Modular Smart Molecules for PSMA-Targeted Chemotherapy

New targeted chemotherapeutics are urgently needed to minimize off-target toxicity and reduce the high mortality rate associated with metastatic prostate cancer (PCa). Herein, we report on the modular synthesis, pharmacokinetics, and efficacy of two small-molecule drug-conjugates (SMDCs) targeted to prostate-specific membrane antigen (PSMA) incorporating either: (1) a cathepsin-B cleavable valine-citrulline, or (2) an acid-cleavable phosphoramidate linker. Crucial components utilized in the design of the conjugates include: (1) CTT1298, a nanomolar affinity ligand that binds irreversibly to PSMA and has proven in past studies to rapidly internalize and shuttle payloads into PSMA-expressing PCa cells, (2) MMAE, a known potent cytotoxic payload, and (3) an albumin-binder, proven to improve residence time of drug conjugates. At dose of 0.8 mg/kg (~250 nmol/kg), the two SMDCs showed significant efficacy in a PSMA(+)


INTRODUCTION
Prostate cancer (PCa) is the most diagnosed cancer in American men and one of the major causes of cancer-related deaths, second only to lung cancers. 1 Generally, patients with localized disease can be treated with radical prostatectomy and/or radiation therapy. 2 Patients with metastatic PCa can be temporarily treated with androgen deprivation strategies; 3 however, many progress to metastatic disease, which can evolve into castrationresistant cancer that does not respond to androgen deprivation therapy (mCRPC). Only 26-30% of patients with mCRPC survive 5 years. 1,4,5 Chemotherapy (e.g., docetaxel, cabazitaxel) is the standard treatment given following anti-androgen therapy failure, which often results in marginal responses accompanied by off-target toxicity. [6][7][8] Improved therapeutic approaches are critically needed for late-stage PCa patients that both improve efficacy and minimize toxicity.
Prostate-specific membrane antigen (PSMA) is a type II transmembrane protein that is highly overexpressed in the majority of all prostate cancers. 9,10 PSMA expression is more highly upregulated in poorly differentiated, metastatic, hormone-refractory carcinomas, and in cancer cells from mCRPC patients. [9][10][11][12][13] Furthermore, PSMA is constitutively internalized from the cell surface 14,15 making it an ideal target for imaging and therapy. Indeed, PSMA-targeted radiopharmaceutical therapies have shown considerable promise in clinical trials, [16][17][18][19] with 177 Lu-PSMA-617 recently receiving FDA breakthrough therapy designation. 20 Several other PSMA-targeted radioligand therapies and antibody-drug conjugate (ADCs) [21][22][23][24] have also demonstrated promising efficacy in preclinical studies. However, problems associated with relapse, off-target toxicity, immunogenicity of antibodybased therapies, and the stringent requirements imposed on the facilities that produce and manage radiopharmaceutical production represent significant challenges for drug development, FDA-approval, and use in the oncology community for these classes of agents to treat mCRPC. accurate localization to the prostate cancer lesions; however, the considerably lower molecular weight and simpler molecular characterization will be associated with lower manufacturing costs, flexibility in determining the optimal dose regimen, and higher tolerated doses. [32][33][34] Importantly, the toxicity profile of PSMA-SMDCs should be much lower than ADCs, as they have a shorter residence time and undergo more rapid and uniform diffusion into the tumor mass compared to normal organs such as kidneys, lacrimal glands, and salivary glands. 33,34 In this proof-of-concept study, we describe the design, efficacy and pharmacokinetics of two novel PSMA-targeted SMDCs for treatment of mCRPC in mouse model ( Figure 1).

