A Theranostic Small-Molecule Prodrug Conjugate for Neuroendocrine Prostate Cancer

After androgen deprivation therapy, a significant number of prostate cancer cases progress with a therapy-resistant neuroendocrine phenotype (NEPC). This represents a challenge for diagnosis and treatment. Based on our previously reported design of theranostic small-molecule prodrug conjugates (T-SMPDCs), herein we report a T-SMPDC tailored for targeted positron emission tomography (PET) imaging and chemotherapy of NEPC. The T-SMPDC is built upon a triazine core (TZ) to present three functionalities: (1) a chelating moiety (DOTA: 1,4,7,10-tetraazacyclododecane-1,4,7,10-tetraacetic acid) for PET imaging when labeled with 68Ga (t1/2 = 68 min) or other relevant radiometals; (2) an octreotide (Octr) that targets the somatostatin receptor 2 (SSTR2), which is overexpressed in the innervated tumor microenvironment (TME); and (3) fingolimod, FTY720—an antagonist of sphingosine kinase 1 that is an intracellular enzyme upregulated in NEPC. Polyethylene glycol (PEG) chains were incorporated via conventional conjugation methods or a click chemistry reaction forming a 1,4-disubstituted 1,2,3-triazole (Trz) linkage for the optimization of in vivo kinetics as necessary. The T-SMPDC, DOTA-PEG3-TZ(PEG4-Octr)-PEG2-Trz-PEG3-Val-Cit-pABOC-FTY720 (PEGn: PEG with n repeating ethyleneoxy units (n = 2, 3, or 4); Val: valine; Cit: citrulline; pABOC: p-amino-benzyloxycarbonyl), showed selective SSTR2 binding and mediated internalization of the molecule in SSTR2 high cells. Release of FTY720 was observed when the T-SMPDC was exposed to cathepsin B, and the released FTY720 exerted cytotoxicity in cells. In vivo PET imaging showed significantly higher accumulation (2.1 ± 0.3 %ID/g; p = 0.02) of [68Ga]Ga-DOTA-PEG3-TZ(PEG4-Octr)-PEG2-Trz-PEG3-Val-Cit-pABOC-FTY720 in SSTR2high prostate cancer xenografts than in the SSTR2low xenografts (1.5 ± 0.4 %ID/g) at 13 min post-injection (p.i.) with a rapid excretion through the kidneys. Taken together, these proof-of-concept results validate the design concept of the T-SMPDC, which may hold a great potential for targeted diagnosis and therapy of NEPC.


Introduction
A lethal disease with median survival of less than 1 year from the time of detection, neuroendocrine prostate cancer (NEPC) accounts for approximately 25% of all therapy-and castration-resistant prostate cancer (tCRPC) [1,2]. Since prostatic adenocarcinoma (ADPC) is a typical androgen-dependent (AD) cancer, androgen deprivation therapy is the standard of care for this disease. However, over the course of treatment, cancer cells can adapt and restore their growth signaling through mutated androgen receptors (ARs) or AR variants, necessitating the next therapeutic option-total androgen blockade. Nevertheless, the disease can further progress to become therapy-and castration-resistant, often exhibiting neuroendocrine features commonly known as NEPC-a subset of CRPC [2]. Without AR expression, the cells of tCRPC present low-to-zero prostate-specific membrane antigen It has been reported that chronic androgen deprivation therapy triggers the overexpression of Sphk1, which is implicated in the progression of NEPC [24,25]. A cytoplasmic kinase, SphK1 is a lipid mediator that is critical for tumor cell growth, survival, and therapeutic resistance. Therefore, SphK1 inhibitors have been used to suppress tumor growth [26,27]. The mechanism of action of FTY720-a SphK1 inhibitor-involves its phosphorylation in vivo to form fingolimod phosphate, with a similar structure to that of sphingosine 1-phosphate (S1P), which is the natural ligand of the sphingosine 1-phosphate recep- The T-SMPDC, DOTA-PEG 3 -TZ(PEG 4 -Octr)-PEG 2 -Trz-PEG 3 -Val-Cit-pABOC-FTY720, is built upon a 1,3,5-triazine core (TZ) to present three functionalities: (1) a chelating moiety (DOTA) for PET imaging when labeled with 68 Ga (t 1/2 = 68 min) or other rele-vant radiometals; (2) an octreotide (Octr) that targets somatostatin receptor 2 (SSTR2), which is overexpressed in the innervated tumor microenvironment (TME) [17,[22][23][24]; and (3) FTY720 coupled with a cathepsin-B-cleavable linker comprising a valine-citrulline (Val-Cit) motif and a p-amino-benzyloxycarbonyl (pABOC) moiety. Polyethylene glycol (PEG) chains with a varying number of ethyleneoxy repeating units (n = 2, 3, or 4) were incorporated via conventional conjugation methods or a click chemistry reaction forming a 1,4-disubstituted 1,2,3-triazole (Trz) linkage for the optimization of in vivo kinetics as necessary.
