<<

PSMA-1- CONJUGATES FOR TARGETED THERAPY OF PROSTATE

by

NATALIE WALKER

Submitted in partial fulfillment of the requirements for the degree

of Master of Science

Biomedical Engineering

CASE WESTERN RESERVE UNIVERSITY

May, 2019 CASE WESTERN RESERVE UNIVERSITY SCHOOL OF GRADUATE STUDIES

We hereby approve the thesis of

Natalie Walker

candidate for the degree of Master of Science.

Committee Chair

Efstathios Karathanasis, PhD

Committee Members

James Basilion, PhD, Research Advisor

Christopher Hoimes, DO

Xinning Wang, PhD

Date of Defense

14 January 2019

*We also certify that written approval has been obtained for any proprietary material contained therin. Contents

List of Tables ...... iv

List of Figures ...... v

Abstract ...... 1

Introduction ...... 2

Table 1: Review of PSMA Expression in Nonprostate Malignancies...... 5

Figure 1: Outline of Structures of Three PSMA-1-Doxorubicin Prodrug

Conjugates...... 8

Figure 2: Proposed Mechanism for Prodrug Release of Free Doxorubicin ...... 8

Methods...... 9

Materials ...... 9

General Comments...... 9

Synthesis of PSMA-1(Cys) (Glu-CO-Glu’-Amc-Ahx-Glu-Glu-Glu-Cys-NH2)...... 9

Synthesis of PSMA-1-Doxorubicin Conjugates ...... 11

Synthesis of PSMA-1-SMCC-Dox ...... 11

Synthesis of PSMA-1-MC-Val-Cit-PABC-Dox ...... 12

Synthesis of PSMA-1-MMCCH-Dox ...... 13

Cell Culture ...... 13

In Vitro Cellular Uptake Studies ...... 13

i

Cytotoxicity Assay ...... 14

Mouse Tumor Xenograft Studies ...... 15

Statistical Analysis ...... 15

Results ...... 16

Chemistry ...... 16

Figure 3: Structure and Characterization of PSMA-1(Cys) ...... 17

Figure 4: Synthesis and Characterization of SMCC-Dox ...... 18

Figure 5: Synthesis and Characterization of Non-cleavable PSMA-1-SMCC-Dox

...... 19

Figure 6: Synthesis and Characterization of MC-Val-Cit-PABC-Dox...... 20

Figure 7: Synthesis and Characterization of Cathepsin-Cleavable PSMA-1-MC-

Val-Cit-PABC-Dox...... 21

Figure 8: Synthesis and Characterization of MMCCH-Dox...... 23

Figure 9: Synthesis and Characterization of Acid-Labile PSMA-1-MMCCH-Dox

...... 24

Figure 10: Absorption Spectra of Free Doxorubicin and PSMA-Targeted

Conjugates...... 25

In Vitro Cellular Uptake Studies ...... 25

Figure 11: In Vitro Drug Uptake and Cellular Localization by Fluorescence

Microscopy ...... 26

ii

In Vitro Cytotoxicity ...... 27

Figure 12: 48 Hour Cytotoxicity of Free and PSMA-Targeted Doxorubicin ...... 28

Figure 13: Cytotoxicity of Dox and PSMA-MMCCH-Dox, Alternate Conditions

...... 29

Mouse Tumor Xenograft Response ...... 30

Figure 14: Tumor Response and Mouse Weight ...... 30

Discussion ...... 31

Figure 15: Test of PSMA-1-MMCCH-Dox Cleavage in Cell Culture Media ...... 36

Acknowledgements ...... 41

Bibliography ...... 42

iii

List of Tables

Table 1: Review of PSMA Expression in Nonprostate Malignancies...... 5

iv

List of Figures

Figure 1: Outline of Structures of Three PSMA-1-Doxorubicin Prodrug Conjugates ...... 8

Figure 2: Proposed Mechanism for Prodrug Release of Free Doxorubicin ...... 8

Figure 3: Structure and Characterization of PSMA-1(Cys) ...... 17

Figure 4: Synthesis and Characterization of SMCC-Dox ...... 18

Figure 5: Synthesis and Characterization of Non-cleavable PSMA-1-SMCC-Dox ...... 19

Figure 6: Synthesis and Characterization of MC-Val-Cit-PABC-Dox...... 20

Figure 7: Synthesis and Characterization of Cathepsin-Cleavable PSMA-1-MC-Val-Cit-

PABC-Dox ...... 21

Figure 8: Synthesis and Characterization of MMCCH-Dox...... 23

Figure 9: Synthesis and Characterization of Acid-Labile PSMA-1-MMCCH-Dox ...... 24

Figure 10: Absorption Spectra of Free Doxorubicin and PSMA-Targeted Conjugates ... 25

Figure 11: In Vitro Drug Uptake and Cellular Localization by Fluorescence Microscopy

...... 26

Figure 12: 48 Hour Cytotoxicity of Free and PSMA-Targeted Doxorubicin ...... 28

Figure 13: Cytotoxicity of Dox and PSMA-MMCCH-Dox, Alternate Conditions ...... 29

Figure 14: Tumor Response and Mouse Weight ...... 30

Figure 15: Test of PSMA-1-MMCCH-Dox Cleavage in Cell Culture Media ...... 36

v

PSMA-1-Doxorubicin Conjugates for Targeted Therapy of

Abstract

by

NATALIE WALKER

Metastatic-castration resistant prostate cancer poses a serious clinical problem with poor outcomes, and targeted treatments that would widen the therapeutic window are desired.

Prostate-specific membrane antigen (PSMA) is a membrane-bound glycoprotein that is overexpressed in prostate cancer and found in the neovasculature of many other tumors.

Our previously developed ligand, PSMA-1, has affinity in the nM range. The aim of this study is to synthesize and evaluate three PSMA-1-Doxorubicin conjugates using different chemical linkers: non-cleavable SMCC, protease-cleavable MC-Val-Cit-PABC-PNP, and the acid-labile hydrazone MMCCH. PSMA-1-Doxorubicin conjugates are selective for

PSMA(+) cell lines by fluorescence microscopy, but only PSMA-1-MMCCH-Dox showed evidence of linker cleavage, doxorubicin release to the nucleus of the cell, and cytotoxic activity. It was also more effective and less toxic than free doxorubicin in a flank tumor mouse model. In the future, these targeted agents may aid in treatment of patients with prostate cancer or other PSMA-expressing tumors.

1

Introduction

Prostate cancer is the third most common non-skin cancer in the United States,

after breast and lung . In 2018, there were an estimated 164,690 new cases

diagnosed and 29,430 deaths, up from 161,360 and 26,730 respectively in 2017 (1).

Screening with serum prostate specific antigen (PSA) allows for 78% of prostate cancers to be diagnosed at the localized stage, facilitating therapy with radical prostatectomy or

radiation therapy (1) (2) (3). Patients with metastatic disease, however, may see initial

response to suppression of gonadal testosterone, but the vast majority of them progress to

metastatic castration-resistant prostate cancer (mCRPC), which remains a deadly disease.

Several agents have demonstrated survival benefit for mCRPC including combination

with , sipuleucel-T, cabazitaxel with prednisone, arbiraterone with

prednisone, enzalutamide, and radium-223. However, in randomized controlled trials,

generally comparing the treatment group to placebo or , survival benefit for

all of these agents was less than four months (2). Median survival on combination docetaxel/prednisone survival is only 18.9 months (4). New molecularly targeted therapies are urgently needed.

Prostate-specific membrane antigen (PSMA) is an 84 kDa, class II transmembrane glycoprotein with hydrolyzing and endocytic functions, first characterized by Horoszewicz in 1986 (5). PSMA is highly overexpressed in prostate cancer compared with benign prostate tissue, on the order of 100 to 1000-fold increased

(6). High expression correlates with tumor grade, pathological stage, aneuploidy, and biochemical recurrence (7). PSMA expression has also long been described in the

2

neovasculature of many different nonprostatic malignancies, as outlined in Table 1

below. Additionally, multiple groups have more recently reported on PSMA expression

in the tumor cells in a subset of endometrial carcinoma, primary ovarian carcinoma, and

non-small cell lung cancer. The majority of these studies are based on histopathological

staining for PSMA expression, but with the rise of PSMA-PET, case reports of PSMA

expression detected incidentally in nonprostatic cancers via imaging have begun to

circulate, too. As with prostate cancer, PSMA expression correlated with worse outcomes

for several other tumors, including pancreatic, oral cavity squamous cell carcinoma, and osteosarcoma (8) (9) (10). In normal tissues, detectable PSMA expression is limited to prostate epithelium, proximal renal tubules, and duodenum (11). PSMA is therefore an attractive biomarker for targeting therapies and imaging agents for prostate cancer, but

also potentially for a wide array of neoplasms with poor prognosis.