SMDC Synthesis and Characterization.
All starting materials and reagents were of commercial quality and were used without further purification. All anhydrous solvents used in reactions were obtained from commercial sources or freshly distilled over calcium hydride. 1 H, 13 C, and 31 P NMR spectra were recorded on a Varian 400, Brucker Avance Neo 500, or Varian 600 MHz spectrometer. 1 H NMR chemical shifts are relative to CDCl 3 (δ = 7.26 ppm), CD 3 OD (δ = 3.31 ppm) or D 2 O (δ = 4.79 ppm). 13 C NMR chemical shifts were relative to CDCl 3 (δ = 77.23 ppm) or CD 3 OD (δ = 49.15 ppm). High-resolution mass spectrometry (HRMS) spectra were obtained on an Applied Biosystems 4800 MALDI-TOF/TOF mass spectrometer (Applied Biosystems, Foster City, CA). CTT1700 (Figure 1) was obtained from Cancer Targeted Technology. The synthesis of SMDCs 1 and 2 is outlined in Figure 2. The synthesis of azides 10 and 11 are outlined in the Supporting Information and from a previously published study, respectively. 35 (5).-A solution of acid, 4 (2 g, 6.59 mmol) and HBTU (2.5 g, 6.59 mmol) was stirred in anhydrous DMF (11 mL) under Ar at room temperature for 30 min. A solution of 3 (2.6 g, 7.91 mmol) and anhydrous Et 3 N (2.5 mL, 16.5 mmol) in DMF (11 mL) was added to the reaction solution, and the reaction was stirred for 3 h under Ar. The reaction was diluted with EtOAc (150 mL) and washed sequentially with 1 N HCl (75 mL, 3x), sat. NaHCO 3 (aq) (75 mL, 2x) and Brine (75 mL, 1x). Organic layer was dried on Na 2 SO 4 , volatiles were removed under reduced pressure, resulting residue was dried under high vacuum and carried on to the next step without further purification. Isolated yield was 93%. 1

Author Manuscript
Author Manuscript

Author Manuscript
Author Manuscript ( Standard curves were linear from 10-10,000 ng/mL for CTT1700, 2 and 1 while the MMAE standard curves were linear from 0.01 to 10 ng/mL. A 1/y 2 weighted linear regression was used for all standard curves. PK parameters were estimated using standard noncompartmental methods with Phoenix WinNonlin (Certara, Princeton, NJ). Plasma parent drugs were derived using the IV bolus approach, while all other data were derived using the extravascular approach.

In-vivo efficacy
To evaluate efficacy of SMDCs 1 and 2 in a mouse model of human PCa, we employed PSMA+ PC3-PIP tumor cells implanted subcutaneously in the right rear flank of NCr nude mice (6 -8 weeks old). Three weeks after injection of the cells, when the tumor volume reached an average tumor size of ~130 mm 3 , mice were treated with 0.8 mg/kg of each SMDC (253 nmol/kg of SMDC 1; 244 nmol/kg of SMDC 2; 265 nmol/kg of CTT1700; n = 8 for each SMDC) via lateral tail vein injection (100 μL bolus) once weekly for six weeks. Tumor sizes were measured twice weekly for the duration of the study and mice were monitored for 90 days after first administration. Mice were euthanized if the tumor size or mouse health reached IACUC standards for euthanasia (tumor burden of >1.5 cm in any axis, ulceration, >20% mouse weight loss). Blood draws were taken via superficial temporal vein at 0, 4, 7 and 10 weeks, and complete blood count (CBC) analysis was performed with a 3-part differential measurement.

Data Availability
The data generated in this study are available within the article and its supplementary data files.
Olatunji et al.

Page 7
Mol Cancer Ther. Author manuscript.

Author Manuscript
Author Manuscript

Author Manuscript
Author Manuscript

SMDC Synthesis
The synthesis of SMDCs 1 and 2 were achieved through copper-free click reactions ( Figure  2). The key common precursor to both 1 and 2 is compound 9, which contains the strained dibenzocyclooctyne for click chemistry, CTT1298 as the targeting molecule, and the albumin-binding moiety into its molecular structure. Azides 10 and 11 incorporate the Val-Cit and phosphoramidate linkers, respectively, in addition to the cytotoxic payload. The click-chemistry reactions were monitored by HPLC and found to be complete in less than 30 minutes. SMDCs 1 and 2 were purified by preparative C18-HPLC and further desalted to give isolated yields of 30% and 49%, respectively; they were fully characterized by HRMS, HPLC and 31 P NMR (see Supporting Information). Once characterized, IC 50 values for PSMA binding were determined as described previously and found to be in the nanomolar range for 1 (IC 50 = 37.9 ± 7.5 nM) and 2 (IC 50 = 1.67 ± 0.14 nM), consistent with other PSMA inhibitors of this class. CTT1700 was generously provided by Cancer Targeted Technology.