It has been reported that chronic androgen deprivation therapy triggers the overexpression of Sphk1, which is implicated in the progression of NEPC [24,25]. A cytoplasmic kinase, SphK1 is a lipid mediator that is critical for tumor cell growth, survival, and therapeutic resistance. Therefore, SphK1 inhibitors have been used to suppress tumor growth [26,27]. The mechanism of action of FTY720-a SphK1 inhibitor-involves its phosphorylation in vivo to form fingolimod phosphate, with a similar structure to that of sphingosine 1phosphate (S1P), which is the natural ligand of the sphingosine 1-phosphate receptors (S1PRs). However, while such inhibitory effects are desirable in cancer therapy, they might be detrimental to the fundamental biological processes regulated by S1PRs in normal cells. For instance, treatments of multiple sclerosis with FTY720 have been reported to show cardiovascular toxicity [28,29]. For this reason, targeted delivery and controlled release of FTY720 is of clinical significance to achieve an effective NEPC therapy while minimizing its toxicity to normal cells. Our design strategy is to introduce a cathepsin-B-cleavable linker between FTY720 and the conjugate platform.
To date, several lines of studies have reported upregulated levels of SSTR2 expression in NEPC [30]. For instance, the expression level of SSTR2 was found to be elevated following hormone depletion therapy [22], and high uptake of 68 Ga-DOTATATE was observed in biochemically relapsed prostate cancer [1,31]. Most recently, targeting of SSTR2 was reported as a practical strategy to develop a polymeric nanoparticle system for NEPC therapy [32]. Therefore, we chose SSTR2 as a model NEPC target for the design of our T-SMPDC, DOTA-PEG 3 -TZ(PEG 4 -Octr)-PEG 2 -Trz-PEG 3 -Val-Cit-pABOC-FTY720. In this work, we present the T-SMPDC design along with its modular synthesis and the proofof-concept biological evaluations of its potential application for NEPC-targeted diagnosis and chemotherapy.

