions. While this therapy is not intended to replace or supplant antibody‐based therapies, it may be useful to consider lectins in cancers where cancer‐specific receptors or antibodies are not available. Finally, while we did not test all imaging modalities (EUS, MRCP, etc.) against our system, the ability to modify lectins with specific radiosynthons, small molecules, and fluorescent moieties to support these screening methods remains a strong point of our system and a topic for future studies.

In conclusion, our ¹⁸F‐labeled rBC2LCN lectin offers a novel class of tumor‐specific probes for PET that are based on targeting cell surface glycans. [¹⁸F]FB‐rBC2LCN demonstrated high‐contrast PET imaging of pancreatic cancer xenografts in animals that warrants further clinical development to improve the accuracy and sensitivity of early‐stage pancreatic cancer detection.

FUNDING INFORMATION

JSPS KAKENHI, Grant/Award Numbers: JP20KA20467, JP19H05557.

CONFLICT OF INTEREST STATEMENT

KH received research funding from Advanced Imaging Center Tsukuba. TO is an editorial board member of Cancer Science. All the other authors have no conflict of interest.

ETHICS STATEMENT

Approval of the research protocol by an Institutional Reviewer Board. All procedures were approved by the Institutional Review Board of the University of Tsukuba Hospital (IRV code: H28‐90).

Informed Consent. Pancreatic cancer tissue samples were taken only from patients who gave informed, written consent.

Registry and the Registration No. of the study/trial. N/A.

Animal Studies. All animal experiments performed in the study were approved by the University of Tsukuba Animal Ethics Committee (Authorization Number: 21‐211).

Supporting information

Figure S1.

Figure S2.

Kuroda Y, Oda T, Shimomura O, et al. Novel positron emission tomography imaging targeting cell surface glycans for pancreatic cancer: ¹⁸F‐labeled rBC2LCN lectin. Cancer Sci. 2023;114:3364‐3373. doi: 10.1111/cas.15846