Pharmacokinetic studies
Plasma and tumor levels of parent compounds SMDC 1, SMDC 2, and CTT1700, along with released MMAE are shown in Figure 3 and Table S1. The plasma levels of SMDCs and released MMAE give an indication of overall exposure, while the levels in the kidneys and lacrimal gland (see Figure 4) represent the off-target organs typically affected negatively by PSMA-targeted agents. 36 Parent SMDCs 1 and 2 exhibited a distribution volume of 0.106 and 0.122 L/kg, respectively, a value in between plasma volume and extracellular water volume . This is indicative of the albumin binding motif enhancing half-life and retention of the SMDCs in plasma. Released MMAE was undetectable in the plasma of mice administered with compound 1, likely due to the more rapid clearance of this compound compared to CTT1700 and 2 ( Figure 3A, B, and C). In the tumor, the concentration of MMAE released from 2 increased between days 3 and 5, whereas MMAE from 1 leveled off at 3 days ( Figure 3E and F). CTT1700, which employs a linker that slowly releases MMAE, reaches a concentration in the tumor that is approximately 5-fold lower than that of 1 or 2 ( Figure 3D). MMAE concentrations upon release from CTT1700, 1 and 2 were comparable in prostate, kidneys, and lacrimal glands and were all 2-3 orders of magnitude lower than that in the tumor ( Figure 4A-I).
In addition, we investigated the stabilities of all three SMDCs in mouse plasma at 37°C for 24 h (see Figure S1). No degradation was observed for CTT1700 and SMDC 2. In sharp contrast, SMDC 2 showed rapid degradation following apparent first-order kinetics with a half-life of 3.5 h and approximately 60% of associated MMAE released was observed over 24 h.

In-vivo efficacy.
Tumor growth and survival data for mice administered with controls and SMDCs 1 and 2 are shown in Figure 5 and S2. Animals in the saline and MMAE control groups reached tumor size end point criteria within 6 weeks of their first administration ( Figure 5A, B).
This was expected for MMAE administration, since the cytotoxic agent is known to be non-specific, with minimal accumulation in the tumor. Animals that were treated with CTT1700, a PSMA-targeted SMDC with a slowly cleavable phosphoramidate linker due to steric congestion around the labile P-N bond, showed moderate efficacy, with all animals reaching end point tumor burden within 12 weeks of dosing initiation ( Figure 5C). In sharp contrast, SMDCs 1 and 2 showed considerable efficacy compared to the saline and MMAE controls (SMDC 1: p < 0.02, SMDC 2: p < 0.0001). For animals treated with SMDC 1, tumor regression was observed following initiation of administration up until 2 weeks after the last dose, with subsequent tumor growth in five of the eight mice. Five of eight mice survived 90 days after initial dosing ( Figure 5D). Animals treated with SMDC 2 showed durable tumor growth regression, with no detectable tumor in any of the mice by day 35 of first dosage ( Figure 3E), and growth in four of seven mice commencing five weeks after the last drug administration. The survival of mice in the SMDC 2 group were statistically greater than the other groups, including SMDC 1 (p < 0.0001), with all animals surviving at 90 days ( Figure 3F). It is important to note that in all groups of mice, body weights were stable or increased over the duration of the study, there were no overt visual signs of toxicity (scruffy coat, diarrhea, lethargy), and complete blood count (CBC) results were all in the normal range (see Supporting Information, Figures S3-S15).

