**Overview of Radiolabeled Somatostatin Analogs for Cancer Imaging and Therapy**

**Romain Eychenne 1,2,3, Christelle Bouvry 4,5, Mickael Bourgeois 2,3, Pascal Loyer <sup>6</sup> , Eric Benoist <sup>1</sup> and Nicolas Lepareur 4,6,\***


Received: 28 July 2020; Accepted: 1 September 2020; Published: 2 September 2020

**Abstract:** Identified in 1973, somatostatin (SST) is a cyclic hormone peptide with a short biological half-life. Somatostatin receptors (SSTRs) are widely expressed in the whole body, with five subtypes described. The interaction between SST and its receptors leads to the internalization of the ligand–receptor complex and triggers different cellular signaling pathways. Interestingly, the expression of SSTRs is significantly enhanced in many solid tumors, especially gastro-entero-pancreatic neuroendocrine tumors (GEP-NET). Thus, somatostatin analogs (SSAs) have been developed to improve the stability of the endogenous ligand and so extend its half-life. Radiolabeled analogs have been developed with several radioelements such as indium-111, technetium-99 m, and recently gallium-68, fluorine-18, and copper-64, to visualize the distribution of receptor overexpression in tumors. Internal metabolic radiotherapy is also used as a therapeutic strategy (e.g., using yttrium-90, lutetium-177, and actinium-225). With some radiopharmaceuticals now used in clinical practice, somatostatin analogs developed for imaging and therapy are an example of the concept of personalized medicine with a theranostic approach. Here, we review the development of these analogs, from the well-established and authorized ones to the most recently developed radiotracers, which have better pharmacokinetic properties and demonstrate increased efficacy and safety, as well as the search for new clinical indications.

**Keywords:** somatostatin analogs; radiolabeling; radiopharmaceuticals; radionuclide therapy; imaging

#### **1. Introduction**

Somatostatin (SST), also called somatotropin release inhibiting factor (SRIF), is a cyclic peptide hormone, first isolated in 1968 from an ovine hypothalamus, and actually identified in 1973 [1]. It was originally discovered as a growth hormone inhibitor, but is now known to be involved in the inhibition of numerous metabolic processes relating to neurotransmitters, endocrine secretions (e.g., growth hormone, insulin, glucagon, and gastrin) but also modulating exocrine secretions (e.g., gastric acid and pancreatic enzymes). In the body, its synthesis takes place in the form of an inactive precursor of 116 amino acids (AA), preprosomatostatin, which is then converted by the action of proteases into prosomatostatin (96 AA). Depending on where it is produced in the body, enzymes do not cleave the pro-peptide on the same amino acid motif, resulting in two distinct active forms, SRIF-28 and SRIF-14. Although SRIF-14 is predominant in the central nervous system and SRIF-28 in the digestive tract, the distribution of these two biologically active forms is similar.

In the early 1990s, concomitantly to studies on the binding properties and mechanisms of action of somatostatin, five receptor subtypes were discovered (SSTR1 to SSTR5) [2]. These subtypes belong to the family of receptors coupled to G-proteins, and their length varies from 364 to 418 AA. They all exhibit seven α helices with transmembrane domains and most of the differences between subtypes are found in the extracellular (N-terminal) and intracellular (C-terminal) ends. SSTR-1, -3, -4, and -5 have a single subtype, while two variants exist for SSTR2, called SSTR2A and SSTR2B. SSTR1 to 4 link SRIF-14 and -28 with a very high affinity (in the nanomolar order), whereas SSTR5 shows an affinity 5 to 10 times higher, but for SRIF-28 only. seven α helices with transmembrane domains and most of the differences between subtypes

Somatostatin receptors are widely distributed in healthy tissues, with distinct expression throughout the body (Figure 1). It is quite possible to find several subtypes in the same tissue. Each of the SSTRs is involved in the regulation of the various processes: (i) SSTR1 is involved in the antisecretory effects of growth hormone, prolactin (a peptide hormone involved in lactation, reproduction, growth, and immunity) and calcitonin (regulation of calcemia); (ii) SSTR2 also inhibits the secretion of growth hormone and adrenocorticotropin (hormone that stimulates the adrenal glands), glucagon, insulin, interferon-γ (protein produced by immune cells), and stomach acid; (iii) SSTR5 has the same inhibiting effect on growth hormone, adrenocorticotropin, insulin, and inhibits the secretion of amylase (digestive enzyme constituting saliva and pancreatic juice); (iv) SSTR3 reduces cell proliferation and causes cell apoptosis; (v) the functions of SSTR4 are not yet well defined [3]. γ (protein produced by immune cells)

**Figure 1.** Somatostatin receptors (SSTRs) biodistribution in the body (from The Human Protein Atlas https://www.proteinatlas.org/).

The effects of somatostatin are expressed through different signaling pathways [4,5]. After a cascade of reactions, this leads on the one hand to the inhibition of tumor growth (action on the secretion of hormones) and blocking proliferation via the activation of different tyrosine phosphatases (anti-proliferative and pro-apoptotic action), but also to the inhibition of the secretion of growth factors such as growth hormone or IGF-1 having a major role in the inhibition of tumor growth (anti-angiogenic) (Figure 2) [6,7].

**Figure 2.** Schematic representation of the signaling pathways induced by somatostatin receptors activation. Green arrows: activated pathways; red arrows: inhibited pathways. Adapted from [8].

– Over the past 20 years, our understanding of the phenomena due to the activation of SSTRs has increased thanks to numerous translational and clinical studies, leading to the development of new therapeutic options [3]. The use of SST analogs has demonstrated real effectiveness in the treatment of various pathologies: acromegaly (production of an excess of growth hormone), pancreatitis, complications linked to diabetes and obesity (e.g., retinopathy or nephropathy), action on inflammation and pain in some cases [5,9]. However, SSTRs and SST analogs are mainly known for their presence and role in the detection and treatment of some solid tumors. Tumor cells and peritumoral vessels express receptor subtypes whose density depends on the type of tumors (Table 1) [10–13]. For those overexpressing SSTRs, such as pituitary adenomas, gastroentero-pancreatic neuroendocrine tumors (GEP-NET), or other cancers (e.g., lymphomas, small cell lung cancers, etc.), targeting with SST analogs becomes possible [14]. Many therapeutic protocols based on these analogs (classic octreotide or with a longer release time (octreotide LAR), Lanreotide, Vapreotide, Pasireotide, etc.) have been the subject of phase II and III clinical trials. The majority of results were generally disappointing and did not provide clear evidence of a significant antitumor effect on solid tumors, probably due to the existence of other pathways of tumor progression [15,16].

**Table 1.** SSTRs expression in different tumor types.



**Table 1.** *Cont.*

Bold +, receptors with particularly high density and incidence. Subtypes preferentially expressed are listed in parentheses, only when compelling evidence is available (immunohistochemistry or autoradiography). Adapted from [11] and [17].

For example, regarding liver tumors, such as hepatocellular carcinoma (HCC), in vitro studies clearly demonstrated (i) the lack of SSTRs expression in healthy liver cells; (ii) overexpression in tumors and metastases of HCC, even though their density is less than in neuroendocrine tumors [21,22]. On the other hand, the results show a heterogeneous expression and strong inter-individual differences. In fact, according to studies, HCCs express high levels of SSTR2 [21,23,24] or SSTR5 [13,19], or even SSTR1 [22] or SSTR3 [25]. In general, around 40% of HCCs studied express somatostatin receptors. These differences could be due to the different methodologies used during the measurements, by studying different stages of the disease or even by heterogeneous behaviors of HCC. Further studies have also found a correlation between the density of SSTRs expression, disease aggressiveness [26], and the rate of tumor recurrence after treatment with octreotide LAR [27]. In a study by Nguyen-Khac et al. [23], 41.2% of extrahepatic metastases express SSTR2. Preclinical tests on HCC cell lines have shown an antiproliferative effect of SST analogs [25,28]. In addition, a real decrease in invasion and cell migration of HCC cells after stimulation of SSTR1 by a specific agonist has also been demonstrated [22]. This action has also been confirmed in vivo [29], with the demonstration of a similar effect on metastatic dissemination [23,30]. These initial results paved the way for clinical trials on patients with HCC, but their conclusions are quite contradictory, [31] showing rather positive effects in the advanced stages [32,33] and others quite negative [34,35]. These outcome discrepancies could come from heterogeneity in the choice of patients, but available data are still insufficient to truly conclude on the effectiveness of analogs of SST alone in the control of HCC tumors [6,31,36]. Cholangiocarcinoma, the other main primary liver tumor, might also be a potential target [18,37].

On the other hand, in certain cases, and in particular for neuroendocrine tumors (a category of tumors where SSTRs are the most expressed), a benefit has been proven via two Phase III studies, which have greatly contributed to the fact that SST analogs are now used in clinical routine [38,39].

#### **2. Somatostatin Analogs**

Somatostatin has a short half-life in the body (between one and three minutes), because it is rapidly degraded by peptidases found in plasma and tissues [40]. Therefore, the amount present in the bloodstream is extremely low (between 14 and 32.5 pg/mL). This very short half-life has been considered a limiting factor for possible clinical applications, thus many analogs with better metabolic properties (longer half-life between 1.5 h and 12 h) have been rapidly developed [2,5,9]. These are most often hexapeptide or octapeptide molecules which incorporate the biologically active core of native somatostatin (see some examples in Figure 3). Indeed, studies on the structure–activity correlation have shown that the Phe 7 , Trp 8 , Lys 9 , and Thr 10 sequence in the form of a β-sheet is necessary for biological activity. The residues Trp <sup>8</sup> and Lys <sup>9</sup> are essential for this activity, whereas Phe <sup>7</sup> and Thr <sup>10</sup> may undergo some substitutions. Among somatostatin analogs, there are two main categories: the agonists (substances capable of activating somatostatin receptors) and the antagonists (molecules that interact with somatostatin receptors and block or reduce the physiological effect of an agonist). It is also important to note that somatostatin analogs have different affinities for the different receptor subtypes [2].

**Figure 3.** Chemical structures of SRIF-14, SRIF-28, and selected examples of somatostatin analogs.

to stabilize the β The first agonist peptide analog to be approved by the FDA was octreotide (SMS 201-995), marketed under the name Sandostatin ®. From a structural point of view, it has a d-Trp and a d-Phe, to stabilize the β-sheet and a disulfide bridge closer to the active core, for a better metabolic

stability. Its pharmacodynamics is highly similar to native SST, which has made it widely used in clinical trials for the treatment of GEP (gastro-entero-pancreatic) tumors [41,42]. Next, Lanreotide (BIM 23014, tradename Somatuline®), whose structure is similar to that of octreotide (Phe and Thr having been replaced by Tyr and Val respectively), showed comparable characteristics and is also widely used in the treatment of neuroendocrine tumors [43]. In 2005, another analog, Vapreotide (RC160), was marketed under the name Sanvar®, with properties close to those of the two previous analogs, and is also used for the treatment of esophageal varices. More recently, Pasireotide (SOM-230 or Signifor®) was one of the first analogs to show a strong affinity for most of the somatostatin receptor subtypes (pansomatostatin analog). Marketed by Novartis, it is used for the treatment of Cushing's disease [44]. Many other analogs have been developed, from "ultra-short" peptides, such as SDZ 222-100 (an adamantine cyclopeptide), to longer ones, such as KE-108 or CH-275 [5]. Regarding antagonist peptide analogs, the wide variety of compounds that the octapeptide model can offer has allowed the discovery of several structures that can block this kind of receptors. The first antagonist that has been described in the literature is CYN-154806, followed by PRL-2970, sst3-ODN-8 or even non-cyclic models such as BIM-23056 and BIM-23627. New non-peptide compounds have also emerged [45]. These agonists and antagonists (selective or not) constitute a very promising field in the chemistry of somatostatin analogs, in particular because of their pharmacological, pharmacokinetic, and physicochemical properties. This type of compound may have a stronger affinity and/or selectivity for certain subtypes of somatostatin receptors than the majority of peptide analogs. They can thus provide additional information on the exact role of each of these subtypes [5,9].

#### **3. Targeting of Somatostatin Receptors with Radiopharmaceuticals**

In the field of medicine, much research is focused on finding methods to achieve earlier detection of pathologies to allow treatment at early stages of the disease, to increase the chances of total recovery. For this purpose, nuclear medicine, through the use of radiopharmaceuticals, is a very powerful tool. Its application can have two different aims: imaging, with the visualization of a radioactive element's distribution in the body, or therapy, with specific irradiation of abnormal cells, thereby reducing damages to nearby healthy tissue. Having a broad range of potential biological targets and desirable pharmacokinetic characteristics—such as high uptake in target tissue and fast blood and non-target tissue clearance—peptides can also be easily chemically modified for incorporation into a radiopharmaceutical, making them a very potent targeting vector for nuclear medicine. Research in that domain has thus gained widespread interest [46–49]. These compounds can be directly labeled with a radionuclide, such as a halogen radioisotope, but they are generally based on a triple structure involving: (i) a radiometal, the radiation of which allows either the localization (γ and β <sup>+</sup> emitters) or the destruction (β <sup>−</sup>, α or Auger electron emitters) of the targeted cells; (ii) a bifunctional chelating agent (BFCA), the dual role of which is not only to bind the radiometal in a very stable manner to minimize its dissociation in vivo, but also to allow its conjugation with targeting moiety (or vector) via a functionalized arm; (iii) a targeting moiety (the peptide analog), which aims to convey this set in a specific way to a well-defined target. To limit the influence of the chelating moiety, a linker (or spacer) is usually inserted between the BFCA and the biomolecule (Figure 4).

The choice of the radiometal is crucial, since it deeply influences the design of the chelating structure [50–53]. Several criteria govern the choice of radionuclide: (i) the nature of the radiation emitted, depending on the intended application (diagnosis or therapy); (ii) the half-life, which must be long enough to allow effective fixation of the radiotracer on the target cells, but relatively short to avoid irradiation of the organism (neighboring healthy tissues) and more specifically non-targeted organs; (iii) the isotope decay profile. By emitting its radiation, the nuclide disintegrates into a daughter nuclide, which must be non-radioactive to avoid any additional harmfulness to the organism; (iv) the means of production. Most of the radioelements used in nuclear medicine are artificial. They can be produced in three different ways: from a nuclear reactor, a cyclotron or via a generator. Generator production remains the most convenient way for clinical application, as it can provide in-house radionuclides

's distribution

localization (γ and β emitters) or the destruction (β<sup>−</sup>

when a cyclotron is not available nearby, but cyclotron production still remains the cheapest and most used. As an example, Table 2 shows some of the characteristics of radioactive nuclides among the most used today for the radiolabeling of peptides.