Percentage Samples Expressing Sample Tumor Type PSMA in Reference Number Neovasculature / Tumor Cells (if reported)

Wernicke, Arch Pathol Lab Med, 2011 Glioblastoma 100.00% 32 (12)

Wernicke, App Immunohistochem Mol Endometrial Carcinoma 100 % / 70% 23 Morph, 2017 (13)

Wernicke, App Immunohistochem Mol Primary Ovarian Carcinoma 100% / 48% 21 Morph, 2017 (13)

Ovarian Carcinoma Wernicke, App Immunohistochem Mol 100% / 8% 25 Metastases Morph, 2017 (13)

Breast Metastases to Brain 100% 14 Wernicke, APMIS, 2014 (14)

3

Malignant Melanoma 100% 5 Chang, Cancer Research, 1999 (15)

Bladder Urothelial 100% / 3% 96 Samplaski, Mod Pathology, 2011 (16) Carcinoma

Non-small Cell Lung Cancer 85% / 54% 87 Wang, Plos One, 2015 (17)

Cervical Squamous Cell Wernicke, App Immunohistochem Mol 85% / 0% 20 Carcinoma Morph, 2017 (13)

Colorectal 85% 130 Haffner, Hum Pathol, 2009 (18)

Pancreas 84% 147 Ren, Medical Oncology, 2014 (8)

Clear Cell Renal Cell 76% 21 Baccala, Urology, 2007 (19) Carcinoma

Squamous Cell Carcinoma 75% 96 Haffner, Mod Pathol, 2012 (9) of the Oral Cavity

Primary Breast 74% 92 Wernicke, APMIS, 2014 (14)

Small Cell Lung Cancer 70% / 0% 30 Wang, Plos One, 2015 (17)

Gastric 66% 119 Haffner, Hum Pathol, 2009 (18)

Thyroid 57% 63 Heitkotter, Oncotarget, 2018 (20)

Wernicke, App Immunohistochem Mol Cervical Adenocarcinoma 50% / 0% 8 Morph, 2017 (13)

Osteosarcoma 47% 45 Zeng, Med Oncol, 2012 (10)

Synovial Sarcoma 35% 16 Heitkotter, Oncotarget, 2017 (21)

Wernicke, App Immunohistochem Mol Vulvar Carcinoma 25% / 0% 20 Morph, 2017 (13)

Peripheral Nerve Sheath 19% 21 Heitkotter, Oncotarget, 2017 (21) Tumor

Rhabdomyosarcoma 15% 20 Heitkotter, Oncotarget, 2017 (21)

Case report of incidental Schwannoma Dias, Clin Nuc Med, 2018 (22) detection on PSMA-PET

4

Case report of Neck Squamous Cell incidental Lawhn-heath, Clin Nuc Med, 2017 (23) Carcinoma detection on PSMA-PET

Table 1: Review of PSMA Expression in Nonprostate Malignancies. PSMA expression is widespread in the neovasculature of solid tumors. More rarely, expression is also seen in the tumor cells themselves. Cancers presented here in order of decreasing percentage of tumor samples that expressed PSMA in the neovasculature.

Multiple PSMA-targeted therapies and diagnostics are under investigation.

Antibody-drug conjugates with the payloads monomethyl auristatin E (PSMA ADC) and

maytansinoid 1 (PSMA-DM1) are in clinical trials (24) (25). PSMA-PET imaging is

showing promising success in combination with multiparametric MRI for cancer

localization within the prostate or in combination with CT or MRI for metastatic disease

detection. The most widely studied agent is the 68Ga-labelled PSMA inhibitor Glu-NH-

CO-NH-Lys(Ahx)-HBED-CC (26).

Our group has developed a highly negatively-charged peptide-based PSMA

receptor ligand, PSMA-1, with affinity in the nM range. We have successfully used this

to selectively target and near-infrared intraoperative imaging agents for prostate cancer in preclinical studies (27) (28). This project aims to capitalize on the targeting ability of our PSMA-1 ligand, in order to improve the efficacy and widen

the therapeutic window of the commonly used chemotherapeutic agent, doxorubicin.

Doxorubicin is an attractive drug choice for PSMA-targeting for several reasons.

Though docetaxel is first line therapy for prostate cancer, doxorubicin has shown a benefit in reducing pain in the palliative setting for prostate cancer patients who are generally afflicted with osteoblastic bone metastases (29). It has been shown to have synergistic effects with docetaxel in combination therapy in vitro (30). It is effective

5 against many metastatic cancers including PSMA-expressing breast and ovarian cancers, a key point if this drug may eventually have use for cancers other than prostate (31). It fluoresces with a peak emission wavelength of about 600 nm and can therefore be easily followed in preclinical models (32) (33). Finally, its structure is amenable to conjugation reactions via an amino group and a carbonyl group off two of its six member rings (34).

Clinically, doxorubicin’s use is not optimal due to significant toxicities including cardiotoxicity and myelosuppression, effects which may be decreased with improved tumor-targeting (31) (35).

Doxorubicin works by intercalation of DNA base pairs, inhibiting topoisomerase

II and blocking DNA and RNA synthesis (31) (36). Jayaprakash et al. have previously synthesized a bioconjugate comprising doxorubicin and a different PSMA inhibitor. This retained high affinity for PSMA but exhibited poor in vitro cell killing (37). It is likely that the poor cytotoxicity was related to inactivation of the drug by conjugating it to the ligand. Therefore, in this application we will modify the conjugation chemistry and attach

PSMA-1 to doxorubicin with different linkages that can be cleaved either via proteases or the acidic environment found at/within most tumors (34) (38) (39). With this prodrug strategy, demonstrated in Figure 1 and Figure 2, doxorubicin can dissociate from PSMA-

1 inside of a cancer cell or in the extracellular milieu and exert its anti-proliferative effects freely. Further, we will exploit the inactivity of conjugated doxorubicin to decrease off target toxicity typically found with the free or liposome-incorporated drug.

In summary, we hypothesize PSMA-1-Doxorubicin conjugates will be selectively taken up by PSMA-expressing prostate cancer cells, and these drugs will exhibit equal or

6 superior antineoplastic effects in vitro and in vivo compared to free doxorubicin with minimal off-target toxicity.

Presented in Figure 1 below is our reaction schema, in brief. We use a series of three linkers to join the PSMA-1 ligand to doxorubicin:

1. Non-cleavable linker, SMCC (succinimidyl 4-(N-maleimidomethyl)cyclohexane-

1-carboxylate). Use of this linker does not yield a true prodrug as with the other

two linkers below.

2. Protease-cleavable linker, MC-Val-Cit-PABC-PNP (Maleimidocaproyl-L-Valine-

L-Citrulline-p-Aminobenzyl Alcohol p-Nitrophenyl Carbonate) (40). This is the

linker utilized in the FDA approved antibody-drug conjugate brentuximab vedotin

for treatment of Hodgkin’s disease, and consists of MC and PABC spacers

sandwiching the cathepsin B-sensitive dipeptide, valine-citrulline (41) (42).

Cathepsin B cleaves the valine-citrulline bond, the PABC spacer is self-imolative,

and intact doxorubicin in its native state should be released.

3. Acid-labile hydrazone linker, MMCCH (4-(4-maleimidomethyl)cyclohexane-1-

carboxyl hydrazide) (43). The hydrozone bond cleaves releasing doxorubicin in

its native state.

7

Figure 1: Outline of Structures of Three PSMA-1-Doxorubicin Prodrug Conjugates

Figure 2: Proposed Mechanism for Prodrug Release of Free Doxorubicin 1.) PSMA-1 ligand will bind with extracellular domain of PSMA on prostate cancer cell. 2.) Endocytosis of the PSMA- 1-Doxorubicin conjugates will expose prodrug to endosomal/lysosomal environment where either cathepsin B (represented with red X) or the low pH will promote breakage of the linker joining the PSMA-1 and doxorubicin. In the tumor microenvironment, secreted cathepsins and a low pH may result in some extracellular cleavage, too. 3.) Free doxorubicin will accumulate in the nucleus of the cell, inhibit DNA synthesis, and promote cell death.

8

Methods

Materials

Fmoc rink amide MBHA resin and Fmoc amino acids were purchased from

Peptide International Inc, Novabiochem, and Advanced Biochemical Compounds.

Doxorubicin HCl was purchased from LC Laboratories. SMCC was purchased from

ThermoFisher. MC-Val-Cit-PABC-PNP was purchased from MedChem Express.

MMCCH TFA was purchased from Chem-Impex International Inc. Basic solvents and laboratory supplies were purchased from ThermoFisher and Sigma Aldrich.

