The therapeutic and diagnostic potential of the prostate specific membrane antigen/glutamate carboxypeptidase II (PSMA/GCPII) in cancer and neurological disease

Abstract

Prostate specific membrane antigen (PSMA) otherwise known as glutamate carboxypeptidase II (GCPII) is a membrane bound protein that is highly expressed in prostate cancer and in the neovasculature of a wide variety of tumours including glioblastomas, breast and bladder cancers. This protein is also involved in a variety of neurological diseases including schizophrenia and ALS. In recent years, there has been a surge in the development of both diagnostics and therapeutics that take advantage of the expression and activity of PSMA/GCPII. These include gene therapy, immunotherapy, chemotherapy and radiotherapy. In this review, we discuss the biological roles that PSMA/GCPII plays, both in normal and diseased tissues, and the current therapies exploiting its activity that are at the preclinical stage. We conclude by giving an expert opinion on the future direction of PSMA/GCPII based therapies and diagnostics and hurdles that need to be overcome to make them effective and viable.

Abbreviations

2‐MPPA

2‐(3‐mercaptopropyl)pentanedioic acid

2‐PMPA

2‐(phosphonomethyl)pentanedioic acid

aa

amino acids

ADC

antibody‐drug conjugate

AR

androgen receptor

BBB

blood–brain barrier

CAR

chimeric antigen receptor

CTL

cytotoxic T‐lymphocyte

DCs

Dendritic cells

DHT

dihydrotestosterone

dsRNA

double stranded RNA

FBP

folate‐binding proteins

FLNa

filamin A

FOLH1

folate hydrolase

hNIS

human sodium iodide symporter

huPSMA

expressing human PSMA

KD

knockdown

mGLU

metabotropic glutamate receptors

NAA

N‐acetylaspartate

NAAG

N‐acetyl‐aspartyl‐glutamate

NFATc1

transcriptional activator of PSME

NSCLC

non‐small cell lung cancers

PCa

prostate cancer

PLA

polylactic acid

PSA

prostate specific antigen

PSME

PSMA/GCPII enhancer

QD

quantum dots

RFC

reduced folate carriers

RIT

radioimmunotherapy

RNAi

RNA interference

SCLC

small cell lung cancer

Tfr

transferrin receptor

Tables of Links

TARGETS
Enzymes a	G protein‐coupled receptors c
Folate hydrolase (prostate‐specific membrane antigen) 1	mGlu3 receptor
Kallikrein related peptidase 3	Transporters d
Catalytic Receptors b	Reduced folate transporter 1, SLC19A1
Epidermal growth factor receptor	NIS, sodium iodide symporter
Fibroblast growth factor receptor 1	Nuclear hormone receptors e
Type I receptor serine/threonine kinases	AR, androgen receptor

LIGANDS
DHT, dihydrotestosterone	Insulin‐like growth factor 1
EGF	L‐glutamic acid
GnRH I	Matrix metalloproteinase 2
IL‐6	NAAG, N‐acetyl‐aspartyl‐glutamate

These Tables list key protein targets and ligands in this article which are hyperlinked to corresponding entries in http://www.guidetopharmacology.org, the common portal for data from the IUPHAR/BPS Guide to PHARMACOLOGY (Southan et al., 2016), and are permanently archived in the Concise Guide to PHARMACOLOGY 2015/16 (^a,b,c,d,eAlexander et al., 2015a,b,c,d,e).

Introduction

Prostate specific membrane antigen (PSMA) also known as glutamate carboxypeptidase II (GCPII), N‐acetyl‐L‐aspartyl‐L‐glutamate peptidase I (NAALADase I) or N‐acetyl‐aspartyl‐glutamate (NAAG) peptidase, is an enzyme that is encoded by the folate hydrolase (FOLH1) gene in humans (O'Keefe et al., 1998). PSMA/GCPII plays many different roles and is expressed in different tissues such as the prostate, kidney, small intestine, central and peripheral nervous system and thus is recognized by different names. This review focuses on the biological role of PSMA/GCPII, with the main emphasis on prostate cancer (PCa) and neurological diseases. PSMA/GCPII's exact function in PCa is unknown; however, many studies have linked its role to tumour progression and carcinogenesis (Yao et al., 2008). In the brain, PSMA/GCPII metabolizes the neurotransmitter NAAG. PSMA/GCPII has now been identified as a target for therapeutic interventions and diagnostics in various neurological disorders and in PCa (Bařinka et al., 2012).

PSMA/GCPII was first characterized by the murine monoclonal antibody 7E11, derived from mice immunized with partially purified, cell membrane fractions, isolated from the human prostate adenocarcinoma (LNCap) cell line (Horoszewicz et al., 1986). Immunohistochemical analysis revealed high expression of PSMA/GCPII in the epithelial cells of the prostate with an intense overexpression in the cancerous tissue, compared with normal or hyperplastic prostates. Other tissues have also shown to express lower amounts of PSMA/GCPII, for example epithelia of small bowel and the proximal tubules of the kidney (Chang et al., 2000). PSMA/GCPII is encoded by a gene that consists of 19 exons, spanning 60 kilobases (Kb) of genomic DNA. The cDNA encoding PSMA/GCPII is 2.65 Kb in length and is mapped to chromosome 11 (Israeli et al., 1993; O'Keefe et al., 1998). It is a class II transmembrane glycoprotein, with a short N‐terminal cytoplasmic tail of 1–18 amino acids (aa), a single membrane‐spanning helix of 19–43 aa and an extracellular part, consisting of 44–750 aa with an approximate molecular weight of 84 kDa (Barinka et al., 2004). The short N‐terminal cytoplasmic tail interacts with membrane scaffold proteins that govern the endocytosis of some of the PSMA/GCPII bound substrates, such as clathrin, clathrin adaptor protein 2, filamin A (FLNa), and caveolin‐1 (Anilkumar et al., 2003; Rajasekaran et al., 2003; Anilkumar et al., 2006; Goodman et al., 2007). The bulk of this transmembrane protein is the extracellular part, which is further divided in three domains, namely, the protease (57–116 aa and 352–590 aa), apical (117–351 aa) and the C‐terminal domain or the dimerization domain (591–750 aa) and collectively performs the substrate/ligand recognition role (Davis et al., 2005; Mesters et al., 2006; Bařinka et al., 2012) (Figure 1). The 3‐dimensional extracellular structure of PSMA/GCPII closely resembles the transferrin receptor (Tfr) with 54% homology with Tfr and 60% with Tfr2 (Kawabata et al., 1999; Lawrence et al., 1999). This homology has been found both at the levels of aa (PSMA: 276–592 aa and Tfr: 316–606 aa) and domain organization (Mahadevan and Saldanha, 1999). In the small intestine, PSMA/GCPII is located in the brush border of the proximal jejunum, where it acts as a hydrolase on poly‐γ‐glutamated folate and actively transports the mono‐glutamylated folates into the blood stream (Pinto et al., 1996; Luthi‐Carter et al., 1998; Zhao et al., 2009; Navrátil et al., 2014). It thus exhibits both hydrolysing and endocytic functions. However, it is to be noted that PSMA/GCPII mediated poly‐γ‐glutamated folate hydrolysis is only observed in pigs and humans whereas, in rats, the same hydrolysing activity requires an intracellular enzyme called γ‐glutamyl hydrolase FOLH1 that is abundant in the postprandial pancreatic secretion (Shafizadeh and Halsted, 2007). The endocytic internalization is through clathrin coated pits, which are specific for receptor‐mediated endocytosis. In the nervous system, PSMA/GCPII, also termed as NAAG‐Hydrolase, catalyses the hydrolysis of NAAG (a widely distributed neurotransmitter in the mammalian brain) to glutamate and N‐acetylaspartate (NAA) (Robinson et al., 1987; Coyle, 1997; Neale et al., 2000).

Figure 1.

Schematic representation of PSMA/ GCPII transmembrane protein (homodimer). Adapted and modified from Bařinka et al., (2012).

The various roles of PSMA/GCPII in different tissues have enabled the exploration of various therapeutic approaches to target the delivery of drugs and small molecules specifically to PSMA/GCPII‐expressing cells. As mentioned above, PSMA/GCPII expression levels are higher in the malignant tissues of different origin than the normal/healthy tissues (Chang et al., 1999a,b; Milowsky et al., 2007; Morris et al., 2007; Haffner et al., 2009). This directly implies a role of PSMA/GCPII in cancer progression and invasion (Conway et al., 2006; Yao and Bacich, 2006; Yao et al., 2010b). Thus, this makes PSMA/GCPII highly desirable for imaging and treatment of solid tumours or be used to develop targeted drug inhibitors that block its enzymatic activity.

Physiological function of PSMA/GCPII and its role in prostate cancer

In PCa, the expression of PSMA/GCPII is negatively regulated by androgens (Israeli et al., 1993) and is favoured by other growth factors such as, basic fibroblast growth factor, TGF and EGF. The increased PSMA/GCPII expression in PCa tissues leads to an increased ability to process folate. However, the direct role of PSMA/GCPII in PCa metastasis is still unknown.

Prostate cancer involves a range of different molecules that participate in different signalling pathways. The prostate gland is regulated by the androgen hormone (testosterone), and this hormone plays an important role in the development and maintenance of the organ. The androgen‐signalling axis is directly involved in PCa. Under normal circumstances, testosterone is converted to an active metabolite dihydrotestosterone (DHT) via 5‐α reductase enzyme in the cytoplasm. DHT binds and activates the androgen receptor (AR), which translocates to the nucleus to bind its target gene and regulate gene expression (Zhou et al., 1995; Wright et al., 1996) (Figure 2). However, the AR can also be activated via insulin growth factor 1, EGF and IL‐6 signalling pathways (Lonergan and Tindall, 2011). Targeting the androgen axis has now become the prime therapeutic target for PCa. Hormonal therapy, known as androgen ablation therapy, with luteinizing‐hormone‐releasing hormone agonists, is mostly used for advanced stages of the disease. However, this therapy fails after 18–24 months, indicated by rising prostate specific antigen (PSA) levels in the blood (Pienta and Bradley, 2006). The increased PSA levels along with the resurgence in the expression of AR‐regulated genes lead to an aggressively malignant, metastatic and non‐treatable type of PCa, called castration‐resistant PCa. On the other hand, hormonal therapy on androgen sensitive cells has been shown to increase PSMA/GCPII levels. Interestingly, in this study, it was observed that the increased expression of PSMA/GCPII makes the cells less invasive in nature. This was indirectly confirmed by PSMA/GCPII knockdown (KD) in LNCap cells (which endogenously express PSMA/GCPII), which resulted in a fivefold increase in invasive activity (Ghosh et al., 2005b). However, other studies have contradicted this finding and have shown that reduced expression of PSMA/GCPII leads to reduction in invasiveness (Guo et al., 2011; Conway et al., 2013).

Figure 2.

Schematic representation of PSMA/GCPII regulation in prostate cancer cells Adapted and modified from Ghosh and Heston (2005a). Up‐regulation and down regulation of PSMA by Ca^{2 +} ions and AR, respectively, is shown. (A) PSMA upregulation: NAAG and polyglutamated folates (PGF) are enzymically cleaved to folates and glutamates. The folates are taken up by the RFC or FBP present on the cell membrane. The glutamates activate the metabotropic glutamate receptors, which on activation, lead to the efflux of Cl⁻ ions and influx of Ca^{2 +} ions. Ca^{2 +} ions further alter the expression of PSMA in two ways. First by activating the inactive transcription factor NFATc1 (which is a transcriptional activator of PSMA enhancer [PSME]), or cause activation of calpain, which cleaves FLNa. The truncated FLNa binds to AR and localizes to the nucleus and suppresses AR‐mediated transactivation. This effect leads to the upregulation of PSMA expression. (B) PSMA downregulation. Under normal conditions the cells do not express PSMA. Testosterone, an androgen is taken up by cells and is converted by 5‐α reductase enzyme to the active metabolite DHT. AR binds to the DHT and translocates it to the nucleus, where they activate the androgen‐regulated gene, thus down regulating PSMA expression. AR would also interact and sequester the transcription factor AP1 or tissue‐specific transcription factors, such as, SRY and SOX, which suppress the transcription of PSME.