—

–

, α or Auger electron emitters) of the targeted

—

" "

'

**Figure 4.** Schematic design of a radiometallated bioconjugate.

**Table 2.** Some of the main radionuclides studied for imaging and therapy (SPECT—Single-Photon Emission Computed Tomography; PET—Positron Emission Tomography).


From a structural point of view, each radiometal has its own properties such as polarizability, degree of oxidation, or coordination number. These features have a direct impact on the choice of the bifunctional chelating agent, in particular in terms of denticity and nature of the donor atoms (most often *O*-, *N*-, or *S*-donors) [54,55]. The BFCA makes it possible to link the biomolecule and the radiometal; its choice is a crucial step in the construction of a radiopharmaceutical. As indicated above, this structure plays a double role: the first is to complex the radioelement in a very stable manner. Several criteria can be evaluated to truly attest to the stability of the complex formed. First of all, the formed radiocomplex must be thermodynamically stable, i.e., the metal-ligand affinity must be as strong as possible. Then it must be kinetically inert. Many metalation reactions take place in the body and the complex formed must be stable enough to avoid any in vivo degradation (e.g., demetallation or transchelation). In addition, radiolabeling conditions with low concentrations are required, ideally with efficient complexation kinetics (high labeling yield) and fast and mild reaction conditions. Beside chemistry considerations, the radiotracer must have: (i) a strong affinity for the target receptor; (ii) a high accumulation for the target and low for the non-target organs; (iii) relatively rapid clearance in the organism; (iv) preferably a mainly renal route of excretion.

Chelating ligands used for the design of radiotracers are usually classified into two categories: macrocyclic and acyclic compounds (Figure 5). Generally, acyclic ligands are less kinetically inert than macrocycles, although some may have shown very good characteristics. On the other hand, these ligands generally have faster metal-chelate binding kinetics compared to macrocyclic analogs, which represents a huge advantage for working with isotopes that have a short lifespan. Despite the

coordination properties specific to each metal, some chelating agents—such as polyaminopolycarboxylic acids—are considered to be 'universal' because they can complex different radiometals. — ' '

—

**Figure 5.** Representative (but not exhaustive) examples of acyclic and macrocyclic polyamino and polyaminocarboxylic chelator families and their derivatives.

′′ Among acyclic ligands, the first BFCAs developed were EDTA (ethylenediaminetetraacetic acid) and DTPA (diethylenetriaminepentaacetic acid). They have been widely used in the chemistry of radiopharmaceuticals, in particular with radioelements such as 111 In, <sup>90</sup>Y or <sup>177</sup>Lu, and even 99mTc [54]. Later on, DTPA derivatives such as CHX-A′′ -DTPA with a cyclohexyl moiety bringing more rigidity to the DTPA backbone (allowing a pre-organization of the system) showed better kinetic inertia [56]. Regarding cyclic compounds, cyclen derivatives such as DOTA (1,4,7,10-tetraazacyclododecane-1,4,7,10-tetraacetic acid) and triaza analogs—NOTA (1,4,7-triazacyclononane-1,4,7-triacetic acid)—are among the most studied ligands. NOTA has the smaller chelating cavity of the two, and is generally used for Ga (III) or Cu (II) because it has a

particular attraction for these metals, which results in mild radiolabeling conditions and good in vivo stability of the complexes formed. DOTA (which is considered as the gold standard chelator) and its derivatives play an important role in clinical applications because they form very stable complexes with a wide range of trivalent radiometals such as Ga (III), Y (III), In (III), Lu (III), or even divalent such as Cu (II) [57,58]. For DOTA or NOTA, the introduction of a functionalized arm offers the possibility of coupling a biomolecule (NODASA/NODAGA and DOTASA/DOTAGA). Similarly, TETA (1,4,8,11-tetraazacyclotetradecane-1,4,8,11-tetraacetic acid), has mainly been studied with Cu (II) and have shown a stability similar to DOTA [59].

Whether on the side of macrocyclic ligands, derivatives or variations of DOTA (e.g., p-SCN-Bn-DOTA, DOTAGA, CB-DO2A, TCMC . . . ), NOTA (e.g., p-SCN-Bn-NOTA, NETA . . . ), or TETA (e.g., CB-TE2A, p-NH2-Bn-TE3A . . . ), or on the side of acyclic ligands, derivatives or variations of DTPA (e.g., CHX-A′′-DTPA . . . ), a large number of ligands have been developed so far. A wide choice of ligands is available for the design of new agents, and numerous journals have described and carefully classified all the structures that can be used in the design of a radiopharmaceutical, whatever the intended application [46,51,53,54,57,60].

BFCA's second role is to allow the conjugation of the complex with a biomolecule. The nature of this link is very important, because it is essential for it to be stable, and above all, for it to not interfere in any way with the binding to the receiver. The slightest structural modification of the ligand and/or of the biomolecule can have a very marked effect on the affinity to the targeted receptors. To minimize this impact as much as possible, that sometimes a 'spacer' or 'linker' can be used between these two entities. Biomolecules are often functionalized through a primary amine, which provides an ideal conjugation site for a coupling reaction, most often with peptide or thiourea type links. Other links based on thioether, triazole, oxime, or more recently via a copper-free click-chemistry with tetrazine/cyclooctyne may prove to be interesting, in particular, because they have very good stability in vivo [51,54,61].

Many somatostatin analogs have already been labeled with various radioelements, whether for imaging, with probes used today in clinical applications, or for therapy, with many compounds in clinical studies [17,61–63]. These analogs were obtained from modifications in the sequence of amino acids that make up the peptide. For example, replacing Phe<sup>3</sup> in octreotide (OC) with Tyr<sup>3</sup> (TOC) improves the affinity for SSTRs (in particular SSTR2) and introduction of a Thr (TATE) instead of Thr(ol) (TOC) further improves this. By following this procedure, many analogs have been developed and studied, often with the same chelating cavity to be able to compare their properties (Table 3) [64,65].


**Table 3.** Peptidic sequences of the main somatostatin agonist analogs. Differences towards Octreotide (OC) are highlighted in red.

#### *3.1. Radiolabeled Somatostatin Analogs for Imaging*

β

The very first proof of concept for the visualization of tumors expressing SSTRs was carried out with [ 123 I-Tyr 3 ]-octreotide, obtained from an iodination reaction (electrophilic substitution) of tyrosine [66,67]. This compound demonstrated biological activity and an affinity for receptors similar to those of native SST [68]. Despite the obvious interest of this probe, several factors such as the difficult radiolabeling procedure, the significant cost, and particularly, the clearance via the liver and the hepatobiliary system (which makes it difficult to interpret the obtained images) were the main drawbacks of its application [67]. To overcome all of these disadvantages, iodine-123 has been replaced with indium-111, which, through the chelating agent DTPA, has been coupled to octreotide (Figure 6) [69]. In vivo studies of [ 111 In-DTPA<sup>0</sup> ]-octreotide ([ 111 In]-pentetreotide) have shown that it is possible to visualize tumors expressing SSTRs and their metastases, even 24 h after injection. In comparison with the compounds coupled to antibodies, this reveals a relatively rapid clearance via the kidneys, which represents a huge advantage compared to [ 123 I-Tyr 3 ]-octreotide [70,71]. This compound was the first radiopharmaceutical targeting SSTRs to be approved by the FDA (Octreoscan ® marketed in 1994). It has been widely used, and has long been considered a 'gold standard' for the visualization of neuroendocrine tumors. It still has a few limits: in fact, it requires a high tumor/noise intensity ratio, shows low spatial resolution, has a moderate affinity for receptors and finally, and possesses a high γ energy which results in a high dose of radioactivity received by the patient. For all these reasons, research in the field of radiopharmaceuticals has focused on other radioelements such as technetium-99m for SPECT and gallium-68 for PET. In addition to having excellent physical properties, these two elements are available from a commercial clinical-grade generator, an important advantage for clinical applications. ' ' a high γ energy which results in a high dose of radioactivity received by

**Figure 6.** Structure of [ 111 In]-pentetreotide (Octreoscan ®).

#### 3.1.1. Gallium-68 and Indium-111

DOTATOC analog was the first to be radiolabeled with indium-111, and its comparative study with Octreoscan ® showed similar diagnostic accuracy, but with better biodistribution and clearance [72]. Although DOTATATE alone showed better affinity for SSTRs, the two analogs [ 111 In]-DOTATOC and [ <sup>111</sup>In]-DOTATATE showed relatively similar pharmacokinetic properties [73]. SSTR2 receptors—and to a lesser extent, SSTR5—are most often overexpressed in tumors. Consequently, the majority of the radiotracers described have a strong affinity for these two SSTRs subtypes. Systems such as DOTANOC were designed to develop a probe capable of targeting all subtypes. Compared to DOTATOC and DOTATATE, it has a similar affinity for SSTR2 and SSTR5 subtypes, but a much higher affinity towards SSTR3. Their high internalization rate results in interesting biodistribution data, with a greater accumulation of the probe in the tumor and in the organs or tissues expressing SSTRs (e.g., pancreas and adrenal glands), ending with excretion mainly by kidneys [74].

These three systems, similarly labeled with gallium-68 (Figure 7), have proven to be very good radiotracers, and are currently routinely used in clinical applications [75]. These three radiopharmaceuticals have slightly different pharmacokinetic properties, but their diversity is mainly due to the variation in affinity for certain subtypes. This feature is even more marked depending on the radioelement chosen ( <sup>68</sup>Ga or 111 In). This can be explained by the differences in the geometry of the complexes. [ <sup>68</sup>Ga]-DOTATOC is very affine for SSTR2 and more moderate for SSTR5, [ <sup>68</sup>Ga]-DOTATATE

is specific to SSTR2 and finally, [ <sup>68</sup>Ga]-DOTANOC binds with great affinity to SSTR2, SSTR3, and SSTR5 [76–78]. –

— —

**Figure 7.** Structures of the three main systems radiolabeled with gallium-68.

A study with DOTANOC aimed at determining the impact of the introduction of a spacer on the pharmacokinetic properties of the formed radiotracer. The aim was to insert polyethyleneglycol (PEG) moieties or sugars between the chelating cavity (DOTA) and the biomolecule (NOC), which resulted in the modification of the lipophilicity or the charge of the final compound. As a result, the hydrophilicity of the system seems to be involved only in the affinity phenomenon towards the receptor, and the overall charge of the compound influences the excretion profile [79].

DOTA is not the only macrocycle to have been coupled to somatostatin analogs. Knowing the attraction of Ga (III) for NOTA, the latter has been the subject of comparative studies. Conjugated with octreotide (NODAGATOC), the compound showed a strong affinity for SSTR2 (similar to that of DOTATOC). Once marked with 111 In, affinity was even stronger for SSTR2, with even a gain on SSTR3 and SSTR5 (compared to <sup>68</sup>Ga-NODAGATOC), which confirms the influence that the geometry of the complex can have on affinity. In terms of stability, as expected, that of [ <sup>68</sup>Ga]-NODAGATOC was higher than that of [ 111 In]-NODAGATOC. The biodistribution of [ <sup>68</sup>Ga]-NODAGATOC was similar to that of [68Ga]-DOTATOC, but showed a better accumulation in the tumor than [ 111 In]-DOTATOC. This is probably due to the strong agonist character, and the high rate of internalization of the NODAGATOC derivative [80].

A large variety of derivatives have also been investigated, such as DOTALAN, DOTABOC, DOTAGA [81], DOTANOCATE or DOTABOCATE (all derivatives of DOTANOC) [82,83], or THP-TATE (comparison of the overall behavior of the tris chelating system (hydroxypyridinone) with DOTATATE) [84]. New generation analogs with broader affinity profiles or pan-somatostatin analogs have been developed. For instance, AM3 (DOTA-Tyr-cyclo(DAB-Argcyclo(Cys-Phe-d-Trp-Lys-Thr-Cys))), a bicyclic somatostatin analog demonstrated affinity to SSTR2, 3, and 5, when labeled with <sup>68</sup>Ga. It showed a fast background clearance coupled with a high tumor/non-tumor ratio. [85] KE108 was coupled with DOTA and labeled with 111 In and <sup>68</sup>Ga, giving [ 111 In/ <sup>68</sup>Ga]-KE88 (DOTA-d-Dab-Arg-Phe-Phe-d-Trp-Lys-Thr-Phe), which bound to all five SSTRs with high affinity. [86] However, in an in vitro study, it had a low SSTR2 uptake, but was very effective for SSTR3-expressing tumors. More recently, a Pasireotide derivative, DOTA-PA1

(DOTA-cyclo-[HyPro-Phe-d-Trp-Lys-Tyr(Bzl)-Phe]) was labeled with <sup>68</sup>Ga and was investigated in three human lung cancer models, where it demonstrated superiority compared to [ <sup>68</sup>Ga]-DOTATATE [87]. In parallel, the group from Demokritos Institute, in Athens, developed pansomatostatin radiopeptides based on native somatostatin (SRIF-14 and SRIF-28). Both were derivatized with DOTA chelator and labeled with 111 In. Subsequent radiotracers exhibited high affinity and internalization profiles. SRIF-14 derivatives unfortunately demonstrated low in vivo stability. [ 111 In]-DOTA-LTT-SS28, on the contrary, demonstrated a much higher stability and showed more promise [88,89].

#### 3.1.2. Technetium-99m

A wide range of chelating agents have been used to prepare somatostatin analogs labeled with technetium-99m: peptide moieties [90,91], propyleneaminooxime [92], tetraamines [93,94] or a cyclopentadienyl group [95]. Macrocyclic ligands have also been investigated [96]. Three systems stand out for the radiolabeling of somatostatin analogs: HYNIC-TOC and Demotate scaffolds, and P829 (Figure 8).