General Comments

Preparative high-performance liquid chromatography (HPLC) was used for all

purifications, unless otherwise noted. A Shimadzu HPLC system was utilized with a

UV/visible detector monitored at 220 nm. A Luna 5µ C18(2) 100A column

(250mm×10mm×5µm, Phenomenex, Torrance, CA) was used at a flow rate of 2.5ml/min.

Gradients and buffers as noted below.

Synthesis of PSMA-1(Cys) (Glu-CO-Glu’-Amc-Ahx-Glu-Glu-Glu-Cys-NH2)

PSMA-1(Cys) was synthesized as PSMA-1 has been described before, with a

substitution of cysteine in place of the terminal lysine to provide the –SH group necessary

for reaction with the maleimide group of each linker (27). PSMA-1(Cys) was synthesized

manually using standard Fmoc chemistry. Generally, peptide was synthesized at 0.2

mmol scale starting from C-terminal Fmoc-rink amide MBHA resin.

9

Resin was allowed to soak in dimethylformamide (DMF) overnight before

beginning synthesis. Fmoc deprotection initially and for each coupling cycle was carried

out using 20% piperidine in DMF, applied to resin and shaken for 5 minutes. Coupling

reactions were carried out using 3.3 equivalents of Fmoc amino acids in DMF activated

with 3.3 equivalents of 2-(6-chloro-1H-benzotriazole-1-yl)-1,1,3,3-tetramethylaminium hexafluorophosphate (HCTU) and 5 equivalents of diisopropylethylamine (DIPEA) in

DMF, with this mixture applied to the resin and shaken for 30 minutes. These alternating steps (deprotection, new peptide, repeat) were used to build the peptide sequence Fmoc-

Glu'-Amc-Ahx-Glu-Glu-Glu-Cys(Mtt) on the resin.

The Fmoc group of N-terminal amino acid Glu' was deprotected by 20% piperidine. A chloroform solution containing 3 equivalents of H-Glu(OtBu)-OtBu mixed with 2.5 equivalents of DIPEA was prepared. The solution was then added slowly to 0.25 equivalents triphosgene in chloroform over 10 minutes at room temperature. After 15- minute incubation, the reaction blend was mixed with Glu'-Amc-Ahx-Glu-Glu-Glu-Lys on rink amide resin preswollen in chloroform with 2.5 equivalents of DIPEA. After the reaction was complete, the resin was washed with DMF and then dichloromethane and dried. The peptide was cleaved from resin by TFA/water/triisopropylsilane (950:25:25).

The cleaved peptide was purified by preparative high-performance liquid chromatography (HPLC). The gradient was 5-55% acetonitrile against 0.1% trifluoroacetic acid over 30 minutes, followed by 15 minutes of 55% acetonitrile. Peptide retention time was 21 minutes. Expected structures were confirmed by molecular weight using electrospray mass spectrometry (MS). Peptide was lyophilized and stored at -20C.

10

Synthesis of PSMA-1-Doxorubicin Conjugates

Generally, each linker was reacted with an excess of doxorubicin in DMF and

purified by HPLC, resulting in a maleimide-functionalized doxorubicin. Next, the

cysteine tail of the PSMA ligand was reacted with the maleimide group on the linker to

form the final conjugate. (Doxorubicin is light-sensitive and was handled with special

care to wrap tubes in aluminum foil to prevent light exposure.) The final conjugate was again purified by HPLC and lyophilized for storage at -20C. The structures at each reaction step were confirmed using either electrospray MS or MALDI-MS. Our electrospray MS machine generally produced cleaner images and was used when available. MALDI was used in other cases or if MW of the compound exceeded 2200, the maximum electrospray capacity. Our PSMA-targeted conjugates do make use of the modified PSMA-1(Cys), but for simplicity they will be referred to as PSMA-1-

Doxorubicin conjugates.

Synthesis of PSMA-1-SMCC-Dox

To produce the maleimide-functionalized doxorubicin, SMCC was dissolved in

DMF with the addition of triethylamine (TEA) to adjust the solution pH to approximately

8, measured roughly with pH paper. To this, 1.1 equivalents of doxorubicin HCl was added. The mixture was allowed to mix in the dark overnight at room temperature before purification by HPLC using a gradient of 30-90% acetonitrile against 0.1% trifluoroacetic acid over 40 minutes. Retention time for unreacted doxorubicin was 13 minutes, unreacted SMCC was 23.5 minutes, and the product SMCC-Dox was 24.5 minutes.

11

To produce PSMA-1-SMCC-Dox, SMCC-Dox was reacted with 1.1 equivalents

PSMA-1(Cys) in DMF with pH adjustment to 8 by TEA. The solution was allowed to

mix overnight in dark conditions at room temperature. The same HPLC gradient was

used resulting in a retention time for the final product of 17 minutes. Unreacted PSMA-

1(Cys) elutes with the solvent DMF. The reaction generally went to completion with no

unreacted SMCC-Dox.

Synthesis of PSMA-1-MC-Val-Cit-PABC-Dox

To produce the maleimide-functionalized doxorubicin, MC-Val-Cit-PABC-PNP

was reacted with 1.1 equivalents doxorubicin HCl and 1.1 equivalents DIPEA in DMF

(40). It was left to react on a spinner in the dark for two days at room temperature. This is

the one reaction that was purified without using HPLC. Instead, ~10x volume excess

CH2Cl2 was added to the reaction mixture and this was left to sit overnight at -20 to

precipitate the product MC-Val-Cit-PABC-Dox. The resultant solid was isolated by

centrifugation.

To produce PSMA-1-MC-Val-Cit-PABC-Dox, the MC-Val-Cit-PABC-Dox was reacted with 1.1 equivalents of PSMA-1(Cys) in DMF with pH adjusted to 7-8 using

TEA. This was allowed to spin in the dark overnight. The product was purified with

HPLC using a gradient of 30-90% acetonitrile against 0.1% trifluoroacetic acid over 40 minutes. Retention time was 17 minutes. Unreacted PSMA-1(Cys) elutes with the solvent

DMF. This reaction generally went to completion with no unreacted MC-Val-Cit-PABC-

Dox.

12

Synthesis of PSMA-1-MMCCH-Dox

To produce the maleimide-functionalized doxorubicin, MMCCH TFA was

reacted with 1.1 equivalents doxorubicin HCl in DMF at 50C for 2 hours in a dark fume hood (43). The product was purified with HPLC using a gradient of 30-90% acetonitrile against 2 mM TEAA. A basic buffer was chosen to avoid lysis of the acid-labile linkage.

Retention time of the uncreated MMCCH was 8 minutes, unreacted doxorubicin was 15 minutes, and MMCC-Dox was 19 minutes.

To produce PSMA-1-MMCCH-Dox, MMCCH-Dox was reacted with 1.1 equivalents PSMA-1(Cys) in DMF with pH adjusted to 8 using TEA. This was mixed overnight and the product purified using an HPLC gradient of 10-90% acetonitrile against

2 mM TEAA. Product retention time was 19 minutes. Unreacted PSMA-1(Cys) elutes with the solvent DMF. This reaction generally went to completion with no unreacted

MMCCH-Dox.

Cell Culture

Retrovirally transfected PSMA-positive PC3pip cells and control PC3flu cells were used for all experiments. Cells were cultured at 37C in a 5% CO2, humidified atmosphere. Cultures were maintained in RPMI-1640 medium supplemented with 10% fetal bovine serum (FBS).

In Vitro Cellular Uptake Studies

PC3Pip and PC3flu cells (1.5 x 105) were plated on coverslips and allowed to adhere for 36-48 hours. Cells were incubated in 1 µM of free doxorubicin or the various

PSMA-1-Doxorubicin conjugates in RPMI media. Similar experiments are reported

13

widely in the literature, with doses usually in the µM range because of doxorubicin’s relatively dim fluorescence, at least compared with some of the commercially available dyes used in our lab (44) (45). This dose was chosen to balance that relatively dim fluorescence with a goal of attempting to maintain selectivity.

The PSMA-1-Doxorubicin conjugates were not soluble in water. Stock solutions were prepared in dimethyl sulfoxide, DMSO, for short term storage. Experimental dilutions were prepared in RPMI media. An extinction coefficient of 10410 liter/(mol cm) was used, as reported in the literature, for dilution calculations using the Beer-

Lambert Law after absorption was measured at 480 nm (46).

Next, cells were washed 3 times with RPMI media, fixed with 4% paraformaldehyde for 10 minutes, and washed twice with phosphate-buffered saline

(PBS). The cells were counterstained with DAPI nuclear staining, mounted on microscope slides with a 50% mounting solution mixture of PBS and glycol, and the coverslips were sealed with clear nail polish. Fluorescent imaging was performed with

Leica DM4000B fluorescence microscopy. DAPI imaging used the Blue filter with excitation/emission at 410/455 nm. Doxorubicin was imaged with the Texas Red filter with Ex/Em 587/612 nm.