A possible signalling pathway that leads to increased expression of PSMA/GCPII in PCa cells has been shown to be regulated by the PSMA/GCPII enhancer (PSME) (1.2 Kb, located within the third intron of FOLH1) (O'Keefe et al., 2000; Watt et al., 2001). As mentioned before, PSMA/GCPII expression is negatively regulated by androgen, but it is also positively regulated by Ca^{2 +} ions. The enzymatic activity of PSMA/GCPII hydrolyses the polyglutamated folates into folates and glutamates. The folates are internalized by the cells via reduced folate carriers (RFC) and folate‐binding proteins (FBP). The released glutamates bind and activate the metabotropic glutamate mGlu3receptors, leading to efflux of Cl⁻ and influx of Ca^{2 +} ions, via a CaT‐like calcium channel (Shuba et al., 2000; Wissenbach et al., 2001). The influx of Ca^{2 +} further modulates the expression of PSMA/GCPII by two pathways. The first is by activating the transcriptional activator of PSME (NFATc1). The NFATc1 protein undergoes dephosphorylation via calcineurin and is translocated to the nucleus and activates transcription of PSMA/GCPII. The second pathway is by activating calpain, which cleaves FLNa (a membrane scaffold protein, which facilitates internalization of substrates bound to PSMA/GCPII). Cleaved FLNa binds to the AR and localizes to the nucleus and suppresses AR‐mediated transactivation. Under normal circumstances, AR would repress PSME by binding to AP1 or tissue specific transcription factors, such as SRY and SOX in the nucleus, which would explain why PSMA/GCPII levels are negatively regulated by androgens (Figure 2) (Ghosh and Heston, 2005a).

In PCa, the degree of PSMA/GCPII expression is positively correlated with the Gleason score and disease progression (i.e. the more advanced the stage of the disease, the greater the level of PSMA/GCPII expression in the prostate tissue). In addition, the rapid internalization and recycling of this receptor means that high concentrations of a targeted drug can be accumulated in PSMA/GCPII positive cells (Behnam Azad et al., 2015).

PC3 cells were transfected to express PSMA/GCPII. These cells showed a significant increase in proliferation levels when compared with non‐transfected counterparts. This increase was attributed to the enzymatic activity of PSMA/GCPII, as the effect was reversed in the presence of a PSMA/GCPII enzyme inhibitor (Yao et al., 2010a). Interestingly, PSMA/GCPII has been shown to be expressed in two AR‐negative canine cell lines of PCa; Leo and Ace‐1. Leo cells form brain metastases in a xenograft mouse model, while Ace‐1 cells metastasise to the bone to form osteoblastic and osteolytic lesions that mimic human metastases (Wu et al., 2014).

Many studies have assessed the role of PSMA/GCPII in the initiation, progression and metastasis of PCa cells. The effect of inhibiting PSMA/GCPII in PCa has been studied, using lentiviral delivery of PSMA/GCPII shRNA in the PSMA/GCPII positive cell line, LNCaP (Guo et al., 2011). The results from this study indicated that reduced PSMA/GCPII expression suppressed cell growth, induced significant cell cycle arrest at G₀/G₁ and reduced invasiveness. It has further been hypothesised that this reduction in cell proliferation and invasion is due to the deactivation of the p38 MAPK pathway that was observed following PSMA/GCPII silencing (Zhang et al., 2013). As well as the p38 MAPK pathway, the KD of PSMA/GCPII has been shown to down‐regulate p‐Akt in the P13K/Akt signalling pathway, and Akt has been proposed as a downstream effector gene of PSMA/GCPII (Guo et al., 2013b).

Recently, a relationship has been established between matrix metalloproteinase 2 (MMP2) and PSMA/GCPII (Conway et al., 2013). Increased MMP expression has been widely associated as a characteristic of tumour progression and metastasis (Evans et al., 2015). Sequential proteolysis of laminin (a component of the extracellular matrix) by MMP2 and PSMA/GCPII produced small peptide fragments that increased the rate of endothelial cell migration. This two‐step degradation pathway highlights the possible role that PSMA/GCPII activation has on the induction of angiogenesis and metastasis (Conway et al., 2013).

PSMA/GCPII in other cancers

In a study of 92 patients with invasive breast cancer, 98% had PSMA/GCPII staining that was confined to the neovasculature of the tumour. In all of the cases examined, normal breast tissue as well as carcinoma cells were PSMA/GCPII negative. In the same study, the level of PSMA/GCPII expression in 14 patients with brain metastases as a result of breast cancer was also determined. All 14 samples stained strongly for PSMA/GCPII, and interestingly, ten of the brain metastasis patients had a similar PSMA/GCPII staining score to their respective primary breast carcinoma (Wernicke et al., 2014).

Samples from 150 lung cancer patient samples were examined for the presence of PSMA/GCPII in the tumour cells and surrounding neovasculature (Wang et al., 2015). Approximately 85% of non‐small cell lung cancers (NSCLC) and 70% of small cell lung cancer (SCLC) samples showed PSMA/GCPII expression in the neovasculature but was absent from normal blood vessels. Interestingly, while there was strong PSMA/GCPII staining within the tumour cells of NSCLC (over 50% of samples exhibited staining), this was absent from the SCLC tumours. It is not surprising that the PSMA/GCPII expression differs between these cancer types as they have their own unique biology and genetic aberrations (Oser et al., 2015).

Evaluation of PSMA/GCPII expression in patients with colorectal cancer revealed two interesting correlations. There were positive relationships between PSMA/GCPII expression and metastasis to distant sites as well as to invasion of surrounding vasculature (Abdel‐Hadi et al., 2014). PSMA/GCPII mRNA expression levels are elevated in pancreatic adenocarcinomas compared with normal tissues. In this study, those patients with high expression of PSMA/GCPII had a significantly shorter overall survival than those with low expression levels, and PSMA/GCPII expression was shown to be correlated to Tumour‐Node‐Metastasis stage (Ren et al., 2014).

In 167 bladder cancer patients of various subtypes (adenocarcinoma, small cell, urothelial and squamous cell), samples tested positive for PSMA/GCPII staining in the tumour neovasculature, with small cell tumour vasculature showing the greatest intensity. As with PSMA/GCPII expression in lung cancer samples, there were discrepancies in PSMA/GCPII staining within the tumour between the different cancer cell types (Samplaski et al., 2011).

In a recent study, PSMA expression in four grades of glioma (astrocytoma) was determined. The surrounding vasculature of grade IV gliomas (glioblastoma multiforme) stained heavily for PSMA, while grades II and III showed little staining of the surrounding blood vessels but some staining of tumour parenchyma cells. The level of PSMA expression for grade IV tumours was three times greater than that seen in normal brain tissue (Nomura et al., 2014).

Physiological function of PSMA/GCPII and its role in neurological diseases

In the brain, PSMA/GCPII acts as an enzyme and performs NAAG‐hydrolyzing activity, as illustrated in Figure 3. Using Mab GCP‐04, which binds the extracellular portion of PSMA/GCPII, approximately 50–300 ng of PSMA/GCPII per mg of total protein depending on the region was detected, with astrocytes showing GCPII expression in all parts of the brain (Šácha et al., 2007). In astrocytes, the drug valproic acid increased the stability of GCPII due to the acetylation of lysine residues (Choi et al., 2014).

Figure 3.

Schematic representation of the role of PSMA/GCPII in the brain. PSMA/GCPII is expressed on the astrocytes and can lead the cells towards a neurodegenerative or a neuroprotective outcome. (A) in the neuroprotective role, to reduce the glutamate excitotoxicity, NAAG inhibitors can be used to competitively bind to the PSMA or to the glutamate receptor, avoiding the hydrolysis of NAAG into NAA and glutamate. Due to this effect, NAAG interacts with the metabotropic receptors on astrocytes, which leads to the production of growth factors and thus acts as a neuroprotective agent. NAAG and glutamate are also regularly recycled into the presynaptic neuron via G‐protein coupled pathway and glutamate‐glutamine pathway, respectively. (B) in the neurodegenerative role, NAAG, a neuropeptide released from the presynaptic neuron in the brain interacts with the PSMA/GCPII transmembrane protein. The interaction causes hydrolysis of NAAG into NAA and glutamate. The excess of glutamate leads to glutamate excitotoxicity in the synaptic cleft and further activates the ionotropic and metabotropic receptors on the post synaptic terminal of the neuron, leading to degeneration of cells.

According to Figure 3, NAAG is synthesized in neurons from glutamate and NAA with the help of NAAG‐synthetase I and is stored in synaptic vesicles of presynaptic axon terminals (Williamson and Neale, 1988; Becker et al., 2010; Neale et al., 2011). A second structurally related protein, NAAG synthetase II has been discovered with similar NAAG synthesizing activity. The physiological role of this protein remains to be determined (Lodder‐Gadaczek et al., 2011). On depolarisation, NAAG is released from the synaptic vesicles into the extra‐synaptic space, in a calcium‐dependent manner. After being released, NAAG can interact with its target receptors, mainly the metabotropic glutamate mGlu3 receptor located on both presynaptic nerve terminals and astrocytes. The interaction with mGlu3 receptors leads to the signalling of G‐protein coupled pathway, resulting in the reduced release of neurotransmitters such as GABA from pre‐synaptic nerve terminals (Zhao et al., 2001; Sanabria et al., 2004; Zhong et al., 2006; Romei et al., 2013). At the astrocyte level, the NAAG and mGlu3 receptor interaction stimulates the secretion of TGFs and peptides (Bruno et al., 1998; Thomas et al., 2001). Thus, the interaction of NAAG with mGlu3 receptors on either presynaptic nerve terminals or astrocytes are both considered to be neuroprotective in nature (Slusher et al., 1999; Gurkoff et al., 2013; Cao et al., 2016). Another fate of NAAG in the extra‐synaptic space involves the inactivation of NAAG in the presence of PSMA/GCPII (NAAG peptidase) into NAA and glutamate and the reaction products are transported to astrocytes and oligodendrocytes (Riveros and Orrego, 1984). About 80–90% of the glutamate in the neurons is recycled via the glutamate‐glutamine cycle and thus a regulated recycling of glutamate maintains the healthy functioning of the brain.

Decades of research has shown that in order to be effective at inhibiting PSMA/GCPII activity, potential compounds must meet two criteria: a glutarate moiety to bind the glutamate recognition site and a zinc chelating group for the zinc atoms at the PSMA/GCPII active site (Bařinka et al., 2012). At present, there are several classes of PSMA/GCPII inhibitor, which are phosphonate‐based, urea‐based, thiol‐based and also hydroxamate derivatives (Tsukamoto et al., 2007; Bařinka et al., 2012; Ferraris et al., 2012). It is important to note that these inhibitors do not have an effect on glutamate function at normal levels of activity (Slusher et al., 1999).

One of the first successful inhibitors of GCPII, 2‐(phosphonomethyl)pentanedioic acid (2‐PMPA), a phosphonate analogue of glutamate, was designed in 1996 and is extremely potent, with an IC₅₀ of 300 pM (Jackson et al., 1996). 2‐PMPA (NAALADase inhibitor) has shown efficacy in inhibiting PSMA/GCPII activity in over 20 various in vivo models of neurological disease, including schizophrenia, ischemic brain injury and neuropathic pain (Table 1). Unfortunately, 2‐PMPA has exhibited poor pharmacokinetics due to its highly polar nature, which has led to structure–activity relationship studies, yielding more potent inhibitors, such as GPI5232 and VA‐033 (Zhong et al., 2014). In a recent study, 2‐PMPA was administered to the brain intranasally in non‐human primates and, 30 min later, the level of 2‐PMPA was below the limit of detection in the plasma (<50 nM), but reached 1.5 μM in the cerebrospinal fluid. This study highlights a direct pathway for delivering 2‐PMPA to the CNS via the olfactory nerve (Rais et al., 2015).

Table 1.