**Figure 8.** [ 99mTc]-labeled somatostatin analogs.

Initially, the HYNIC core (hydrazinonicotinamide) was designed for the radiolabeling of antibodies and proteins with technetium-99m [97], then this was transposed to peptides and more specifically to octreotide. This ligand can complex the metal in a monodentate or bidentate way, therefore, it is necessary to use one or more co-ligands to complete the coordination of the [ 99mTc]-HYNIC core. Among the most commonly used co-ligands are tricin, nicotinic acid, or EDDA (ethylenediaminodiacetic acid). Each co-ligand has its own influences on the properties of the complex obtained (e.g., lipophilicity and biodistribution) [98]. The first studies were carried out using tricin as a co-ligand ([ 99mTc]-HYNIC-TOC), but quickly EDDA demonstrated a very favorable influence on the pharmacokinetics of the radiotracers [99]. Compared to Octreoscan ®, [ 99mTc]-EDDA/HYNIC-TOC showed better accumulation in the tumor and a weaker accumulation in the kidneys. The improved spatial resolution, the reduction in the radiation dose received by the patient and the better availability of 99mTc made it a possible alternative to Octreoscan ® [99,100]. Finally, its conjugation with the octreotate analog ([ 99mTc]-EDDA/HYNIC-TATE) has shown significantly similar behavior to its octreotide counterpart [101]. [ 99mTc]-EDDA/HYNIC-TOC (Tektrotyd ®) was granted marketing authorization in Europe in adult patients with gastro-enteropancreatic neuroendocrine tumors (GEP-NET) for localizing primary tumors and their metastases.

The second radiotracer, based on the tetraamine motif 6-R-1,4,8,11-tetraazaundecane, is available in a series with [ 99mTc]-Demotate 1 ([ 99mTc-N<sup>4</sup> 0 , Tyr 3 ]-octreotate) and 99mTc-Demotate 2 ([ 99mTc-N<sup>4</sup> 0–1 , Asp<sup>0</sup> , Tyr<sup>3</sup> ]-octreotate). The first version of this probe demonstrated excellent pharmacokinetic properties, including faster accumulation in the tumor compared to Octreoscan® [102]. The objective of the second version was to improve the qualities of [99mTc]-Demotate 1, by modifying the overall charge of the complex and adding an Asp residue. In the end, [99mTc]-Demotate 2 showed overall behavior similar to [111In]-DOTATATE, even if the latter has a faster clearance and a better retention time in the tumor [103]. The last of the main analogs based on technetium-99m is [99mTc]-P829 (99mTc-Depreotide), marketed in 2000 by the company CISBio International under the name of NeoSpect®, but recently withdrawn from the market. The P829 peptide (directly radiolabeled with 99mTc) showed results similar to the other SST analogs [104]. Its use for the detection of neuroendocrine tumors appeared to be less precise than with Octreoscan® [105]. On the other hand, its affinity for SSTR3, subtype which may be the origin of cross-competition from other types of receptors (notably VIP receptors), gave it the ability to bind to a larger number of primary tumors [104]. In particular, it was used clinically for the diagnosis of malignant lung tumors [106–108], for which it got its market authorization [109], and also demonstrated some interest in breast cancer, but it was never confirmed in a larger series of patients [110].

The question that now remains to be answered is that of the clinical interest of a SPECT tracer among the wide choice of PET SSTRs imaging agents [111,112].

#### 3.1.3. Copper-64

Due to the short half-life of <sup>68</sup>Ga (T1/<sup>2</sup> = 67.7 min.) each center willing to perform <sup>68</sup>Ga PET imaging must purchase a currently expensive <sup>68</sup>Ge/ <sup>68</sup>Ga generator and a specifically shielded hot-cell. For this reason and despite the FDA and EMA market authorizations for [68Ga]-DOTATATE and [ <sup>68</sup>Ga]-DOTATOC and the better diagnostic performances for these two radiopharmaceuticals products, the use of <sup>68</sup>Ga appears to be under the dependence of an economic choice for many hospitals and only a few large centers are making the financial investment to perform <sup>68</sup>Ga-radiolabeling. In this context, the use of a PET-emitter with a longer half-life such as copper-64 (T1/<sup>2</sup> = 12.7 h) appears to be an interesting alternative to remove the financial hindrance of gallium-68 [113]. This physical parameter allows for a centralized radiolabeling site with a large multicentric supply of ready-to-use <sup>64</sup>Cu-radiolabeled compounds. The chemistry of copper is also well known, which is a real asset in the design of new radiotracers. Many systems already presented before, such as DOTATOC/TATE or NODAGATOC/TATE, or others more copper-specific BFCAs, such as TETA (1,4,8,11-tetraazacyclotetradecane-*N*,*N*′ ,*N*′′ ,*N*′′′-tetraacetic acid) [114], and its more stable derivatives such as cross-bridge CB-TE2A (4,11-bis(carboxymethyl)-1,4,8,11-tetraazabicyclo [6.6.2]hexadecane) [115], and CPTA (4-[(1,4,8,11-tetraazacyclotetradec-1-yl)methyl]benzoic acid]) [116] or sarcophagine derivatives [117] have been studied. A review on the development of copper radiolabeled somatostatin analogs was recently published by Marciniak et al. [118].

To validate the clinical interest of [64Cu]-somatostatin analogs, various clinical studies have been conducted around the world. Among the different somatostatin analogs, [64Cu]-DOTATATE was one of the first used. In 2015, [64Cu]-DOTATATE was compared head-to-head to [111In]-DTPA-octreotide in 112 patients and showed that the PET <sup>64</sup>Cu-compound was far superior to SPECT <sup>111</sup>In compound performances [119]. In 2017, [64Cu]-DOTATATE was challenged to [68Ga]-DOTATOC according to an identical PET/CT imaging modality [120]. The results of this study, where 59 patients were injected with [68Ga]-DOTATOC followed by an injection of [64Cu]-DOTATATE one week later, concluded that the two radiopharmaceuticals had the same sensitivity. Nevertheless, in this cohort of neuroendocrine tumors, [64Cu]-DOTATATE had a substantially better lesion detection rate. The patient follow-up revealed that these additional lesions detected by [64Cu]-DOTATATE were true positives. To evaluate the benefits of this better detection of lesions with [64Cu]-DOTATATE than with [68Ga]-DOTATOC, the correlation between PET image [64Cu]-DOTATATE uptake (expressed in maximal standardized uptake value - SUVmax) and overall (OS)/progression free survival (PFS) was studied during 24 months after [64Cu]-DOTATATE PET/CT acquisition. The conclusion of this study claimed a

good correlation/prognostic between SUVmax and PFS but not with OS [121]. The major drawback of these preliminary human studies consist of the affinities differences for the five SSTRs subtypes between DOTATOC and DOTATATE compounds. To circumvent these discrepancies, an in vitro study in a mouse model was conducted and compared [64Cu]-DOTATATE to [68Ga]-DOTATATE. The results showed a similar pharmacokinetic and absolute uptake between both compounds 1 h post-injection [122]. In Europe, where the PET radiopharmaceutical approved is [68Ga]-DOTATOC, it could be interesting to perform some PET imaging with [64Cu]-DOTATOC to compare the performance of the two tracers. A first-in-human retrospective study was recently conducted and seems to present same results than [64Cu]-DOTATATE with high detection rate of suspected lesion associated to a high target-to-background contrast [123]. A recent first-in-human study also demonstrated potential interest for [64Cu]-SARTATE analog [124].

In conclusion, despite a higher dosimetric impact for copper-64 (only 17.6% of radioactive decay lead to positron emission), copper-64 somatostatin analogs appear to be an advantageous alternative to gallium-68 radiopharmaceuticals. Compared to <sup>68</sup>Ga, in addition to economic advantages, <sup>64</sup>Cu has a lower positron range which leads to a better PET intrinsic resolution and a higher half-life which allows for a more flexible scanning window. The better patient care management and outcomes remain to be proven and the work is in progress to establish these points [121,125]. In parallel, at present, a radiopharmaceutical industrial company submitted a market authorization from FDA for [ <sup>64</sup>Cu]-DOTATATE and thus confirms the interest of copper-64 in SSTRs imaging.

#### 3.1.4. Other Radiometals

Other radionuclides have also been investigated for SSTRs imaging. Cobalt-55 seems to be a possible alternative to gallium-68 and copper-64 compounds, with similar behavior and lifespan (17.5 h vs. 12.7 h) to the latter, but with a higher positron yield (75.9% vs. 17.6%). Preliminary complexation tests of DOTATOC with the isotope <sup>57</sup>Co as a surrogate for <sup>55</sup>Co showed a higher affinity for SSTR2 than [68Ga]-DOTATOC, implying a rate of internalization among the highest of all derivatives of SST and thus, a strong accumulation in targeted tissues. Despite similar structures, the analogs of cobalt and gallium have different biological behaviors. This confirms the fact that the physical characteristics of radioactive elements influence the affinity, biodistribution, and pharmacokinetics of radiolabeled peptides [126]. The properties of cobalt-based compounds have been further investigated with the comprehensive evaluation of other octreotide analogs such as DOTANOC and DOTATATE [127]. Furthermore, [55Co]-DOTATATE compared favorably with [68Ga]-DOTATATE and [64Cu]-DOTATATE in an animal model [122]. Associated with the Auger-emitting 58mCo, it could represent a potentially interesting theranostic pair [128].

Scandium and terbium are two metals that recently emerged as possibly useful for theranostic applications, as both possess imaging and therapeutic radionuclides [129]. DOTATOC was radiolabeled with scandium-44 (T1/<sup>2</sup> = 3.97 h, Eβ<sup>+</sup> = 632 keV) [130] and terbium-152 (T1/<sup>2</sup> = 17.5 h, Eβ<sup>+</sup> = 1140 keV) [131] and rapidly injected in patients in proof-of-concept studies [132,133]. No adverse effects were observed during follow-up periods and images proved suitable for diagnosis. With DOTATATE, it seems the affinity to SSTR2 receptors is lower with scandium than with gallium, thus limiting its interest [134]. In a study comparing the labeling and stability of DOTANOC and NODAGANOC with <sup>44</sup>Sc and <sup>68</sup>Ga, it was observed that [44Sc]-NODAGANOC labeling was more challenging and less stable than [44Sc]-DOTANOC [135]. The opposite was observed with <sup>68</sup>Ga. Recently, a new chelator was proposed, AAZTA (1,4-bis (carboxymethyl)-6-[bis (carboxymethyl)]amino-6-methylperhydro-1,4-diazepine), which enables fast and easy labeling at room temperature. AAZTA-TOC labeled with <sup>44</sup>Sc demonstrated high in vitro stability [136]. Affinity tests are now necessary to assess its potential utility. DOTATATE has also been labeled with <sup>155</sup>Tb (T1/<sup>2</sup> = 5.32 days, E<sup>γ</sup> = 87 keV (32%), 105 keV (25%)) for SPECT imaging [137]. Though a potentially promising radionuclide for theranostic applications, availability of <sup>155</sup>Tb is currently the main limitation for further development.

At the turn of the millennium, yttrium-86 (T1/<sup>2</sup> = 14.74 h, 32% β <sup>+</sup>) was thought to be a potential radionuclide of interest, particularly for pretherapeutic dosimetry of <sup>90</sup>Y-radiotracers, and notably <sup>90</sup>Y-labeled somatostatin analogs [138]. Thus, several octreotide analogs were developed [139,140]. [ <sup>86</sup>Y]-DOTATOC even reached the clinics [139,141]; however, <sup>86</sup>Y properties are less than optimal, and availability is limited, so interest soon faded out.

#### 3.1.5. Fluorine-18

Radiometals' production is currently still limited, even for the most advanced ones [142–144]. Fluorine-18, on the contrary, can be mass-produced and distributed daily, thanks to a worldwide network of cyclotrons. Because of this availability, and favorable decay characteristics (T1/<sup>2</sup> = 110 min, 97% β <sup>+</sup>), it thus should be noted that some radiotracers based on fluorine-18 have been described (Figure 9) [145]. The first generations such as 2-[ <sup>18</sup>F]fluoropropionyl-d-Phe 1 -octreotide [146] or 4-[18F]fluorobenzoyl-d-Phe 1 -octreotide [147] generally showed unfavorable biokinetic properties (low accumulation and low retention in the tumor). The probes developed subsequently contained hydrophilic or charged moieties to reduce the lipophilicity of the radiotracer. In particular, several carbohydrate derivatives of octreotide/octreotate have been developed [148,149]. A disadvantage of fluorine-labeling compared to radiometal labeling is the use of generally long and tedious multi-step procedures. To circumvent this, innovative strategies, enabling fast and purification-less labeling, have been developed, such as the formation of <sup>18</sup>F-boron or <sup>18</sup>F-silicon bonds, or the use of click-chemistry [150–152]. Another elegant method to label somatostatin analogs is the use of [ <sup>18</sup>F]-aluminum fluoride with radiotracers previously developed for radiometals, such as NOTATOC [153]. These new generation analogs demonstrated general properties (affinity for the targeted receptors, metabolic stability, biodistribution and clearance) which are much more interesting, and some of them have been investigated in patients, where they gave results comparable to [ <sup>68</sup>Ga]-DOTATOC [154,155]. In addition, [ <sup>18</sup>F]F-FET-βAG-TOCA and [ <sup>18</sup>F]-IMP466 ([Al <sup>18</sup>F]-NOTATOC) are currently being evaluated in phase I clinical trials (EudraCT number 2013-003152-20 and NCT03511768, respectively). Recently published results with [ <sup>18</sup>F]-IMP466 demonstrated it was safe and well-tolerated, with a physiologic uptake pattern similar to [ <sup>68</sup>Ga]-DOTATATE [156]. Besides cost and availability, another advantage of fluorine-18 is its shorter positron range compared with gallium-68, leading to an improved spatial resolution, and thus, better quantification of uptake [157].