Cytotoxicity Assay

PC3Pip and PC3Flu cells (2000 cells/100 µL/well) were plated in 96 well plates and rested for 24 hours to adhere to the plate surface. To each well, we added 10 µL of an

11X concentrated solution of doxorubicin or conjugate. Each drug dose was tested in groups of six wells. Dojindo CCK8 assays were used to determine relative cell survival

14 of treated wells compared with untreated controls. This is a calorimetric assay for determination of cell viability and proliferation. A water-soluble tetrazolium salt, WST-8, is reduced by dehydrogenase activity in the cell which produces an orange dye with relative amounts quantified by absorbance (47). Briefly, 10 µL assay solution was added to each well, the plate was left to incubate while orange color was produced, the absorption of the solution in the plate was read at 450 nm, and the cell viabilities were calculated.

Mouse Tumor Xenograft Studies

All animal procedures were performed according to the protocols approved by the

Institutional Animal Care and Use Committee (IACUC). Our approval is Case Western

Reserve University IACUC #2015-0033. Athymic nude mice aged 6-8 weeks were implanted subcutaneously with 1x106 PC3Pip cells in 100 µL Matrigel on the right flank.

Tumors were allowed to grow for two weeks. Mice were then treated with three doses of free doxorubicin or PSMA-1-MMCCH-Dox in PBS on day 0, 7, and 14 of the study via tail vein injection. Dose used was 2 mg/kg doxorubicin, or the equivalent molar dosage of the PSMA-targeted prodrug. Tumor dimensions were measured using calipers twice per week, and mouse weights were recorded on these days, as well. The study duration was

21 days.

Statistical Analysis

Statistical analysis was performed using student t-test with a p<0.05 meeting criteria for statistical significance. Analysis was performed using JMP Pro 13 software.

15

Results

Chemistry

PSMA-1(Cys) is a modified version of our group’s PSMA-1 peptide ligand. It is a

urea-based ligand with three glutamic acid residues providing a negative charge that has been previously shown to reduce non-specific background binding without negatively

influencing PSMA affinity (48). The terminal cysteine allows for easy coupling of the –

SH group to the maleimide in each of the linkers chosen. PSMA-1(Cys) was successfully synthesized by Fmoc chemistry by hand, with an electrospray MS peak showing 1062.4 g/mol. This was consisted with the calculated molecular weight using Chem Draw

Professional of 1062.1 g/mol, as demonstrated in Figure 3. Note in all structures, the

PSMA-1 ligand is shown in blue, doxorubicin is in red, and the linker portion has been left black.

16

OH OH OH

O C O C O C

H 2C CH2 CH2 SH

O H 2C O CH2 O CH2 O CH2 O H H N C N C C N HC C HN C C HN HC C NH2 O H H H

PSMA-Cys, MW 1062.11 g/mol

NH

C

O CH2

CH2 CH NH C O C NH HO O O CH C H C 2 OH

H2C

C O HO

M+

Figure 3: Structure and Characterization of PSMA-1(Cys)

The non-cleavable PSMA-1-SMCC-Dox conjugate was synthesized by allowing the NHS ester in SMCC to react with the –NH2 present in doxorubicin. Expected molecular weight for this compound was 762.3. MALDI-MS of SMCC-Dox showed a

17 strong peak at 785.4, consistent with MW+Na, and confirming successful conjugation.

The full PSMA-1-SMCC-Dox conjugate was also successfully produced when the –SH of cysteine in the linker combined with the maleimide end of the SMCC linker. The expected molecular weight for this compound was 1824.9, and electrospray MS contains this peak, as well as several fragmentation peaks at lower molecular weights. See Figure

4 and Figure 5 below.

M+Na 4700 Reflector S pec #1 MC[BP = 650.2, 44141]

100 785.3844 4.3E+4

90

80 861.2177

70

60

50 801.3594 % Intensity 877.1912

40 855.2007 30

20 845.2369 839.2258 803.3630 879.1888 871.1750 893.1666 885.1820 10 901.1600 807.3742 854.1888 719.2238 847.2448 784.3766 841.2254 825.2550 725.3481 889.1783 881.1782 703.2454 867.2112 814.1582 858.2105 830.4408 917.1339 789.3848 905.1577 817.2981 771.3630 875.1951 712.0966 741.3260 850.2154 897.1514 870.4336 799.3461 834.2158 795.2750 767.3368 909.1817 900.2209 822.3351 737.3780 748.1526 0 703.0 747.2 791.4 835.6 879.8 924.0 Mass (m/z)

Figure 4: Synthesis and Characterization of SMCC-Dox

18

M+Na M+

Figure 5: Synthesis and Characterization of Non-cleavable PSMA-1-SMCC-Dox

Similarly, the production of the protease-cleavable PSMA-1-MC-Val-Cit-PABC-

Dox was achieved by reaction of the NHS ester group in the MC-Val-Cit-PABC-PNP linker with the amine in doxorubicin. The expected MW was 1142.2, and MALDI-MS demosntrated a peak at 1164.2, consistent with MW+Na. The –SH in PSMA-1(Cys)

19 reacts nicely with the maleimide group of the linker to produce the final product, with

MALDI-MS showing a peak at 2225.6, again consistent with MW+NA (expected MW

2204.3). As with the non-cleavable drug, some fragmentation peaks are also present. See

Figure 6 and Figure 7 for reaction and MS data.

M+Na

Figure 6: Synthesis and Characterization of MC-Val-Cit-PABC-Dox

20

M+Na

Figure 7: Synthesis and Characterization of Cathepsin-Cleavable PSMA-1-MC-Val-Cit-PABC-Dox

21

The final conjugate, acid-labile PSMA-1-MMCCH-Dox, was produced by first generating MMCCH-Dox. This is another maleimide-functionalized doxorubicin, but with a hydrazone bond at the carbonyl group of doxorubicin rather than an NHS ester to amine reaction. Electrospray MS shows a peak at the expected molecular weight of 777 and a strong peak at 799, consistent with MW+Na. There are also peaks at 1553.3 consistent with a dimer and 1575.4 consistent with the dimer+Na. As above, the –SH in the PSMA ligand joins the maleimide in the MMCCH linker to produce the final product, and the electrospray MS showed a very pure peak at the expected molecular weight. See

Figure 8 and Figure 9 below.

22

2x M+Na

M+Na 2x M+

M+

Figure 8: Synthesis and Characterization of MMCCH-Dox

23

M+

Figure 9: Synthesis and Characterization of Acid-Labile PSMA-1-MMCCH-Dox

24

Conjugation of the PSMA-targeting ligand to doxorubicin via the three linkers tested did not influence the absorption spectra of doxorubicin. Peak absorption was seen at 480 nm for free doxorubicin and each conjugate, as demonstrated in Figure 10.

1

0.9

0.8

0.7

0.6 PSD 0.5 PCD

Absorbance 0.4 PMD 0.3 Dox

0.2

0.1

0 0 200 400 600 800 1000 Wavelength

Figure 10: Absorption Spectra of Free Doxorubicin and PSMA-Targeted Conjugates Note, absorption spectra was measured from stock dilutions and concentration is not uniform. While the absorbance values are not identical for this reason, the shape of the curve and peak absorption wavelength remains unchanged by conjugation to PSMA-1(Cys). For brevity, PSMA-1-SMCC-Dox is labeled PSD, PSMA-1-MC-Val-Cit-PABC-Dox is labeled PCD, and PSMA-1-MMCCH-Dox is labeled PMD in this figure.

In Vitro Cellular Uptake Studies

To determine whether the PSMA-1-Doxorubicin conjugates might be preferentially uptaken by PSMA-expressing PC3Pip cells compared to PSMA-negative

PC3Flu cells, both of these cell lines were treated with 1 µM of doxorubicin or PSMA-1-

Doxorubicin conjugate for either 90 minutes or 48 hours. Doxorubicin fluoresces and can therefore be visualized in cells by fluorescent microscopy. Qualitative results are shown in Figure 11. Free doxorubicin was equally seen in PC3Pip and PC3Flu cell lines, and the

25 drug was seen primarily in the nucleus, overlying the blue DAPI nuclear staining in these cells. PSMA-1-SMCC-Dox demonstrated more signal in PC3Pip compared to PC3Flu cells, and as was the case with all PSMA-conjugates, the selectivity was more obvious at short cell treatment times. However, unlike free doxorubicin, the signal was seen in a spotty, endosomal type pattern. The same observations were true of PSMA-1-Val-Cit-

PABC-Dox. PSMA-1-MMCCH-Dox showed similar selectivity for PC3Pip cells, but the red doxorubicin signal did colocalize with the blue DAPI nuclear staining for this conjugate.