List of GPCII/NAALADase inhibitor (2‐PMPA) studies focussing on neurological disorders

Neurological disorder	Dose	Route	Animal model	End‐point comments	References
Neuropathic pain	50 and 100 mg·kg−1	IV	Sprague–Dawley rats	Significant increase in paw‐withdraw threshold in rats with partial ligation of the left from less than 5 g at 45 min to ~14 g and 19 g, respectively.	(Chen et al., 2002)
				Significant decrease in the spontaneous ectopic discharges recorded from injured sciatic afferent nerves from ~15 imp.·s‐1 for the control group to ~10 imp.·s‐1 and ~5 imp.·s‐1, respectively, for treatment groups
				Significant improved allodynia induced by sciatic nerve ligation
Chronic pain	50 mg·kg−1	IP	Wistar rats/Sprague–dawley rats	After 2‐PMPA dose administration, the maximum concentration achieved in brain ECF was 30 μM.	(Nagel et al., 2006)
				Following 2‐PMPA administration, a steady increase of NAAG in brain ECF was observed, reaching a peak concentration of 3 μM.
				The study concludes that the low concentration of NAAG in the brain ECF is unlikely to have an effect on any known target, and so they say that the effect of inhibitor is unlikely to be mediated by either NAAG or mGluR3.
Diabetic neuropathy	n/a	n/a	n/a	2‐PMPA rescued the dorsal root ganglion neurons from glucose induced programmed cell death	(Berent‐Spillson et al., 2004)
				mGluR3 antagonist EGLU reversed the 2‐PMPA protection profile in three assays testing for apoptosis; caspase‐3 and −9 cleavage, and TUNEL
				This reversal was not seen with either the group I mGluR antagonist AIDA, nor the group 3 mGluR antagonist MSOP.
				At high glucose concentrations, 2‐PMPA was able to maintain neurite growth and viability.
Inflammatory pain	100 μg	IT	Sprague–Dawley rats	In the rat formalin test (model for inflammatory pain), 2‐PMPA attenuated the phase 1 and 2 biphasic flinching when rats were pre‐treated with inhibitor prior to the formalin test.	(Yamamoto et al., 2001b)
				In the post‐treatment test, 2‐PMPA had no effect on biphasic flinching.
				2‐PMPA did not have an effect on the reaction time of the rats in the hot plate test
Inflammatory pain	10 μg	ICV	Sprague–Dawley rats	Significant reduction in the phase 1 and 2 biphasic flinching in the rat formalin test with 2‐PMPA.	(Yamamoto et al., 2008)
				These effects were attenuated following the administration of group II mGluR antagonist LY341495
Pain	100–400 μg	IT	Sprague–Dawley rats	2‐PMPA improved the level of allodynia induced by paw carrageenan injection.	(Yamamoto et al., 2001a)
				However, this improvement was not observed for other models of pain (i.e. skin incision or mild thermal injury.)
Peripheral pain	1–100 μg	SC	Sprague–Dawley rats	Group II mGluR agonists (SLX‐3095‐1 and APDC) and 2‐PMPA reduced the pain response in the rat formalin test.	(Yamamoto et al., 2007)
				Group II mGluR agonists (SLX‐3095‐1, NAAG and APDC) and 2‐PMPA reduced the pain response in the carrageenan model of induced allodynia
				These effects were attenuated following the administration of group II mGluR antagonist LY341495
Brain injury from soman	50 mg·kg−1	IP	Sprague–Dawley rats	The nerve agent soman was shown to reduce NAAG levels in regions of the brain following exposure	(Guo et al., 2015)
				The administration of 2‐PMPA reduced neuronal cell death in some (but not all) regions of the brain following exposure to soman; these regions include the entorhinal cortex, piriform cortex and the amygdala
Spinal cord injury	1–4 μM	subarachnoid space	Sprague–Dawley rats	The co‐administration of somatostatin with 2‐PMPA and dynorphin A (causes ischemic spinal cord injury) along with 2‐PMPA as a spinal subarachnoid injection resulted in improved hindlimb motor scores by 24 h post injection.	(Long et al., 2005)
				Significant reduction of the extracellular concentrations of glutamate in cerebrospinal fluid was observed when 2‐PMPA was co‐administered with dynorphin A, in comparison to mice treated with dynorphin A alone
Ketamine‐induced neurotoxicity	n/a	n/a	n/a	2‐PMPA attenuated the decrease in cell viability seen with 2 mM of ketamine in neuron–glia mixed culture by decreasing the loss of nodes/cell from <0.5 with 50 μM of 2‐PMPA to >1.00 with 100 μM of 2‐PMPA but the same effect was not observed in cultures of neurons alone.	(Zuo et al., 2014)
				The authors conclude that glia cells must be involved in the neuroprotective effect that 2‐PMPA offers.
Neurotoxicity	n/a	n/a	n/a	The degree of neuroprotection offered by 2‐PMPA in an in vitro model of primary neurons (neurotoxicity) from rat embryos was 100%, 46%, 16% and 0% for hypoxia, glutamate, NMDA injury and veratridine‐induced injury	(Tortella et al., 2000)
Alcohol	50, 100 and 200 mg·kg−1	IP	Alcohol preferring P rat	Significant reduction in the consumption of ethanol by alcohol preferring rats by ~25% during their daily 1‐hour access to 10% (v/v) ethanol following treatment with 2‐PMPA	(McKinzie et al., 2000)
Cocaine	0.01–100 μM	Local exposure	Planarians (Dugesia dorotocephala)	2‐PMPA attenuated the C‐like hyperkinesias (i.e., stereotypical counts) induced in planarians by four compounds; glutamate, NMDA, pilocarpine and cocaine	(Tallarida et al., 2012)
Cocaine	100 mg·kg−1 and 30 mg·kg−1	IP / PO	Sprague–Dawley rats	The acquisition of the conditioned place preference (CPP) response was blocked by either 100 mg·kg−1 2‐PMPA i.p. or 30 mg·kg−1 p.o. of GPI 5693; an orally bioavailable NAALADase inhibitor. Administration of either drug in the absence of cocaine did not significantly alter the time spent in each chamber.	(Slusher et al., 2001)
				2‐PMPA did not significantly reduce the CPP response that has been shown to be induced by food.
Cocaine	10, 30 and 100 mg·kg−1	IP	Long‐Evans rats	Inhibition of intravenous self‐administration of cocaine.	(Xi et al., 2010b)
				Dose‐dependent reduction in extracellular dopamine and glutamate after 2‐PMPA administration.
				Successful prevention of cocaine‐induced reinstatement of drug‐seeking behaviour
Cocaine	1, 10, 30 and 100 mg·kg−1	IP	Long‐Evans rats	Significant inhibition of cocaine self‐administration under progressive‐ratio (PR) reinforcement conditions	(Xi et al., 2010a)
				Significant inhibition of cocaine‐enhanced brain‐stimulation reward (BSR) in rats.
				Dose‐dependent reduction in cocaine‐induced extracellular dopamine in the nucleus accumbens (NAc).
Cocaine	100 mg·kg−1	IP	Sprague–Dawley rats	Increased locomotor activity of the rats following 2‐PMPA administration	(Shippenberg et al., 2000)
				No significant difference in the distance travelled between the control (saline) and the mice pre‐treated with 2‐PMPA, following cocaine administration after 15 min
				2‐PMPA prevents the sensitization that develops to the locomotor activity, following cocaine use.
Cocaine‐kindled seizures	10, 30 and 100 mg·kg−1	IP	Swiss‐Webster mice	2‐PMPA was shown to prevent the expression and development of cocaine kindling.	(Witkin et al., 2002)
				Mice pretreated with 2‐PMPA showed a lower % of mice convulsing when challenged with 60 mg·kg−1 of cocaine following 10 days of treatment with 40 mg·kg−1 of cocaine.
				The inhibitor dose did not alter the behaviour of the mice in the inverted screen test and locomotor activity.
Ischemic brain injury	500 mg·kg−1	IV	Sprague–Dawley rats	Reduction in the extent of injury (magnitude of protection was 54%) in an established focal ischemia model.	(Jackson et al., 1996)
				Using microdialysis, it was shown that the levels of glutamate in the mice treated with inhibitor showed no increase following middle cerebral artery occlusion (MCAO).
				In non‐ischemic mice, the inhibitor did not alter basal levels of glutamate
Ischemic strokes	n/a	n/a	n/a	Ischemic strokes result in an increase in the extracellular concentration of glutamate, which tends to overactivate the N‐methyl‐D‐aspartate receptors (NMDARs)	(Khacho et al., 2015)
				The inhibitor was shown to increase the amplitude of the synaptic NMDAR excitatory postsynaptic currents (EPSCs) while decreasing the extrasynaptic NMDAR ESPCs.
Schizophrenia	50–150 mg·kg−1	IP	C57BL/6 and DBA/2 mice	Group II agonist LY354740 was able to moderate the effects of PCP on prepulse inhibition (PPI) of acoustic startle in DBA/2 but not C57BL/6 mice	(Profaci et al., 2011)
				2‐PMPA was unable to affect the PCP‐mediated PPI in either strain.
Schizophrenia	100 mg·kg−1	IP	C57BL/6 mice, DBA/2 mice and Sprague–Dawley rats	2‐PMPA reduced the motor activation in PCP or d‐amphetamine induced models of schizophrenia	(Olszewski et al., 2012)
				These effects were not observed in mice that are homozygous for a deletion in GCPII
				2‐PMPA attenuated the effect of dizocilpine (MK‐801) in an object recognition test; a model for the cognitive impairment seen in schizophrenia
Memory Process	150 mg·kg−1	IP	Swiss‐Webster mice	No negative impact of the dose on long term memory in the passive avoidance task.	(Łukawski et al., 2008)
				Increased latency to enter the dark box on training day from ~24 s in control mice to ~103 s in treated mice.
				The dose impaired spontaneous alternation in the Y‐maze task.
				NAALADase inhibition may impair alternation behaviour
Memory Process	0.2–100 mg·kg−1	IP	C57BL/6 mice	In comparison to the control mice, the mice treated with 2‐PMPA spent significantly more time exploring the novel object compared with the familiar object on day two.	(Janczura et al., 2013)
				Mice that were homozygous for a deletion in GCPII displayed the same results as those treated with the inhibitors.
Neurodegenerative disorders	30 mg·kg−1 for each dose	IN/IP	Wistar rats, Cynomolgus monkey	i.n. administration showed higher levels of 2‐PMPA accumulation in olfactory bulb and the cerebellum compared with i.p. administration, but had similar plasma profiles.	(Rais et al., 2015)
				In a non‐human primate study, 2‐PMPA achieved micromolar concentrations following i.n. delivery.

In an attempt to reduce the highly polar nature of the phosphonate‐based inhibitors, researchers substituted the phosphorus group with thiol, leading to the design of 2‐(3‐mercaptopropyl) pentanedioic acid (2‐MPPA) (Majer et al., 2003). 2‐MPPA was the first orally bioavailable PSMA/GCPII inhibitor with studies conducted in humans (Majer et al., 2003; Van der Post et al., 2005). Similar to 2‐PMPA, 2‐MPPA exhibited efficacy in a range of animal models of neurological disease, including neuropathic pain, neuropathy and ALS (Ghadge et al., 2003; Potter et al., 2014). Although, no adverse CNS effects were reported, further studies were halted as a result of the potential for immune toxicity, common with thiol‐containing drugs, as well as a relatively low therapeutic potency (Bařinka et al., 2012).

A further class of PSMA/GCPII inhibitors comprises the urea‐based inhibitors (Kozikowski et al., 2001). These inhibitors were initially developed by medicinal chemist Alan Kozikowski from the NAAG‐based mimic 4,4′‐phosphinicobis(butane‐1,3‐dicarboxylic acid) (Nan et al., 2000). ZJ43 is a well‐studied urea‐based analogue developed by this group that reduced PCP‐induced motor activation in a rodent model of schizophrenia (Olszewski et al., 2004). In the rat formalin test of neuropathic pain, IV administration of ZJ43 suppressed both phases of agitated behaviour and attenuated allodynia following sciatic nerve ligation (Yamamoto et al., 2004). In a rat model of traumatic brain injury, ZJ43 significantly reduced the number of ipsilateral degenerating neurons and significantly reduced ipsilateral astrocyte loss (Zhong et al., 2005). There are several advantages to using urea‐based inhibitors including ease of their synthesis as well as subsequent modification and conjugation (Ferraris et al., 2012). Molecules based on these early stage urea‐based prototypes have been conjugated to various imaging agents and therapeutic payloads to diagnose and treat PCa at the preclinical stage (Tables 2, 3, 4). However, in terms of treating neurological disorders, these molecules exhibit poor pharmacokinetics, as they tend to have low oral bioavailability and are very limited in their ability to cross the blood–brain barrier (BBB) due to their hydrophilic nature (Zhong et al., 2014).

Table 2.

