eFAP-51, a novel FAPI dimer with enhanced tumor uptake and retention for theranostic applications

Abstract

Background

Overexpression of fibroblast activation protein (FAP) in cancer-associated fibroblasts enables a highly selective targeting strategy using FAP inhibitors. However, radiolabeled small-molecule FAP inhibitors (FAPIs) show limited tumor retention, restricting therapeutic applications. Multimerization represents a promising strategy to enhance tumor uptake and retention through polyvalent binding. Here, we report the rational design and evaluation of dimeric FAP-targeted radiopharmaceuticals for theranostic applications.

Results

Based on the (4-quinolinoyl)-glycyl-2-cyanopyrrolidine (QCP) scaffold, a series of monomeric FAPIs were assembled into dimeric radioligands using three distinct tripod-based platforms and equipped with a DOTA-GA chelator for efficient radiolabeling with indium-111 and lutetium-177; systematic in vitro and in vivo evaluation of binding affinity, stability, cellular uptake, and tumor targeting identified eFAP-51 and eFAP-52 as lead candidates, both showing high radiochemical yield and purity, favorable stability, and sub-nanomolar affinity for FAP with strong selectivity over DPP4. Further studies demonstrated superior tumor targeting and pharmacokinetics for eFAP-51 compared with the clinical reference FAPI-46, with [¹¹¹In]In-eFAP-51 exhibiting more than 4-fold higher tumor uptake at 48 h post-injection (3.2 ± 0.5 vs. 0.71 ± 0.1%ID/g) and prolonged retention. eFAP-51 was subsequently radiolabeled with lutetium-177 to yield a stable therapeutic candidate, and dose-dependent studies revealed that intermediate molar doses (0.3–0.6 nmol) provided an optimal balance between tumor uptake and tumor-to-background contrast.

Conclusions

The FAPI dimer eFAP-51 showed favorable tumor targeting and pharmacokinetics, supporting its potential for clinical translation in theranostic oncology.

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1186/s41181-026-00481-9.

Introduction

The tumor microenvironment (TME) plays a critical role in cancer progression, metastasis, and resistance to therapy (Wang et al. 2025, Erdogan and Webb 2017). Among the stromal components, cancer-associated fibroblasts (CAFs) represent a major cellular population that promotes tumor growth through extracellular matrix remodeling, angiogenesis, and immunosuppression (Scanlan et al. 1994, Hanahan and Weinberg 2011, Kalluri 2003, Kalluri and Zeisberg 2006). A defining characteristic of CAFs is their high-level expression of fibroblast activation protein (FAP), a membrane-bound serine protease with dipeptidyl peptidase and endopeptidase activity (Scanlan et al. 1994, Sahai et al. 2020). Notably, FAP is highly expressed in over 90% of epithelial tumors, including breast, pancreatic, colorectal, and lung cancers, while its expression in normal tissues is negligible, making it an attractive and selective target for cancer imaging and therapy (Santos et al. 2009).

In recent years, significant progress has been made to design small-molecule inhibiting FAP, predominantly based on the quinoline scaffold of the progenitor compound UAMC-1110. These advancements enabled the development of a new generation of FAP-targeted radiopharmaceuticals radiolabeled with isotopes, such as gallium-68, indium-111 or lutetium-177, for positron emission tomography (PET), single-photon emission computed tomography (SPECT), and targeted radionuclide therapy (TRT), respectively (Loktev et al. 2018, Giesel et al. 2019, Kratochwil et al. 2019). Among these candidates, monomeric FAP-targeting tracers such as FAPI-04 and the clinically validated FAPI-46 have demonstrated rapid tumor uptake and high-contrast imaging profiles (Lindner et al. 2018, Giesel et al. 2021). However, rapid clearance from the tumor compromised their therapeutic efficacy (Moon et al. 2021). To overcome this limitation, several structural modifications were applied to monomeric FAPIs. It includes the optimization of the chelator moiety, variation of the linker to modulate flexibility and hydrophilicity, introduction of lipophilic side chains to enhance affinity (Wang et al. 2021, Wu et al. 2025), conversion of the nitrile group to α-ketoamide to enhance interaction with serine residues (Verhulst et al. 2026, Brzeminski et al. 2025), and incorporation of a fluorosulfate group to bind covalently to FAP via SuFEx reaction (Cui et al. 2024),as well as the introduction of albumin-binding moieties to prolong blood circulation and enhance tumor uptake (Saito et al. 2025, Saito et al. 2025).

To prolong tumor retention, molecular design strategies beyond conventional small-molecule FAP inhibitors have been explored. For example, FAPI-2286 and FAP-3940, two FAP-targeting cyclic peptides, showed promising tumor retention, and [¹⁷⁷Lu]Lu-FAP-2286 recently entered early-phase clinical trials (Banihashemian et al. 2024, Zboralski et al. 2026). Another complementary strategy is multivalency, in particular dimerization, which has emerged as a promising approach to enhance binding avidity, increase tumor uptake, and prolong retention in target tissues (Mammen et al. 1998, Ekerljung et al. 2008). By linking two pharmacophores, bivalent ligands can benefit from cooperative binding and increased local concentration at the target site (Notni et al. 2010, Barinka et al. 2008, Galbiati et al. 2022, Galbiati et al. 2025, Lappchen et al. 2024, Bilinska et al. 2025, Bilinska et al. 2026).

In this study, we report the rational design and synthesis of novel dimeric FAP-targeted radioligands from structurally diverse tripod platforms, as well as their systematic evaluation in preclinical models. The FAPI motifs included in the novel dimers were based on the (4-quinolinoyl)-glycyl-2-cyanopyrrolidine substituted at the 8-position (8-QCP scaffold), derived from eFAP-6, eFAP-9, and eFAP-12 (Fig. 1). In our previous studies, eFAP-6 (Heide et al. 2024) and eFAP-9 showed rapid tumor accumulation, while eFAP-12 demonstrated prolonged tumor retention and improved tumor-to-background ratios (unpublished results). These dimers were subsequently radiolabeled with indium-111 and lutetium-177, and comprehensively characterized in vitro and in vivo to assess their binding affinity, selectivity, stability, and tumor-targeting performance. Our findings demonstrate that these novel dimeric radioligands offer notably improved tumor uptake and retention compared to FAPI-46, making them promising candidates for theranostic applications.

Fig. 1.