A

B

Figure 11: In Vitro Drug Uptake and Cellular Localization by Fluorescence Microscopy PSMA(+) PC3Pip cells and PSMA(-) PC3Flu cells were treated with free doxorubicin or one of three PSMA-targeted conjugates. DAPI nuclear staining is shown in blue, and doxorubicin fluorescence signal is seen in red. Specific uptake of PSMA-1-Doxorubicin conjugates is seen in PC3Pip cells, more so at short incubation periods, Figure A.)1 µM, 90 minutes, compared to long periods, Figure B.) 1 µM, 48 hours. Only free doxorubicin and PSMA-1-MMCCH-Dox showed colocalization of the red doxorubicin and blue DAPI nuclear staining (purple color). Note negligible autofluoresnce in the red channel for PC3Flu cells, seen in the untreated control slide.

26

In Vitro Cytotoxicity

To determine whether our PSMA-1-Doxorubicin conjugates demonstrated any cytotoxic effect in vitro, and whether they may be more toxic to PSMA-expressing cells, the commercially available Dojindo CCK8 assay was used to monitor cell viability after treatment with free and PSMA-targeted doxorubicin. First, PC3Pip and PC3Flu cells were treated continuously for 48 hours with free doxorubicin, non-cleavable PSMA-1-

SMCC-Dox, cathepsin-cleavable PSMA-1-MC-Val-Cit-PABC-Dox, and acid-labile

PSMA-1-MMCCH-Dox. Doxorubicin exhibited potent, equal growth inhibition of

PC3Pip and PC3Flu cell lines. PSMA-1-SMCC-Dox and PSMA-1-MC-Val-Cit-PABC-

Dox showed essentially no growth inhibition of either cell line. PSMA-1-MMCCH-Dox, however, was significantly more toxic to PC3Pip cells compared to PC3Flu cells at all doses tested over the range 62.5-2000 nM, as depicted in Figure 12.

27

A

B

Figure 12: 48 Hour Cytotoxicity of Free and PSMA-Targeted Doxorubicin PSMA(+) PC3Pip cells and PSMA(-)PC3Flu cells were treated with doxorubicin or PSMA-1-Doxorubicin conjugates for 48 hours before measuring cell viability relative to untreated controls using Dojindo CCK8 assays. Doses in nM. Same data displayed both by drug (A) and as a single graph with all drugs overlayed (B) for easier visual comparison. For brevity, PSMA-1-SMCC-Dox is labeled PSD, PSMA-1-MC-Val-Cit-PABC-Dox is labeled PCD, and PSMA-1- MMCCH-Dox is labeled PMD in this figure. Note maximum dose tested for PMD was 2000 nM instead of 4000 nM for the other drugs. Stats performed on variation within six wells tested.

28

The fluorescence microscopy data indicated that some PSMA-1-MMCCH-Dox does enter PC3Flu cells after long incubation periods, so the cytotoxicity assay was also repeated for free doxorubicin and the acid-labile conjugate at a slightly shorter incubation period of 24 hours in an attempt to optimize the cytotoxicity difference in the two cell lines. The cells were then provided fresh media and allowed to continue in culture for an additional 48 hours to allow for sufficient time for growth inhibition and/or cell death.

The increased toxicity to PC3Pip cells compared to PC3Flu cells was again evident, at about the same magnitude, for these conditions. The drug conjugate, in vitro, was generally less toxic than free drug, especially at lower doses, as seen in Figure 13B below

which shows ratios of PSMA-1-MMCCH-Dox toxicity to free doxorubicin viability. At doses of 500 nM or more, the PSMA-1-MMCCH-Dox conjugate showed equal or superior cytotoxicity compared to free drug for PC3Pip cells, and at 1000 nM or more the targeted drug was also toxic to PC3Flu cells similarly to free doxorubicin.

B

A Viability PMD/Dox

Figure 13: Cytotoxicity of Dox and PSMA-MMCCH-Dox, Alternate Conditions PSMA(+) PC3Pip cells and PSMA(-) PC3Flu cells were treated with free doxorubucin and PSMA-1-MMCH-Dox for 24 hours, followed by 48 hours of growth in fresh media and determination of cell viability with Dojindo CK8 Assay. Doses in nM. A.) Cell viability data, showing decreased viability of PC3Pip cells compared to PC3Flu cells when treated with PSMA-1-MMCCH-Dox B.) Cell viability after treatment with PSMA-1-MMCCH-Dox relative to cell viability after treatment with free doxorubicin, generally demonstrating decreased efficacy of PSMA-1-MMCCH-Dox in culture for both cell lines. Asterisks represent a statistically significant difference from a ratio of 1.

29

Mouse Tumor Xenograft Response

Mice implanted with PSMA(+) PC3Pip flank tumors were treated with three

weekly doses of 2 mg/kg of free doxorubicin, or the molar equivalent of PSMA-1-

MMCC-Dox, 6.35 mg/kg. Tumor size was monitored with calipers, and toxicity was

measured by mouse weight over time. Measurements were taken two times per week, day

0, 4, 7, 11, etc. until day 21. At these tested doses, all tumors grew in size with time. The

mice treated with free doxorubicin exhibited more rapid tumor growth than those treated

with PSMA-1-MMCCH-Dox, with growth difference reaching statistical significance

through day 14 and trending toward significance through the end of the study.

Additionally, mice treated with the targeted drug exhibited less weight loss, an indication

of drug toxicity, over the duration of the study (1.8% vs. 17.3%, p=0.009), as depicted in

Figure 14 below.

A B

Figure 14: Tumor Response and Mouse Weight A.) Tumor growth was monitored by volume using calipers and is expressed here as a ratio of initial Day 0 volume. Treatment days indicated with vertical arrows. While all tumors increased in size, tumors of mice treated with PMD (PSMA-1-MMCCH-Dox) grew more slowly than mice treated with free doxorubicin. N=3 per group. B.) Mice treated with PMD experienced less weight loss by the end of the 21 day trial period.

30

Discussion

The concept of the “magic bullet” has existed since as long ago as the year 1900,

popularized by Dr. Paul Ehrlich. This describes treatments that target only the diseased

tissue without any harm to healthy parts of the body (49). Molecularly-targeted cancer

therapy has blossomed into a huge field of research since this time, and there are now

four antibody-drug conjugates approved by the FDA and commercially available in the

United States: Brentuximab vedotin for Hodgkin lymphoma, ado-trastuzumab entansine

for breast cancer, inotuzumab ozogamicin for refractory B-cell acute lymphoblastic

leukemia, and gemtuzumab ozogamicin for acute myeloid leukemia (50). Many more are

being tested in clinical trials. Targeted nanoparticle drug delivery methods, including

PSMA-targeted BIND014 (docetaxel payload), are also showing promise in trials but

none have yet been FDA approved (51).

We have developed a series of PSMA-1-Doxorubicin conjugates that target

PSMA, a molecule highly overexpressed on the surface of prostate cancer cells. Other groups have previously devised PSMA-targeted doxorubicin delivery strategies, though all but one utilize complex nanoparticles. Xu et al. developed an A10 (RNA) aptamer- targeted micelle that showed increased doxorubicin accumulation in mouse xenograft tumor tissue compared to non-targeted micelles (52). Baek et al. similarly designed an A9

(RNA) aptamer-functionalized liposome that inhibited prostate cancer tumor growth in mice (44). Pearce et al. developed a urea-based peptide ligand-targeted hyperbranched polymer theranostic delivering doxorubicin and the dye Cy5.5 that reduced tumor volume compared to free doxorubicin (53). Only Bagalkot et al. have proposed a small molecule

31

therapy, consisting of A10 aptamer-doxorubicin physical conjugate (no true bonds

between the ligand and drug) that proved more toxic to PSMA(+) LNCaP cells compared

to PSMA-PC3 cells but has not yet been tested in vivo (54).

The PSMA-1-Doxorubicin conjugates presented here have distinctive benefits

over these technologies. Our targeted agents remain small molecules, eliminating many

of the challenges for good manufacturing practice associated with nanoparticle

production like controllable and reproducible synthesis, scalable manufacturing, and

regulatory barriers. Further, nanoparticles have challenges associated with clinical

translation compared to small molecule drugs, such as an inability to realistically model

the physical environment of tumors in vitro, like the leaky vasculature on which

nanoparticle treatments rely for their enhanced permeability and retention effect (55). The

physical conjugate proposed by Bagalkot would likely not provide the stability in the

blood stream that our linkers should. Compared to antibody-drug conjugates, our peptide

PSMA-1 ligand has several benefits, too. Peptides are produced using chemical synthesis

methods rather than in vivo cell culture or animal conditions, which is both faster and

more cost-effective. The structure can be modified to alter properties like hydrophobicity

or to add groups for chemical conjugation to other substances (56). Peptide and aptamer-

based therapies have shorter circulating times than antibodies, which is especially useful

for imaging applications but may also be helpful for therapeutics to decrease side effects.