General Materials and Procedures
All chemical reagents and solvents were purchased from commercial sources (Sigma-Aldrich, St. Louis, MO, USA; BroadPharm, San Diego, CA, USA; Fisher Scientific, Hampton, NH, USA). They were used as received unless stated otherwise. Milli-Q water was supplied by a Millipore Gradient Milli-Q water system (Burlington, MA, USA) and used for the preparation of all aqueous solutions. The characterization of the synthesized compounds, intermediates, and conjugates was performed on an Agilent 6540 Accurate-Mass Quadrupole Time-of-Flight Liquid Chromatography-Mass Spectrometry (LC-MS) system, which was coupled with a 1290 ultra-performance liquid chromatography (UPLC) system (Santa Clara, CA, USA). Unless otherwise noted, all LC-MS measurements were conducted in the positive ionization mode. Matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF MS) was performed on a Voyager-DE PRO biospectrometry workstation (Applied Biosystems, Waltham, MA, USA) using 2,5-dihydroxybenzoic acid as the matrix. Nuclear magnetic resonance (NMR) spectra were recorded on a Varian 400 MHz spectrometer (Palo Alto, CA, USA). Purification of synthesized compounds was carried out using an Agilent 1260 Infinity preparative high-performance liquid chromatography (HPLC) system, which was equipped with a 1260 photodiode array detector (PDA) and an Agilent Prep-C18 column (150 × 21.2 mm, 5 µm) (Santa Clara, CA, USA). The radiolabeled compounds were characterized by an Agilent 1220 Infinity II analytical HPLC system equipped with a 1220 LC diode array detector (DAD) (76337 Waldbronn, Germany), an in-line Eckert & Ziegler radio detector (Eckert & Ziegler Radiopharma, Inc. Hopkinton, MA, USA), and a C18 Atlantis ® T3 column (4.6 × 250 mm, 5 µm) (Waters, Ireland).

Cell Culture and Animal Models
All animal experiments presented in this work were approved by the Institutional Animal Care and Use Committee (IACUC) at the University of Texas Southwestern Medical Center (Dallas, TX, USA). Tumor-bearing animal models were established in the Hsieh Lab by following the animal protocol number (APN 2017-102131; three-year renewal approval date: 29 December 2022; expiration date: 29 December 2025). The mouse imaging procedures were performed according to APN 2020-102851 (approval date: 26 May 2020; expiration date: 26 May 2023) in the Sun Lab. The isogenic cell lines PC3_VC (SSTR2 high ) and PC3_sgPTP1B (SSTR2 low ) were generated in the Hsieh Lab. The parental PC3 cell line was purchased from ATCC (Manassas, VA, USA, CRL-1435); PC3_VC and PC3_sgPBP1B denote cells that were generated by CRISPR-vector and CRISPR-PTP1B gene knockout, respectively. Both cell lines were cultured in RPMI medium containing 10% fetal bovine serum, 1% penicillin/streptomycin, and 2 mM L-glutamine. The cells were harvested to establish the tumor model once reaching 75-80% confluency at 37 • C under 5% CO 2 . The dual-tumor murine model was developed by subcutaneously injecting 1.0 × 10 6 cells in 100 µL of Hank's Buffered Salt Solution containing 50% Matrigel into the shoulders of male severe combined immunodeficient (SCID) mice (NOD.CB17-Prkdcscid/NCrHsd, 6-8 weeks). The mice were allowed unrestricted access to autoclaved water and commercial diets and monitored closely for general observations and tumor burden assessment [20] throughout the entire period of the experiments.
For the cell-binding and internalization experiments, AR42J (SSTR2 high ) and PC3-Flu (SSTR2 low ) cells were used. AR42J cells were purchased from ATCC (Manassas, VA, USA, CRL-1492), while PC3-Flu cells were obtained from the laboratory of Prof. Martin G. Pomper at John Hopkins University (Baltimore, MD, USA). The subline PC3-Flu, from the androgen-independent PC3 human prostate cancer cell line derived from an advanced androgen-independent bone metastasis, was engineered to maintain no expression of PSMA according to reported procedures [33,34].

In Vitro Cathepsin-B-Mediated Release Assay
A comparative cathepsin-B-mediated release assay was performed with compound 13 in the presence vs. absence of human cathepsin B (Sino Biological Inc., Chesterbrook, PA, USA). Briefly, compound 13 (150 µg) was dissolved in 60 µL of 100 mM citrate buffer (pH 5.5), followed by adding 300 µg of cysteine. Then, 20 µg of human cathepsin B in 40 µL of 100 mM citrate buffer was added to the mixture and incubated at 37 • C for 1 h. At the end of the incubation, the solution (20 µL) was sampled for HPLC and LC-MS analysis. Controls of this assay included compound 13 without being exposed to cathepsin B, as well as cathepsin B itself processed under the same conditions.