DISCUSSION
Herein, we describe the development of two PSMA-targeted SMDCs (1 and 2; Figure 1A), distinguishable by their cleavable linkers, and assessment of their in vivo pharmacokinetics and efficacy in mice bearing PSMA(+) tumor xenografts. Compound 1 contains a valinecitrulline (Val-Cit) dipeptide linker, which is currently a component of FDA-approved ADC drugs 37-41 (e.g., Adcetris® and Polivy®). This linker is cleavable by lysosomeabundant cathepsin B proteases to release the cytotoxic payload. 42 In compound 2, we used a novel acid-labile phosphoramidate linker recently developed in our lab. 35,43 This linker exhibits stability at neutral pH in plasma, with rapid release of payloads under the acidic conditions of endosomal and lysosomal compartments. Monomethyl auristatin E (MMAE), an anti-tubulin agent that kills cells at sub-nanomolar concentrations by inhibiting the assembly of microtubules followed by subsequent induction of mitotic arrest, was used as the cytotoxic payload for both SMDCs. CTT1700, a compound provided by Cancer Targeted Technology, employs a linker that slowly releases MMAE. 44 The phosphoramidate-based inhibitor, CTT1298, was selected as the PSMA-targeting molecule based upon its documented high-affinity, irreversible binding to PSMA (IC 50 = 19 nM) and high selectivity for PSMA-expressing cancer cells. 45 Once bound to extracellular PSMA, CTT1298 derivatized with radiolabeled pendant groups has been shown to undergo rapid and extensive internalization in PSMA-expressing tumor cells. 45 Moreover, conjugation of chemical moieties to CTT1298 does not diminish the binding affinity of this ligand to PSMA. 36,45 One of the major pharmacokinetic challenges for many of the small-molecule PSMA inhibitors with respect to delivering therapeutic payloads is their rapid renal clearance. To overcome this, 46 we outfitted the two SMDCs with an albumin-binding motif 47 known to reduce blood clearance while increasing tumor uptake of drug-conjugates. 36,48 Generally, the observed plasma MMAE and parent drug profiles exhibited parallel and roughly monoexponential profiles. Furthermore, we observed drug-released MMAE elimination half-lives (t 1/2 ) from plasma of 13 h and 24 h for SMDC 1 and 2 respectively, much higher than the t 1/2 of 4.1 h reported for MMAE when dosed directly. 49 These observations are indicative of formation rate limited elimination, with the pharmacokinetics (PK) of the parent drug determining and prolonging MMAE exposure.
Plasma concentrations of the MMAE were found to be 5 orders of magnitude lower than the parent drugs indicating low systemic release of MMAE payload. In tumors, concentrations of the parent drugs (1, 2, and CTT1700) were roughly in the same order of magnitude suggesting similar efficiency in targeting, while MMAE concentrations varied between compounds according to linker stability.
We observed the slow decline of CTT1700 and 2 in all tissues over the course of 5 days. Surprisingly, SMDC 1 could not be quantified in any other tissue other than the tumor. This anomaly can be partly attributed to the reported low hydrolytic stability of valine-citrulline linkers in mouse plasma 50,51 and the rapid clearance and distribution of released-MMAE to highly perfused tissues like the liver, heart, lungs, and spleen. 49  The reported plasma clearance of MMAE of 60 mL/h (corresponds to 47 L/d/kg), allows us to use the apparent MMAE formation clearance we observed in our data to calculate the fraction of parent compound that is converted to MMAE. The fraction of drug that is estimated to release MMAE for CTT1700, SMDCs 1 and 2 was 1%, 38%, and 36%, respectively. This 40-fold difference in MMAE-releasing ability between CTT1700 relative to 1 and 2 was reflected in MMAE tumor concentrations, arguably, the most important metric for interpreting our efficacy data.
Based on the efficacy data and CBC analyses ( Figure S3-15), the dose regimen uniformly utilized for this comparative study is below the maximum tolerated dose (MTD) in immunecompromised nude mice. Other PSMA-targeted MMAE-loaded SMDCs with the Val-Cit linker have been reported in the literature and investigated in the PC3-PIP model. 11,52 Although direct comparisons are not possible as the doses and regimens are different, agents 1 and 2 compare favorably with lower 11 and fewer 52 doses to achieve tumor growth inhibition and survival out to 90 days. A highly significant finding is that SMDC 2, with the pH sensitive linker, is more effective than SMDC 1 that has the Val-Cit linker similar to previously published PSMA-targeted SMDCs. In addition, the improved performance of SMDC 2 could also, in part, be attributed to its 15-fold greater binding to PSMA compared to SMDC 1.
In summary, the pharmacokinetics and properties of the pH-responsive linker in 2 can be correlated with its superior efficacy when compared to the more stable phosphoramidate linkage in CTT1700. 35 The data indicate that efficient release of the payload from 2 reflects an optimal design of its versatile acid-labile linker. In addition, the in vivo performance of SMDC 2 compared to that of SMDC 1 demonstrates that the acid-labile phosphoramidate linker produced significantly greater efficacy than the Val-Cit linker that is prevalent in ADCs and recently reported SMDCs 11,48 . Although the efficacy and biodistribution experiments conducted are proof-of-concept studies, the results obtained warrant further exploration of the novel acid-labile phosphoramidate linker technology in the context of the controlled release of other clinically relevant payloads, as well as additional rigorous preclinical studies prior to advancement into the clinic.

Supplementary Material
Refer to Web version on PubMed Central for supplementary material.