**Figure 9.** [ <sup>18</sup>F]-labeled somatostatin analogs.

more suited to small tumors. In addition, the energy of its γ radiation is sufficient to allow detection

90, a pure high energy β<sup>−</sup> βmax

medium energy β<sup>−</sup> βmax = 0.5 MeV) with a γ

'

eutic application via the use of β<sup>−</sup>

#### *3.2. Radiolabeled Somatostatin Analogs for Therapy*

Concerning radionuclide therapy and more particularly peptide receptor radionuclide therapy (PRRT), radioactivity is used to destroy the targeted cells. Radiopharmaceuticals used in therapy are designed in the same way as those used in imaging, only the nature of the radioelement being modified. Contrary to imaging, which uses radioelements having very penetrating but little ionizing radiations, PRRT privileges the use of radionuclides that have little penetrating and more energetic and thus more ionizing radiations. Brought directly to the cancer cell, the radiation emitted by the radioactive decay causes irreversible ionization of the cell's DNA, which induces its apoptosis. The main isotopes used today are iodine-131, yttrium-90, lutetium-177 and, to a lesser extent, rhenium-188 [158]. As mentioned earlier, the purpose of the DOTA-SSA design was to work with a chelating cavity capable of complexing radioelements for imaging or therapy. Consequently, most of the platforms discussed above have been transposed for therapeutic application via the use of β <sup>−</sup> emitters [64,74,81,82].

#### 3.2.1. Yttrium-90 and Lutetium-177

Yttrium-90, a pure high energy β <sup>−</sup> emitter (T1/<sup>2</sup> = 64 h, Eβmax = 2.28 MeV), and lutetium-177, a medium energy β <sup>−</sup> emitter (T1/<sup>2</sup> = 6.7 d, Eβmax = 0.5 MeV) with a γ component (208 keV), are currently the most used in PRRT. Each of these two elements has its own advantages for targeted therapy. The particles emitted by <sup>90</sup>Y are more energetic and more penetrating; they are able to diffuse on a thicker layer of cells, which is an advantage for the treatment of large tumors. However, even if high energy radiation allows a more uniform irradiation of the tumor, the risk of imposing an excessive dose of radiation on the adjacent tissues is very present. For its part, the <sup>177</sup>Lu emits less energetic radiation, more suited to small tumors. In addition, the energy of its γ radiation is sufficient to allow detection by scintigraphy and establish dosimetry during the therapy sequences [159].

The first analog to be studied was [90Y]-DOTATOC (Octreother®), and the first treatment sessions quickly showed good results, stopping the progression of the tumor [72,160,161]. Many studies on this long-used treatment have made it possible to observe a good tolerance for this radiotracer, with fairly mild side effects (fatigue) and in very rare cases a little more severe ones (nausea). However, it also showed some toxicity for the kidneys and the bones, these two aspects being the dose-limiting factors for the patient. In vitro, a greater affinity for SSTR2 has been demonstrated for [90Y]-DOTATATE compared to [90Y]-DOTATOC [64]. However, for the diagnosis in humans, a better contrast between the kidneys and the tumor was found for [111In]-DOTATOC compared to [111In]-DOTATATE [73], which may explain the wider use of DOTATOC analog. Despite this, these two analogs have relatively similar properties and have proven to be effective treatment methods that improve survival in some patients with neuroendocrine tumors (approximately 50 months vs. 18 months without treatment) [162]. In a Phase IIA study with [ <sup>90</sup>Y]-DOTALAN (MAURITIUS trial), this one demonstrated lower tumor uptake in neuroendocrine tumors compared to <sup>90</sup>Y-DOTATOC, but could be of potential interest for other tumors, such as HCC or lung cancers [163]. With the perspective of several years of clinical use, PRRT with <sup>90</sup>Y-labeled somatostatin analogs appears to be well-tolerated with favorable long-term outcome. Unfortunately, Phase III studies are still lacking [164,165].

The same analogs have also been radiolabeled with lutetium-177. Initially, [177Lu]-DOTATOC was used in cases of relapse of neuroendocrine tumors after treatment with [90Y]-DOTATOC. Despite satisfactory results [166], its subsequently developed analog [177Lu]-DOTATATE has shown more promise, mainly due to a more significant retention time in the tumor. For this reason, octreotate analog (TATE) is being preferred to octreotide (TOC) for labeling withlutetium [164,167]. It is also important to note that, unlike <sup>90</sup>Y, no cases of nephrotoxicity after treatment with <sup>177</sup>Lu have been reported. In 2005, the possibility of combining these two β <sup>−</sup> emitters for therapy in cases where tumors of variable sizes are detected, was demonstrated [168]. From there, different treatment combinations between the four main systems ([90Y]-DOTATOC, [ <sup>90</sup>Y]-DOTATATE, [177Lu]-DOTATOC, and [177Lu]-DOTATATE) have proven to be interesting and sometimes even more effective than using a single treatment modality [169,170]. Similarly, combination treatments with non-labeled somatostatin analogs, chemotherapy, targeted therapy, and/or

radiosensitizers might further improve the efficacy and/or tolerability [171,172]. [ <sup>177</sup>Lu]-DOTATATE has been investigated in a phase III trial, in well-differentiated, unresectable or metastatic, progressive midgut neuroendocrine tumors (Netter 1 trial). Treatment with [ <sup>177</sup>Lu]-DOTATATE resulted in a significant tumor response rate of 18% compared with 3% in the high-dose octreotide LAR group, coupled with a 79% risk reduction for disease progression or death [173]. Following these positive findings, [ <sup>177</sup>Lu]-DOTATATE was granted marketing authorization in this indication, both in Europe and in the US (Lutathera ®) [174]. Coupled with <sup>68</sup>Ga-imaging (Figure 10), it represents a powerful theranostic tool for the management of neuroendocrine tumors (NETs) [175]. Current research with [ <sup>177</sup>Lu]-DOTATATE aims to improve the safety and efficacy of this procedure, enlarge possible indication, notably in advanced, poorly-differentiated, GEP-NETs, [176,177] or other NETs, such as pheomochromocytoma or paraganglioma [178,179].

**Figure 10.** (**A**) [ <sup>68</sup>Ga]-DOTATOC (Somakit ®) and (**B**) [ <sup>177</sup>Lu]-DOTATATE (Lutathera ®, cures 1, 2, and 3) imaging of a patient treated for progressive metastatic midgut NET (images courtesy of Centre Eugene Marquis, Rennes, France).

#### 3.2.2. Rhenium-188 and Other β-Emitting Radionuclides

ther β – Despite equally interesting characteristics, rhenium-188 remains widely less used than <sup>90</sup>Y and <sup>177</sup>Lu [180]. This is mainly due to more difficult chemistry and the unavailability of a pharmaceutical-grade <sup>188</sup>W/ <sup>188</sup>Re generator, as compared to the other two. Vapreotide and Lanreotide analogs have been described in the literature with <sup>188</sup>Re. They have been investigated in experimental cancer models (e.g., pancreas, colorectal, lungs and cervical) to reduce tumor growth [181–184]. [ <sup>188</sup>Re]-Lanreotide notably demonstrated favorable pharmacokinetics and distribution profiles (tumor-to-liver ratio) in HCC-bearing rats compared to healthy ones [185]. Another example is an equivalent to Depreotide (P829). After the development of 99mTc-Depreotide for imaging, the idea was to label this compound with <sup>188</sup>Re, to assess its potential in vivo. Although the radiolabeling proceeded successfully, the study showed unacceptable toxicity to non-target organs. To improve its properties, structural modifications of the peptide sequences close to the chelating moiety were tested. This optimization led to P2045, which showed better accumulation in the tumor, weaker retention in the kidneys, and faster urinary excretion than [ 99mTc]-depreotide [186]. This new rhenium-based analog of depreotide, [ <sup>188</sup>Re]-P2045 (Figure 11), went up to phase I in therapy for small cell lung cancer [187] and

ther β

has shown promising in vivo results in the treatment of pancreatic tumors in mice [188]. To the best of our knowledge, no HYNIC-TOC/TATE or demotate derivatives have yet been radiolabeled with rhenium. Recent research with rhenium isotopes has been focusing on tricarbonyl core derivatives for the labeling of NOTA-SSAs [96].

–

**Figure 11.** [ <sup>188</sup>Re]-P2045.

In a theranostic perspective, other β-emitting nuclides could have a potential interest—such as <sup>47</sup>Sc (T1/<sup>2</sup> = 3.35 d, Eβmax = 600.8 keV), <sup>67</sup>Cu (T1/<sup>2</sup> = 2.58 d, Eβmax = 577 keV), and <sup>161</sup>Tb (T1/<sup>2</sup> = 6.91 d, Eβmax = 593 keV)—to be coupled with <sup>44</sup>Sc, <sup>64</sup>Cu, and <sup>152</sup>Tb/ <sup>155</sup>Tb respectively [129,158,189]. To date, no <sup>67</sup>Cu-labeled somatostatin analogs have been described so far, and only very preliminary studies have been described with [ <sup>161</sup>Tb]-DTPA-Octreotide and [ <sup>47</sup>Sc]-DOTATOC [190,191].

#### 3.2.3. Alpha and Auger Emitters

Recently, alpha emitters have attracted particular attention for radionuclide therapy. Long confined to hematological tumors, they are now being considered for the potential treatment of solid tumors [192]. In vitro, α-labeled somatostatin analogs (DOTATOC and DOTATATE) demonstrated a significantly higher killing effect compared to <sup>177</sup>Lu [193–195]. [ <sup>213</sup>Bi]- and [ <sup>225</sup>Ac]-labeled DOTATOC ( <sup>213</sup>Bi: T1/<sup>2</sup> = 45.6 min, E<sup>α</sup> = 5.88 MeV; <sup>225</sup>Ac: T1/<sup>2</sup> = 9.92 d, E<sup>α</sup> = 5.83 MeV) have demonstrated promising therapeutic effects in pre-clinical animal studies [196,197]; whereas [ <sup>213</sup>Bi]-DOTATATE, investigated in human small cell lung carcinoma and rat pancreatic tumor models, demonstrated a great therapeutic effect in both small (50 mm<sup>3</sup> ) and large (200 mm<sup>3</sup> ) tumors, but with a higher probability for stable disease in small tumors [198]. First, and, to date, the only clinical experience with [ <sup>213</sup>Bi]-DOTATOC, was published by Kratochwil et al., and included seven patients with advanced NETs with liver metastases refractory to treatment with [ <sup>90</sup>Y]-DOTATOC or [ <sup>177</sup>Lu]-DOTATOC [199]. It demonstrated specific tumor binding, lower toxicity than with β-irradiation and partial remission of metastases. Two years after intra-arterial injection of [ <sup>213</sup>Bi]-DOTATOC, all seven patients were still alive. Regarding <sup>225</sup>Ac, a first-in-human study included 10 patients with progressive NETs after β-PRRT. As with <sup>213</sup>Bi, [ <sup>225</sup>Ac]-DOTATOC was well tolerated and effective [200]. A recent study with [ <sup>225</sup>Ac]-DOTATATE confirmed the potential of these radiotracers as an additional, and valuable, treatment option for patients who are refractory to [ <sup>177</sup>Lu]-DOTATATE therapy. 32 patients with previous [ <sup>177</sup>Lu]-DOTATATE therapy were treated with [ <sup>225</sup>Ac]-DOTATATE (100 kBq/kg body weight). The response was assessed in 24 patients, with 9 stabilized diseases and 15 partial remissions [201].

Though not stricto sensu an α-emitter, lead-212 (T1/<sup>2</sup> = 10.6 h) eventually decays to stable 2 <sup>08</sup>Pb through a cascade chain with two α-emissions of potential therapeutic interest. A somatostatin analog, DOTAMTATE (Figure 12), has been labeled with <sup>212</sup>Pb and investigated in a murine model of neuroendocrine tumor. Results showed a promising safety index with a 3.2-fold increase in median survival and one-third of the animals being tumor-free. A combination with 5-FU (Fluorouracyl) was able to durably cure approximately 80% of the animals. [202] Given these promising outcomes, a Phase I dose-escalation clinical trial has recently been started with [ <sup>212</sup>Pb]-DOTAMTATE (AlphaMedix™) including 50 patients with unresectable or metastatic neuroendocrine tumors (NCT03466216). Preliminary results (nine patients enrolled) demonstrated a favorable safety profile at the tested doses [203].

™

—

α

α α

α α

β —

βmax βmax βmax

–

β

β

**Figure 12.** [ <sup>212</sup>Pb]-DOTAMTATE.

<sup>α</sup> α Cyclotron-produced astatine-211 (T1/<sup>2</sup> = 7.2 h, E<sup>α</sup> = 5.87 MeV) is another very promising α-emitting radionuclide. Astatine is the heaviest halogen with a behavior somehow similar to iodine, but, in certain circumstances, it also displays significant metallic characteristics [204]. Direct astatination of somatostatin analogs is feasible, through tyrosine residues, but it led to poor stability of the resulting analogs, therefore different prosthetic groups have been developed [205–207]. Although *N*-(3-[ <sup>211</sup>At]astato-4-guanidinomethylbenzoyl)-Phe 1 -octreotate ([ <sup>211</sup>At]-AGMBO) and *N*α -(1-deoxy-d-fructosyl)-*N*<sup>ε</sup> -(3-[ <sup>211</sup>At]astatobenzoyl)-Lys 0 -octreotate ([ <sup>211</sup>At]-GABLO) showed disappointing biodistribution results, with poor tumor uptake, [ <sup>211</sup>At]-SPC-octreotide displayed a more favorable biodistribution profile, and a dose-dependent apoptosis in an NSCLC murine model.