Many antibody-drug conjugates still have narrow therapeutic windows (57).

Our three prodrug conjugates have been designed around the hypothesis that

doxorubicin will need to be freed from the targeting peptide in order to exert its antitumor

effects with the greatest efficiency. The goal of doxorubicin release must be balanced

32

with a desire to create a conjugate that is very stable in the blood stream to prevent

release of free drug before it reaches the tumor, which would likely result in some of the

classic toxicities to the heart and bone marrow. We have therefore used a non-cleavable

SMCC linker (this conjugate is therefore not a prodrug), a protease-labile valine-citrulline linker, and an acid-labile hydrazone. These three linkers are listed in order of decreasing stability. Valine-citrulline linkers have been proved to be over 100 times as stable as linkers in human plasma (58).

All three of the proposed conjugates were successfully synthesized using a two- step reaction scheme, where doxorubicin is first joined to the chemical linker to create a maleimide-functionalized doxorubicin. That product was then purified, before addition of

PSMA-1(Cys) via the maleimide linker end. The drug conjugates maintained the same absorption and fluorescence (not shown) wavelength pattern as free doxorubicin.

Using these drug properties, selective uptake of the three PSMA-1-Doxorubicin conjugates by PSMA(+) PC3Pip cells compared to PSMA(-) PC3Flu cells was confirmed qualitatively using fluorescence microscopy (Figure 11). Fluorescent signal in the images was not quantified because fluorescence intensity of doxorubicin does not correlate in a linear fashion with drug concentration, like absorbance measurements do. Further, microenvironment and interaction with DNA can alter fluorescence intensity (59).

Doxorubicin is a hydrophobic drug and crosses cell membranes easily. Free doxorubicin entered both cell lines and accumulated in the nucleus of the cells. The full

PSMA-1-Doxorubicin conjugates do not freely pass cell membranes, but were internalized, assumingly, via the internalization function of the cell surface PSMA. As expected, non-cleavable PSMA-1-SMCC-Dox was not visualized in the nucleus of the

33

cell, suggesting no doxorubicin was released. Instead, doxorubicin signal could be seen in

a punctate (potentially endosomal) pattern. Disappointingly, this same observation was

made when cells were incubated with protease-labile PSMA-1-MC-Val-Cit-PABC-Dox.

It was selective for PC3Pip cells compared to PC3Flu cells, but little to no drug was seen in the nucleus of the cells. This may be explained by low cathepsin B expression in vitro, a finding published by Podgorski (60). Her group found that PC3 cells grown in vitro expressed low levels of cathepsin B but that PC3 bone tumors had high levels of cathepsin B activity, probably promoted by tumor-stromal interactions in the tumor microenvironment. In contrast, PSMA-1-MMCCH-Dox showed doxorubicin signal strongly overlaying the DAPI nuclear signal, even after a short treatment time of 90 minutes, suggesting efficient release of free doxorubicin from the PSMA-1 ligand. These cellular localization patterns held true for long treatment periods extending to 48 hours.

Though drug the selectivity for PC3Pip cells compared to PC3Flu cells was maintained, the PC3Flu cells did begin to internalize some drug over time, perhaps via pinocytosis.

These observations were, on the whole, consistent with the results of the cell viability experiments. Presumably because there was no doxorubicin in the nucleus, neither PSMA-1-SMCC-Dox nor PSMA-1-MC-Val-Cit-PABC-Dox exhibited any significant cytotoxic effects to either PC3Pip or PC3Flu cells at the doses tested (Figure

12). Liang et al. designed a Tat-SMCC-Dox drug designed to increase the cell- penetration ability of doxorubicin in multidrug resistant breast cancer cell lines (45).

Their conjugate did not show drug in the nucleus, either. However, in contrast to our results, they did see some cell killing which was attributed to generation of hydrogen peroxide (an alternate doxorubicin mechanism of action) (61). This was seen only when

34

doses exceeded 5-10 µM, beyond the doses we attempted. It remains possible that

PSMA-1-MC-Val-Cit-PABC-Dox will be effective in vivo when PC3 cells exhibit higher

cathepsin B activity. The acid-labile conjugate, PSMA-1-MMCCH-Dox, did show cytotoxic activity, and more so for PC3Pip cells compared to PC3Flu cells, excitingly.

The conjugate was less toxic than free doxorubicin, but often molecularly-targeted drugs are less toxic in vitro when the free drugs have easy access and close proximity to the cells. In animal or human models, the targeted drug should accumulate in tumors and be more toxic than free drug which disperses throughout the body and is rapidly excreted as a small molecule.

PSMA-1-MMCCH-Dox did not protect PC3Flu cells completely. We know that at long incubation times, some PSMA-1-MMCCH-Dox can enter PC3Flu cells. This may be due to pinocytosis of the conjugate with subsequent doxorubicin release intracellularly. Alternately, it was thought doxorubicin may be released from the full conjugate in the mildly acidic cell culture medium. To test this hypothesis, pH of fresh versus used culture medium was roughly measured using pH paper. PSMA-1-MMCCH-

Dox solution was made in fresh and used culture media and allowed to sit overnight at room temperature. Thin layer chromatography (TLC) using a silica plate and ethanol as the mobile phase was used to test whether free doxorubicin was released from the conjugate in these conditions. This was not observed, giving more credence to the pinocytosis theory rather than extracellular cleavage. A full doxorubicin release study over time at low pH was not performed, but at physiologic conditions the conjugate is relatively stable.

35

Figure 15: Test of PSMA-1-MMCCH-Dox Cleavage in Cell Culture Media Used cell culture media becomes mildly acidic. It was hypothesized that this causes extracellular cleavage of PSMA-1- MMCCH-Dox in cell culture. However, no cleavage was seen, as depicted by TLC. A solution of free doxorubicin was plated in the left lane. The PSMA-1-MMCCH-Dox in used culture media overnight was plated to the right. No doxorubicin was seen in the PSMA-1-MMCCH-Dox lane, orange arrow, suggesting none cleaved.

Based on promising in vitro results, PSMA-1-MMCCH-Dox was advanced to in vivo experiments using a nude mouse flank tumor xenograft model with PSMA(+)

PC3Pip cells. The groups mentioned previously who are using PSMA-targeted nanoparticle delivery systems for doxorubicin have used a wide range of doses, from 0.3 mg/kg weekly on the low end to 4.5 mg/kg given every 4 days on the high end. Common doxorubicin doses used in humans clinically are 60 mg/m2 given every three weeks, or 20 mg/m2 given weekly for a gentler side-effect profile (31). The weekly dose of 20 mg/m2

is the equivalent of a dose of 6.67 mg/kg in mice (62). A decision was made to use a dose

of 2 mg/kg of doxorubicin, or the equivalent molar dose of the PSMA-1-MMCCH-Dox conjugate. This is about one third of the equivalent dose used in people. This choice was made both for practical reasons, to save time in drug synthesis, and because the PSMA-1-

Doxorubicin conjugate should accumulate in the tumor and produce good growth inhibition even at low doses.

The tumors in the mice treated with both free doxorubicin and PSMA-1-

MMCCH-Dox grew over time. However, the tumors treated with the targeted PSMA-1- 36

MMCCH-Dox grew more slowly. This reached statistical significance for days 4-14, and trended towards significance through the end of the trial at day 21. The narrower, nonsignificant difference in volume ratios at the late time points is probably due to the fact that the tumors in the mice treated with doxorubicin grew to be quite large, and large tumors generally grow more slowly due to poor perfusion. Mice treated with the PSMA-

1-MMCCH-Dox lost less body weight, a marker of drug toxicity, compared to mice treated with doxorubicin (Figure 14). Tumor volumes at the end of the 21 day study period ranged from 1000-2000 mm3, and mice maintained mobility and normal behavior,

suggesting the toxicity is drug-related rather than being caused by excessive tumor

burden. These results support the hypothesis that PSMA-targeting will increase efficacy

and decrease toxicity of doxorubicin, and the minimal toxicity of this tested dose would

allow for a trial of a higher dose that could exert a stronger antitumor effect.

Perhaps the biggest challenge and bottleneck of this study is the time-intensive

nature of purifying the PSMA-1-Doxorubicin conjugates. Synthesis and purification are

being performed on the scale of a few mg per batch, currently. In antibody-drug

conjugate production, it is easy to purify the final product quickly by size-

exclusion/filtration methods. However, our PSMA-1-Doxorubicin conjugates are only

1.5-2X the molecular weight of the individual reactants, making this method unreliable

and necessitating HPLC use.