Small Animal PET/CT Imaging
The small animal imaging studies were performed when the size of the tumor xenografts reached the range 150-500 mm 3 on a Mediso NanoScan PET/CT system (Mediso Medical Imaging Systems, Budapest, Hungary) equipped with a 4-mouse bed and a heating chamber to maintain the body temperature of the mice. The scans with [ 68 Ga]Ga-DOTA-PEG 3 -TZ(PEG 4 -Octr)-PEG 2 -Trz-PEG 3 -Val-Cit-pABOC-FTY720 and [ 68 Ga]Ga-PSMA-11 were arranged two days apart to ensure no signal interference between the two radiotracers. Before each scan, tumor-bearing mice were anesthetized and maintained under 1.5-2% isoflurane for the duration of the injection and scan. Immediately after intravenous injection of 3-4.4 MBq of each radiotracer, which was prepared in 100 µL of phosphate-buffered saline (PBS), a 60-min dynamic data acquisition was performed, followed by an anatomical reference CT scan. Using the Tera-Tomo 3D PET iterative algorithm with real-time Monte-Carlo-based physical modeling, we reconstructed the PET images into 10 frames of 1 min, 5 frames of 2 min, and 8 frames of 5 min. Quantitative imaging analysis was performed after the PET and CT images were co-registered to define regions of interest (ROIs) manually. The uptake of [ 68 Ga]Ga-DOTA-PEG 3 -TZ(PEG 4 -Octr)-PEG 2 -Trz-PEG 3 -Val-Cit-pABOC-FTY720 or [ 68 Ga]Ga-PSMA-11 in ROIs and tumor xenografts was calculated at each specific time of interest as the percentage of injected dose per gram of tissue (%ID/g).

Western Blot Analysis
Cell lysates of PC3_VC and PC3_sgPTP1B were prepared by lysing the cells on ice for 30 min using freshly prepared buffer, which contained 50 mM Tris-HCl (pH 7.5), 150 mM NaCl, 0.1% Triton X-100, 1 mM sodium orthovanadate, 1 mM sodium fluoride, 1 mM sodium pyrophosphate, 10 mg/mL aprotinin, 10 mg/mL leupeptin, 2 mM phenylmethylsulfonyl fluoride, and 1 mM EDTA. After centrifugation of the cell lysates at 14,000 rpm for 30 min at 4 • C, the protein extracts were loaded for sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) using Bolt 4-12% NuPAGE gels (Life Technologies, Carlsbad, CA, USA) and then blotted onto a nitrocellulose membrane using the Trans-Blot Turbo Transfer system (BIORAD, Hercules, CA, USA). After the membrane was incubated in the solution of primary antibodies against PTP1B (ProteinTech, Rosemont, IL, USA) and SSTR2 (Abcam, Cambridge, UK) at 4 • C for 16-18 h, it was washed and then incubated with horseradish-peroxidase-conjugated secondary antibodies at r.t. for 2 h. The blotting results were read using an Advansta ECL chemiluminescent detection system (San Jose, CA, USA). The loading control was β-actin (Santa Cruz Biotechnology, Dallas, TX, USA).

Immunohistochemistry (IHC)
Tumors were excised after the experiments and immediately fixed in 10% neutralbuffered formalin for 48 h. The tissue samples were processed and embedded in paraffin blocks. For IHC, the formalin-fixed, paraffin-embedded sections of the tumors were deparaffinized, rehydrated, and subjected to heat-induced antigen retrieval (citrate buffer, pH 6.0). After incubation with primary antibodies against SSTR2 (Abcam, Cambridge, UK) or PSMA (Abcam, Cambridge, UK), the sections were developed with 3,3 -diaminobenzidine chromogen, followed by counterstaining with hematoxylin and eosin. The representative photograph was taken with a Nikon microscope.