Auger electron emitters are also very potent for specific tumor cell killing, sparing surrounding cells, with a highly localized energy deposition. Indium-111 emits Auger electrons (EAe- = 19 keV, 16%), and, as such, has been investigated for therapy. Several clinical trials have been undertaken with high doses of [ 111 In]-Pentetreotide. A first study with 20 patients that had neuroendocrine progressive tumors demonstrated stabilization of the disease in 5 patients, and tumor shrinkage in 5 others. All of them had received a cumulated dose higher than 20 GBq [208]. In a study with 50 SSTR-positive patients treated with cumulated doses from 20 to 160 GBq, of which 40 were evaluable, there was a stabilization in 14 patients, minor remission in 6 and partial remission in 1, with mild bone marrow toxicity [209]. However, half of the patients receiving more than 100 GBq developed a myelodysplastic syndrome or leukemia. A dose of 100 GBq was thus considered the maximal tolerated activity. Another study with 27 patients with GEP-NETs found that two doses of 6.6 GBq (180 mCi) were safe and well-tolerated, demonstrating a clinical benefit in 62% of patients [210]. Benefit of 111 In-Pentetreotide treatment was shown to last at least 6 months for 70% of patients, while only 31% of them still had sustained benefit after 18 months [211]. Efficacy in large tumors and end-stage patients is limited, mainly because of heterogeneous radiopharmaceutical uptake due to poor tumor vascularity and central necrosis [212]. This has been demonstrated by Capello et al. in a rat tumor model, with different sizes of tumors [213]. Effects were much more pronounced in small (≤ 1 cm<sup>2</sup> ) tumors than in large (≥8 cm<sup>2</sup> ). They also found a significant increase in tumor receptor density after tumor regrowth, indicating repeated injections would probably be more efficient than single-dose treatment. It could also be worth using PRRT with Auger emitters in an adjuvant setting after surgery, to destroy occult metastases. A final example is [ 58mCo]-DOTATOC. This radiotracer presented for potential use in Auger-based therapy, particularly for disseminated tumor cells and micrometastases, appears to have more beneficial in vitro properties than those of [ <sup>177</sup>Lu]-DOTATATE, with a significantly more efficient cell killing effect per cumulated decay, which has to be confirmed in vivo [127].

#### **4. Antagonists vs. Agonists**

Pharmacomodulation around the synthetic somatostatin analogs has led to a change of chirality in the first amino-acid (from d to l form) and in cysteine number 2 (from l to d form). These modifications have given a new class of SSTR specific compounds with antagonist effects (Table 4). From a pharmacological point of view, the biological and molecular mechanisms responsible for their targeting effectiveness in vivo are completely different. After binding to an SST receptor, an agonist analog is internalized into the cell as a ligand-receptor complex. This internalization allows it to accumulate in the cell, and to increase the amount of radiation emitted. This very powerful and specific internalization mechanism enables efficient in vivo targeting of receptors. This phenomenon does not occur (or very little) for somatostatin antagonists, and they do not stimulate the G-protein coupled to the SSTR with an associate blockage of the agonist-induced activity. Surprisingly, it has been shown that targeting receptors can also be effective without internalization of the ligand-receptor complex, and some antagonist analogs can sometimes behave better than agonists (e.g., better accumulation in tumor, poor kidney retention, and rapid clearance) [214,215]. This high tumor uptake appears to be a consequence of a greater number of target binding sites for antagonists and a more slowly dissociation than for agonists, which allows for a longer accumulation of radiation [216,217]. The hypothesis of a ligand rebinding mechanism has been put forward, but this still requires some investigation before it can be validated. These first results were confirmed by preclinical studies and by preliminary clinical trials and seems to show superior results for antagonist-based tracers than agonists [218–221]. The first comparative study of antagonists with Octreoscan® confirmed the good characteristics of the [111In]-DOTA-BASS analog, and better accumulation at the level of the tumor and better visualization of metastases. It was truly the first proof of the concept of antagonist SSTRs imaging [222].

**Table 4.** Main somatostatin antagonist analogs. Differences towards octreotide (OC) are highlighted in red.


Concerning the affinity for each SSTR subtype, it turned out that the nature of the chelator and the radiometal is of great importance for the in vivo pharmacokinetic fate (mainly for the tumor uptake and retention time) [223]. Ultimately, copper-64 based radiotracers seem to be more interesting, especially when comparing their contrast ratio between the tumor and normal tissues which increases over time—a direct consequence of their higher half-life. The influence of radiometals (111In, <sup>90</sup>Y, <sup>177</sup>Lu, <sup>64</sup>Cu, and <sup>68</sup>Ga) and chelates (DOTA and NODAGA) on three antagonist families (LM3, JR10, and JR11) were also studied. On the radiometric side, the overall affinity of [68Ga]-DOTA was found to be much lower than for the other elements, which is the opposite of the results obtained with the agonists. For the chelate, the substitution of DOTA by NODAGA seems to greatly improve the affinity of the antagonist analogs. During this study, two particularly promising platforms emerged, DOTA-JR11 and NODAGA-JR11 [224]. Another example highlighting the influence of the chelate is 406-040-15 (cyclo (2–11) H-Cpa-DCys-Asn-Phe-Phe-DTrp-Lys-Thr-Phe-Thr-Cys-2NalNH2), a pansomatostatin analog, with an SSTR3 antagonist behavior. Chelation to DOTA turned this analog to an agonist [225]. Note that the first antagonist labeled via a [99mTc]-tricarbonyl core has been described. 99mTcL-sst2-ANT (with L = tridentate ligand type N, S, N) has shown very promising in vivo behavior, but requires some modifications to improve its pharmacokinetics [226].

As for imaging, antagonists are also an interesting alternative for therapy. As discussed above, the first proof of the feasibility of imaging using antagonists was highlighted by comparing Octreoscan® and [111In]-DOTA-BASS. However, this analog has shown only a very modest affinity for the SSTR2 receptor subtype targeted in the therapy of neuroendocrine tumors [214]. To overcome this problem, the second generation of somatostatin antagonists was synthesized to improve affinity for this receptor. DOTA-JR11 showed the highest affinity for SSTR2 and was selected for use in targeted therapy [218]. A pilot study to assess the possibility of treatment with [177Lu]-DOTA-JR11, by comparing it to [177Lu]-DOTATATE, was carried out. This new antagonist has shown favorable properties, such as better accumulation in the tumor and a higher dose received by the tumor, thanks

to a longer retention time [227]. Further developments led to a theranostic pair with JR11: one with a NODAGA chelator (satoreotide trizoxetan, OPS-202) and one with DOTA chelator (satoreotide tetraxetan, OPS-201) [228,229]. Satoreotide trizoxetan is currently radiolabeled with <sup>68</sup>Ga and used in PET imaging clinical trials (Figure 13) [230,231]. Satoreotide tetraxetan radiolabeled with <sup>177</sup>Lu has been evaluated in a therapeutic clinical trial [232]. First clinical results for this somatostatin antagonist theranostic pair seem to be promising with high sensitivity for neuroendocrine tumors and require further studies in larger patient population.

**Figure 13.** Comparison between [ <sup>68</sup>Ga]-OPS202 (**A**,**B**) and [ <sup>68</sup>Ga]-DOTATOC (**C**,**D**) PET/CT images of the same patient with ileal neuroendocrine tumours, showing bilobar liver metastases (from Rangger et al. [233]).

#### **5. Future Prospects**

α Regarding clinically established somatostatin analogs, the development of kit-based <sup>68</sup>Ga radiotracers, as well as cyclotron production of gallium-68 should improve their availability and worldwide dissemination. Further clinical translation of <sup>64</sup>Cu- and <sup>18</sup>F-based somatostatin SSAs could also represent an attractive alternative. For therapy, current research focuses on optimizing the dose received by the tumor while sparing healthy tissues. Fractionation, as well as combination of <sup>90</sup>Y and <sup>177</sup>Lu, have demonstrated their interest [168,234]. The same approach with other treatment modalities, such as external-beam radiotherapy or chemotherapy could enhance treatment response [235,236]. Targeted α-therapy also seems to hold promises and is currently attracting much interest, notably from the industry.

Recent developments showed a switch from agonist to antagonist derivatives, demonstrating higher efficacy. With the advent of new promising radionuclides and somatostatin analogs with better pharmacokinetic properties and binding profiles, the future looks bright for radiolabeled somatostatin analogs, expanding their use for wider indications, than just GEP-NETs. With peptide derivatives with improved targeting, tumors with lower SSTR expression might nonetheless be clinically relevant. In this context, as already demonstrated with some analogs, use of somatostatin-based radiopharmaceuticals might be of interest in pulmonary or hepatic cancers, warranting further studies. The development of bivalent radiotracers to target several receptors concomitantly expressed could be of interest to

improve targeting [237]. Similarly, improved detection and sensitivity could be achieved using bimodal agents [238]. Besides, the clinical success for radiolabeled somatostatin analogs both with diagnostic and therapeutic radionuclides paved the way for new promising peptide derivatives, such as bombesin, neurotensin, or CXCR4 ligands, and, in a similar way, PSMA ligands, for cancer theranostics [49,233,239,240].

**Author Contributions:** All authors contributed to the writing of the manuscript. All authors have read and agreed to the published version of the manuscript.

**Funding:** This work was partly supported with a funding from Ligue Contre le Cancer (R.E.), a grant from Ligue 22 Contre le Cancer, and Labex IRON (grant no. ANR-11-LABX-0018).

**Acknowledgments:** The authors thank Sophie Laffont for providing the [68Ga]-DOTATOC and [177Lu]-DOTATATE pictures.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

*Article*

## **Evaluation of Met-Val-Lys as a Renal Brush Border Enzyme-Cleavable Linker to Reduce Kidney Uptake of <sup>68</sup>Ga-Labeled DOTA-Conjugated Peptides and Peptidomimetics**

**Shreya Bendre <sup>1</sup> , Zhengxing Zhang <sup>1</sup> , Hsiou-Ting Kuo <sup>1</sup> , Julie Rousseau <sup>1</sup> , Chengcheng Zhang <sup>1</sup> , Helen Merkens <sup>1</sup> , Áron Roxin <sup>1</sup> , François Bénard 1,2,3 and Kuo-Shyan Lin 1,2,3,\***


#### Academic Editor: Krishan Kumar

Received: 29 July 2020; Accepted: 21 August 2020; Published: 25 August 2020

**Abstract:** High kidney uptake is a common feature of peptide-based radiopharmaceuticals, leading to reduced detection sensitivity for lesions adjacent to kidneys and lower maximum tolerated therapeutic dose. In this study, we evaluated if the Met-Val-Lys (MVK) linker could be used to lower kidney uptake of <sup>68</sup>Ga-labeled DOTA-conjugated peptides and peptidomimetics. A model compound, [68Ga]Ga-DOTA-AmBz-MVK(Ac)-OH (AmBz: aminomethylbenzoyl), and its derivative, [ <sup>68</sup>Ga]Ga-DOTA-AmBz-MVK(HTK01166)-OH, coupled with the PSMA (prostate-specific membrane antigen)-targeting motif of the previously reported HTK01166 were synthesized and evaluated to determine if they could be recognized and cleaved by the renal brush border enzymes. Additionally, positron emission tomography (PET) imaging, ex vivo biodistribution and in vivo stability studies were conducted in mice to evaluate their pharmacokinetics. [68Ga]Ga-DOTA-AmBz-MVK(Ac)-OH was effectively cleaved specifically by neutral endopeptidase (NEP) of renal brush border enzymes at the Met-Val amide bond, and the radio-metabolite [68Ga]Ga-DOTA-AmBz-Met-OH was rapidly excreted via the renal pathway with minimal kidney retention. [68Ga]Ga-DOTA-AmBz-MVK(HTK01166)-OH retained its PSMA-targeting capability and was also cleaved by NEP, although less effectively when compared to [68Ga]Ga-DOTA-AmBz-MVK(Ac)-OH. The kidney uptake of [ <sup>68</sup>Ga]Ga-DOTA-AmBz-MVK(HTK01166)-OH was 30% less compared to that of [68Ga]Ga-HTK01166. Our data demonstrated that derivatives of [68Ga]Ga-DOTA-AmBz-MVK-OH can be cleaved specifically by NEP, and therefore, MVK can be a promising cleavable linker for use to reduce kidney uptake of radiolabeled DOTA-conjugated peptides and peptidomimetics.

**Keywords:** radiopharmaceuticals; kidney uptake; cleavable linkers; neutral endopeptidase (NEP); renal brush border enzymes; prostate-specific membrane antigen (PSMA); cancer imaging and therapy

#### **1. Introduction**

The use of low molecular weight radiolabeled peptides and antibody fragments for applications in oncology is rapidly gaining momentum [1–5]. Such a site-directed radiation delivery involves targeting of certain specific receptors overexpressed on the surface of cancer cells for the purpose of targeted imaging and radionuclide therapy [6]. While these oncophilic molecules serve as biological targeting vectors, they commonly exhibit very high and sustained renal uptake [7]. This is caused by either a high renal expression of cancer markers targeted by these oncophilic molecules, megalin-cubilin mediated endocytosis and transcellular transport, or lysosomal proteolysis following glomerular filtration and renal reabsorption [7–9]. This reduces detection sensitivity for lesions adjacent to kidneys and lowers the maximum tolerated dose for radiotherapy.

Arano et al. reported an effective strategy to reduce kidney uptake of radiopharmaceuticals by incorporating specific cleavable linkages into them, with the Met-Val-Lys (MVK) sequence found to be the most effective thus far [1]. Use of this strategy is attributable to recognition and cleavage of MVK between Met-Val residues by a metalloendopeptidase enzyme called neutral endopeptidase or neprilysin (NEP) [10]. This enzyme is found in abundance on the renal brush border membrane lining, particularly of the proximal convoluted tubules [11]. Arano et al. exploited this renal brush border enzyme and designed a NOTA(1,4,7-triazacyclononane-1,4,7-triacetic acid)-conjugated <sup>67</sup>/68Ga-labeled antibody fragment bearing this MVK linker sequence to successfully lower renal uptake by 80% at 3 h post-injection (p.i.) without loss of tumor uptake [12].

By adopting this design strategy Zhang et al. recently reported the comparison of two radiolabeled Exendin 4 derivatives for imaging the expression of the glucagon-like peptide-1 receptor (GLP-1R) with positron emission tomography (PET) [13]. GLP-1R is highly expressed in insulinomas. However, the very high and sustained uptake of radiolabeled GLP-1R-targeting Exendin 4 derivatives in kidneys hinders the application of these tracers for detecting insulinomas. Compared with [68Ga]Ga-NOTA-Cys40-Leu14-Exendin 4, the derivative bearing the MVK linker ([68Ga]Ga-NOTA-MVK-Cys40-Leu14-Exendin 4) had similar tumor uptake values but only one third of kidney uptake at 2 h p.i., greatly enhancing the tumor-to-kidney contrast and detection sensitivity [13].