List of PSMA based therapeutics developed with focus on Chemotherapy and Gene therapy

Therapeutic payload	Delivery vector	Targeting ligand	In vitro model	In vivo model	Concluding remarks	References
Chemotherapy
Methotrexate	PAMAM dendrimer	Glutamate urea	LnCap, PC3 cells	N/A	Nanoparticle binding and uptake observed in LnCap cells (PSMA+) and not in PC3 cells (PSMA‐).	(Huang et al., 2014)
CpG and Doxorubicin	Dendrimer	DNA‐A9 PSMA (RNA aptamer hybrid)	RAW264.7, LNCaP, 22RV1, DU145, and PC3 cells	BALB/c mice (22RV1 xenograft tumour model)	Seventy‐eight per cent decrease in tumour volume observed in animals treated with targeted nanoparticles compared with controls.	(Lee et al., 2011)
Mono methylauristatin E	N/A	PSMA Monoclonal antibody	LNCaP, C4–2, CWR22rv1, and PC3, DU145, MDA PCa2b cells	Athymic nude mice (C4–2 xenograft tumour model)	1000‐fold better cell death observed in cells that expressed PSMA.	(Wang et al., 2011)
					Effective in treating tumours in a C4–2 xenograft model that had progressed following treatment with docetaxel.
Mono methylauristatin E	N/A	PSMA monoclonal antibody	LuCap 96CR PDM model	SCID mice	Complete tumour regression in LuCaP 96 CR xenograft mouse model, when treated with drug‐antibody conjugate.	(DiPippo et al., 2015)
					The PDX model displayed a high level of PSMA expression.
Docetaxel	PLA–PEG/PLGA‐PEG	ACUPA (small molecule)	N/A	Athymic nude mice (LnCap xenograft model) / Male Sprague–Dawley rats / Cynomolgus monkeys	Targeted nanoparticles had increased tumour accumulation and suppressed tumour growth in tumour bearing mice.	(Hrkach et al., 2012)
Epirubicin	PEG	RNA aptamer (A10)	PC3 and LnCap cells	N/A	Reduced toxicity observed in PSMA‐ cell line when treated with PEG‐Apt‐Epi formulation versus Epirubicin alone.	(Taghdisi et al., 2013)
					No difference in cell viability observed in PSMA+ cell line, when treated with epirubicin alone or PEG‐Apt‐Epi
Doxorubicin	PLA–PEG micelles	RNA aptamer (A10)	CWR22Rν1 cells	Athymic nude mice (CWR22Rν1 cells xenograft model)	Doxorubicin‐loaded micelles showed increased uptake in PSMA+ 22Rv1 cells via PSMA‐mediated endocytosis.	(Xu et al., 2013)
Doxorubicin	Liposome	RNA aptamer	LnCap and PC3 cells	BALB/c athymic nude mice (LnCap xenograft model)	Targeted liposomes containing docetaxel were significantly more toxic to PSMA + cell line (LNCaP) than compared with PC‐3 cell line (PSMA ‐) and also showed significant tumour regression in comparison untargeted liposomes or drug alone.	(Baek et al., 2014)
Circumin	PLGA/PLL/PVA	PSMA antibody	LnCap, DU145, PC3 and C4–2 cells	Athymic nude mice	PSMA antibody tagged PLGA curcumin nanoparticles showed specific targeting in C4–2 cells (PSMA +) in a xenograft mouse model	(Yallapu et al., 2014)
Indenoisoquinoline topoisomerase I inhibitor	N/A	DUPA	22RV1 cells	Athymic nude mice (22RV1 cells xenograft model)	After 40 days, complete regression of tumour growth was observed in the animals treated with DUPA‐drug conjugate.	(Roy et al., 2015)
					DUPA conjugate exibited PSMA receptor‐mediated uptake.
Toremifene	PLGA‐PEG nanoparticles	PSMA antibody	PC3M cells	Athymic nude mice	Incorporation of oestrogen receptor‐α (ERα) blocker, toremifene, into PLGA nanoparticle resulted in a 15‐fold increase in its uptake when compared with free toremifene.	(Hariri et al., 2015)
					The model used was modified PC‐3 cells to express PSMA (PC3M) orthotopically implanted into mouse prostate.
Doxorubicin	PEGylated PLA nanoparticle	RNA aptamer (A10)	LNCaP and PC‐3 cells and canine prostatic adenocarcinoma cell line (cHSA) cell line	BALB/c mice	Mice treated with aptamer‐conjugate showed a twofold increase in the amount of necrotic tissue in the tumour compared with the non‐targeted nanoparticles.	(Tang et al., 2015)
					A significant reduction in tumour burden in the aptamer group after 7 days (~70%) was observed, when compared with the non‐targeted group (14% reduction).
Docetaxel	HPMA co polymer	DUPA	C4–2 and PC3 cell line	Athymic nude mice (C4–2 xenograft model)	Lower IC50 of targeted nanoparticles with DTX (3.18 ± 0.42 nM) compared with untargeted counterparts (4.14 ± 1.03 nM).	(Peng et al., 2013)
					Significantly (p < 0.05) reduction in tumour volume with targeted nanoparticles vs. untargeted nanoparticles (<800 mm3 vs. >1500 mm3 at 42 days post injection, respectively).
Thapsigargin	Drug conjugated with peptide (Asp‐Glu‐Glu‐Glu‐Glu‐Glu)	N/A	LNCaP, MDA‐PCa2b, TSU, SN12C, and MCF‐7 cells, CWR22R‐H xenograft	Nude mice (Harlan Sprague Dawley)	This delivery vector takes advantage of the glutamate carboxypeptidase activity of PSMA to cleave the glutamate residues from the peptide to form 8‐O‐(12‐aminododecanoyl)‐8‐O‐debutanoyl thapsigargin (12ADT)‐βAsp.	(Denmeade et al., 2012)
					This prodrug is over 60‐fold more toxic to PMSA + cells than PSMA – cells.
					Delivery of thapsigargin in the prodrug form is 150‐fold more effective than free thapsigargin.
					Tumour regression and growth inhibition was observed in six in vivo tumour xenograft models with the prodrug.
Suicide enzyme yCD	N/A	N/A	LnCap and PC3 cells	N/A	Significant reduction of cell viability in PSMA+ cell lines (~ 60% reduction), in comparison to PSMA‐ cell line (PC3)	(Martin et al., 2014)
RNA aptamer A9 g	N/A	RNA apatmer A9 g	CT26, PC3 (PSMA +/−) and 22Rv1	SCID mice	A9 g binds and inhibits the enzymatic activity of PSMA.	(Dassie et al., 2014)
					*A9 g demonstrated the ability to reduce cell invasion (~ 75% reduction) and migration (~ 70% reduction) of prostate cancer cells engineered to express PSMA (PC‐3)
P13k Inhibitor ZSTK474 and immunotoxin J591PE	N/A	N/A	LnCap, C4–2 and BT‐549	BALB/c nude mice	The combination of the 2 drugs resulted in increased apoptosis and cell deaths in PSMA+ cell lines (LNCaP and C4–2) compared with PSMA‐ (BT549).	(Baiz et al., 2013)
TGX‐221	PEG‐PCL micelles	RNA aptamer A10	PC3, DU145 and LnCap cells	BALB/c nude mice	TGX‐221 is a selective inhibitor of the P13K p110β catalytic subunit.	(Zhao et al., 2012)
					Micelles that were targeted to PSMA resulted in a significant increase in uptake of drug in PSMA+ cells.
TGX‐221	PEG‐PCL micelles	–	LAPC‐4, LNCaP, C4–2 and 22RV1 cells	Nude mice (LAPC‐4, LNCaP, C4–2 and 22RV1 cells used for xenograft models)	*Complete inhibition of tumour growth in four mouse xenograft models (LAPC‐4, LNCaP, C4–2 and 22RV1).	(Chen et al., 2015)
					Induction of apoptosis and reduction of PSA levels by approximately 80% in the four models tested.
Cisplatin	PLGA‐b‐PEG NPs	RNA apatmer	LnCap cells	Sprague Dawley rats and Swiss Albino mice, BALB/c mice	Achieved greater maximum tolerated dose (MTD) for cisplatin was greater when encapsulated in NP than for free cisplatin.	(Dhar et al., 2011)
					Enhanced residence time in blood.
					Reduction in tumour growth at much lower platinum dose than than cisplatin.
Epigallocatechin 3‐gallate (EGCG)	PLGA‐PEG	DCL	LnCap cells and HUVEC cells	N/A	Significantly greater reductions in LNCaP cell viability when EGCG was complexed in targeted vs non‐targeted NP.	(Sanna et al., 2011)
Genetherapy
siRNA targeting Notch1	Two fusion proteins and a truncated protamine	PSMA antibody	LnCap adn DU145 cells	BALB/c nude mice	Significant KD of Notch1 in LNCaP mouse xenograft model	(Su et al., 2013)
siRNA targeting choline kinase (ChK)	PLL‐PEG‐PEI	PSMA inhibitor	PC3 cells	SCID and BALB/c mice	Theranostic vector contains I111 for imaging, bacterial cytosine deaminase and siRNA targeting Chk.	(Chen et al., 2012)
					Specific uptake of targeted nanoparticles into PSMA+ PC3‐PIP tumours was observed.
					Nanoparticles were well tolerated with no liver or kidney toxicity and no immune response.
siRNA targeting PSMA	Lentivirus	N/A	LnCap, DU145 and HEK 293 T‐cells	N/A	Suppressed growth, migration and invasion of PSMA+ LNCap and DU145 cell lines	(Guo et al., 2011)
siRNA targeting mTOR	Lentivirus	N/A	RWPE1, LNCap, C4–2b and HEK293 cells	SCID mice (C4–2b xenografts)	Tumour growth was significantly retarded in SCID mice treated with LV‐PSMA‐shmTOR.	(Du et al., 2013)
ShRNA targeting DNAPK	N/A	Aptamer A10–3	DU145, LnCap and PC3 cells	Athymic nude mice	25% reduction in DNAPK expression in human prostate tissue by immunostaining compared with control aptamers.	(Ni et al., 2011)
					60% cell death in LNCaP cells treated with aptamer‐targeted shRNA.
					Combination of A10–3‐DNAPK and radiation treatment (6 Gy) resulted in a significant and extended tumour response in LNCaP tumours but not PC‐3 tumours.
MiR‐15a and miR‐16‐1	PAMAM‐PEG	Aptamer A10–3.2	PC3 ad LnCap cells	N/A	Targeted nanoparticles showed sixfold higher transfection efficiency in LNCaP cells (in relation to untargeted NP).	(Wu et al., 2011)
					Targeted nanoparticles also showed a 5.47 fold lower IC50 compared with non‐targeted control in LNCaP cells in vitro.
HSV‐TK and connexin43 gene	PAMAM dendrimer	Folate	LnCap, PC3, NIH3T3, HepG2,	Nude mice	HSV‐TK and connexion43 gene expression was driven by PSMA promoter.	(Chen et al., 2013)
					Expression of these genes was confined to LNCaP cells (but absent from PC‐3 cells).
Bcl‐xL shRNA (in combination with docetaxel)	PEI‐PEG	Aptamer	LnCap cells	N/A	Three‐fold higher transfection efficiency in LnCap cells with targeted nanoparticles.	(Kim et al., 2010)
					LNCaP cell viability fell from 81% to 21% 48 h after treatment with NPs, with virtually no toxicity in PC‐3 cells observed.
					IC50 value for the NPs loaded with docetaxel + shRNA was 17‐fold lower than for the drug and shRNA/Lipofectamine.
PLK1 siRNA	Liposome	Folate	PC3 and 22Rv1	BALB/c nude mice	Dual targeted liposomes with folate (for PSMA targeting) and a cleavable peptide that reveals a cell penetrating peptide in the presence of PSA.	(Xiang et al., 2013)
					twofold greater KD of PLK1 mRNA by targeted NPs in 22Rv1 xenograft mouse model compared with the non‐targeted control.
MiR‐15a and miR‐16‐1	Atelocollagen	Aptamer 10–3.2	PC3 and LnCap cells	BALB/c nude mice	5.4 fold higher uptake efficiency with targeted formulation in LnCap cells	(Hao et al., 2016)
					miRNA complexed with targeted formulations four‐fold lower IC50 value compared with untargeted forulation
					In a human PCa bone metastasis model, mean survival time for mice treated with miRNA/ATE was 38 days, while those treated with miRNA/ATE‐APT had an average survival time of 57 days.

Table 3.