The multistep synthesis started with the preparation of each individual component, to enable their incorporation to TZ at later steps. A DOTA chelator (Scheme 1a) was added to a PEG 3 diamine linker through a condensation reaction. For the scope of this work, the chelator was incorporated for radiolabeling the T-SMPDC with 68 Ga or 64 Cu for PET imaging or in vitro assays. The triazine core was functionalized with a propargyl-PEG 2amine group (Scheme 1b) through an SN2 reaction that functioned as the connecting site for the FTY720-cathepsin B conjugate. A succinimidyl-ester-activated PEG 4 linker was synthesized to serve as a linker between the octreotide molecule and the triazine core (Scheme 1c). For the chemotherapeutic functionality, the FTY720 molecule was attached to the cathepsin-B-cleavable linker through an SN2 reaction, followed by its addition to the triazine core through the propargyl-PEG 2 linker using click chemistry (Scheme 1d). For the imaging functionality, the DOTA-PEG 3 (compound 2) was then attached to the triazine core using an SN2 reaction between the chlorine of the triazine and the amine on the N-terminal of compound 2. Lastly, to add the SSTR2-targeting functionality, the octreotide-PEG 4 molecule was added to the triazine core through a substitution reaction. In the last step, the tertbutyl groups from the DOTA chelator were hydrolyzed to yield the designed T-SMPDC, DOTA-PEG 3 -TZ(PEG 4 -Octr)-PEG 2 -Trz-PEG 3 -Val-Cit-pABOC-FTY720 (Scheme 1d). Detailed reaction conditions and compound characterization for each step are provided in the Supplemental Materials ( Figures S1-S12).

SSTR2-Selective Cell Uptake and SSTR2-Mediated Internalization of [ 64 Cu]Cu-DOTA-PEG 3 -TZ(PEG 4 -Octr)-PEG 2 -Trz-PEG 3 -Val-Cit-pABOC-FTY720
With the functionality of PEG 4 -Octreotide, the construct was designed to maintain the desired SSTR2-selective binding and SSTR2-mediated internalization. The comparative in vitro assays were performed using the SSTR2 high AR42J cell line [40] and SSTR2 low PC3-Flu cells. As shown in Figure 2a, the normalized uptake of [ 64 Cu]Cu-DOTA-PEG 3 -TZ(PEG 4 -Octr)-PEG 2 -Trz-PEG 3 -Val-Cit-pABOC-FTY720 demonstrated a six-fold higher accumulation in SSTR2 high AR42J cells than that in SSTR2 low PC3-Flu cells after 1 h of incubation, confirming the desired SSTR2-selective binding of the conjugate. Furthermore, incubation with SSTR2 high AR42J cells showed that approximately 40% of [ 64 Cu]Cu-DOTA-PEG 3 -TZ(PEG 4 -Octr)-PEG 2 -Trz-PEG 3 -Val-Cit-pABOC-FTY720 became internalized after 60 min (Figure 2b). To assess whether the construct was capable of controlled release of FTY720 upon cathepsin-B-catalyzed cleavage, a release assay was performed using a previously reported procedure [41]. After 1 h of incubation at 37 °C, a sample was removed from the mixture, and analysis with LC/MS showed a molecular weight of 366.0488 [M + 2H2O + Na] + , which represents the molecular weight of FTY720 (307.2511 g/mol), two water molecules, and a sodium ion (Figures S15 and S16). The cathepsin-B-mediated release of FTY720 was further confirmed with a control assay, where the peak of 366.0488 was absent when the experiment was conducted under the same conditions in the absence of cathepsin B ( Figure S17).

The Cytotoxicity of FTY720 Was Masked by Its Conjugation with the Cathepsin B Prodrug Linkage, Val-Cit
Cell viability assays were performed using different concentrations of Val-Cit-pABOC-FTY720 and free FTY720 to compare their cytotoxicity. To that end, four different concentrations of the two compounds were used (0, 1, 5, and 10 µM), and the cells were incubated for either 48 or 72 h. The results showed that the EC50 for FTY720 was ~ 5 µM after 48 h (Figure 3a, blue bars) and 72 h (Figure 3b, blue bars) of incubation, while the EC50 for Val-Cit-pABOC-FTY720 was found to be >15 µM after 48 h (Figure 4a, red bars) and 72 h of treatment (Figure 4b, red bars), indicating that the prodrug Val-Cit-pABOC-FTY720 moiety had become three times less toxic than FTY720 itself. This implies that the side effects of free FTY720 can be mitigated by the prodrug design [42].