Despite successful application of this strategy to lower kidney uptake of radiolabeled peptides, the use of NOTA as a radiometal chelator unfortunately excludes application for using therapeutic isotopes like <sup>177</sup>Lu [14]. <sup>177</sup>Lu emits both β- and γ-radiation and is widely used as a theranostic pair with the positron-emitter <sup>68</sup>Ga. The preferred radiometal chelator for potential theranostic applications is 1,4,7,10-tetraazacyclododecane-1,4,7,10-tetraacetic acid (DOTA), a macrocyclic bifunctional chelating agent which forms stable complexes with a variety of radiometals including <sup>68</sup>Ga and <sup>177</sup>Lu.

In this study, we evaluated the MVK sequence as a cleavable tripeptide linker to reduce renal uptake of radiolabeled DOTA-conjugated peptides and peptidomimetics. We modified the original design, NOTA-MVK(Targeting vector)-OH, reported by Uehara et al. [12], and replaced the NOTA chelator with DOTA for potential radiolabeling with both imaging and radiotherapeutic isotopes (Figure 1). We also inserted an aminomethylbenzoyl (AmBz) group between DOTA and the MVK sequence to mimic the aminobenzyl group in the original design. The maleimide-thiol linkage at the Lys side chain was replaced with an amide linkage to enable the facile synthesis of the DOTA-conjugated peptides on solid phase.

**Figure 1.** DOTA-AmBz-MVK(Targeting vector)-OH inspired by the reported design of NOTA-MVK (Targeting vector)-OH. The amide bond between Met-Val (pointed by an arrow) is recognized and cleaved by NEP.

We first synthesized the model compound, [68Ga]Ga-DOTA-AmBz-MVK(Ac)-OH (Figure 2), and confirmed that it could be recognized by the renal brush border enzymes and cleaved at the Met-Val amide bond. We conducted PET imaging and in vivo stability studies in mice and confirmed that the expected radio-metabolite [68Ga]Ga-DOTA-AmBz-Met-OH was rapidly excreted through the urinary pathway with minimal uptake in kidneys. We then replaced the acetyl group of [68Ga]Ga-DOTA-AmBz-MVK(Ac)-OH with the pharmacophore of HTK01166 (Figure 2), a peptidomimetic targeting the prostate-specific membrane antigen (PSMA), which is highly expressed in prostate cancer [15]. The targeting pharmacophore of [68Ga]Ga-HTK01166 was selected as it was shown previously to have very high renal uptake (147 %ID/g, 1 h p.i.) in mice. Since Met is prone to oxidation to generate methionine sulfoxide (Met(O) or M(O)) [16] and the resulting M(O)VK sequence might not be recognized and cleaved by the renal brush border enzyme, we also synthesized [68Ga]Ga-DOTA-AmBz-M(O)VK(HTK01166)-OH for comparison. Here we report the syntheses of [68Ga]Ga-DOTA-AmBz-MVK(Ac)-OH, [68Ga]Ga-DOTA-AmBz-MVK(HTK01166)-OH and [68Ga]Ga-DOTA-AmBz-M(O)VK(HTK01166)-OH, and the results of enzyme assay, PET imaging, ex vivo biodistribution, and in vivo stability studies to evaluate the applicability of using the MVK linker to reduce the renal uptake of radiolabeled DOTA-conjugated peptides and peptidomimetics.

**Figure 2.** Chemical structures of (**A**) HTK01166, (**B**) DOTA-AmBz-MVK(Ac)-OH, (**C**) DOTA-AmBz-MVK (HTK01166)-OH, (**D**) DOTA-AmBz-Met-OH, and (**E**) DOTA-AmBz-M(O)VK(HTK01166)-OH.

#### **2. Results**

#### *2.1. Chemistry and Radiochemistry*

Synthesis of Fmoc-Lys(pentynoyl)-O*t*Bu (**2**) is shown in Scheme 1. Fmoc-Lys(pentynoyl)-OH (**1**) prepared from literature procedures [17] was reacted with 2,2,2-trichloroacetimidate (2.2 equiv) in CH2Cl2. After overnight incubation at room temperature and purification by flash column chromatography, the desired product **2** was obtained in 79% yield.

The DOTA-conjugated peptides and peptidomimetics including DOTA-AmBz-MVK(Ac)-OH, DOTA-AmBz-MVK(HTK01166)-OH, DOTA-AmBz-Met-OH, and DOTA-AmBz-M(O)VK(HTK01166)- OH (Figure 2) were assembled on solid phase (Schemes 2 and 3). After cleavage/deprotection with TFA and HPLC purification, these DOTA-conjugated peptides and peptidomimetics were obtained in 1–35% isolated yields. Their nonradioactive Ga-complexed standards were obtained by incubating the DOTA-conjugated peptides and peptidomimetics with excess GaCl<sup>3</sup> in acetate buffer (0.1 M, pH 4.2) at 80 ◦C. After HPLC purification, the nonradioactive Ga-complexed standards were obtained in 16–78% isolated yields. Detailed HPLC conditions and retention times for the purification of the DOTA-conjugated peptides and peptidomimetics and their nonradioactive Ga-complexed standards are provided in Supplemental Tables S1 and S2 (see the Supplemental Materials). The identities of the DOTA-conjugated peptides and peptidomimetics and their nonradioactive Ga-complexed standards were confirmed by MS analysis.

<sup>68</sup>Ga labeling of the DOTA-conjugated peptides and peptidomimetics was conducted in HEPES buffer (2 M, pH 5.0) with microwave heating for 1 min. After HPLC purification, [68Ga]Ga-DOTA-AmBz-MVK(Ac)-OH, [68Ga]Ga-DOTA-AmBz-MVK(HTK01166)-OH and [68Ga]Ga-DOTA-AmBz-M(O)VK(HTK01166)-OH were obtained in 40–78% decay-corrected radiochemical yields with >2 GBq/µmol molar activity and >93% radiochemical purity. Detailed HPLC conditions and retention times for the purification and quality control of the <sup>68</sup>Ga-labeled DOTA-conjugated peptides and peptidomimetics are provided in Supplemental Table S3 (see the Supplemental Materials).

**Scheme 1.** Synthesis of Fmoc-l-Lys(pentynoyl)-O*t*Bu (**2**).

● **Scheme 2.** Synthesis of (**A**) DOTA-AmBz-MVK(Ac)-OH and (**B**) DOTA-AmBz-Met-OH. •●= resin.

● **Scheme 3.** Synthesis of DOTA-AmBz-MVK(HTK01166)-OH and DOTA-AmBz-M(O)VK(HTK01166)- OH. • = resin.

#### *2.2. In Vitro Enzyme Assays*

In vitro enzyme assays revealed very efficient cleavage (>95%) of the <sup>68</sup>Ga-labeled DOTA-conjugated linker, [68Ga]Ga-DOTA-AmBz-MVK(Ac)-OH, with the expected fragment [68Ga]Ga-DOTA-AmBz-Met-OH as the dominant radio-metabolite (>90%) (Figure 3A). The presence of the NEP inhibitor phosphoramidon significantly inhibited cleavage of [68Ga]Ga-DOTA-AmBz-MVK(Ac)-OH, with 85% of the recovered radioactivity as the intact tracer (Figure 3B).

[ <sup>68</sup>Ga]Ga-DOTA-AmBz-MVK(HTK01166)-OH, a derivative of [68Ga]Ga-DOTA-AmBz-MVK(Ac)-OH by replacing the acetyl group with the PSMA-targeting motif Lys-urea-Glu and the lipophilic linker of HTK01166, was cleaved less effectively under the same assay conditions. After 1-h incubation, ~80% of [68Ga]Ga-DOTA-AmBz-MVK(HTK01166)-OH remained intact and ~16% of the recovered radioactivity was present as the expected radio-metabolite [68Ga]Ga-DOTA-AmBz-Met-OH (Figure 4A). An unidentified radio-metabolite with retention time at ~4.3 min accounted for ~4% of the recovered radioactivity. In the presence of phosphoramidon, the intact [68Ga]Ga-DOTA-AmBz-MVK(HTK01166)-OH fraction increased to ~95% and the formation of [68Ga]Ga-DOTA-AmBz-Met-OH was completely inhibited (Figure 4B).

[ <sup>68</sup>Ga]Ga-DOTA-AmBz-M(O)VK(HTK01166)-OH, the oxidized version of [68Ga]Ga-DOTA-AmBz-MVK(HTK01166)-OH was fairly stable under the same assay conditions with 94% and 99.5% remaining intact without and with the presence of phosphoramidon, respectively (Figure 5).

**Figure 3.** Radio-HPLC chromatograms of in vitro enzyme assay samples of [68Ga]Ga-DOTA-AmBz-MVK(Ac)-OH (**A**) without and (**B**) with the presence of phosphoramidon. HPLC conditions were 82/18 A/B at a flow rate of 2 mL/min; A: H2O containing 0.1% TFA; B: CH3CN containing 0.1% TFA; HPLC column: Luna C18, 5 µm particle size, 100 Å pore size, 250 × 4.6 mm.

**Figure 4.** Radio-HPLC chromatograms of in vitro enzyme assay samples of [68Ga]Ga-DOTA-AmBz-MVK(HTK01166)-OH (**A**) without and (**B**) with the presence of phosphoramidon. HPLC conditions were 72/28 A/B at a flow rate of 2 mL/min; A: H2O containing 0.1% TFA; B: CH3CN containing 0.1% TFA; HPLC column: Luna C18, 5 µm particle size, 100 Å pore size, 250 × 4.6 mm.

**Figure 5.** Radio-HPLC chromatograms of in vitro enzyme assay samples of [68Ga]Ga-DOTA-AmBz-M(O)VK(HTK01166)-OH (**A**) without and (**B**) with the presence of phosphoramidon. HPLC conditions were 76/24 A/B at a flow rate of 2 mL/min; A: H2O containing 0.1% TFA; B: CH3CN containing 0.1% TFA; HPLC column: Luna C18, 5 µm particle size, 100 Å pore size, 250 × 4.6 mm.

#### *2.3. PET*/*CT Imaging and Ex Vivo Biodistribution Studies*

The pharmacokinetics of [68Ga]Ga-DOTA-AmBz-MVK(Ac)-OH was first evaluated in mice via PET/CT imaging studies. As shown in Figure 6, radioactivity from the injected [68Ga]Ga-DOTA-AmBz-MVK(Ac)-OH was excreted rapidly from blood pool and all background organs/tissues predominately via the renal pathway. At 1 h p.i., only kidneys and urinary bladder were clearly visualized in PET images, with low radioactivity (<2.5 %ID/g) retained in kidneys.

**Figure 6.** A representative maximum-intensity-projection PET image of [68Ga]Ga-DOTA-AmBz-MVK(Ac)-OH showing its rapid excretion predominantly via the renal pathway with minimal kidney retention (<2.5 %ID/g). The range of color bar is 0–5 %ID/g. k: kidney; b: bladder.

Next, we conducted the PET/CT imaging study of [68Ga]Ga-DOTA-AmBz-MVK(HTK01166)-OH in mice bearing PSMA-expressing LNCaP tumor xenografts to evaluate its pharmacokinetics and PSMA-targeting capability. As shown in Figure 7A, [68Ga]Ga-DOTA-AmBz-MVK(HTK01166)-OH was excreted quickly from background organs/tissues predominately via the renal pathway. At 1 h p.i., only urinary bladder and the PSMA-expressing kidneys and LNCaP tumor were clearly visualized in the PET images (Figure 7A). Co-injection of the PSMA inhibitor, 2-PMPA (2-(phosphonomethyl)-pentanedioic acid), blocked most of the uptake into the tumor and kidneys (Figure 7B).

**Figure 7.** Representative maximum-intensity-projection PET images of [68Ga]Ga-DOTA-AmBz-MVK(HTK01166)-OH acquired at 1 h p.i. from LNCaP tumor-bearing mice (**A**) without and (**B**) with the co-injection of 2-PMPA (0.2 mg). The range of color bar is 0–5 %ID/g. t: tumor; k: kidney; b: bladder.

For comparison, the PET/CT imaging study was also conducted using [68Ga]Ga-DOTA-AmBz-M(O)VK(HTK01166)-OH, the sulfoxide analog of [68Ga]Ga-DOTA-AmBz-MVK(HTK01166)-OH. Similar to [68Ga]Ga-DOTA-AmBz-MVK(HTK01166)-OH, the excretion of [68Ga]Ga-DOTA-AmBz-M(O)VK(HTK01166)-OH was fast and predominately via the renal pathway (Figure 8A). At 1 h p.i., only urinary bladder, LNCaP tumor xenograft and kidneys were clearly visualized in PET images. Co-injection of 2-PMPA blocked most of the uptake in tumors and kidneys (Figure 8B), demonstrating the uptake was PSMA-mediated.

**Figure 8.** Representative maximum-intensity-projection PET/CT images of [68Ga]Ga-DOTA-AmBz-M(O)VK(HTK01166)-OH acquired at 1 h p.i. from LNCaP tumor-bearing mice (**A**) without and (**B**) with the co-injection of 2-PMPA (0.2 mg). The range of color bar is 0–5 %ID/g. t: tumor; k: kidney; b: bladder.

The ex vivo biodistribution studies were conducted for [68Ga]Ga-DOTA-AmBz-MVK(HTK01166)-OH and [68Ga]Ga-DOTA-AmBz-M(O)VK(HTK01166)-OH in LNCaP tumor-bearing mice, and the results are shown in Table 1. The ex vivo biodistribution data acquired at 1 h p.i. were consistent with the observations from their PET images (Figures 7A and 8A, and Table 1) with low background in blood pool and non-target tissues/organs, good uptake in LNCaP tumor xenografts, and high uptake in kidneys. [ <sup>68</sup>Ga]Ga-DOTA-AmBz-MVK(HTK01166)-OH and [68Ga]Ga-DOTA-AmBz-M(O)VK(HTK01166)-OH had similar uptake values (%ID/g) for the collected tissues/organs (blood: 0.55 ± 0.09 vs. 0.56 ± 0.13; pancreas: 0.60 ± 0.22 vs. 0.56 ± 0.24; spleen: 8.87 ± 2.63 vs. 11.5 ± 2.70; kidneys: 104 ± 12.2 vs. 89.7 ± 19.9; heart: 0.25 ± 0.12 vs. 0.23 ± 0.05; lung: 0.96 ± 0.12 vs. 0.88 ± 0.10; muscle: 0.21 ± 0.05 vs. 0.22 ± 0.05) and comparable tumor-to-background (blood, muscle and kidney) contrast ratios. Compared with the previously reported [68Ga]Ga-HTK01166 [15], [68Ga]Ga-DOTA-AmBz-M(O)VK (HTK01166)-OH showed 39% reduction in average kidney uptake (147 vs. 89.7 %ID/g, *p* = 0.025). [ <sup>68</sup>Ga]Ga-DOTA-AmBz-MVK(HTK01166)-OH also showed 30% reduction in average kidney uptake (147 vs. 104 %ID/g), but the difference is not statistically significant (*p* = 0.055).