Modular design of 8-QCP-based dimeric FAP radioligands from the monomeric FAPIs (eFAP-6, eFAP-9, and eFAP-12). Each dimer consists of a central tripod platform (P₁-P₇, in gray), two 8-QCP FAP-binding motifs (8-QCP, 8-QCP-AMCHA, and 8-QCP-pipa, in blue), and a chelator (TCO-DOTA-GA, BCN-DOTA-GA, or DOTA-GA anhydride, in green)

Results

Chemistry

The eFAP dimer library was designed based on the previously established 8-QCP scaffold (Heide et al. 2024, Ma et al. 2025). Among the initial monomeric eFAP analogs, eFAP-6, eFAP-9, and eFAP-12 exhibited the most favorable pharmacokinetic profiles and were therefore selected as FAP targeting building blocks for dimerization and the construction of the eFAP dimer library (Fig. 1). Three distinct tripod cores were conjugated to two FAP-targeting moieties via amide bonds. Each platform was subsequently functionalized with a DOTA-GA chelator for radiolabeling (Scheme 1–3; characterization data are provided in Fig. S1-S34). Collectively, these complementary platforms provided modular access to structurally diverse bivalent eFAP ligands with varying linker architectures and chelator conjugation strategies.

Scheme 1.

Syntheses of dimeric eFAPs based on a tetrazine core (eFAP-20 and eFAP-31). Reagents and conditions: a 3-mercaptopropanoic acid, DMF, DIPEA, rt, 30 min, 71%; b 8-QCP, DMF, HATU, DIPEA, rt, 1 h, 45%; c TCO-DOTA-GA, ACN/H₂O (1:1), 37 °C, 30 min, 34%; d 6-mercaptohexanoic acid, DMF, DIPEA, rt, 30 min, 78%; e 8-QCP, DMF, HATU, DIPEA, rt, 1 h, 51%; f TCO-DOTA-GA, ACN/H₂O (1:1), 37 °C, 30 min, 64%

The first platform utilized 1,4-dichlorotetrazine (Wang et al. 2025) as the central core. Two chlorides were first displaced by thiol-bearing linkers of different lengths, namely 3-mercaptopropionic acid and 6-mercaptohexanoic acid, affording the corresponding tetrazine intermediates 2 and 4 in 71 and 78% yield. The carboxylic acid groups of these intermediates were subsequently coupled to the primary amine of the 8-QCP FAP binding motif via amide bond formation, yielding intermediates 3 and 5 in 44 and 51% yield, respectively. Final functionalization was achieved through an inverse electron-demand Diels–Alder (IEDDA) click reaction with commercially available TCO-DOTA-GA, as described in our previous work (Murce et al. 2023), providing the final dimeric conjugates eFAP-20 and eFAP-31 in 34 and 64% yield, respectively.

The second platform, namely N₃-TATB (Kalluri and Zeisberg 2006), was chosen based on the azide-functionalized polar hinge scaffold reported by van de Langemheen et al. (Langemheen et al. 2017), which provides improved solubility and a convenient handle for a strain-promoted azide-alkyne cycloaddition (SPAAC) of the chelator. The two bromomethyl groups of 6 were first substituted with linkers (i.e., 3-mercaptopropionic acid, 6-mercaptohexanoic acid, and NH₂-PEG₂-tBu) affording the corresponding linker-functionalized intermediates 7, 9, and 11 in 35, 46, and 50% yield, respectively. For the intermediate 11, an additional deprotection step was required to generate the free carboxylic acid. Subsequently, the carboxylic acids 7, 9, and 11 were coupled to the primary amine of the 8-QCP motif via standard amide bond formation, yielding the key dimeric intermediates 8, 10, and 12 in 60, 31 and 44% yield, respectively. The remaining azide functionality enabled final conjugation with commercially available BCN-DOTA-GA through a SPAAC reaction, providing the corresponding chelator-containing dimeric precursors eFAP-34, eFAP-35, and eFAP-36 in 31, 54 and 34% yield, respectively.

The third platform employed was 2,4,6-trichloro-1,3,5-triazine (Giesel et al. 2021). The two chlorines were first substituted with NH₂-PEG₂-tBu via nucleophilic aromatic substitution. Subsequent acidic deprotection afforded the corresponding diacid intermediate 14. The remaining chlorine atom was then displaced by two different amine-terminated linkers, namely 3-azidopropan-1-amine and N-Boc-1,6-diaminohexane, yielding intermediates 15 and 17, respectively, in 54 and 48% yields. The lower yield of intermediate 17 was attributed to steric constraints during nucleophilic substitution. The free carboxylic acid groups on intermediates 15 and 17 were subsequently coupled to FAP-targeting motifs derived from eFAP-6 (8-QCP), eFAP-9 (8-QCP-AMCHA), and eFAP-12 (8-QCP-pipa) via amide bond formation to afford the corresponding triazine-based dimers 16, 18, 19 and 20 in 40–60% yield. AMCHA and pipa correspond to 4-(aminomethyl)cyclohexanecarboxylic acid and 2-(piperazin-1-yl)acetic acid, respectively. For the azide-functionalized intermediate (Wu et al. 2025), final conjugation with commercially available BCN-DOTA-GA was achieved through a SPAAC reaction, yielding eFAP-37 in 39% yield. In contrast, the diaminohexane-containing intermediates (18, 19, and 20) enabled direct coupling with commercially available DOTA-GA anhydride, providing the final chelator-functionalized dimers eFAP-50, eFAP-51, and eFAP-52 in 33, 30 and 45% yield, respectively. (Scheme 1, 2 and 3)

Scheme 2.

Syntheses of dimeric eFAP-34, eFAP-35 and eFAP-36 based on a hinge core (6). a 3-mercaptopropanoic acid, DMF, DIPEA, rt, 30 min, 35%; b 8-QCP, DMF, HATU, DIPEA, rt, 1 h, 60%; c BCN-DOTA-GA, ACN/H2O (1:1), 37 °C, 30 min, 31%; d 6-mercaptohexanoic acid, DMF, DIPEA, rt, 30 min, 46%; e 8-QCP, DMF, HATU, DIPEA, rt, 1 h, 31%; f BCN-DOTA-GA, ACN/H2O (1:1), 37 °C, 30 min, 54%; g NH2-PEG2-tBu, DMF, DIPEA, rt, 30 min; DCM, TFA, TIPS, rt, 30 min, 55% over two steps; h 8-QCP, DMF, HATU, DIPEA, rt, 1 h, 44%; i BCN-DOTA-GA, ACN/H2O (1:1), 37 °C, 30 min, 34%.

Scheme 3.