Cathepsin-B-Mediated Release of FTY720 from DOTA-PEG 3 -TZ(PEG 4 -Octr)-PEG 2 -Trz-PEG 3 -Val-Cit-pABOC-FTY720
To assess whether the construct was capable of controlled release of FTY720 upon cathepsin-B-catalyzed cleavage, a release assay was performed using a previously reported procedure [41]. After 1 h of incubation at 37 • C, a sample was removed from the mixture, and analysis with LC/MS showed a molecular weight of 366.0488 [M + 2H 2 O + Na] + , which represents the molecular weight of FTY720 (307.2511 g/mol), two water molecules, and a sodium ion (Figures S15 and S16). The cathepsin-B-mediated release of FTY720 was further confirmed with a control assay, where the peak of 366.0488 was absent when the experiment was conducted under the same conditions in the absence of cathepsin B ( Figure S17).

The Cytotoxicity of FTY720 Was Masked by Its Conjugation with the Cathepsin B Prodrug Linkage, Val-Cit
Cell viability assays were performed using different concentrations of Val-Cit-pABOC-FTY720 and free FTY720 to compare their cytotoxicity. To that end, four different concentrations of the two compounds were used (0, 1, 5, and 10 µM), and the cells were incubated for either 48 or 72 h. The results showed that the EC 50 for FTY720 was~5 µM after 48 h (Figure 3a, blue bars) and 72 h (Figure 3b, blue bars) of incubation, while the EC 50 for Val-Cit-pABOC-FTY720 was found to be >15 µM after 48 h (Figure 4a, red bars) and 72 h of treatment (Figure 4b, red bars), indicating that the prodrug Val-Cit-pABOC-FTY720 moiety had become three times less toxic than FTY720 itself. This implies that the side effects of free FTY720 can be mitigated by the prodrug design [42].
The same mice were imaged using [ 68 Ga]Ga-PSMA-11, and the results showed low uptake in both SSTR2 high PC3-VC and SSTR2 low PC3_sgPTP1B tumors (Figure 4a-right). In addition, [ 68 Ga]Ga-PSMA-11 failed to differentiate the two tumor phenotypes (Figure 4b-lower panel) that are implicated in the progression of prostate cancer to NEPC. Taken together, these results suggest the potential of using [ 68 Ga]Ga-DOTA-PEG 3 -TZ(PEG 4 -Octr)-PEG 2 -Trz-PEG 3 -Val-Cit-pABOC-FTY720 for noninvasive imaging detection of NEPC, and the T-SMPDC system could deliver FTY720 selectively to NEPC for targeted therapy. As shown in Figure 4c, the levels of SSTR2 expression in PC3_VC and PC3_sgPTP1B cells were assessed by Western blot (note: the original raw Western blots are presented in Figure S18). In addition, the excised PC3_VC and PC3_sgPTP1B tumors were confirmed to express SSTR2 high and SSTR2 low (Figure 4d

Discussion
Targeted therapy serves as the foundation of precision medicine. According to the latest information provided by the National Cancer Institute [43], most targeted cancer therapies currently utilize either small-molecule drugs or monoclonal antibodies. Small molecule drugs are typically developed for intracellular targets, as they are small and can be readily engineered to cross the plasma membrane. As such, for actions on molecular targets that are highly upregulated in cancer cells, targeted cancer therapies have demonstrated significant progression-free survival (PFS) benefits over conventional chemotherapies, along with reduced off-target toxicity and side effects [44]. However, many of the molecular targets or pathways that have been explored for targeted therapies are not absent in normal cells; instead, they are tightly regulated, such as Sphk1. To achieve the goal of NEPCtargeted delivery and controlled intracellular release of an SphK1 inhibitor to suppress the function of Sphk1, which is implicated in the progression of CRPC to NEPC [20,35,45], our design strategy of T-SMPDC takes advantage of the chemistry of 2,4,6-trichloro-1,3,5triazine for modular assembly of different functional components.