List of PSMA based therapeutics developed with focus on Immunotherapy and radiotherapy

Therapeutic payload	Delivery vector	Targeting ligand	In vitro model	In vivo model	Concluding remarks	References
Immunotherapy
CD‐40 Targeted adenovirus	CD‐40 Targeted adenovirus	N/A	RM1 cells (PSMA+)	N/A	Only combination therapy of Ad5‐huPSMA and Ad5‐IFNγ resulted in delayed growth of RM1‐PSMA tumours by CTL.	(Williams et al., 2012)
					Tumours that were PSMA‐ showed a weak inhibition of tumour growth.
DCs treated with recombinant survivin and recombinant PSMA	N/A	N/A	N/A	Patient study	Response Evaluation Criteria In Solid Tumours was 72.7% for the DC group and 45.4% in the control group.	(Xi et al., 2015)
					The median overall survival of patients in the DC vaccine group improved by 11 months compared with the control group
DCs	Adenovirus	N/A	RM1 cells (PSMA+)	N/A	DCs were transduced with Ad‐tPSMA, Ad‐m4‐1BBL and Ad‐tPSMA‐m4‐1BBL.	(Youlin et al., 2013)
					Highest level of T‐cell cytotoxicity in RM‐1 cells (~60%) was seen in the group transduced with Ad‐tPSMA‐m4‐1BBL.
T‐cells	Lentivirus expressing CAR (J591)	N/A	MS1PSMA,H5VPSMA	NOD/SCID mice	*The CAR T‐cells were shown to be capable of killing the PMSA+ cells while sparing the PSMA‐ lines in vitro.	(Santoro et al., 2014)
					In vivo;full regression of PMSA+ MS1 tumours following dosing of CAR T‐cells 24 days post tumour inoculation.
T‐cells	Lentivirus expressing anti‐human PSMA	N/A	LNCaP, PC‐3	NOD/SCID, SCID	Specific CTL activity against PC‐3 cells manipulated to express PSMA (PC‐3‐PIP), but not in wild type PC‐3.	(Zuccolotto et al., 2014)
					In vivo; Detectable metastatic disease 21 post T‐cell injection in control group, with no visible signs of PC‐3 PIP cells after this point in the treated group.
T‐cells	Retrovirus expressing scFv of anti‐human PSMA	N/A	PC‐3 (PSMA+/−)	BALB/c nude mice	Specific cell lysis of PC‐3PSMA, which was not observed in WT PC‐3 cells.	(Ma et al., 2014)
					In vivo; In a PC‐3 PSMA xenograft mouse model, 75% of mice treated with T‐cells expressing anti‐PSMA CARs were tumour free 27 days post inoculation.
T‐cells	Bispecific antibody targeting PSMA and PSCA	N/A	PC‐3, LNCap	N/A	*Antibody constructs were able to elicit a T‐cell response and resulted in subsequent lysis of PC‐3PSMA, PC‐3PSCA, PC‐3PSMA/PSCA when co‐cultured with the constructs and inactivated t‐cells.	(Arndt et al., 2014)
Genes encoding ttk and sFLT3L	Adenovirus	N/A	LNCaP, PC‐3, DU‐145, CWR22rv	N/A	Expression of “therapeutic” and viral replication genes and could only take place in cells expression PSMA and PSA.	(Kim et al., 2013)
					LNCaP/CWR22rv had higher levels of viral accumulation and cell death when compared with DU‐145/PC‐3.
T‐cells	Bispecific diabody targeting PSMA and CD3	N/A	LNCaP, C4–2, DU‐145	N/A	Only in the presence of PMSA+ cells was there T‐cell activation and expression of CD25 and CD69 in T‐cells.	(Fortmüller et al., 2011)
					C4–2 mice treated with T‐cells + PSMA x CD3 diabody showed a lower tumour burden (<200mm3) compared with the control groups (>1500 mm3).
T‐cells	Electroporation of plasmid DNA vaccines encoding PSA and PSMA	N/A	N/A	BALB/C nude	Vaccines were delivered IM followed by electroporation.	(Ferraro et al., 2011)
					Plasmid vaccines induced a strong IFNγ and IL‐2 response by CD4+ and CD8+ T‐cells, with CD4+ also exhibiting a strong TNFα response.
T‐cells	DNA fusion vaccines w/PSMA/FrC of tetanus vaccine	N/A	TRAMP‐C1	HHD mice	Fusion vaccines stimulated over 75% specific lysis of cells endogenously expressing PSMA.	(Vittes et al., 2011)
					In vivo; Specific lysis of PSMA + cells was stimulated by the fusion vaccines in three quarters of mice tested.
Radiotherapy
225Ac	Liposomes	J591 antibody/ A10 Aptamer	LnCaP, Mat‐Lu, HUVEC, BT474, and MDA‐MB‐231	N/A	J591‐liposomes were more effective at binding to all cell lines compared with A10‐liposomes.	(Bandekar et al., 2014)
					J591‐liposomes loaded with 225Ac was also the most cytotoxic liposomal construct tested.
177Lu	N/A	DKFZ‐617 PSMA inhibitor	N/A	Human clinical study	10 patients with chemotherapy resistant and/or hormone refractory prostate cancer were treated with 177Lu‐DKFZ‐617 PSMA.	(Ahmadzadehfar et al., 2015)
					70% of patients had a PSA decline, with 5 patients showing PSA decline >50%.
					177Lu‐DKFZ‐617 PSMA also exhibited a low early side effect profile.
177Lu/68Ga	–	DOTAGA‐ffk(Sub‐KuE)	LnCap cells	CD‐1 nu/nu mice	This theranostic allows for combined diagnosis and therapy. The uptake of imaging and therapeutic agent was greatest in the tumour and kidneys (both PSMA expressing) 1 hour post injection in an LNCaP xenograft model (4.95% ID/g for 68Ga‐PSMA and 7.96% ID/g for 177Lu‐PSMA).	(Weineisen et al., 2015)
					Treatment of two patients with metastatic disease was shown to be effective with no side effects detected.
177Lu	N/A	DKFZ‐11 PSMA inhibitor	N/A	Human clinical study	Following treatment, patient showed a radiological response and in addition, PSA levels fell from 38 to 4.6 ng/ml.	(Kratochwil et al., 2015)
124I/131I	N/A	MIP‐1095 PSMA inhibitor	N/A	Human clincal study	16 patients underwent 124I/131I‐MIP‐1095 treatment. PSA levels fell by >50% in 60.7% of patients treated.	(Zechmann et al., 2014)
					84.6% of patients that reported bone pain indicated either a complete or moderate pain relief. There were no significant haematological or renal toxicities reported.
177Lu	N/A	PSMA antibody J591	N/A	Human clinical study	47 hormone refractory patients were treated in a Phase II clinical trial with 177Lu‐J591. 10.6% AND 36.2% of patients received a ≥ 50% ≥ 30% decrease in PSMA levels respectively.	(Tagawa et al., 2013), (Simone and Hahn, 2013)
					There was a significantly longer survival (P = 0.03) in those patients receiving the maximum tolerated dose (21.8 vs 11.9 months). Patients with low PSMA expression were less likely to respond.
177Lu	N/A	DOTA‐3/F11 (PSMA antibody)	C4–2 and DU145 cells	SCID mouse (C4–2 xenograft model)	Imaging agents showed T1/2 of >7 days in vivo. A single dose of 177Lu‐DOTA‐3/F11 significantly slowed growth in C4–2 xenograft tumour model (<250 mm3 tumour volume in treated group at day 20 versus >1500 mm3 in the control group).	(Behe et al., 2011)
125I	N/A	DCIBzL PSMA inhibitor	PC3 PIP (PSMA+) and PC3 flu (PSMA‐) and LnCap cells	Athymic nude mice	Uptake of 125I‐DCIBzL was significantly higher in PSMA PC‐3 PIP (PSMA+) compared with PC‐3 flu (PSMA‐) cells.	(Kiess et al., 2015)
					PC‐3 PIP tumours showed significantly (p = 0.002) delayed growth when treated with 125I‐DCIBzL compared with PC‐3 flu tumours.
131I	Adenovirus	hNIS	LnCap cells	BALB/c nude mice	A significantly higher uptake (p = 0.0075) of iodide was observed by xenograft tumours transfected with Ad.PSMApro‐hNIS vs. Ad.CMV (2846.03 ± 188.29 c.p.p.mg −1 vs. 9.19 ± 1.00 c.p.p.mg −1 respectively).	(Gao et al., 2014)
90Y/177Lu	N/A	PSMA antibody J591	N/A	Human clinical study	The single maximum tolerated dose (MTD) of 90Y‐J591 and 177Lu‐J5916. was 17.5 mCi/m2 and 70 mCi/m2.	(Vallabhajosula et al., 2005)
					Several administrations over 4–6 months was well tolerated in patients where thrombocytopenia was manageable.
177Lu	N/A	PSMA antibody J591	N/A	Human clinical study	35 patients received 177Lu‐J591. 177Lu‐J591 was able to target all skeletal and soft tissue metastasis (determined by MRI).	(Bander et al., 2005)
					*There was no immunogenic response to J591. Four patients had a > 50% decline in PSMA with 16 showing PSA stabilization for median 60 days.
90Y/131I	N/A	PSMA antibody J591	LnCap cells	BALB/c nude mice	15–90% reduction in LNCaP xenograft tumour volume following a single dose of either 90Y‐J591 or 131I‐J591 was observed.	(Vallabhajosula et al., 2005)
					*An increase in survival time two to threefold compared with untreated controls was also observed.

Table 4.

List of PSMA based diagnostics developed for prostate cancer and neurological diseases