Syntheses of dimeric eFAPs (eFAP-37, eFAP-50, eFAP-51 and eFAP-52) based on a triazine core. a NH2-PEG2-tBu, DMF, DIPEA, rt, 30 min; DCM, TFA, TIPS, rt, 30 min, 72% over two step; b 3-azidopropan-1-amine, DMF, DIPEA, rt, 30 min, 54%; c 8-QCP, DMF, HATU, DIPEA, rt, 1 h, 52%; d BCN-DOTA-GA, ACN/H2O (1:1), 37 °C, 30 min, 39%; e N-Boc-1,6-hexanediamine,DMF, DIPEA, rt, overnight, 48%; f 8-QCP, DMF, HATU, DIPEA, rt, 1 h; DCM, TFA, TIPS, rt, 30 min,55% over two step; g DOTA-GA anhydride, DMF, DIPEA, rt, 30 min, 33%; h 8-QCP-AMCHA, DMF, HATU, DIPEA, rt, 1 h;DCM, TFA, TIPS, rt, 30 min, 47% over two step; i DOTA-GA anhydride, DMF, DIPEA, rt, 30 min, 30%; j 8-QCP-pipa, DMF, HATU, DIPEA, rt, 1 h; DCM, TFA, TIPS, rt, 30 min,57% over two step; k DOTA-GA anhydride, DMF, DIPEA, rt, 30 min, 45%.

In vitro screening of ¹¹¹In-labeled dimers

Radiochemistry, stability and lipophilicity

All eFAP compounds were efficiently radiolabeled with indium-111, achieving radiochemical yields (RCY) above 95% and radiochemical purities (RCP) of 67–99% (Table 1, Fig. S35–S43). Stability analyses revealed clear differences between the compounds. eFAP-20 showed pronounced instability in PBS, with only 41.9% of intact radioligand remaining after 24 h. Therefore, stability evaluation in murine serum and LogD measurement were not further pursued due to the limited stability of this compound. Similarly, eFAP-31 exhibited limited stability, with 52.3 and 43.4% intact compound in PBS and murine serum at 24 h. In contrast, eFAP-34 and eFAP-35 demonstrated high stability in PBS (91.0 and 92.4% intact at 24 h, respectively), but they both underwent substantial degradation in murine serum (45.1 and 17.8% intact at 24 h, respectively). Notably, eFAP-36 and eFAP-37 displayed improved serum stability compared to eFAP-34 and eFAP-35, retaining 78.2 and 67.8% intact radioligand at 24 h, respectively. Furthermore, incorporation of a hydrophilic PEG₂ linker in eFAP-50, eFAP-51, and eFAP-52 enhanced their overall stability profiles (67.7, 64.8, and 79.3% intact radioligand remaining in murine serum at 24 h), as previously observed for eFAP-36 and eFAP-37. Subsequently, lipophilicity was determined by the shake-flask method. LogD_7.4 values ranged from − 1.3 to −3.8, indicating varying degrees of hydrophilicity (Table 1).

Table 1.

Radiochemical proprieties of the dimeric [¹¹¹In]In-eFAPs

Compound	RCY (%)	RCP (%)	LogD7.4 (n = 3)	PBS (%)*			Murine serum (%)*
				1 h	4 h	24 h	1 h	4 h	24 h
[111In]In-eFAP-20	> 95	67.6	–	64.8	61.4	41.9	–	–	-
[111In]In-eFAP-31	> 95	77.3	−1.3 ± 0.02	55.9	55.6	52.3	53.6	53.5	43.4
[111In]In-eFAP-34	> 95	97.3	−3.2 ± 0.1	97.0	97.1	91.0	90.3	87.3	45.1
[111In]In-eFAP-35	> 95	98.7	−2.9 ± 0.1	97.1	92.4	92.4	83.7	75.2	17.8
[111In]In-eFAP-36	> 95	98.5	−3.4 ± 0.6	98.0	97.0	85.5	95.5	91.6	78.2
[111In]In-eFAP-37	> 95	95.6	−3.1 ± 0.23	94.7	94.0	91.6	95.0	88.3	67.8
[111In]In-eFAP-50	> 95	96.8	−3.7 ± 0.01	96.8	96.8	94.4	95.2	87.8	67.7
[111In]In-eFAP-51	> 95	95.9	−2.8 ± 0.1	97.7	96.1	90.4	93.5	77.0	64.8
[111In]In-eFAP-52	> 95	95.9	−3.8 ± 0.1	98.4	98.0	96.8	95.9	88.6	79.3

^* Results are expressed as the percentage of intact labeled ligand after incubation at 37 °C (n = 1)

Affinity for recombinant huFAP, muFAP, and DPP4

The affinity of the eFAP dimers (eFAP-34, 35, 36, 37, and eFAP-50, 51, 52) for human FAP (huFAP), murine FAP (muFAP), and the related protease DPP4 was determined using recombinant enzymatic assays (Table 2). Due to their insufficient stability, eFAP-20 and eFAP-31 were not advanced to this binding affinity assessment. All compounds exhibited nanomolar-range binding affinity to huFAP and muFAP, while showing a markedly lower affinity for DPP4 (in the micromolar range). It demonstrated the high selectivity of the dimeric eFAPs for FAP over DPP4. Among them, eFAP-51 showed the highest inhibitory activity for both FAP proteins (IC_{50, huFAP} = 0.3 ± 0.1 nM, IC_{50, muFAP} = 0.3 ± 0.2 nM), corresponding to a 5-fold higher affinity compared to FAPI-46 (IC_{50, huFAP} = 1.5 ± 0.5 nM, IC_{50, muFAP} = 1.4 ± 0.5 nM). Notably, eFAP-52 also showed comparable high affinity, with subnanomolar IC₅₀ values (0.4 ± 0.2 nM for huFAP and 0.5 ± 0.4 nM for muFAP).

Table 2.

IC₅₀ values of the dimeric eFAPs for huFAP, muFAP and DPP4

Compounds	Inhibition
	IC50, huFAP nM	IC50, muFAP nM	IC50, DPP4 µM
FAPI-46	1.5± 0.5	1.4 ± 0.5	20.5 ± 2.5
eFAP-34	1.6 ± 0.3	1.1 ± 0.6	10.6 ± 2.6
eFAP-35	1.3 ± 0.4	0.7 ± 0.4	13.7 ± 6.4
eFAP-36	0.6 ± 0.3	0.4 ± 0.1	14.8 ± 8.3
eFAP-37	1.6 ± 0.2	0.8 ± 0.1	12.5 ± 1.9
eFAP-50	0.9 ± 0.2	0.8 ± 0.1	7.5 ± 3.8
eFAP-51	0.3 ± 0.1	0.3 ± 0.2	18.2 ± 3.9
eFAP-52	0.4 ± 0.2	0.5 ± 0.4	24.7 ± 13.3

Overall, eFAP-51 and eFAP-52 displayed the most favorable IC₅₀ values within this series, justifying their advancement into subsequent uptake and in vivo pharmacokinetic studies. The closely related analog eFAP-50 was also retained considering that it could help elucidate linker-dependent effects on the biological properties of the triazine-based dimeric scaffold.