Similar to the synthesis of our recently reported T-SMPDC [20], the modular assembly on the trifunctional platform started from stepwise chemical modifications of each functional component. To conjugate the prodrug moiety, Val-Cit-pABOC-FTY720, we found that it was necessary to add an extra spacer between the prodrug moiety and the 1,3,5triazine core to minimize the steric hindrance, thereby facilitating the reaction. Among the many reported strategies available, we took a click chemistry route by functionalizing 2,4,6-trichloro-1,3,5-triazine with propargyl-PEG2-amine and adding an azido-PEG to Val-Cit-pABOC-FTY720. The click chemistry reaction was straightforward to form TZ-PEG 2 -Trz-PEG 3 -Val-Cit-pABOC-FTY720. In addition, it minimized the potential interferences of various functional groups in following the modular assembly of the T-SMPDC. The yields of the four major synthetic steps towards DOTA-PEG 3 -TZ(PEG 4 -Octr)-PEG 2 -Trz-PEG 3 -Val-Cit-pABOC-FTY720 were~45% (for compound 10),~54% (compound 11),~11% (compound 12), and~33% (compound 13), which are higher than what we reported for the synthesis of our first T-SMPDC [20]. This demonstrates the synthetic capability that we have developed to scale up the synthesis of various T-SMPDC systems when translational studies are warranted to proceed.
The radiolabeling of DOTA-PEG 3 -TZ(PEG 4 -Octr)-PEG 2 -Trz-PEG 3 -Val-Cit-pABOC-FTY720 with 68 Ga or 64 Cu was straightforward. However, the radiochemical yield was low, because it was calculated based on the radioactivity of 68 Ga or 64 Cu added to a fixed amount of the conjugate. Under these radiochemical conditions, we were able to achieve an optimal molar activity (38 GBq/µmol) for [ 68 Ga]Ga-DOTA-PEG 3 -TZ(PEG 4 -Octr)-PEG 2 -Trz-PEG 3 -Val-Cit-pABOC-FTY720 to be used for in vivo evaluations in tumor-bearing mouse models, where the highest achievable molar activity is desired. However, we acknowledge that the radiochemical reaction conditions can be further optimized to improve the radiochemical yield as well as the molar activity. It should be noted that the incorporation of DOTA into the T-SMPDC can also potentiate the use of the prodrug conjugate systems for developing targeted radiotheranostic agents when loaded with βor α-emitters.
To achieve NEPC-targeted delivery and controlled release of an intracellular target therapy, we chose an octreotide molecule to target SSTR2, whose expression is arguably elevated as CRPC progresses to NEPC. In addition, SSTR2 can mediate the desired internalization of the T-SMPDC upon its ligand binding of octreotide. Although debates on the correlation of SSTR2 with NEPC exist in the literature, we believe that the main reason perhaps arises from the heterogeneous and dynamic nature of the innervated tumor microenvironment (TME) [23], which makes biopsy-based IHC results either inaccurate or unrepresentative [46]. The choice of SSTR2 as the target for our T-SMPDC design does not reflect our position in these debates; instead, it reflects the fact that the innervated TME remains to be further explored. Even if it is ultimately confirmed that SSTR2 is not among the specific markers of NEPC, the design concept of T-SMPDC and its modular synthesis presented in this work can be readily adapted to accommodate the rapid advances in the field of cancer innervation.