**Table 1.** Biodistribution data and uptake ratios of <sup>68</sup>Ga-labeled PSMA-targeted tracers in LNCaP tumor-bearing mice acquired at 1 h p.i. (\*\*\* *p* < 0.001).

#### *2.4. Quantification of Radio-Metabolites in Blood and Urine*

Radio-metabolites of [68Ga]Ga-DOTA-AmBz-MVK(Ac)-OH and [68Ga]Ga-DOTA-AmBz-MVK (HTK01166)-OH in mouse blood and urine samples were analyzed by HPLC. The blood samples were collected at 5 min p.i. due to the fast excretion nature of the small radio-metabolites, whereas the urine samples were collected at 15 min p.i. to allow sufficient time for the accumulation of enough radio-metabolites for analysis.

As shown in Figure 9A, there were mainly unidentified polar metabolites (retention time 1.5–3.0 min) present in the blood samples, but no intact [68Ga]Ga-DOTA-AmBz-MVK(Ac)-OH and only minimal amounts of the expected metabolite [68Ga]Ga-DOTA-AmBz-Met-OH (<5%) was detected. On the contrary, >90% of the radioactivity presented in the urine samples was [ <sup>68</sup>Ga]Ga-DOTA-AmBz-Met-OH (Figure 9B).

Analysis of blood samples from mice injected with [68Ga]Ga-DOTA-AmBz-MVK(HTK01166)-OH showed only 9% remaining intact and 35% of the tracer was metabolized to the expected fragment [ <sup>68</sup>Ga]Ga-DOTA-AmBz-Met-OH (Figure 10A). However, a major unidentified radio-metabolite, accounting for 56% of the recovered radioactivity, was also observed. Analysis of the urine samples revealed that only 20% of the recovered radioactivity was presented as the intact tracer and the remaining 80% was the expected fragment [68Ga]Ga-DOTA-AmBz-Met-OH (Figure 10B).

**Figure 9.** Representative radio-HPLC chromatograms of [68Ga]Ga-DOTA-AmBz-MVK(Ac)-OH obtained from mouse (**A**) blood and (**B**) urine samples collected at 5 and 15 min p.i., respectively. HPLC conditions were 82/18 A/B at a flow rate of 2 mL/min; A: H2O containing 0.1% TFA; B: CH3CN containing 0.1% TFA; HPLC column: Luna C18, 5 µm particle size, 100 Å pore size, 250 × 4.6 mm.

**Figure 10.** Representative radio-HPLC chromatograms of [68Ga]Ga-DOTA-AmBz-MVK(HTK01166)-OH obtained from mouse (**A**) blood and (**B**) urine samples collected at 5 and 15 min p.i., respectively. HPLC conditions were 72/28 A/B at a flow rate of 2 mL/min; A: H2O containing 0.1% TFA; B: CH3CN containing 0.1% TFA; HPLC column: Luna C18, 5 µm particle size, 100 Å pore size, 250 × 4.6 mm.

ε

#### **3. Discussion**

The potential of NEP to recognize and cleave specific sequences has been exploited by various groups in the recent years to reduce kidney uptake of radiopharmaceuticals [1,13,18]. This is attributable to the abundant expression of NEP in the brush border membrane lining primarily of the proximal convoluted tubules of the juxtamedullary nephrons [11]. The mechanism underlying degradation of certain specific linker sequences has been well elucidated [1]. The type of amino acid in the radio-metabolite(s) as well as the radiometal chelate used plays a critical role in deciding the kidney residence time of the generated radio-metabolite(s).

NEP is a type-II integral membrane glycoprotein with metalloendopeptidase activity, and also presents an even better carboxydipeptidase activity when the two situations are possible [19]. To apply this strategy to reduce the kidney uptake of radiolabeled peptides and peptidomimetics, we modified the original design as shown in Figure 2. We followed the same design with the targeting vector conjugated to the cleavable linker at the Lys side chain. Such design preserves the free carboxylic group of Lys and enhances the cleavage of the MVK linker by NEP via its carboxydipeptidase activity.

We first synthesized the model tracer [68Ga]Ga-DOTA-AmBz-MVK(Ac)-OH with an acetyl group coupled to the ε-amino group of Lys to provide the amide linkage which would be present when a targeting vector is coupled to this linker. Enzyme assays using the brush border membrane vesicles (BBMVs) extracted from mouse kidneys confirmed that [68Ga]Ga-DOTA-AmBz-MVK(Ac)-OH can be efficiently cleaved (>95%) by renal brush border enzymes (Figure 3), and the expected [ <sup>68</sup>Ga]Ga-DOTA-AmBz-Met-OH was identified as the major radio-metabolite (>90%). Co-incubation with the NEP inhibitor phosphoramidon greatly enhanced the stability of [68Ga]Ga-DOTA-AmBz-MVK(Ac)-OH against renal brush border enzymes (~85% remaining intact), indicating the cleavage of [ <sup>68</sup>Ga]Ga-DOTA-AmBz-MVK(Ac)-OH into [68Ga]Ga-DOTA-AmBz-Met-OH was mediated by NEP. These data confirmed the success of our modifications on the original design (NOTA→DOTA, aminobenzyl→aminomethylbenzoyl, and maleimide-thiol linkage→amide linkage, Figure 1), and the resulting [68Ga]Ga-DOTA-AmBz-MVK(Ac)-OH was still recognized and cleaved by NEP at the same Met-Val amide bond.

We then conducted PET imaging and in vivo stability studies in mice to confirm that [ <sup>68</sup>Ga]Ga-DOTA-AmBz-MVK(Ac)-OH can be metabolized in vivo into [68Ga]Ga-DOTA-Met-OH, and low retention of [68Ga]Ga-DOTA-AmBz-Met-OH in kidneys. This is vital for the success of this modified strategy as it depends on the generation of the expected radio-metabolite [ <sup>68</sup>Ga]Ga-DOTA-AmBz-Met-OH and most importantly [68Ga]Ga-DOTA-AmBz-Met-OH needs to have low kidney retention too. As shown in Figure 6, [68Ga]Ga-DOTA-AmBz-MVK(Ac)-OH was excreted rapidly and predominately via the renal pathway with low kidney retention (<2.5 %ID/g at 1 h p.i.). The blood samples of [68Ga]Ga-DOTA-AmBz-MVK(Ac)-OH revealed no presence of the intact tracer and a small fraction of the expected radio-metabolite [68Ga]Ga-DOTA-AmBz-Met-OH (<5%, Figure 9A). This indicates that [68Ga]Ga-DOTA-AmBz-MVK(Ac)-OH was cleared rapidly from the blood pool either as its intact form or as radio-metabolite(s). Besides kidneys, NEP is also present in some tissues although in a much lower expression level. Therefore, the cleavage of [68Ga]Ga-DOTA-AmBz-MVK(Ac)-OH by NEP present in other tissues cannot be ruled out. Analysis of the urine samples revealed no presence of [68Ga]Ga-DOTA-AmBz-MVK(Ac)-OH and [68Ga]Ga-DOTA-AmBz-Met-OH was presented as the major radio-metabolite (>90%). These data suggest that [68Ga]Ga-DOTA-AmBz-Met-OH had low retention in kidneys and any intact [68Ga]Ga-DOTA-AmBz-MVK(Ac)-OH excreted through kidneys was metabolized presumably by renal brush border enzymes mainly into [68Ga]Ga-DOTA-AmBz-Met-OH.

After confirming the in vivo cleavage of [68Ga]Ga-DOTA-AmBz-MVK(Ac)-OH into [68Ga]Ga-DOTA-AmBz-Met-OH, and the low kidney retention of the [68Ga]Ga-DOTA-AmBz-Met-OH, we next tested this modified design with a PSMA-targeting vector coupled to the Lys side chain. PSMA has become a very promising imaging and therapeutic target for the management of prostate cancer. However, due to a high expression level of PSMA in kidneys, high and sustained renal uptake of PSMA-targeting radioligands are constantly observed, leading to suboptimal detection sensitivity for

lesions adjacent to kidneys, and concerns for renal toxicity when radiotherapeutic agents are used. The idea of incorporating the MVK cleavable linker to a radiolabeled DOTA-conjugated PSMA-targeting radioligand is to have the radioligand cleaved by the renal brush border enzymes when it is excreted through kidneys. Since the expected radiometal-complexed DOTA-AmBz-Met-OH is not retained in kidneys, the overall renal uptake of the PSMA-targeting radioligand containing the MVK cleavable linker will be reduced.

Enzyme assays revealed that unlike the instability of [68Ga]Ga-DOTA-AmBz-MVK(Ac)-OH against renal brush border enzymes, [68Ga]Ga-DOTA-AmBz-MVK(HTK01116)-OH was relatively stable with 80% remaining intact and 16% converted to the expected radio-metabolite [ <sup>68</sup>Ga]Ga-DOTA-AmBz-Met-OH under the same assay conditions (Figure 4). The enhanced stability could be due to the steric hindrance introduced by replacing the acetyl group with the much bulkier HTK01166 motif. Further enhancement in stability achieved by co-incubation with the NEP inhibitor phosphoramidon indicates that [68Ga]Ga-DOTA-AmBz-MVK(HTK01116)-OH was cleaved mainly by NEP into [68Ga]Ga-DOTA-AmBz-Met-OH.

PET imaging studies showed that [68Ga]Ga-DOTA-AmBz-MVK(HTK01166)-OH and its oxidized version [68Ga]Ga-DOTA-AmBz-M(O)VK(HTK01166)-OH had very similar distribution patterns: low background, high kidney retention and good tumor visualization (Figures 7 and 8). The observed PET images were consistent with the ex vivo biodistribution data showing comparable uptake for all collected organs/tissues (Table 1). The brain uptake (≤0.03 %ID/g) of both tracers was negligible indicating that they cannot freely cross the blood-brain barrier. Their fast blood clearance (~0.55 %ID/g at 1 h p.i.) suggests that the [68Ga]Ga-DOTA complex was stable as free <sup>68</sup>Ga would be captured by transferrins leading to prolonged retention in blood pool [20]. Minimal uptake (≤0.60 %ID/g) of both tracers in liver and intestines indicates that they were excreted mainly by the renal pathway. Higher uptake was observed in PSMA-expressing LNCaP tumors (~4 %ID/g), kidneys (90–104 %ID/g) and spleen (8.9–11.5 %ID/g) suggesting that the uptake of both tracers in these tissues was PSMA-mediated [15]. This was further confirmed by PET imaging studies as co-injection of the PSMA inhibitor, 2-PMPA (0.2 mg), reduced the tumor uptake of both tracers to the background level (Figures 7 and 8).

The average kidney uptake of [68Ga]Ga-DOTA-AmBz-MVK(HTK01166)-OH was lower than that of the previously reported [68Ga]Ga-HTK01166 (104 vs. 147 %ID/g), suggesting the insertion of the cleavable MVK linker might be a useful strategy to reduce kidney uptake of radiopharmaceuticals. However, although statistically not significant (*p* = 0.15), a slightly higher kidney retention was observed for mice injected with [68Ga]Ga-DOTA-AmBz-MVK(HTK01166)-OH (104 ± 12.2 %ID/g) than for mice injected with [68Ga]Ga-DOTA-AmBz-MV(O)K(HTK01166)-OH (89.7 ± 19.3 %ID/g). Since [68Ga]Ga-DOTA-AmBz-MVK(HTK01166)-OH contains a cleavable MVK linker, whereas [ <sup>68</sup>Ga]Ga-DOTA-AmBz-M(O)VK(HTK01166)-OH contains a non-cleavable M(O)VK linker, we would expect a much lower kidney uptake from [68Ga]Ga-DOTA-AmBz-MVK(HTK01166)-OH. These discrepant data suggest that there might be some unknown mechanism(s) causing the higher kidney retention of [68Ga]Ga-DOTA-AmBz-MVK(HTK01166)-OH. In addition, we noticed that while the kidney retention (1.3 %ID/g) in the mouse injected with [68Ga]Ga-DOTA-AmBz-M(O)VK(HTK01166)-OH and 2-PMPA was minimal (Figure 8B), there was significantly higher kidney uptake (4.2 %ID/g) in the mouse injected with [68Ga]Ga-DOTA-AmBz-MVK(HTK01166)-OH and 2-PMPA (Figure 7B).

The retained kidney uptake was unlikely to have resulted from the intact [68Ga]Ga-DOTA-AmBz-MVK(HTK01166)-OH, as the mouse was co-injected with excess 2-PMPA to block PSMA. The retained uptake was unlikely to have resulted from the expected radio-metabolite [68Ga]Ga-DOTA-AmBz-Met-OH either, as we have shown that no significant kidney retention was observed in the mouse injected with [68Ga]Ga-DOTA-AmBz-MVK(Ac)-OH (Figure 6). The only possibility for the kidney retention would be unidentified radio-metabolite(s), which cannot bind PSMA but can be retained in kidneys. Therefore, the in vivo stability study was further conducted for [68Ga]Ga-DOTA-AmBz-MVK(HTK01166)-OH to discern the cause of its higher kidney retention in mice.