Radiotracer uptake in FAP-expressing cells

Cellular uptakes of [¹¹¹In]In-FAPI-46, [¹¹¹In]In-eFAP-50, [¹¹¹In]In-eFAP-51, and [¹¹¹In]In-eFAP-52 in HT-1080.huFAP cells are illustrated in Fig. 2. [¹¹¹In]In-eFAP-51 demonstrated a pronounced time-dependent increase in total cellular uptake, increasing from 15.8 ± 3.9% AD at 0.5 h to 23.7 ± 3.4% AD at 1 h, and reaching a peak of 31.9 ± 3.4% AD at 2 h. In contrast, the uptake of [¹¹¹In]In-FAPI-46 exhibited a comparatively modest increase over the same period, from 12.0 ± 1.3% AD at 0.5 h to 13.9 ± 1.7% AD at 2 h. Therefore, at the 2 h time point, the cellular accumulation of [¹¹¹In]In-eFAP-51 was 2.3-fold higher than that of [¹¹¹In]In-FAPI-46, making it the most promising candidate for further evaluation.

Fig. 2.

Uptake studies of [¹¹¹In]In-FAPI-46 and [¹¹¹In]In-eFAPs (n = 3) using HT-1080.huFAP cells at different time points. Values are expressed as % of the added dose (% AD) per 150,0000 cells

[¹¹¹In]In-eFAP-50, however, showed a modest increase in uptake compared to [¹¹¹In]In-FAPI-46 (12.8 ± 1.2% AD at 0.5 h, 15.8 ± 1.6% AD at 1 h, and 18.3 ± 1.1% AD at 2 h). Unexpectedly, [¹¹¹In]In-eFAP-52 exhibited a relatively lower cellular uptake compared to the other compounds (11.3 ± 1.2% AD at 0.5 h, 12.2 ± 1.3% AD at 1 h, and 11.9 ± 1.2% AD at 2 h), which remained fairly constant over time. Despite its lower cellular uptake, its monomeric analog, eFAP-12, exhibited comparable tumor uptake and better tumor-to-kidney ratio than [¹¹¹In]In-FAPI-46. This could turn [¹¹¹In]In-eFAP-52 into an attractive candidate for further evaluation. The uptake specificity of [¹¹¹In]In-FAPI-46 and [¹¹¹In]In-eFAPs was confirmed by co-incubation with an excess of FAPI-46, which resulted in negligible uptake for all radiotracers (Table S1).

In vivo evaluation of ¹¹¹In-labeled candidates

In vivo pharmacokinetic studies by SPECT/CT imaging were performed in xenografted NMRI-Foxn1 nu/nu mice (n = 4 per group) bearing both FAP positive HT-1080.huFAP and FAP negative HT-1080.WT at 1, 4, 24, and 48 h after the administration of [¹¹¹In]In-FAPI-46, [¹¹¹In]In-eFAP-51, and [¹¹¹In]In-eFAP-52. HT-1080.huFAP tumors were clearly visible on SPECT images for all radioligands after injections (Fig. 3A), while HT-1080.WT tumors, serving as a negative control, were not visible, confirming the specificity of the radioligands for FAP. All radiotracers were primarily excreted through the bladder. [¹¹¹In]In-FAPI-46 showed moderate tumor uptake, which was nearly completely washed out at 24 h p.i., whereas [¹¹¹In]In-eFAP-51 and [¹¹¹In]In-eFAP-52 showed higher tumor uptake and prolonged tumor retention, while keeping minimal off-target uptake. Quantification of the imaging data confirmed that [¹¹¹In]In-eFAP-51 had nearly twice the tumor uptake of [¹¹¹In]In-FAPI-46 at 1 h p.i. (6.5 ± 0.4% ID/cc vs. 3.5 ± 1.7% ID/cc), and continued to increase at 4 h p.i. (7.0 ± 0.9% ID/cc). For [¹¹¹In]In-eFAP-52, tumor uptake at 1 h p.i. was slightly higher than [¹¹¹In]In-FAPI-46 (3.7 ± 0.4% ID/cc vs. 3.5 ± 1.7% ID/cc), and remained stable at 4 h p.i. (3.6 ± 0.2% ID/cc) (Table S2, Fig. S44A). Correspondingly, the tumor half-life of [¹¹¹In]In-eFAP-51 was notably longer than that of [¹¹¹In]InIn-FAPI-46 (33.63 h vs. 10.54 h), indicating enhanced tumor retention, while [¹¹¹In]In-eFAP-52 had a half-life of 17.50 h (Fig. S44B).

Fig. 3.

A SPECT/CT images of mice bearing HT-1080.huFAP (right shoulder) and HT-1080.WT tumors (left shoulder) at 1, 4, 24 and 48 h after injection of [¹¹¹In]In-FAPI-46, [¹¹¹In]In-eFAP-51 and [¹¹¹In]In-eFAP-52. Red arrows point toward HT-1080.huFAP tumors, and white circles highlight the HT-1080.WT tumors. B Ex vivo biodistribution of [¹¹¹In]In-FAPI-46, [¹¹¹In]In-eFAP-51 and [¹¹¹In]In-eFAP-52 in HT-1080.huFAP and HT-1080.WT tumor bearing mice at 48 h p.i.; uptake in tissues is expressed as the percentage of the injected dose per gram of tissue (% ID/g), and WT represents HT-1080.WT tumor

Tumor-to-background ratios in the liver and kidneys favored [¹¹¹In]In-FAPI-46 at 1 and 4 h p.i., but the higher tumor uptake of [¹¹¹In]In-eFAP-51 resulted in more favorable tumor-to-background ratios at the later time points (Fig. S44C). Ex vivo biodistribution at 48 h p.i. further confirmed the imaging results (Fig. 3B, Table S3), with [¹¹¹In]In-eFAP-51 showing over 4-fold higher tumor uptake in HT-1080.huFAP tumors compared to [¹¹¹In]In-FAPI-46 (3.2 ± 0.5% ID/g vs. 0.71 ± 0.1% ID/g). A low but detectable activity was observed in HT-1080.WT tumors, which may be attributed to nonspecific binding, residual blood-pool activity, or low-level murine FAP expression from stromal cells within the xenograft microenvironment(Hintz et al. 2019, Heide et al. 2024, Wang et al. 2023). Notably, [¹¹¹In]In-eFAP-51 showed low retention in non-target organs, particularly in kidneys (0.49 ± 0.10% ID/g). In comparison, [¹¹¹In]In-eFAP-52 exhibited increased accumulation in clearance organs (i.e., kidneys and liver), indicating less favorable pharmacokinetics. Overall, these results corroborate the imaging data and emphasize eFAP-51 as the lead compound, exhibiting the most advantageous combination of high tumor uptake, prolonged retention, and minimal off-target accumulation.(FIg. 3)

Molar dose optimization of [¹⁷⁷Lu]Lu-eFAP-51

In vitro stability and dose-dependent cell uptake

Following the identification of eFAP-51 as the lead candidate, the compound was subsequently radiolabeled with ¹⁷⁷Lu to enable further evaluation toward potential therapeutic application. Radiolabeling proceeded with RCY and RCP exceeding 95%. Importantly, [¹⁷⁷Lu]Lu-eFAP-51 demonstrated excellent in vitro stability, with radiochemical integrity exceeding 88% in PBS and murine serum at 24 h (Table 3, Fig. S45).