The proof-of-concept biological studies demonstrated that the T-SMPDC can initiate a substantial extent (~40%) of SSTR2-mediated internalization for intracellular delivery of the whole conjugate. The prodrug design was able to release FTY720 intact via cathepsin B. Importantly, the cytotoxicity of FTY720 can be masked when conjugated to mitigate the potential side effects of using FTY720 for cancer therapy.
The two tumor models used in the small animal imaging evaluation were established by wild-type PC3 cells with high expression of SSTR2, and by PTP1B-silenced PC3 cells (PC3_sgPTP1B) with low expression of SSTR2. It should be noted that PTP1B, protein tyrosine phosphatase non-receptor 1, is a soluble endoplasmic-reticulum-associated tyrosinespecific phosphatase. It plays a well-established role in insulin signaling [47,48]. To date, it has been considered as a therapeutic target for type 2 diabetes and cancer [24,25]. The dynamic PET imaging with [ 68 Ga]Ga-DOTA-PEG 3 -TZ(PEG 4 -Octr)-PEG 2 -Trz-PEG 3 -Val-Cit-pABOC-FTY720 clearly showed that the radiolabeled agent was capable of differentiating the two phenotypes of PC3 xenografts by their distinct expressions of SSTR2, further validating our design concept of the T-SMPDC. Encouragingly, the T-SMPDC displayed a desirable in vivo kinetic profile, with clearance from the kidneys. With regards to the use of the T-SMPDC for therapeutic evaluation, we can further optimize its structure by varying the linkages to enhance the tumor uptake and retention of the conjugate, as well as improving the contrast ratios of target/non-target tissues.
To date, we have witnessed the paradigm-shifting roles of PSMA-targeted radiopharmaceuticals in prostate cancer patient care, as represented by [ 68 Ga]Ga-PSMA-11 and [ 18 F]F-DCFPyl for PET imaging and [ 177 Lu]Lu-PSMA-617 for radiotherapy [6]. However, not every prostate cancer phenotype has PSMA expression [49]. Indeed, during the progression of CRPC to NEPC, the expression of PSMA becomes attenuated, limiting the benefits of PSMA-targeted radiopharmaceuticals to the patient populations with non-PSMA-expressing tumors or metastases. As clearly demonstrated in our small animal imaging evaluation (Figure 4), [ 68 Ga]Ga-PSMA-11 failed to detect either SSTR2 high PC3-VC or SSTR2 low PC3_sgPTP1B tumors, due to their lack of PSMA expression. This fact necessitates the development of radio-and chemotheranostic agents by targeting other molecular mechanisms, such as the T-SMPDC presented in this work. Indeed, many other genes and proteins have been reported in association with NEPC [50], and recently we have validated the synaptic vesicle glycoprotein 2 isoform A (SV2A) [23] as a marker of NEPC.
It is noteworthy that FTY720 is an Sphk1 inhibitor that is approved by the United States Food and Drug Administration (US-FDA) for the treatment of multiple sclerosis. Repurposing of FTY720 for cancer therapies is now only reported in preclinical studies [24,51]. It remains to be seen whether SphK1 could serve as a practical target to develop NEPC therapies. However, we are not limited to Sphk1. As new druggable intracellular targets will certainly emerge in future, we will be able to readily incorporate them into the modular design of our T-SMPDC system.

Conclusions
We successfully designed, synthesized, and evaluated a T-SMPDC system, DOTA-PEG 3 -TZ(PEG 4 -Octr)-PEG 2 -Trz-PEG 3 -Val-Cit-pABOC-FTY720, for NEPC-targeted delivery and controlled intracellular release of an Sphk1 inhibitor to suppress the function of Sphk1, which is implicated in the progression of CRPC to NEPC. The T-SMPDC system offers advantages over ADCs in terms of in vivo pharmacokinetics, solid tumor penetration, definitive chemical structure, and ease of synthesis. In addition, the modular assembly feature of our accomplished synthesis allows a rapid adaption of the T-SMPDC platform to target other intracellular oncotargets or to deliver other potent chemotherapy molecules as necessary.