Imaging agent	Targeting ligand	In vitro model	In vivo model	Concluding remarks	References
Prostate cancer
123I	MIP‐1072 and MIP‐1095 (PSMA inhibitors)	N/A	Patient study	Both imaging agents were able to effectively detect lesions in bone, soft tissue and the prostate gland within 1 hour.	(Barrett et al., 2013)
Iron oxide nanoparticles w/MRI	PSMA antibody J591	LnCap, DU145	N/A	Significantly higher uptake of targeted iron oxide nanoparticles in LNCap cells (PSMA+) (>1.4 mM of Fe) compared with DU145 (PSMA‐) cells (<0.3 mM of Fe).	(Abdolahi et al., 2013)
18F	DCFBC (PSMA inhibitor)	N/A	Patient study	18F‐ DCFBC identified 32 suspected metastatic lesions, with 21 identified by conventional imaging. All of the 11 additional sites identified were mostly in the bone.	(Cho et al., 2012)
18F	PSMA inhibitor 2‐PMPA	N/A	BALB/c nude mice (LnCap xenograft model)	Introduction of 18F did not affect binding affinity of 2‐PMPA.	(Graham et al., 2012)
				Specific binding of imaging agent to LnCaP xenograft tumours was observed
68Ga	Urea‐based PSMA inhibitor	LnCap and PC3 cells	BALB/c nude mice	The imaging agent was formed by the dimerization of the urea‐based inhibitor coupled with 68Ga. Dimer had better binding properties than the monomer (IC50 of 3.9 ± 1.8 nM for dimer vs. 12.1 ± 2.1 nM for monomer).	(Schäfer et al., 2012)
				There was a higher accumulation of agent in LnCaps (PSMA+) compared with PC3 (PSMA‐) tumours; (8.22 ± 1.78% injected dose per gram 1 h post injection in LNCaP tumours versus 0.93 ± 0.53% in PC‐3 tumours).
IR800 and Cy5.5	Anionic PSMA ligand	PC3‐PIP(PSMA+) and PC3‐flu (PSMA‐)	Athymic nude mice	High binding affinity in xenograft tumours derived from PC‐3 cells transfected to express PSMA (PC3 PIP)	(Wang et al., 2014)
111In	Anti‐PSMA nanobody JVZ‐007	LNCaP, PC346C, VCaP, and MDA‐PCa‐2b, B16‐PSMA	BALB/c nude mice	Effective labelling of LNCaP cells and PSMA expressing PDX models with no binding to PSMA‐negative PDX models and kidneys.	(Chatalic et al., 2015)
111In	PSMA antibody J591	N/A	Patient study	In a 20 patient study, 111In‐J591 identified 74% of skeletal lesions, 53% of nodal lesions and 64% of other soft tissue lesions.	(Pandit‐Taskar et al., 2015b)
111In/IRDye800CW	PSMA antibody D2B	LnCap, PC3, LS174T‐PSMA	BALB/c nude mice	11In‐DTPA‐D2B‐IRDye800CW accumulated to a significantly greater level in LNCaP (PSMA+) tumours (45.8 ± 8.0% injected dose per gram at 7 days post injection) compared with PC‐3 tumours (6.6 ± 1.3% injected dose per gram at 7 days post injection).	(Lütje et al., 2014)
125I/111In	PSMA antibody capromab pendetide	LnCap and PC3 cells	BALB/c nude mice	Capromab binds to the intracellular domain of PSMA. There was less accumulation of 125I in healthy organs compared with 111In (1.6 fold lower in the colon and 2.3 fold in the bone).	(Tolmachev et al., 2014)
18F	Phosphoramidate PSMA inhibitor	LNCaP, CWR22Rv1, PC‐3 cells	Athymic NCr‐nu/nu mice	Increased uptake of imaging agent in PSMA + cells (2.2% in CWR22Rv1 and 12.1% in LNCaP cells) in comparison with PSMA – cells (0.08% in PC‐3)	(Ganguly et al., 2015)
18F	DCFBC PSMA inhibitor	N/A	Patient study	Significantly lower uptake of 18F‐DCFBC by benign prostatic hypertrophy than primary tumours (p = 0.004). Able to detect higher grade (Gleason score 8 or 9) and larger tumours more effectively than MRI	(Rowe et al., 2015)
18F	2‐PMPA PSMA inhibitor	22RV1, PC3, HEK293, V79 cells	BALB/c nude mice	High accumulation of imaging agent in LNCaP xenograft tumours. Safety studies indicated no off‐target activity, effect on vital organ functions or dose‐dependent adverse effects.	(Lesche et al., 2014)
64Cu	Binding domains of PSMA mAb 3/F11	LnCap, DU145, HEK293 cells	SCID mice	Two recombinant Anti‐PSMA antibodies were conjugated to human IgG3 CH3 or Fc domain. There was accumulation of conjugated IgG3 CH3 (10.7 ± 2.0% injected dose per gram 48 h post injection) and IgG3 Fc (10.6 ± 1.4% injected dose per gram 48 h post injection) antibody fragments in PSMA + C4–2 tumours on xenograft model.	(Wiehr et al., 2014)
				However, uptake was less with the full length antibody (20.9 ± 2.2% injected dose per gram 48 h post injection for full length mAb).
64Cu	Urea‐based PSMA inhibitor	PC3 cells	SCID mice	64Cu‐CB‐TE2A displayed the most favourable kinetics and a high image contrast (29.50 ± 8.1% injected dose per gram 2 h post injection in PC‐3 PIP tumours versus 0.66 ± 0.26% in PC‐3 flu tumours).	(Banerjee et al., 2014)
64Cu	DUPA	PC3 and LnCap cells	SCID mice	Imaging agent was targeted to two biomarkers for prostate cancer, PSMA and gastrin‐releasing peptide (GRPr). Maximum intensity was observed 18 h post injection in PSMA positive (LNCaP) and GRPr positive (PC‐3) xenograft tumours.	(Bandari et al., 2014)
68Ga	Urea‐based PSMA inhibitor	N/A	Patient study	Patient with biochemical relapse underwent lymph node dissection based on the results from imaging agent. Patient exhibited reduction in PSA following surgery.	(Schiavina et al., 2015)
68Ga	HBED‐CC PSMA ligand	N/A	Patient study	68Ga‐PSMA‐11 was able to detect lymph node metastasis in two‐thirds of patients who would have been missed by conventional CT diagnosis	(Giesel et al., 2015)
68Ga/ 177Lu‐	DOTAGA‐ffk(Sub‐KuE)	LnCap cells	CD‐1 nu/nu mice	In combination with radiotherapeutic 117Lu, this theranostic exhibited a more favourable PK profile than DOTA conjugated, was rapidly internalized in PSMA + cells with minimal non‐specific uptake.	(Weineisen et al., 2014)
68Ga	Urea‐based PSMA inhibitor	N/A	Patient study	There was a significantly higher detection rate of positive lesions (59 vs 29) by 68Ga‐PSMA‐HBED in patients with biochemical recurrence compared with 18F‐Fluoromethylcholine (FMC)	(Morigi et al., 2015)
68Ga	PSMA ligand HBED‐CC	N/A	Patient study	In 70 patients with recurrent prostate cancer following primary treatment, 52/70 tested positive with 68Ga; 30 patients had lesions in the pelvis, 8 had distant lesions and 14 had local plus systemic lesions.	(Ceci et al., 2015)
68Ga	PSMA ligand HBED‐CC	N/A	Patient study	222/248 (89.5%) of patients with biochemical recurrence showed pathologic findings with 68Ga‐HBED‐CC.	(Eiber et al., 2015)
				Detection rated increased with increasing PSMA levels. 68Ga‐HBED‐CC exclusively provided pathological findings in 81 (32.7%) patients.
68Ga	PSMA ligand HBED‐CC	N/A	Patient study	82.8% of patients had at least one prostate cancer lesion detected. Detection of tumours was positively associated with PSA level and androgen deprivation therapy.	(Afshar‐Oromieh et al., 2015)
68Ga	PSMA ligand HBED‐CC	LnCap, PC3, AR42J	BALB/c nude mice	Imaging agent was targeted to two biomarkers for prostate cancer, PSMA and gastrin‐releasing peptide (GRPr).	(Eder et al., 2014)
				Reduction in intensity following blocking of the respective receptor (from 5.44 ± 1.54% to 1.06 ± 0.24% for LNCaP xenograft tumours).
68Ga & 18F	PSMA ligands HBED‐CC & DCFPyL	N/A	Patient study	18F‐DCFPyL was able to detect all lesions identified by 68Ga‐HBED‐CC. In addition, 18F‐DCFPyL was able to detect three additional lesions.	(Dietlein et al., 2015)
				In PSMA positive lesions, the maximized standardized uptake value (SUVmax) was significantly higher with 18F‐DCFPyL than 68Ga‐HBED‐CC (0.028)
68Ga & 18F	Urea‐based PSMA inhibitor	N/A	Patient study	68Ga‐HBED‐CC was able to detect all of the lesions detected by 18F‐fluoromethylcholine. In addition, 68Ga‐HBED‐CC had a statistically higher detection rate (p = 0.04).	(Afshar‐Oromieh et al., 2014)
				The maximized standardized uptake value (SUVmax) was greater with 68Ga‐HBED‐CC than18F‐fluoromethylcholine.
86Y	ZJ43 PSMA inhibitor	PC3 cells	Papio Anubis (Baboon)	There was a greater uptake of targeted agents in PSMA positive (PC‐3 PIP tumours) 5 h post injection (32.17 ± 7.99% injected dose per gram) compared with PC‐3 flu tumours.	(Banerjee et al., 2015)
89Zr	PSMA antibody J591	N/A	Patient study	Compared with conventional imaging methods, 89Zr‐J591 detected 99 additional osseous sites. Accuracy of 89Zr‐J591 in prostate cancer detection was 95% for osseous sites and 60% for soft tissue lesions	(Pandit‐Taskar et al., 2015a)
99mTc	Diabody derived from J591	DU145 and LnCap cells	SCID mice	Significantly higher (p < 0.001) of accumulation of tracer in PSMA positive tumours (DU145‐PSMA) (12.1 ± 1.7% injected dose per gram 8 h post injection) versus PSMA negative tumours (DU145) (6.3 ± 0.5% injected dose per gram 8 h post injection).	(Kampmeier et al., 2014)
99mTc	Urea‐based PSMA inhibitor	PC3 cells	NOD/SCID mice	[99mTc]L8 had the most favourable characteristics, with the lowest retention in normal tissues and higher accumulation in PSMA positive tumours compared with PSMA negative tumours (26.29 ± 7.45% injected dose per gram 2 h post injection in PC‐3 PIP tumours versus 0.19 ± 0.08% in PC‐3 flu tumours).	(Ray Banerjee et al., 2013)
99mTc	MIP‐1404 and MIP‐1405 (PSMA inhibitors)	LnCap and PC3 cells	NCr‐nu/nu mice	Both 99mTc‐mip‐1404 and 99mTc‐mip‐1405 were able to rapidly detect soft tissue prostate cancer lesions and more skeletal metastatic lesions compared with standard‐of‐care bone scanning	(Hillier et al., 2013)
Indocyanine green (ICG)/ 125I	Minibody (MB) against PSMA.	PC3 cells	Athymic nude mice	The minibody was conjugated to an activatable fluorophore, ICG. This would only fluoresce following PSMA binding.	(Watanabe et al., 2014)
				There was greater accumulation of tracer in PSMA positive cells (> 20% injected dose per gram 24 h post injection) than in PSMA negative tumours (< 5% injected dose per gram 24 h post injection).
Biotin‐streptavidin system	Anti‐PSMA nanobody	LnCap, C4–2 and MKN45 cells	BALB/c nude mice	Arrival time, peak time, peak intensity and enhanced duration were all significantly different from blank nanobubbles in two PSMA + animal xenograft (LNCaP and C4–2)	(Fan et al., 2015)
IRDye800CW	Urea‐based PSMA inhibitor	U87‐MG and PC3 cels	NOD/SCID mice	Imaging agent was targeted to two biomarkers for prostate cancer, PSMA and integrin‐αvβ3.	(Shallal et al., 2014)
Iron oxide nanoparticles w/MRI	PSMA antibody J591	LNCaP, PC3, DU145, 22RV1, RWPE‐1 and BPH‐1 cells	NOD/SCID mice	Iron oxide nanoparticles were non‐toxic to prostate cancer cells. Conjugation of J591 to iron oxide nanoparticles did not adversely affect its ability to bind PSMA.	(Tse et al., 2015)
				In vivo MRI of tumours was improved using PSMA targeted nanoparticles, but not with non‐targeted nanoparticles.
Iron oxide nanoparticles w/MRI	PSMA‐targeting polypeptide (CQKHHNYLC, C1–C9 disulfide)	LnCap and Pc3 cells	BALB/c nude mice	The relative signal enhancement (RSE) in LNCaP cells treated with targeted iron oxide nanoparticles was significantly higher at 2, 6 and 12 h (19.9 ± 3.4, 34.3 ± 3.6, 30.9 ± 1.4 respectively) compared with PC‐3 cells treated with targeted iron oxide nanoparticles (8.7 ± 2.2, 11.1 ± 3.1, 10.4 ± 2.6 respectively).	(Zhu et al., 2015)
18F	PSMA ligand DCFPyL	PC3 and 22RV1 cells	NOD/SCID mice	Greater accumulation of tracer in PSMA positive tumours (PC‐3 PIP) (39.4 ± 5.4% injected dose per gram 1 hour post injection) versus PSMA negative tumours (PC‐3) (0.11 ± 0.02% injected dose per gram)	(Chen et al., 2011)
Iron platinum nanoparticles w/MRI	PSMA antibody J591	C4–2 and PC3 cells	N/A	Iron platinum nanoparticles were encapsulated by PEGylated phospholipids in order to create stealth nanoparticles that were targeted to PSMA positive cells with J591.	(Taylor et al., 2011)
89Zr	PSMA antibody 7E11	LnCap and PC3 cells	SCID mice	There was high levels of target‐specific uptake of 89Zr‐7E11 in LNCaP cells (20.35 ± 7.50% injected dose per gram 34 h post injection)	(Ruggiero et al., 2011)
CdSe/CdTe	Apatamer A9	LnCap and PC3 cells	N/A	Aptamer‐nanocrystal conjugates showed specific binding of live and fixed cells, and were also able to target PSMA positive cells that were dispersed in a collagen gel matrix.	(Chu et al., 2006)
Cy5	Apatamer A10	CPA, LnCap and PC3 cells	BALB/C nude mice	Aptamer‐targeted nanoparticles were found to only target PSMA positive LNCaP and canine prostate adenocarcinoma cells, but not PC‐3 cells.	(Tong et al., 2010)
Fluorescent quantum dot (QD)	Apatamer A10	LnCap and PC3 cells	N/A	There was specific binding of Aptamer‐targeted QDs to LNCaP cells with no uptake into PC‐3 cells.	(Bagalkot et al., 2007)
				When the particles were loaded with Docetaxel, there was a significant reduction of cell viability in LNCaP cells compared with PC‐3 cells (approximately 50% cell viability versus 75% respectively).
Neurological disorders
68Ga	Urea‐based PSMA inhibitor	N/A	Human brain tissue	Increased accumulation of 68Ga‐HBED‐CC‐PSMA was observed in tumour areas of the brain, with no uptake in unaffected brain regions IHC confirmed increased PSMA expression in tumour vasculature.	(Schwenck et al., 2015)
125I	PSMA ligand DCIT	N/A	Rodent brain tissue	Autoradiography of rodent brain sections was used to detect GCPII.	(Guilarte et al., 2005)
				Imaging of brain with 125I‐DCIT was reduced in knockout mice in a gene‐dose‐dependent manner, thus indicating its binding specificity. The goal was to develop an imaging agent that quantitatively determines GCPII levels in vivo
125I	PSMA ligand DCIT	N/A	Human brain tissue	Autoradiography of human brain sections was used to detect GCPII.	(Guilarte et al., 2008)
				Patients with schizophrenia had significantly lower levels of GCPII in the prefrontal cortex and entorhinal cortex compared with age matched controls. Highlights benefit of GCPII imaging agent in the diagnosis of neurological disease.
11C	Urea‐based PSMA inhibitor	N/A	CD‐1 mice	Brain uptake of 11C‐MCG was low, highlighting that the agent may be inappropriate for monitoring brain GCP II activity. This was thought to be due to the hydrophilic nature of the agent (logP = −0.235)	(Pomper et al., 2002)

The closest homologue to PSMA/GCPII is glutamate carboxypeptidase III (GCPIII) (Hlouchová et al., 2007). While investigating closely related peptidases, it was shown that GCPIII has a similar NAAG hydrolysing activity to PSMA/GCPII (Pangalos et al., 1999). Despite these similar functions, GCPIII has several different motifs in the active site which account for the differing substrate specificities and inhibitor susceptibilities (Hlouchová et al., 2007). In addition, GCPIII has recently been shown to have its own distinct function in the cleavage of β‐citrylglutamate (Navrátil et al., 2016).