Table 3.

In vitro stability of [¹⁷⁷Lu]Lu-eFAP-51 in PBS and murine serum

Compound	RCY (%)	RCP (%)	PBS (%)			Murine serum (%)
			1 h	4 h	24 h	1 h	4 h	24 h
[177Lu]Lu-eFAP-51	> 95	> 95%	96.7	95.7	92.1	95.6	92.4	88.1

To investigate the influence of dose on cellular uptake, the uptake of [¹⁷⁷Lu]Lu-eFAP-51 was evaluated at different molar amounts (Table S4). The highest uptake was observed at 0.3 nmol (28.2 ± 1.1% AD), followed by 0.6 nmol (20.5 ± 0.7% AD) and 1.0 nmol (13.7 ± 0.8% AD), indicating a dose-dependent decrease in cellular accumulation. These findings suggest that molar dose of the radiotracer may influence FAP-mediated uptake, providing a rationale for further evaluation of molar dose effects in vivo. Blocking studies were performed by co-incubation with an excess of unlabeled FAPI-46. As expected, the presence of excess competitor resulted in a pronounced reduction in uptake across all tested doses, confirming the FAP-specific binding of the radioligand.

Dose-dependent in vivo pharmacokinetics

An evaluation of the molar dose effect of [¹⁷⁷Lu]Lu-eFAP-51 was conducted in HT-1080.huFAP and HT-1080.WT xenograft models (n = 4 per group) by administrating four different injected doses (0.1, 0.3, 0.6, and 2.4 nmol per animal). SPECT/CT imaging revealed dose-dependent pharmacokinetic profiles. Lower molar doses were associated with higher background activity, whereas higher doses showed increased renal clearance, as reflected by early kidney signal (Fig. S46). At 4 h p.i., all molar doses, except 0.1 nmol, exhibited rapid washout from non-target tissues and clear tumor visualization. Persistent tumor uptake was observed up to 48 h p.i. for all FAP-positive tumors.

Ex vivo biodistribution at 48 h p.i. (Table S5) confirmed low uptake in HT-1080.WT tumors and non-target organs. Tumor uptake was highest at the lowest injected molar dose of 0.1 nmol (2.49 ± 0.32% ID/g) and decreased with increasing dose, reaching a 60% reduction at 2.4 nmol. Statistical analysis confirmed that increasing the injected molar dose significantly reduced normal tissue uptake at 48 h p.i. (p < 0.05 for liver and kidneys, and p < 0.01 for spleen, for the 0.1 vs. 2.4 nmol comparison; One-way ANOVA with Tukey’s multiple comparisons test; Fig. S47). Liver uptake decreased from 0.49 ± 0.05% ID/g at 0.1 nmol to 0.19 ± 0.03% ID/g at 2.4 nmol, spleen from 0.58 ± 0.10% ID/g to 0.11 ± 0.03% ID/g, and kidneys from 0.16 ± 0.03% ID/g to 0.06 ± 0.01% ID/g. Notably, no significant differences were observed between the 0.3 and 0.6 nmol doses for any of these organs (p > 0.05), indicating that intermediate doses provide an optimal balance between tumor uptake and background clearance.

Based on these findings, the molar doses of 0.3 and 0.6 nmol were selected for a time-dependent biodistribution study (Fig. 4, Table S6).

Fig. 4.

Biodistribution of [¹⁷⁷Lu]Lu-eFAP-51 at two different doses in HT-1080.huFAP and HT-1080.WT tumor bearing mice at 1, 4, 24, 48 h p.i. (WT = HT-1080.WT tumor)

Both doses showed high tumor uptake as early as 1 h p.i., peaking at 4 h p.i. (12.31% ID/g for 0.3 nmol; 11.20% ID/g for 0.6 nmol) and gradually declining thereafter, while retaining > 59% of peak uptake at 24 h and > 25% at 48 h. The 0.3 nmol group exhibited slightly higher uptake in non-target organs at early time points, whereas the 0.6 nmol group showed increased blood, kidney and liver activity, consistent with the dose-dependent effect observed previously on systemic distribution. Pharmacokinetic analysis further corroborates these results. Tumor half-lives were comparable for the 0.3 and 0.6 nmol groups (40.3 vs. 40.2 h), with similar tumor AUC values (360.8 ± 47.5 vs. 332.5 ± 17.6), indicating robust cumulative tumor exposure. Clearance from non-target organs, including blood, kidney and liver, was faster at the higher injected dose, consistent with the imaging observations (Fig. 5).

Fig. 5.

A Pharmacokinetic modeling of [¹⁷⁷Lu]Lu-eFAP-51 clearance in major organs at 0.3 and 0.6 nmol doses; B Area under the curve up to 48 h p.i. (AUC_{0 → 48 h}) derived from the biodistribution data

Discussion

Recent advancements in quinoline-based FAPI variants have demonstrated substantial potential in oncological imaging, whereas FAP-targeted radionuclide therapy has gained considerable attention due to promising results in both preclinical studies and clinical trials (Lindner et al. 2018, Prive et al. 2023, Jansen et al. 2013). Despite these developments, existing FAPIs are limited by their insufficient duration of action on tumors, which restricts their effectiveness in therapeutic applications (Mosessian et al. 2023). A promising strategy to overcome these limitations is the multimerization of targeting vectors, which enhances valency and modifies the pharmacokinetic properties, leading to more effective tumor targeting (Pang et al. 2023, Galbiati et al. 2022, Galbiati et al. 2024). The dimerization approach has been shown to improve tumor uptake and retention in previous studies (Galbiati et al. 2022, Tan et al. 2024). For example, DOTAGA.(SA.FAPi)₂ displayed improved tumor residence time over DOTA.SA.FAPi, and BiOncoFAP similarly highlighted the benefit of bivalent small-molecule designs compared with monovalent analogs (Galbiati et al. 2022, Ballal et al. 2021). However, these reports also emphasize that the performance of multimeric FAP ligands is not solely dictated by valency, but strongly influenced by scaffold architecture, including linker composition, spacing, and metabolic stability. Herein, rather than relying on a single dimeric design, we synthesized and evaluated 9 dimeric FAP-targeting radioligands using three distinct tripod core platforms and our 8-QCP-based FAP ligands. Our results suggest that the dimeric architecture had minor influence on the binding affinity to FAP. All conjugates were successfully radiolabeled with high radiochemical yield and purity, demonstrating the robustness of these modular platforms for both diagnostic (¹¹¹In) and therapeutic (¹⁷⁷Lu) applications. Nevertheless, the stability of these conjugates varied depending on linker chemistry, in particular thioether-containing variants (i.e., eFAP-20 and eFAP-31) appeared to exhibit relatively lower stability in PBS and murine serum. A possible contributing factor is the susceptibility of exposed sulfur-containing moieties to metabolic oxidation or enzymatic transformation, as it was observed for some thioether-containing small molecules in plasma and microsomes (Usmani et al. 2004, Han and Kucukguzel 2022). Although our study did not include a direct head-to-head comparison of identical cores with and without PEG₂ linkers, the PEG₂-linked variants (eFAP-50, eFAP-51, and eFAP-52) maintained comparatively higher stability under the tested conditions (Yuan et al. 2025).