Several further experiments can be done to better characterize the PSMA-1-

Doxorubicin conjugates. Stability of the prodrugs in PBS (pH 7.0), acetate buffer (pH

5.0), and blood plasma over time can be easily tested using HPLC with an analytical

column, now that retention times are known from developing the synthesis and

37

purification methods. This would give some evidence as to whether the drugs are stable

enough to maintain their structure in various conditions or whether free doxorubicin may

be released in circulation, decreasing the selectivity and increasing toxicity of the targeted agents. At this stage, only rough experiments have been performed using TLC, as discussed previously.

To better understand and confirm the proposed mechanism of PSMA-targeting,

competition binding studies should be performed in vitro to confirm that linkage to

doxorubicin does not negatively impact the binding affinity of PSMA-1 to the PSMA receptor, though this has not been the case with conjugation to other agents such as near infrared fluorescent dyes in the lab. Further, co-treating cells in culture with the PSMA-1-

Doxorubicin conjugates and fluorescent endosomal/lysosomal tracers to see whether or not the signal colocalizes using fluorescent microscopy may confirm the assumption that the drugs are being internalized by receptor-mediated endocytosis after binding to PSMA on the PC3Pip cell surface. Last, it would be interesting to grow a cell line resistant to doxorubicin to determine if PSMA-1-Doxorubicin conjugates might be more toxic than free drug in this case because of an endocytosis-mediated uptake.

Several additional mouse experiments would provide more information about the efficacy, pharmacokinetics, and pharmacodynamics and would improve the quality of this study. Capitalizing again on the fluorescent properties of doxorubicin, in vivo fluorescent imaging would test the hypothesis that PSMA-1-Doxorubicin conjugates show targeted delivery to PSMA-expressing tumors and provide some data about bio- distribution and circulation time compared to free doxorubicin. (These studies are ongoing.) To get a more precise measure of half-life and drug stability in the blood

38

stream, blood samples may also be obtained and analyzed for PSMA-1-Doxorubicin or

free doxorubicin levels (if released from the ligand in circulation) by HPLC or MS. To

characterize toxicity, bone marrow and heart histology may be helpful.

At this time, PSMA-1-SMCC-Dox and PSMA-1-MC-Val-Cit-PABC-Dox have not been tested in mice. We would hypothesize that the conjugate with non-cleavable

SMCC linker would be targeted but not very effective, as it was in vitro. The cathepsin- cleavable drug, though, may show better effect in vivo if cathepsin expression is higher in the tumor model than in cell culture. Further, though there is sufficient evidence to conclude that PSMA-1-MMCCH-Dox is more effective than free doxorubicin, dosing studies should be performed to determine maximum-tolerated dose and to see if tumor shrinkage can be achieved. If this data were promising, it would be worthwhile to test the effective PSMA-1-Doxorubicin conjugates in a mouse model of metastatic prostate cancer to test whether we can improve overall survival.

In conclusion, three PSMA-1-Doxorubicin conjugates were synthesized making use of differing linker chemistries. All three were selective for PSMA(+) cell lines, but only acid-labile PSMA-1-MMCCH-Dox promoted release of free doxorubicin, allowing the drug to travel to the nucleus and exert its cytotoxic activity. This drug was tested in a flank tumor prostate cancer model, and resulted in tumor growth inhibition compared to free doxorubicin. Further, it was less toxic to the mice than free doxorubicin as determined by decreased weight loss while on treatment. In the future, it is anticipated that these agents may prove more effective while reducing significant toxicities/side effects compared with free doxorubicin for treatment of metastatic prostate cancer

39 models in nude mice and later for treatment of human patients with prostate cancer or other PSMA-expressing tumors.

40

Acknowledgements

Thank you to Drs. Xinning Wang and James P. Basilion for assistance with study design, development of methodology, and mentorship throughout this research experience.

Thank you to Drs. Xinning Wang, Ramamurthy Gopalakrishnan, Dong Luo, Elizabeth

Kerpan, and Aditi Shirke for their help training in various techniques and in data collection. Thank you to Drs. Stathis Karathanasis and Christopher Hoimes for their time and input in reviewing this project as members of my thesis committee. Thank you to the

Research Education Committee at the Cleveland Clinic Lerner College of Medicine for their support and supervision of this excellent educational experience.

41

Bibliography

1. Prostate Cancer - Cancer Stat Facts. SEER.

https://seer.cancer.gov/statfacts/html/prost.html. Accessed December 8, 2018.

2. Park JC, Eisenberger MA. Advances in the Treatment of Metastatic Prostate Cancer.

Mayo Clinic Proceedings. 2015;90(12):1719-1733.

3. "Prostate Cancer," NCCN Guidelines, 2017.

4. Tannock IF, de Wit R, Berry WR, et al. Docetaxel plus Prednisone or Mitoxantrone

plus Prednisone for Advanced Prostate Cancer. New England Journal of Medicine.

2004;351(15):1502-1512.

5. Evans JC, Malhotra M, Cryan JF, O’Driscoll CM. The therapeutic and diagnostic

potential of the prostate specific membrane antigen/glutamate carboxypeptidase II

(PSMA/GCPII) in cancer and neurological disease. Br J Pharmacol. 2016;173(21):3041-

3079.

6. Evans JD, Jethwa KR, Ost P, et al. Prostate cancer–specific PET radiotracers: A

review on the clinical utility in recurrent disease. Practical Radiation Oncology.

2018;8(1):28-39.

7. Ross JS, Sheehan CE, Fisher HAG, et al. Correlation of primary tumor prostate-

specific membrane antigen expression with disease recurrence in prostate cancer. Clin

Cancer Res. 2003;9(17):6357-6362.

8. Ren H, Zhang H, Wang X, Liu J, Yuan Z, Hao J. Prostate-specific membrane antigen as a marker of pancreatic cancer cells. Med Oncol. 2014;31(3):857.

42

9. Haffner MC, Laimer J, Chaux A, et al. High expression of prostate-specific membrane antigen in the tumor-associated neo-vasculature is associated with worse prognosis in squamous cell carcinoma of the oral cavity. Mod Pathol. 2012;25(8):1079-1085.

10. Zeng C, Ke Z-F, Yang Z, et al. Prostate-specific membrane antigen: a new potential prognostic marker of osteosarcoma. Med Oncol. 2012;29(3):2234-2239.

11. Silver DA, Pellicer I, Fair WR, Heston WD, Cordon-Cardo C. Prostate-specific membrane antigen expression in normal and malignant human tissues. Clin Cancer Res.

1997;3(1):81-85.

12. Wernicke AG, Edgar MA, Lavi E, et al. Prostate-Specific Membrane Antigen as a

Potential Novel Vascular Target for Treatment of Glioblastoma Multiforme. Archives of

Pathology & Laboratory Medicine. 2011;135(11):1486-1489.

13. Wernicke AG, Kim S, Liu H, Bander NH, Pirog EC. Prostate-specific Membrane

Antigen (PSMA) Expression in the Neovasculature of Gynecologic Malignancies:

Implications for PSMA-targeted Therapy. Appl Immunohistochem Mol Morphol.

February 2016.

14. Wernicke AG, Varma S, Greenwood EA, et al. Prostate-specific membrane antigen expression in tumor-associated vasculature of breast cancers. APMIS. 2014;122(6):482-

489.

15. Chang SS, Reuter VE, Heston WD, Bander NH, Grauer LS, Gaudin PB. Five different anti-prostate-specific membrane antigen (PSMA) antibodies confirm PSMA expression in tumor-associated neovasculature. Cancer Res. 1999;59(13):3192-3198.

43

16. Samplaski MK, Heston W, Elson P, Magi-Galluzzi C, Hansel DE. Folate hydrolase

(prostate-specific membrane [corrected] antigen) 1 expression in bladder cancer subtypes

and associated tumor neovasculature. Mod Pathol. 2011;24(11):1521-1529.

17. Wang H, Wang S, Song W, et al. Expression of prostate-specific membrane antigen in lung cancer cells and tumor neovasculature endothelial cells and its clinical significance. PLoS ONE. 2015;10(5):e0125924.

18. Haffner MC, Kronberger IE, Ross JS, et al. Prostate-specific membrane antigen

expression in the neovasculature of gastric and colorectal cancers. Human Pathology.

2009;40(12):1754-1761.

19. Baccala A, Sercia L, Li J, Heston W, Zhou M. Expression of Prostate-Specific

Membrane Antigen in Tumor-Associated Neovasculature of Renal Neoplasms. Urology.

2007;70(2):385-390.

20. Heitkötter B, Steinestel K, Trautmann M, et al. Neovascular PSMA expression is a common feature in malignant neoplasms of the thyroid. Oncotarget. 2018;9(11):9867-

9874.

21. Heitkötter B, Trautmann M, Grünewald I, et al. Expression of PSMA in tumor neovasculature of high grade sarcomas including synovial sarcoma, rhabdomyosarcoma, undifferentiated sarcoma and MPNST. Oncotarget. 2017;8(3):4268-4276.