Targeting PSMA/GCPII as a potential therapeutic approach

Chemotherapy

Traditional chemotherapies work against rapidly dividing cells (such as cancer cells). However, they also attack healthy cells of the body that naturally rapidly divide, such as those of the digestive tract or hair follicles (Sutradhar and Amin, 2014). Targeting chemotherapies to diseased tissues overexpressing a certain receptor (such as PSMA/GCPII) allows for minimal toxicity to normal tissue, while simultaneously allowing the drug to accumulate at a much higher concentration in the tumour tissue compared with the free drug (Sun et al., 2014b; Pérez‐Herrero and Fernández‐Medarde, 2015). There are numerous studies that exploit the PSMA/GCPII for targeted delivery of chemotherapeutic drugs in PCa, either via aptamers, antibodies or small molecules (Table 2).

A PSMA/GCPII‐targeted generation 5 PAMAM dendrimer has been used to deliver methotrexate (Huang et al., 2014). A urea‐based small molecule (glutamate urea) was used as a targeting ligand for PSMA/GCPII. The urea‐based conjugate showed binding to LNCaP cells in a dose‐dependent manner, in contrast to the control which did not yield any significant binding. The targeted conjugate also induced potent cytotoxicity in LNCaP cells, with no observable difference in PSMA/GCPII negative PC3 cells between control and targeted formulations.

Micelles have recently been developed with a H40 polymer core and an amphiphilic copolymer containing the hydrophobic polylactic acid (PLA) facing inside and the hydrophilic polyethylene glycol (PEG) facing outside. The aptamer A10 was covalently linked to the PEG on the outer surface and allows for effective PSMA/GCPII targeting (Xu et al., 2013). Aptamers are nucleic acid ligands (either DNA or RNA) that fold into a specific conformation that confers specificity for the intended target site (Sun et al., 2014a). Hydrophobic doxorubicin ‘sits’ in the PLA component of the micelle. Higher levels of doxorubicin were observed in the nucleus of CWR22Rv1 cells following treatment with the targeted formulations, while doxorubicin from the untargeted controls was mainly confined to the cytoplasm. In a CWR22Rv1 xenograft mouse model of PCa, the highest doxorubicin fluorescence intensity was observed in the tumours from the groups treated with targeted micelles (Xu et al., 2013). Similarly, the small molecule PSMA/GCPII inhibitor was conjugated to PCL‐PEG copolymers to form self‐assembling micelles with the chemotherapeutic agent docetaxel in the core of the formulation. In this case, the targeted therapy had a lower IC₅₀ and a five‐fold higher increase in uptake compared with the non‐targeted therapy (Jin et al., 2014).

Aside from aptamers and small molecules, antibodies have also been used to target therapies to PSMA/GCPII positive cells (Wang et al., 2011; Yallapu et al., 2014; DiPippo et al., 2015; Hariri et al., 2015). Monomethylauristatin E was conjugated to a fully human anti‐PSMA/GCPII antibody to form an antibody‐drug conjugate (ADC) (Wang et al., 2011). ADCs have shown significant clinical activity in patients that have been previously treated with chemotherapy (Krop et al., 2010). Using this PSMA/GCPII ADC, the induction of cell death was shown to be 1000‐fold higher in PSMA/GCPII positive cells. Also, in a docetaxel‐resistant mouse model of PCa, the mean tumour volume was significantly lower in the PSMA/GCPII ADC group then when compared with docetaxel alone, with survival times significantly increased in the former group (Wang et al., 2011).

Gene therapy

In recent years, advances in the field of next generation sequencing has brought the fields of gene and RNA interference (RNAi) therapy closer to the clinic. While gene therapy involves the replacement of a defective or deleted gene in a cell or tissue, RNAi therapy involves the silencing of a target gene (i.e. disease causing gene) in a sequence‐specific manner, using double stranded RNA (dsRNA) (Grimm and Kay, 2007; Guo et al., 2013a). Provided that the dsRNA is appropriately designed, RNAi has the potential to silence (theoretically) any gene within the genome (Jackson and Linsley, 2010). In recent years, the PSMA/GCPII has been used to successfully target siRNA, shRNA, microRNA or therapeutic genes to PSMA/GCPII positive cells (Table 2).

A recent study incorporated a PSA responsive element and a PSMA/GCPII targeted component into a liposome for the treatment of PCa. DSPE‐PEG₅₀₀₀‐folate was used to target the liposome to the PSMA/GCPII receptor in a 22rv1 xenograft model of PCa. This multifunctional liposome showed preferential binding to cells that secreted PSA and expressed PSMA/GCPII. There was a significant reduction in tumour burden and target mRNA (Plk1) in a xenograft tumour model (Xiang et al., 2013). Plk1 is a key regulator of cell mitosis, and its inhibition has been shown to induce apoptosis in rapidly dividing tumour cells (Strebhardt and Ullrich, 2006; Dassie et al., 2009).

In two‐thirds of metastatic breast and PCa patients, there is some spread of the disease to the bone (www.cancer.org). Bone metastases can be painful and debilitating for the patient, and generally result in significant morbidity, with most therapeutic interventions offering only modest benefits (Takeshita et al., 2010). A second generation RNA aptamer (A10–3.2) targeting PSMA/GCPII has been used to deliver microRNA (miR‐15a and miR‐16‐1) to bone metastases using atellocollagen as a delivery vector (Hao et al., 2016). The resulting construct showed significantly higher levels of uptake in PSMA positive LNCaP cells in vitro compared with PSMA/GCPII negative PC3 cells. The construct containing the PSMA/GCPII aptamer also significantly increased the survival times in a mouse model of human PCa bone metastasis.

One approach currently under investigation for PSMA/GCPII‐mediated gene therapy is suicide gene therapy. This is where a gene encoding an enzyme that converts a non‐toxic prodrug into a highly toxic compound is delivered into tumour cells (Dachs et al., 2005). In a recent study, the HSV‐thymidine kinase (TK) and connexin43 gene was delivered to PSMA/GCPII positive LNCaP cells using a PAMAM dendrimer delivery vector with folate as a targeting ligand for PSMA/GCPII (Chen et al., 2013). The expression of TK and connexin43 was driven by the PSMA/GCPII promoter, thus they could only be expressed in PSMA/GCPII positive cells, with no detectable gene expression in PSMA/GCPII negative PC3 cells. This example of the double‐targeting of PSMA/GCPII highlights the therapeutic advantages of targeting this receptor.

Immunotherapy

The role of the immune system to seek out and destroy cancer tumour cells is well documented (Apetoh et al., 2015). The ability of the immune system to control the growth of tumours can be demonstrated in patients with an immunodeficiency such as HIV, in whom the cancer burden is far greater than in the uninfected population (Robbins et al., 2015). Several decades of research into immunotherapy have recently started to show returns with several immunotherapies in clinical trials and approved for use in recent years (Rowdo et al., 2015). Several highly effective immunotherapies at both the preclinical and clinical stage targeting PSMA/GCPII are summarized in Table 3.

Dendritic cells (DCs) are classed as effective antigen presenting cells, initiating and directing a T‐cell response, thus making them ideal candidates as the powerhouses for a cancer vaccine (Palucka and Banchereau, 2012). DCs function by eliciting the activation of T‐cells. This activation occurs when DCs present antigens to naïve CD8+ cells (via MHCI) and CD4+ cells (via MHCII). This results in the maturation of T‐cells and can induce a cytotoxic T‐lymphocyte (CTL) response against the cancer cell carrying the offending antigen. Several studies have highlighted the potential of DCs that have been modified ex vivo to elicit a CTL response in patients (Fuessel et al., 2006; Waeckerle‐Men et al., 2006; Fishman, 2009).

In a recent study, DCs were primed with recombinant survivin and recombinant PSMA (Xi et al., 2015). These DCs were then used in a clinical study of 21 patients with castration‐resistant PCa, with docetaxel and prednisone used as a control. Objective response rate by Response Evaluation in Solid Tumours was 8/11 in patients who received DC therapy, with an increased survival time and 2 patients showing partial remission. Another study using DC based therapy involved the transfection of DCs with adenovirus expressing human PSMA (huPSMA) and IFNγ (Williams et al., 2012). The results of this study showed a delayed growth of RM1‐PSMA/GCPII tumours and a significant increase in cytotoxic T‐cell activity in mice treated with a combination of Ad5‐ huPSMA and Ad5‐ IFNγ, with weak tumour inhibition observed in PSMA/GCPII negative tumours.

Apart from DCs, there has been a move to develop T‐cells bearing a chimeric antigen receptor (CAR) against PSMA/GCPII to elicit CTL activity against PSMA/GCPII positive tumour cells (Ma et al., 2014; Santoro et al., 2014; Zuccolotto et al., 2014). CAR technology has already shown promising efficacy in acute and chronic lymphoid leukaemia (Porter et al., 2011; Grupp et al., 2013). In the case of PSMA/GCPII, these T‐cells are generated by the transduction of naïve T‐cells using a virus that encodes some form of anti‐PSMA/GCPII antibody along with co‐stimulatory signalling domains such as CD‐28 and 4‐1BB which drives T‐cell maturation (Santoro et al., 2014). A NOD/SCID mouse model inoculated with modified PC3 cells to express PSMA/GCPII (PC3‐PIP) was treated using a T‐body (genetically engineered T‐cell carrying a chimeric receptor [Eshhar, 2008)] directed against PSMA/GCPII. These T‐cells had previously shown specific lysis against PSMA/GCPII positive cells only and, in vivo, they were able to inhibit the formation of metastasis, with no visible signs of any PC3‐PIP cells present, 21 days after the injection of the T‐cells (Zuccolotto et al., 2014).

Immunotherapy (while still in the early stages), has shown great therapeutic potential, and as it exploits the host's own immune system to fight the disease it offers long term protection against the relapse of cancer.

Radiotherapy

Radiation therapy (also called ionizing radiation) has been used to treat cancer for over 100 years since the discovery of X‐rays in 1895 (Thariat et al., 2013). Today, nearly half of all cancer patients routinely receive radiotherapy either alone, or else in combination with other treatments such as surgery or chemotherapy (Delaney et al., 2006). In PCa, approximately 25% of patients with localized disease will undergo some form of radiotherapy (Cooperberg et al., 2010). Radiation deposits high levels of energy in tissues and cells as it passes through them, damaging genetic material in the process. In doing so, the aim is that the cancer cells lose their ability to grow and divide (Jackson and Bartek, 2009; Baskar et al., 2012).

The most common radionuclides used in radiotherapy are ⁹⁰Y, ¹³¹I and ¹⁷⁷Lu. Due to the differing physical properties of the different radionuclides, each one may be suited to an optimal tumour size and type (Bouchelouche et al., 2010). For example, ¹⁷⁷Lu is more suited for the treatment of smaller tumours (1–3 mm) while ⁹⁰Y is suited to treat larger tumours (28–42 mm) (O'donoghue et al., 1995). In recent years, traditional radiotherapy has reduced unwanted side effects, allowing an increase in dosage. The addition of a targeting ligand, such as a monoclonal antibody or aptamer, to a radionuclide allows for the targeting of a cancer‐associated cell‐surface antigen. This is known as radioimmunotherapy (RIT) and, while traditionally RIT is more effective in more radiosensitive tumours, such as leukaemias and lymphomas, there is now a move towards using RIT to treat the more radioresistant solid tumours (Larson et al., 2015).

The PSMA/GCPII has been a desirable target for RIT for many years, with a summary of studies involving the delivery of RIT using PSMA/GCPII given in Table 3. The first anti‐PSMA/GCPII antibody that was used for RIT and imaging was capromab (7E11/CYT‐356) (Kahn et al., 1998a; Kahn et al., 1998b). The initial results from this therapy were disappointing, as capromab binds to the intracellular domain of PSMA/GCPII, and so will only bind to dying or permeable cells (Bouchelouche et al., 2010). In addition, in a phase II clinical trial with ⁹⁰Y‐capromab, there was significant toxicity (myelosuppression), and another study was prematurely terminated due to a lack of efficacy (no decline in PSA level or radiological response) (Deb et al., 1996; Kahn et al., 1999).

In response to this, second generation antibodies have been developed that target the extracellular domain of PSMA/GCPII. Such antibodies that have been widely researched in recent years for PCa RIT include 3F/11 (Behe et al., 2011) and J591 (Vallabhajosula et al., 2004; Bander et al., 2005; Vallabhajosula et al., 2005; Simone and Hahn, 2013; Tagawa et al., 2013; Bandekar et al., 2014), which unlike its predecessors can bind living cells (Smith‐Jones et al., 2000). In a phase II clinical trial, a single infusion of ¹⁷⁷Lu‐J591 was administered to two cohorts of patients (n = 47) with metastatic PCa. Of the 47 patients in the trial, 10.6% had a greater than 50% decrease in their PSA levels, 36.2% had greater than 30% reduction and 59.6% showed at least some level of PSA decline. The therapy was well tolerated with all patients showing reversible haematological toxicity and there was also a significant increase in survival time (21.8 vs. 11.9 months) (Simone and Hahn, 2013; Tagawa et al., 2013). The outcome of this trial indicates the benefits that PSMA/GCPII‐mediated RIT can offer patients with metastatic disease.