A further structural observation is that introducing piperazine rings into the FAP ligand (i.e., eFAP-52) did not significantly improve the performance of the dimers, contrary to the initial expectation that the piperazine ring would enhance pharmacokinetics (Köchel et al. 2025, Kochel et al. 2025, Ten et al. 2026). In contrast, our results showed that eFAP-51, containing an AMCHA linker, exhibited improved tumor uptake and retention, both in cell uptake studies and in vivo models. These results are consistent with previous reports on AMCHA-based dimeric FAP ligands showing enhanced tumor accumulation compared to monomeric analogs (Luo et al. 2024). This difference may be attributed to the distinct charge properties of the linkers, as AMCHA is neutral whereas piperazine is positively charged under physiological conditions. The positive charge of piperazine can promote interactions with negatively charged cellular membranes in excretory organs, potentially increasing nonspecific retention and reduced tumor delivery (Raheem et al. 2023, Ailabouni and Prasad 2025). In contrast, the neutral AMCHA linker may minimize such interactions, resulting in more favorable clearance from non-target tissues and superior tumor-to-background contrast for eFAP-51. The cell uptake results also indicated an increase in uptake over time with eFAP-51, a finding that was further supported by the in vivo studies, where tumor uptake at 4 h was notably higher than the uptake at 1 h. However, while dimerization enhanced tumor targeting, we also observed that accumulation in nontarget organs increased at 1- and 4-hours post-injection. This underscores the importance of optimizing the injected molar dose. Indeed, our dose optimization study demonstrated that, although the highest tumor uptake was achieved with the lowest injected dose, higher molar doses partially reduced tumor accumulation but simultaneously decreased uptake in normal organs, such as liver and kidneys. Therefore, intermediate injected molar doses (0.3–0.6 nmol) provided the most favorable balance between high tumor uptake and minimal background activity, resulting in improved tumor-to-background contrast. Similar trade-offs between enhanced multivalent binding and increased normal tissue exposure were also reported for other multimeric FAPI derivatives (Bilinska et al. 2025, Yadav et al. 2024).

Taken together, the in vivo data obtained with [¹⁷⁷Lu]Lu-eFAP-51 are consistent with the favorable tumor retention and clearance profile observed for [¹¹¹In]In-eFAP-51, despite differences in injected molar doses of the two radioligands. The tumor half-life of [¹¹¹In]In-eFAP-51 (33.6 h) was relatively similar to that of [¹⁷⁷Lu]Lu-eFAP-51 (40.2–40.3 h at the two evaluated molar doses), indicating prolonged tumor retention for both radiometal complexes. [¹¹¹In]In-eFAP-51 and [¹⁷⁷Lu]Lu-eFAP-51 were primarily excreted via the renal pathway, and showed low accumulation in non-target organs. These findings support that the favorable pharmacokinetic profile of eFAP-51 is preserved upon replacement of ¹¹¹In with ¹⁷⁷Lu, further substantiating its potential as a theranostic agent. Similar in vivo behavior has also been reported for other FAP-targeting agents in preclinical studies (Mukkamala et al. 2024, Zboralski et al. 2022).

However, although our results highlight eFAP-51 as a very promising candidate for theranostic applications, our study has several limitations that should be considered in the context of translating this dimeric FAP-targeted radioligand to the clinic. First, while the HT-1080.huFAP xenograft provides a controlled model for evaluating FAP-mediated targeting, it does not fully capture the stromal complexity and heterogeneity of human tumors, which can influence tracer distribution and retention. Validation in more relevant models, such as patient-derived xenografts tumors, will be important to confirm the translational applicability of eFAP-51. Second, although we extended the evaluation of eFAP-51 to ¹⁷⁷Lu radiolabeling and performed biodistribution studies and dose optimization, the therapeutic efficacy and long-term toxicity were not directly assessed. Comprehensive dosimetry and therapy studies, including assessment of clearance organs, such as kidneys and liver, are necessary to fully establish its potential as a therapeutic agent. Finally, the choice of radionuclide and combination strategies remain important considerations. While ¹⁷⁷Lu provides a suitable balance of tissue penetration and cytotoxicity, alpha-emitters like ²¹²Pb could offer enhanced potency for stroma-rich tumors (Liu et al. 2022). Combining FAP-targeted radionuclide therapy with other therapeutic modalities, such as immunotherapy (Öztürk et al. 2025, Zhao et al. 2024), may further enhance antitumor effects. Taken together, these considerations define the next steps for translating multivalent FAP-targeted radioligands into effective and safe therapies.

Conclusions

In this study, we designed and synthesized a series of dimeric FAP-targeting radioligands using different tripod-based multimerization platforms. Among the evaluated candidates, [¹¹¹In]In-eFAP-51 demonstrated the most favorable profile, combining high FAP-specific tumor uptake with prolonged tumor retention and limited off-target accumulation compared with the monomeric reference [¹¹¹In]In-FAPI-46. Furthermore, radiolabeling of eFAP-51 with ¹⁷⁷Lu yielded a highly stable therapeutic analog, supporting its potential for radionuclide therapy applications. Dose optimization studies identified intermediate injected molar doses (0.3–0.6 nmol) providing an optimal balance between tumor uptake and background clearance. Overall, these findings highlight eFAP-51 as a promising lead compound for clinical translation in theranostic oncology.

Materials and methods

General information

Chemistry

All chemicals and solvents were obtained from commercial suppliers and used as received, unless otherwise specified. Quality control was carried out via LC-MS using an Agilent 1260 Infinity II LC/MSD XT system (Amstelveen, The Netherlands). Electrospray ionization in positive mode was used to confirm the identity of the synthesized products. Purifications were performed on a preparative high-performance liquid chromatography (HPLC) using an Agilent 1290 Infinity II Preparative LC system (Amstelveen, The Netherlands) equipped with an autosampler and a fraction collector. UV absorbance was measured at 230, 254 and 280 nm. Both the LC-MS and HPLC systems were controlled by Agilent OpenLab CDS Chemstation software. NMR spectra were recorded in DMSO-d₆, CD₃OD, or CDCl₃ in diluted solutions on a 60 MHz Nanalysis NMReady 60Pro (Calgary, Canada), 400 MHz or 600 MHz NMR Bruker systems (Billerica, MA, USA) NMR at ambient temperature. Chemical shifts are given as δ values in ppm and coupling constants J are given in Hz. The splitting patterns are reported as s (singlet), d (doublet), t (triplet), q (quartet), m (multiplet), dd (doublet of doublets), and br (broad signal).