22. Dias AH, Bouchelouche K. Prostate-specific Membrane Antigen Pet/ct Incidental

Finding of a Schwannoma. Clinical Nuclear Medicine. 2018;43(4):267-268.

44

23. Lawhn-heath C, Flavell RR, Glastonbury C, Hope TA, Behr SC. Incidental Detection

of Head and Neck Squamous Cell Carcinoma on 68ga-psma-11 Pet/ct. Clinical Nuclear

Medicine. 2017;42(4).

24. Milowsky MI, Galsky MD, Morris MJ, et al. Phase 1/2 multiple ascending dose trial

of the prostate-specific membrane antigen-targeted antibody drug conjugate MLN2704 in metastatic castration-resistant prostate cancer. Urol Oncol. 2016;34(12):530.

25. A Study of PSMA ADC in Subjects With Metastatic Castration-resistant Prostate

Cancer (mCRPC) - Study Results - ClinicalTrials.gov. https://clinicaltrials.gov/ct2/show/results/NCT01695044. Accessed October 10, 2017.

26. Maurer T, Eiber M, Schwaiger M, Gschwend JE. Current use of PSMA-PET in prostate cancer management. Nat Rev Urol. 2016;13(4):226-235.

27. Wang X, Huang SS, Heston WDW, Guo H, Wang B-C, Basilion JP. Development of targeted near-infrared imaging agents for prostate cancer. Mol Cancer Ther.

2014;13(11):2595-2606.

28. Wang X, Tsui B, Ramamurthy G, et al. Theranostic Agents for Photodynamic

Therapy of Prostate Cancer by Targeting Prostate-Specific Membrane Antigen. Mol

Cancer Ther. 2016;15(8):1834-1844.

29. Petrioli R, Fiaschi A, Francini E, Pascucci A, Francini G, The role of doxorubicin and in the treatment of patients with metastatic hormone-refractory prostate cancer.

Cancer Treatment Reviews. 2008; 34:710-718.

45

30. Tsakalozou E, Eckman AM, Bae Y. Combination Effects of Docetaxel and

Doxorubicin in Hormone-Refractory Prostate Cancer Cells. Biochemistry Research

International. 2012.

31. "Doxorubicin drug information" Lexicomp, Inc. 2017.

32. Motlagh NSH, Parvin P, Ghasemi F, Atyabi F. Fluorescence properties of several

drugs: doxorubicin, and . Biomed Opt Express.

2016;7(6):2400-2406.

33. Dai X, Yue Z, Eccleston ME, Swartling J, Slater NKH, Kaminski CF. Fluorescence

intensity and lifetime imaging of free and micellar-encapsulated doxorubicin in living cells. Nanomedicine. 2008;4(1):49-56.

34. Zhong Y-J, Shao L-H, Li Y. Cathepsin B-cleavable doxorubicin prodrugs for targeted cancer therapy. Int J Oncol. 2012;42(2):373-383.

35. Thorn CF, Oshiro C, Marsh S, et al. Doxorubicin pathways: pharmacodynamics and adverse effects. Pharmacogenet Genomics. 2011;21(7):440-446.

36. Pérez-Arnaiz C, Busto N, Leal JM, García B. New Insights into the Mechanism of the

DNA/Doxorubicin Interaction. J Phys Chem B. 2014;118(5):1288-1295.

37. Jayaprakash S, Wang X, Heston WD, Kozikowski AP. Design and synthesis of a

PSMA inhibitor-doxorubicin conjugate for targeted prostate cancer therapy.

ChemMedChem. 2006;1(3):299-302.

46

38. Sinha AA, Morgan JL, Buus RJ, et al. Cathepsin B Expression is Similar in African-

American and Caucasian Prostate Cancer Patients. Anticancer Res. 2007;27(5A):3135-

3141.

39. Lu, "Linkers Having a Crucial Role in Antibody-Drug Conjugates," Int J Mol Sci.

2016 Apr; 17(4): 561.

40. Dubowchik GM, et al. Cathepsin B-Labile Dipeptide Linkers for Lysosomal Release of Doxorubicin from Internalizing Immunoconjugates: Model Studies of Enzymatic Drug

Release and Antigen-Specific In Vitro Anticancer Activity. Bioconjugate Chem.

2002;13:855.

41. Peters C, Brown S. Antibody–drug conjugates as novel anti-cancer chemotherapeutics. Biosci Rep. 2015;35(4).

42. Jain N, Smith SW, Ghone S, Tomczuk B. Current ADC Linker Chemistry. Pharm

Res. 2015;32(11):3526-3540.

43. Griffiths et al. Cure of SCID Mice Bearing Human B-Lymphoma Xenografts by an

Anti-CD74 Antibody- Drug Conjugate. Clinical Cancer Research. 2003;

9:6567-6571.

44. Baek SE, Lee KH, Park YS, et al. RNA aptamer-conjugated liposome as an efficient anticancer drug delivery vehicle targeting cancer cells in vivo. J Control Release.

2014;196:234-242.

45. Liang JF, Yang VC. Synthesis of doxorubicin-peptide conjugate with multidrug resistant tumor cell killing activity. Bioorg Med Chem Lett. 2005;15(22):5071-5075.

47

46. Tian Y, Tam KC, et al. Titration Microcalorimetry Study: Interaction of Drug and

Ionic Microgel System. J Contr Rel. 2007, 121:137-145.

47. Cell Counting Kit-8 - WST-8 Cell Proliferation Cytotoxicity Assay Kit | Dojindo,

MTT alternative, MTS alternative. https://www.dojindo.com/store/p/456-Cell-Counting-

Kit-8.html.

48. Huang SS, Wang X, Zhang Y, Doke A, DiFilippo FP, Heston WD. Improving the

biodistribution of PSMA-targeting tracers with a highly negatively charged linker.

Prostate. 2014;74(7):702-713.

49. Strebhardt K, Ullrich A. Paul Ehrlich’s magic bullet concept: 100 years of progress.

Nature Reviews Cancer. 2008;8:473.

50. The Expanding Field of Antibody-Drug Conjugates. ADC Review.

https://adcreview.com/news/expanding-field-antibody-drug-conjugates/. Published

January 20, 2018.

51. Autio KA, et al. Safety and Efficacy of BIND-014, a Docetaxel Nanoparticle

Targeting Prostate-Specific Membrane Antigen for Patients With Metastatic Castration-

Resistant Prostate Cancer: A Phase 2 . JAMA Oncol. 2018;4(10):1344-

1351.

52. Xu W, Siddiqui IA, Nihal M, et al. Aptamer-conjugated and doxorubicin-loaded

unimolecular micelles for targeted therapy of prostate cancer. Biomaterials.

2013;34(21):5244-5253.

48

53. Pearce AK, Simpson JD, Fletcher NL, et al. Localised delivery of doxorubicin to

prostate cancer cells through a PSMA-targeted hyperbranched polymer theranostic.

Biomaterials. 2017;141:330-339.

54. Bagalkot V, Farokhzad OC, Langer R, Jon S. An Aptamer–Doxorubicin Physical

Conjugate as a Novel Targeted Drug-Delivery Platform. Angewandte Chemie

International Edition. 2006;45(48):8149-8152.

55. Shi. Cancer nanomedicine: progress, challenges, and opportunities. Nat Rev Cancer.

2017 Jan;17(1):20-37.

56. Mascini, Nucleic Acid and Peptide Aptamers: Fundamentals and Bioanalytical

Aspects, Angew. Chem. Int. Ed., (2012);51:1316 – 1332.

57. de Goeij BE, Lambert JM. New developments for antibody-drug conjugate-based

therapeutic approaches. Current Opinion in Immunology. 2016;40:14-23.

58. McCombs JR, Owen SC. Antibody Drug Conjugates: Design and Selection of Linker,

Payload and Conjugation Chemistry. AAPS J. 2015;17(2):339-351.

59. Mohan P, Rapoport N. Doxorubicin as a molecular nanotheranostic agent: effect of

doxorubicin encapsulation in micelles or nanoemulsions on the ultrasound-mediated

intracellular delivery and nuclear trafficking. Molecular pharmaceutics. 2010;7(6):1959.

60. Podgorski I, Linebaugh BE, Sameni M, et al. Bone Microenvironment Modulates

Expression and Activity of Cathepsin B in Prostate Cancer. Neoplasia. 2005;7(3):207-

223.

49

61. Mizutani, Mechanism of apoptosis induced by doxorubicin through the generation of hydrogen peroxide, Life Sciences 76 (2005) 1439 – 1453.

62. Nair AB, Jacob S. A simple practice guide for dose conversion between animals and human. J Basic Clin Pharm. 2016;7(2):27-31.

50