Over the years, different PSMA/GCPII inhibitors have also been used to target radionuclides to PCa cells. As already mentioned, there are many different classes of inhibitors that have been developed to target the PSMA/GCPII. In terms of radionuclide delivery, these mainly constitute the urea‐based inhibitors (Zechmann et al., 2014; Ahmadzadehfar et al., 2015; Kiess et al., 2015; Kratochwil et al., 2015; Weineisen et al., 2015). In a recent clinical trial, 10 patients with hormone‐refractory and/or chemotherapy‐resistant PCa with confirmed metastasis were treated with ¹⁷⁷Lu, which was conjugated to the urea‐based DKFZ‐617 PSMA/GCPII inhibitor. Eight weeks following treatment, 70% of the patients experienced some level of PSA decline, with five showing a greater than 50% decline. Sixty per cent of patients showed no haematotoxicity following treatment and none exhibited nephrotoxicity. This study shows the therapeutic potential of a single dose of ¹⁷⁷Lu‐DKFZ‐617 in metastatic PCa patients that have no other therapeutic options (Ahmadzadehfar et al., 2015).

Another recent study involved the transfection of PCa cells with an adenovirus containing the gene for the human sodium iodide symporter (hNIS) (Gao et al., 2014). The expression of this hNIS gene was dependent on the expression of PSMA/GCPII in the target cell. Under normal conditions, hNIS is utilized by the thyroid gland to allow for far higher levels of iodide to accumulate in the gland compared with plasma (De la Vieja et al., 2000). This usually allows for effective therapy of thyroid cancer with radioiodine (even in advanced cases) (Mazzaferri and Kloos, 1996). In an LNCaP Xenograft mouse model, there were significantly greater levels of iodide in animals treated with ad.PSMApro‐hNIS compared with the control, and as such, these tumours had the lowest volume. This shows the potential benefits offerred by a combination of gene and radiotherapy in the treatment of advanced PCa.

PSMA/GCPII as a diagnostic tool

PSMA/GCPII‐targeting ligands including PSMA/GCPII antibodies, aptamers and low molecular weight molecules and inhibitors have been used to target effective imaging agents including ⁶⁴Cu, ¹⁸F, ⁶⁸Ga, ¹¹¹In, ^99mTc and others such as iron oxide nanoparticles and IRDye 800. These studies are summarized in Table 4. A significant reduction in mortality from PCa in recent years has been attributed to earlier detection, but this has come at a price, with a number of (in some cases) unnecessary treatments being administered, which can have a negative impact on the patient's quality of life (Resnick et al., 2013). In recent years, there has been a call to move away from the use of PSA as a diagnostic for PCa as it has been responsible for over‐diagnosis of patients who may only require active surveillance (Etzioni et al., 2002; Donovan, 2012). In this regard, the need for new biomarkers for effective diagnosis of lethal forms of PCa only is an unmet clinical need.

The J591 monoclonal antibody has been successfully used to image PCa metastasis in several studies. In a 20 patient study, ¹¹¹In‐J591 was used to successfully identify 74% of skeletal lesions, 53% of nodal lesions and 64% of other soft tissue lesions (Pandit‐Taskar et al., 2015b), while in a 50 patient study of metastatic PCa, ⁸⁹Zr‐J591 detected 99 additional osseous sites compared with conventional imaging modalities (Pandit‐Taskar et al., 2015a), and ^99mTc‐J591 showed a significantly higher accumulation in PSMA/GCPII positive tumours (DU‐145‐PSMA) (12.1 ± 1.7% injected dose per gram 8 h post injection) compared with that in PSMA/GCPII negative tumours (DU‐145) (6.3 ± 0.5% injected dose per gram 8 h post injection) (Kampmeier et al., 2014). Despite the advantages offered by mAb imaging, they are usually associated with a delay in target recognition and background clearance in a suitable timeframe for diagnostic imaging (Osborne et al., 2013).

Low molecular weight agents may provide more suitable pharmacokinetics for imaging (Bouchelouche et al., 2010). Recently, a PSMA/GCPII phosphoramidate‐based inhibitor was conjugated to ¹⁸F for imaging. Uptake of the imaging agent in PSMA/GCPII positive LNCaP cells (12.1%) was much higher than in PSMA/GCPII negative PC3 cells (0.08%) (Ganguly et al., 2015). Low molecular weight PSMA/GCPII inhibitors, MIP‐1404 and MIP‐1405 were linked to ^99mTc, which were capable of detecting more skeletal metastatic lesions when compared with conventional imaging (Hillier et al., 2013). A urea‐based PSMA/GCPII inhibitor conjugated to ⁶⁸Ga was recently used to image 248 patients with biochemical recurrence following radical prostatectomy (Eiber et al., 2015). 222/248 (89.5%) of patients with biochemical recurrence showed pathological findings with ⁶⁸Ga‐HBED‐CC. The rate of detection increased with higher PSA levels and ⁶⁸Ga‐HBED‐CC highlighted an additional 81 pathological findings, not identified by conventional imaging.

Another alternative to using mAb imaging is the use of aptamers. Recently, fluorescent quantum dots (QD) were conjugated to the aptamer A10. These imaging agents showed specific binding to PSMA/GCPII positive LNCaP cells, with no uptake in PC‐3 cells. Also, when docetaxel was loaded onto the QDs, there was a significant reduction of cell viability in LNCaP cells compared with PC‐3 cells (approximately 50% and 25% reductions respectively) (Bagalkot et al., 2007).

In addition to conventional radionuclide imaging, other modalities targeting PSMA/GCPII have been developed in recent years. PSMA/GCPII‐targeted iron oxide nanoparticles combined with magnetic resonance imaging (MRI) have shown encouraging results. J591‐iron oxide nanoparticles show significantly higher uptake of targeted iron oxide nanoparticles in PSMA/GCPII positive (LNCap) cells compared with PSMA/GCPII negative (DU145) cells, and in vivo MRI of tumours was improved using PSMA/GCPII targeted nanoparticles, but not with non‐targeted nanoparticles (Abdolahi et al., 2013; Tse et al., 2015).

While showing enormous potential and efficacy in the case of PCa, evidence of exploiting PSMA/GCPII as a diagnostic tool for neurological disorders has been scarce with relatively few studies published in the area (Table 4). Autoradiography of human brain sections using ¹²⁵I‐DCIT showed that patients with schizophrenia had a significantly lower level of GCPII expression in the prefrontal cortex and entorhinal cortex, compared with age matched controls (Guilarte et al., 2008). This study highlights the benefit of GCPII imaging agents in the diagnosis of neurological disease. However, there are major shortcomings for real time in vivo imaging of GCPII expression in the brain. A urea‐based PSMA/GCPII inhibitor conjugated to ¹¹C showed low brain uptake in primates. This was thought to be due to the hydrophilic nature of the agent decreasing its ability to cross the BBB (Pomper et al., 2002). A study was carried out in 2015 where GCPII was successfully imaged in the neovasculature of glioblastoma in a human patient using ⁶⁸Ga and a urea‐based inhibitor of PSMA/GCPII (Schwenck et al., 2015). Currently, work is ongoing to modify inhibitors to make them more applicable for use in neurological disease (Wang et al., 2010; Feng et al., 2011; Majer et al., 2016).

Conclusions

PSMA/GCPII is a transmembrane protein that is highly expressed in PCa tissues and in the neovasculature of other tumour types, as well as in the brain. PSMA/GCPII plays a key role as a biomarker capable of diagnosing and staging PCa, and thus, many therapeutic and diagnostic agents have been developed that take advantage of recognizing this protein (Akhtar et al., 2011; Demirkol et al., 2015). Due to the multifunctional nature of PSMA/GCPII, its role in different areas of the body can be exploited for various biomedical applications including the treatment of PCa and neurodegenerative disease. Its function as a cell‐surface ligand‐binding receptor has been utilized to develop therapies that exploit the advantage of receptor‐mediated endocytosis of therapeutic agents. This protein also acts as a hydrolysing enzyme to enhance the absorption of substrates from tissue to the blood stream (e.g. folate absorption from small intestine to blood stream) (Chang et al., 2004). PSMA/GCPII has a similar hydrolysing activity in the brain, where it hydrolyzes NAAG into glutamates, which in certain cases causes glutamate toxicity. Glutamate toxicity triggers a set of signalling molecules that lead to degenerative processes causing cell death. Such degenerative effects are characteristic of diseases like dementia, ALS, schizophrenia, multiple sclerosis (Rahn et al., 2012).

Historically, CNS therapies developed tend to be NAAG‐catabolism inhibitors that block the NAAG hydrolysing capacity of the PSMA/GCPII protein. Although such therapies have shown success in preclinical studies, their translation to clinical setting has been a challenge due to the poor passage of the PSMA/GCPII inhibitors through the BBB (Bařinka et al., 2012). Most of the inhibitor molecules developed are hydrophilic in nature and have high molecular weight, which acts as a limitation to delivery and, although certain approaches have been pursued to develop more hydrophobic drugs, very little success has been achieved in this area (Kozikowski et al., 2004).

From an evolutionary point of view, it is interesting to consider what conditions drove the development of a receptor with two very different functions depending on where it is expressed, an area that has received very little or no attention, so far. This multifunctionality of PSMA/GCPII provides opportunities for two different therapeutic approaches: (i) as a receptor ligand and (ii) as a hydrolysing enzyme. For targeted therapeutic purposes, it can be used as a facilitator to deliver drugs or therapies to PCa cells. A number of studies describe the use of PSMA/GCPII specific ligands, such as antibodies, aptamers, folate molecules attached or conjugated to therapeutic molecules (drugs, genes and radiolabelled‐diagnostic molecules) or modified on the surface of nanoparticles for receptor‐specific delivery. This strategy has been heavily exploited for PCa, as discussed above. While this has not been exploited for the treatment of neurological disorders and brain cancers, at present, two clinical trials are underway (results not yet disclosed) that use antibody‐drug conjugates targeting PSMA/GCPII to treat patients with the highly invasive glioblastoma (NCT02067156, NCT01856933). It is hoped that the results of these trials will highlight the usefulness of targeting the PSMA/GCPII in brain malignancies.

Another potential approach for treating neurological disorders would be to knock‐down (KD) the PSMA/GCPII transmembrane protein in the brain regions, instead of blocking it with inhibitors. Such KD of the PSMA/GCPII protein would have clinical effects, similar to those of the inhibitor drugs. The RNAi‐mediated approach to KD of certain receptors and transporters in the brain has previously been successfully applied to other neurological conditions such as depression, wherein the siRNA‐mediated KD of a 5‐HT transporter lead to increased 5‐HT signalling, nullifying the effect of corticosterone‐induced stress in animals and resulting in improved cognitive and behavioural functions (Thakker et al., 2005; Wu et al., 2016). This approach is emerging as a new model of therapy instead of drugs and inhibitors that cannot cross the BBB. The evolving strategies in developing non‐viral nanoparticles or therapeutic conjugate molecules for specific targeting and delivery of RNAi molecules, drugs or radio‐ligands will further enhance the understanding of novel therapies and diagnoses developed for PSMA/GCPII. However, for this approach to be feasible, the dual delivery barrier of transport across the BBB and cell‐specific uptake must be addressed. In order to tackle these delivery challenges, a recent study employed a bispecific antibody which simultaneously targeted the Tfr (allowing for crossing of the BBB) and BACE1. This approach resulted in reduced production of amyloid‐β plaques in the brain of monkeys, thereby offering a promising therapeutic in the treatment of Alzheimer's disease (Yu et al., 2014). If such a targeting mechanism was incorporated into a nanoparticle loaded with siRNA targeting PSMA/GCPII, this would offer a promising therapeutic strategy for various neurological disorders.

In conclusion, while the PSMA/GCPII offers many exciting opportunities for developing diagnostics and therapeutics in the PCa field, the potential for using the PSMA/GCPII for neurological disorders has been under‐explored, due mainly to the complex issues surrounding delivery of drugs across the BBB. However, if these were resolved, targeting the PSMA/GCPII would be a very promising target for neurological disorders.

Conflict of interest

The authors declare no conflicts of interest.

Evans, J. C. , Malhotra, M. , Cryan, J. F. , and O'Driscoll, C. M. (2016) The therapeutic and diagnostic potential of the prostate specific membrane antigen/glutamate carboxypeptidase II (PSMA/GCPII) in cancer and neurological disease. British Journal of Pharmacology, 173: 3041–3079. doi: 10.1111/bph.13576.