Radiochemistry

[¹¹¹In]InCl₃ was provided by Curium (Petten, The Netherlands) and [¹⁷⁷Lu]LuCl₃ was purchased from PI Medical (Raamdonksveer, The Netherlands).Radio-HPLC analyses were carried out using a Waters 2695 system equipped with a diode array detector 2998 and a radioactivity detector from Canberra (Zadik, Belgium). This system was controlled by Empower 3 software. Instant thin-layer chromatography (iTLC) was performed on iTLC-SG plates (Aligent; Folsom, CA, USA), eluted with sodium citrate solution (0.1 M, pH 5.0), and analyzed using a bSCAN radio-chromatography scanner (Brightspec; Antwerp, Belgium) equipped with a sodium iodide detector. Radiochemical yield (RCY) was determined from iTLC, and the radiochemical purity (RCP) of the labeled compounds was measured by radio-HPLC. Activity measurements were performed using a VDC-405 dose calibrator from Comecer (Joure, The Netherlands). Radioactive samples for LogD_7.4 determination, in vitro assays, and in vivo tissue uptakes were counted using a Wizard 2480 gamma counter from Perkin Elmer (Waltham, MA, USA).

HPLC conditions

The LC-MS analyses were performed on a Poroshell 120 EC-C18 column (3 × 100 mm, 2.7 μm) at a flow rate of 0.5 mL/min and with a mobile phase consisting of: A [0.1% formic acid (FA) in water, v/v] and B [0.1% FA in acetonitrile (ACN), v/v]. Elution of the products was performed according to the following gradient: 0–5 min, 5–100% B; 5–8 min, 100% B.

Purification of eFAPs was performed using an Agilent Prep C18 column (50 × 21.2 mm, 5 μm). The mobile phase consisted of solvents A [0.1% trifluoroacetic acid (TFA) in water, v/v] and B [0.1% TFA in ACN, v/v] at a flow rate of 10 mL/min. The following elution method was used: 0–2 min: 5–20% B; 2–10 min: 20–40% B; 10–12 min: 40–100% B.

The analyses of the radioactive products were performed on an analytical Gemini C₁₈ column (5 μm, 250.0 × 4.6 mm) from Phenomenex at a flow rate of 1 mL/min and with a mobile phase consisting of: A [0.1% TFA in water, v/v] and B [0.1% TFA in ACN, v/v]. Elution of the products was performed according to the following gradient: 0–3 min: 5% B; 3–23 min: 5 to 100% B and 23–27 min: 100% B.

Chemical synthesis

The core building block (8-QCP) was synthesized through a straightforward multi-step organic syntheses procedure, as previously reported by our group (Heide et al. 2024). Detailed information on the synthesis of the eFAP precursors from 8-QCP, as well as HPLC purification, mass analyses, and NMR characterizations of intermediates and precursors is provided in the Supplementary Materials (Fig. S1–S34).

Binding affinity

All enzymatic assays were performed in black 384-well plates, and enzymatic activity was measured on a Hidex Sense microplate reader (Turku, Finland), monitoring the fluorescence at λ_EX 350 nm/λ_EM 485 nm. For huFAP and muFAP, 50 µL of reaction mixture consisting of the substrate (Z-Gly-Pro-AMC, 10 µM), huFAP or muFAP protein (10 ng; Biolegend; San Diego, CA, USA), assay buffer (50 mM Tris, 100 mM NaCl, and 1.5 µM BSA, DMSO (1% v/v), pH 7.4), were supplemented with increasing concentrations of eFAPs or FAPI-46 (10^− 11 M to 10^− 5 M). Val-boroPro (10^− 4 M) from Bachem (Bubendorf, Switzerland) and assay buffer served as a positive and negative controls, respectively. The enzymatic assay of DPP4 contained substrate H-Gly-Pro-AMC (100 µM; MedChemExpress; Monmouth Junction, NJ, USA), DPP4 (10 nM; Abcam; Cambridge, UK), assay buffer, and increasing concentrations of the inhibitors (10^− 10 M to 10^− 4M). Sitagliptin (10^− 4 M) from Abcam and assay buffer were used as a positive and negative controls, respectively. H-Gly-Pro-AMC (100 µM; Enzo Life Sciences; Bruxelles, Belgium) in assay buffer was used as a baseline control. The plates were analyzed by microplate reader at 37 °C under dark conditions for 1 h (huFAP), 2 h (muFAP), or 30 min (DPP4). For all assays, the components were kept on ice before plating. Experiments were performed in triplicate.

Radiolabeling

Prior to labeling, the concentrations of the eFAPs were determined by titration, according to a method previously described (Breeman et al. 2014). [¹¹¹In]InCl₃ or [¹⁷⁷Lu]LuCl₃ in 0.5 M HCl was added to a mixture of FAPI-46 or eFAPs, ascorbic acid/gentisic acid (10 µL, 50 mM), ethanol (10 µL), sodium acetate (1 µL, 2.5 M), and H₂O (final volume of 140 µL). For imaging studies, 20 MBq of radionuclide was used per labeling reaction, while for biodistribution studies, 5 MBq was used. The mixture was incubated for 15 min at 90 °C. To complex any remaining free radionuclide, 5 µL of DTPA at a concentration of 3 mg/mL was added.

In vitro stability

[¹¹¹In]In-eFAPs (~ 1 MBq) or [¹⁷⁷Lu]Lu-eFAPs (~ 3 MBq) was incubated in 300 µL of PBS (0.1 M, pH 7.4) or murine serum at 37 °C. The stability of the radiolabeled peptide was assessed at 1, 4 and 24 h post-incubation. For the samples in PBS, no prior treatment was required before analysis by radio-HPLC. However, an equal volume of acetonitrile was added to the samples in murine serum to precipitate the proteins. The solution was then centrifuged for 10 min at 13’000 rpm, and the supernatant was subsequently subjected to radio-HPLC analyses.

LogD_7.4

Distribution coefficients (LogD_7.4) were determined using a shake-flask method. A sample containing [¹¹¹In]In-eFAPs (~ 0.5 MBq) was added to a vial containing 500 µL of PBS (pH 7.4) and 500 µL of n-octanol. The vial was vigorously vortexed and then centrifuged at 10’000 rpm for 10 min to allow phase separation. Samples of 10 µL from both phases were collected and measured in a γ-counter. The LogD_7.4 value was calculated using the following equation: LogD_7.4 = log [(counts in n-octanol phase)/(counts in aqueous phase)]. Measurements were performed in triplicates.

Cell culture

Cell experiments were performed with a human sarcoma HT-1080 cell line, stably transduced with human FAP (HT-1080.huFAP), which was kindly provided by Uwe Haberkorn (University of Heidelberg), Christoph Renner (KHO) and Stefan Bauer (NCT Heidelberg). The wild type HT-1080.WT (ATTC) was used as a negative control. HT-1080.huFAP and HT-1080.WT cells were cultured in DMEM Glutamax© (Gibco; Breda, The Netherlands). Both media were supplemented with 2 mM L-glutamine, 10% FBS (Gibco), 50 UI/mL penicillin, and 50 µg/mL streptomycin (Merck Life Science NV; Amsterdam, The Netherlands). All cell lines were regularly passaged at 80% confluence and cultured at 37 °C, 5% CO₂.

Uptake/internalization studies

Cells were seeded in 12-well plates 24 h before the experiment (1 × 10⁵ cells/well). Adhered cells were incubated with 10^− 9 M of [¹¹¹In]In-FAPI-46 or [¹¹¹In]In-eFAPs in 750 µL of culture medium for 0.5, 1–2 h at 37 °C. Uptake specificity was evaluated with a block containing 1000-fold excess of non-labelled FAPI-46. After incubation, medium was removed, and cells were washed twice with cold PBS. The membrane-bound fraction was collected by incubating cells with an acid buffer (50 mM glycine, 100 mM sodium chloride, pH 2.8) for 10 min at room temperature (rt). Cells were lysed using sodium hydroxide (1 M) for 5 min at rt to obtain the internalized fraction. Both fractions were measured in a γ-counter, and data were analyzed in GraphPad Prism 9.0 and expressed as percentage of added dose.

Animal models

Five-week-old female NMRI-Foxn1 nu/nu mice (Janvier; Le Genest-Saint-Isle, France) were housed in individually ventilated cages with four mice per cage. After one week of acclimatization, mice were subcutaneously inoculated with HT-1080.huFAP on the right shoulder and HT-1080.WT on the left shoulder (5 × 10⁶ cells in Hank’s Balanced Salt Solution from Gibco). Tumors were allowed to grow for 10 days, resulting in tumor volumes of approximately 200 mm³ for both HT-1080.huFAP and HT-1080.WT at the start of the study.

In vivo single photon emission computed topology/computer tomography (SPECT/CT)

Mice were intravenously (i.v.) injected in the tail vein with 100 µL of [¹¹¹In]In-eFAP-51, [¹¹¹In]In-eFAP-52 or [¹¹¹In]In-FAPI-46 (20 MBq/1 nmol, n = 4 per group) diluted in a solution of Kolliphor HS15 (Merck; Haarlerbergweg, The Netherlands) in PBS (0.06 mg/mL). Mice were imaged at 1, 4, 24 and 48 h post injection (p.i.) in a dedicated small-animal SPECT/CT scanner (VECTor⁵-CT, MILabs B.V.; Utrecht, The Netherlands) with a multi-mouse bed. Mice were placed on a heated bed under 2% isoflurane/O₂ anesthesia during the scans. A high-sensitivity pinhole collimator (UHR-RM, 1.0 mm pinhole diameter) was used, and whole-body SPECT images (transaxial field of view, 54 mm) were acquired over 30 min using a spiral scan in normal scan mode with in list-mode acquisition. This was followed by a whole-body CT scan within 5 min (settings: full angle scan, angle step 0.75 degrees, normal scan mode, 50 kV tube voltage, 0.21 mA tube current, 500 μm aluminum filter). For reconstruction of the SPECT images, the similarity-regulated SROSEM method and MLMN method (MILabs Rec 12.00 software) were employed, performing 9 and 128 iterations, respectively, at 0.8 mm³ resolution, using 173 keV ± 10% and 247 keV ± 10% energy windows for indium-111. Two adjacent background windows per photo peak were selected for triple-energy window scatter and for crosstalk correction. Post-filtering of reconstructed volumes of SPECT scans was performed with an isotropic 3-dimensional Gaussian filter with 1 mm full width at half-maximum. The CT was reconstructed at 200 µm³ and analyzed together with the registered attenuation-corrected SPECT images using PMOD 3.9 (Zurich, Switzerland).

Ex-Vivo biodistribution

To determine the biodistribution of the compounds, blood was collected via retro-orbital puncture under isoflurane/O₂ anesthesia, after which the mice were sacrificed via cervical dislocation. Both tumors and organs of interest (heart, lungs, liver, spleen, stomach, large and small intestine, pancreas, kidneys, muscle, skin, and bone) were excised, washed in PBS, blotted dry, and collected into tubes. All collected tissues were weighed and measured in a γ-counter. The percentage of injected dose per gram (% ID/g) was determined for each tissue sample and corrected for the injected activity and the % ID present at the injection site.

Statistics

All statistical analyses were performed with GraphPad Prism version 9.0 (San Diego, CA, USA). IC₅₀ values and curves were determined using a non-linear regression Log(inhibitor) vs. response analysis with variable slope (four-parameters). Tumor and organ clearance graphs were calculated by one-phase decay linear regression, corrected for the grow-in phase when applicable. Data are reported as mean ± sd (standard deviation) for at least three independent replicates. The areas under the curve (AUC) obtained from the biodistribution data sets were compared, using a one-way ANOVA with Bonferroni’s multiple comparison post-test. Separately, for the dose-dependent study (Fig. S47), one-way ANOVA with Tukey’s post-hoc test was used for pairwise comparisons.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

Le Li (L.L.) contributed to investigation, formal analysis, and writing-review & editing. Hanyue Ma (H.M.) contributed to investigation and writing-review & editing. Amber Piet (A.P.) contributed to methodology, software, validation, formal analysis, and writing-review & editing. Corrina de Ridder (C.R.) contributed to investigation and writing-review & editing. Debra Stuurman (D.S.) contributed to investigation and writing-review & editing. Priciana Paraïso (P.P.) contributed to software, validation, formal analysis, and writing-review & editing. Yann Seimbille (Y.S.) contributed to conceptualization, methodology, supervision, project administration, funding acquisition, and writing-review & editing. All authors read and approved the final manuscript.

Funding

Le Li was supported by the China Scholarship Council (CSC).

Data availability

All available data are described in the manuscript and available through the supplementary material. Additional information can be obtained by contacting the authors upon reasonable request.

Declarations

Ethics approval and consent to participate

The study was conducted according to the guidelines of the Declaration of Helsinki. Approval was granted by the Animal Welfare Committee of the Erasmus MC (CCD#: 2216075, SP2400006).

Consent for publication

Not applicable.

Competing interests

The authors declare that they have no competing interests.