As shown in Figure 10A, unlike the good stability observed in enzyme assays, [ <sup>68</sup>Ga]Ga-DOTA-AmBz-MVK(HTK01166)-OH was rapidly metabolized in vivo with only 9% of the tracer remaining intact in the blood at 5 min p.i. However, the expected radio-metabolite [ <sup>68</sup>Ga]Ga-DOTA-AmBz-Met-OH was accounted for only 35% of the recovered radioactivity, while 56% consists of an unidentified radio-metabolite. Interestingly, in urine samples, only [68Ga]Ga-DOTA-AmBz-MVK(HTK01166)-OH (20%) and [68Ga]Ga-DOTA-AmBz-Met-OH (80%) were detected, and there was no presence of the unidentified radio-metabolite (Figure 10B). The absence of the unidentified radio-metabolite in urine samples suggests that it might be retained in kidneys. This would explain the higher kidney uptake of [68Ga]Ga-DOTA-AmBz-MVK(HTK01166)-OH than [ <sup>68</sup>Ga]Ga-DOTA-AmBz-M(O)VK(HTK01166)-OH (Table 1), as well as its higher kidney retention when co-injected with 2-PMPA (Figures 7 and 8). This would also explain the insufficient reduction (~30%) in kidney uptake when compared [68Ga]Ga-DOTA-AmBz-MVK(HTK01166)-OH and [ <sup>68</sup>Ga]Ga-HTK01166 [15], which is far less than the ~80% and ~67% reduction in kidney uptake reported by Uehara et al. [12] and Zhang et al. [13], respectively, using the radiolabeled NOTA-MVK(Targeting vector)-OH design. The insufficient reduction reported here using the DOTA-AmBz-MVK(Targeting vector)-OH design is likely caused by the unidentified radio-metabolite that could be trapped in kidneys. Therefore, further optimization of the design of DOTA-conjugated cleavable linkers should avoid the generation of radio-metabolites that could be trapped in kidneys and cause high and sustained kidney uptake.

The identity of the unidentified fragment remains unknown (Figure 10). This is because the core structure of HTK01166 contains no amide bonds formed by two natural amino acids. Moreover, the data from our [68Ga]Ga-DOTA-AmBz-MVK(Ac)-OH, [68Ga]Ga-DOTA-AmBz-M(O)VK(HTK01166)-OH and the previously reported [68Ga]Ga-NOTA- MVK-conjugated antibody fragment [12] and Exendin 4 [13], did not suggest the cleavage of the Val-Lys amide bond. Therefore, we did not expect cleavage of [68Ga]Ga-DOTA-AmBz-MVK(HTK01166)-OH at locations other than the Met-Val amide bond, and more investigations are needed to verify the identity of this unknown radio-metabolite.

To conclude, we showed that replacing NOTA and the aminobenzyl group in the reported NOTA-MVK linker with DOTA and AmBz, respectively, generated the model compound [ <sup>68</sup>Ga]Ga-DOTA-AmBz-MVK(Ac)-OH, which was still recognized and specifically cleaved at the Met-Val amide bond by NEP. Coupling a bulkier PSMA-targeting vector to the side chain of Lys in DOTA-AmBz-MVK enhanced its stability against NEP, but possibly also rendered its vulnerability against other enzyme(s) as evident by the formation of an unidentified radio-metabolite that could be retained in kidneys. Nevertheless, the renal uptake of the resulting [68Ga]Ga-DOTA-AmBz-MVK(HTK01166)-OH was still lower than that of [68Ga]Ga-HTK01166. These data demonstrated that MVK could be a promising cleavable linker for use to reduce renal uptake of radiolabeled DOTA-conjugated tumor-targeting peptides and peptidomimetics. This strategy can be used to enhance detection sensitivity of the imaging agents for lesions adjacent to kidneys, and improve the tumor-to-kidney absorbed dose ratio for the radiotherapeutic agents.

#### **4. Materials and Methods**

#### *4.1. General Methods*

Fmoc-l-Lys(pentynoyl)-OH (**1**) was synthesized according to the literature procedures [17]. Brush border membrane vesicles (BBMVs) were extracted from mouse kidneys following literature procedures [21]. All other chemicals were procured from commercial sources and used without further purification. All peptides and peptidomimetics were synthesized either on an AAPPTec (Louisville, KY, USA) Endeavor 90 peptide synthesizer or a CEM (Matthews, NC, USA) Liberty Blue™ automated microwave peptide synthesizer. Purification and quality control of radiolabeling precursor, nonradioactive Ga-complexed standards and <sup>68</sup>Ga-labeled peptides and peptidomimetics were performed on Agilent (Santa Clara, CA, USA) HPLC systems equipped with a model 1200

quaternary pump, a model 1200 UV absorbance detector (set at 220 nm), and a Bioscan (Washington, DC, USA) NaI scintillation detector. The operation of Agilent HPLC systems was controlled using the Agilent ChemStation software. HPLC columns used were a semipreparative column (Luna C18, 5 µm particle size, 100 Å pore size, 250 × 10 mm) and an analytical column (Luna C18, 5 µm particle size, 100 Å pore size, 250 × 4.6 mm) from Phenomenex (Torrance, CA, USA). The HPLC solvents were A: H2O containing 0.1% TFA; B: CH3CN containing 0.1% TFA. The collected HPLC eluates containing the desired peptides were lyophilized using a Labconco (Kansas City, MO, USA) FreeZone 4.5 Plus freeze drier. <sup>1</sup>H-NMR spectrum was acquired using an AVANCE Bruker 400 MHz NMR spectrometer equipped with BBI probe with Z gradients. Mass analyses were performed using an AB SCIEX (Framingham, MA, USA) 4000 QTRAP mass spectrometer system with an ESI ion source. C18 Sep-Pak cartridges (1 cm<sup>3</sup> , 50 mg) were obtained from Waters (Milford, MA, USA). <sup>68</sup>Ga was eluted from an iThemba Laboratories (Somerset West, South Africa) generator and purified according to the previously published procedures using a DGA resin column from Eichrom Technologies LLC (Lisle, IL, USA) [22,23]. Radioactivity of <sup>68</sup>Ga-labeled peptides and peptidomimetics was measured using a Capintec (Ramsey, NJ, USA) CRC-25R/W dose calibrator. PET/CT imaging was performed using a Siemens Inveon (Knoxville, TN, USA) micro PET/CT scanner. The radioactivity of mouse tissues collected from biodistribution studies was counted using a PerkinElmer (Waltham, MA, USA) Wizard2 2480 automatic gamma counter.

#### *4.2. Synthesis of Fmoc-*l*-Lys(pentynoyl)-OtBu (2)*

Fmoc-l-Lys(pentynoyl)-OH (**1**) (2.05 g, 4.6 mmol) in CH2Cl<sup>2</sup> (20 mL) was added *t*-butyl 2,2,2-trichloroacetimidate (2.18 g, 10 mmol). The resulting mixture was stirred at RT for 22 h, and purified by flash column chromatography eluted with 1:1 ethyl acetate/hexane to 100% ethyl acetate. The desired product **2** was obtained as a white solid (1.82 g, 79%). <sup>1</sup>H-NMR (400 MHz, CDCl3) δ ppm: 1.39–2.08 (m, 7H), 1.47 (s, 9H), 2.37 (t, *J* = 7.0 Hz, 2H), 2.52 (dt, *J* = 2.2, 7.0 Hz, 2H). 3.28 (m, 2H), 4.22 (t, *J* = 7.0 Hz, 2H), 4.39 (m, 2H), 5.42 (d, *J* = 8.0 Hz, 1H), 5.90 (bs, 1H), 7.31 (dt, *J* = 1.0, 7.4 Hz, 2H), 7.40 (t, *J* = 7.5 Hz, 2H), 7.60 (d, *J* = 7.4 Hz, 2H), 7.77 (d, *J* = 7.5 Hz, 2H). ESI-MS: calculated [M + H]<sup>+</sup> for Fmoc-l-Lys(pentynoyl)-OtBu C30H36N2O<sup>5</sup> 505.3; found 506.0.

#### *4.3. Synthesis of DOTA-Conjugated Precursors*

#### 4.3.1. Synthesis of DOTA-AmBz-MVK(Ac)-OH

For synthesizing the DOTA-conjugated linker DOTA-AmBz-MVK(Ac)-OH, Fmoc-Lys(Mtt) wang resin (0.05 mmol scale, 0.5–0.8 mmol/g loading) was first swollen using DMF. The Mtt protecting group was removed using 2% TFA in DCM for 30 min (5 min, 6 times) and neutralized using DIEA in DMF. The free amine was acetylated using acetic anhydride (20 equiv)/DIEA (20 equiv)/DMF. Fmoc on Lys was then deprotected using 20% piperidine in DMF and coupled with Fmoc-Val-OH (5 equiv) in the presence of activators HATU/HOAt (5 equiv) and DIEA (10 equiv) in DMF. Further elongation of the peptide chain was carried out by repeating the Fmoc deprotection and coupling steps. Fmoc-Met-OH, Fmoc-(4-aminomethyl)benzoic acid, and finally DOTA-tris(*t*-bu)ester (tri-*t*-butyl 1,4,7,10-tetraazacyclododecane-1,4,7,10-tetraacetate) were coupled in that order. The DOTA conjugated peptide was cleaved off the resin and simultaneously deprotected using TFA:H2O:triisopropylsilane (TIS):2,2′ -(ethylenedioxy)diethanethiol (DODT) for 1.5 h at RT, in a 92.5:2.5:2.5:2.5 ratio. The cleaved peptide was filtered and precipitated with cold diethyl ether before purification using the semipreparative HPLC column. HPLC conditions were 84/16 A/B at a flow rate of 4.5 mL/min. The retention time of DOTA-AmBz-MVK(Ac)-OH was 14.1 min. Eluates containing the desired peptide were collected, pooled, and lyophilized. The isolated yield was 12%. ESI-MS: calculated [M + H]<sup>+</sup> for DOTA-AmBz-MVK(Ac)-OH C42H67N9O13S 938.5; found 938.4.

#### 4.3.2. Synthesis of DOTA-AmBz-MVK(HTK01166)-OH

For synthesizing the PSMA-targeting DOTA-AmBz-MVK(HTK01166)-OH the PSMA-targeting motif (Lys-urea-Glu) and the lipophilic linker ((2-indanyl)-Gly-tranexamic acid) of HTK01166 were first assembled on solid phase following our previously published procedures [15]. Briefly, Fmoc-Lys(ivDde) wang resin (0.1 mmol scale, 0.58 mmol/g loading) was swollen using DMF. The isocyanate derivative of di-*tert*-butyl ester of glutamate (5 equiv) was prepared according to literature procedures [24]. The isocyante derivative was then added to Fmoc-deprotected Lys(ivDde) wang resin and the reaction mixture was allowed to shake overnight. Next, the ivDde protecting group was removed using 2% hydrazine in DMF (5 mL, five times, 5 min). Subsequent couplings of Fmoc-(2-indanyl)-Gly-OH, Fmoc-tranexamic acid and 2-azidoacetic acid were conducted using standard Fmoc chemistry. Next, Fmoc-Lys(pentynoyl)-O*t*Bu (**2**), (0.5 mmol) was clicked onto the azido group using CuSO<sup>4</sup> (0.05 mmol) and ascorbic acid (0.25 mmol). After the click reaction, further peptide elongation was continued using Fmoc-Val-OH, Fmoc-Met-OH, Fmoc-(4-aminomethyl)benzoic acid, and finally, DOTA-tris(*t*-bu)ester. Peptide cleavage and deprotection was performed using TFA:H2O:TIS:DODT for 1.5 h at RT in a 92.5:2.5:2.5:2.5 ratio. The cleaved peptide was filtered and precipitated with cold diethyl ether before purification using the semipreparative HPLC column. HPLC conditions were 73/27 A/B at a flow rate of 4.5 mL/min. The retention time of DOTA-AmBz-MVK(HTK01166)-OH was 12.6 min. Eluates containing the desired peptide were collected, pooled, and lyophilized. The isolated yield was 1%. ESI-MS: calculated [M + H]<sup>+</sup> for DOTA-AmBz-MVK(HTK01166)-OH C78H115N17O23S 1690.8; found 1690.9.

#### 4.3.3. Synthesis of DOTA-AmBz-M(O)VK(HTK01166)-OH

DOTA-AmBz-M(O)VK(HTK01166)-OH, the oxidized (sulfoxide) version of DOTA-AmBz-MVK (HTK01166)-OH was synthesized as a control using a similar synthetic and purification strategy as for DOTA-AmBz-MVK(HTK01166)-OH. For this purpose, Fmoc-Met-OH was replaced with Fmoc-Met(O)-OH. The HPLC conditions were 74/26 A/B at a flow rate of 4.5 mL/min. The retention time and isolated yield were 8.3 min and 2%, respectively. ESI-MS: calculated [M + H]<sup>+</sup> for DOTA-AmBz-M(O)VK(HTK01166)-OH C78H115N17O24S 1706.8; found 1706.9.

#### 4.3.4. Synthesis of DOTA-AmBz-Met-OH

For synthesizing the expected cleaved fragment DOTA-AmBz-Met-OH, ClTCP(Cl)ProTide resin (0.1 mmol scale, 0.48 mmol/g loading) was swollen in DMF. Fmoc-Met-OH (5 equiv) in DMF containing 1 M DIEA/0.125M KI was coupled to the resin. After elongation with Fmoc-(4-aminomethyl)benzoic acid and DOTA-tris(*t*-bu)ester, the peptide was cleaved off the resin using TFA:H2O:TIS:DODT for 1.5 h at RT in a 92.5:2.5:2.5:2.5 ratio and purified using the semipreparative HPLC column. The HPLC conditions were 83/17 A/B at a flow rate of 4.5 mL/min. The retention time was 10.2 min, and the isolated yield was 35%. ESI-MS: calculated [M + H]<sup>+</sup> for DOTA-AmBz-Met-OH C29H44N6O10S 669.3; found 669.0.

#### *4.4. Synthesis of Nonradioactive Ga-Complexed Standards*

Ga-DOTA-AmBz-MVK(Ac)-OH, Ga-DOTA-AmBz-Met-OH, Ga-DOTA-AmBz-MVK(HTK01166)-OH and Ga-DOTA-AmBz-M(O)VK(HTK01166)-OH were prepared by incubating the DOTA-conjugated peptides and peptidomimetics with GaCl<sup>3</sup> (5 equiv) in NaOAc buffer (0.1 M, 300–500 µL, pH 4.2) at 80 ◦C for 15 min. The reaction mixtures were directly purified using the semipreparative HPLC column. The eluates containing the desired product were collected, pooled, and lyophilized.

For Ga-DOTA-AmBz-MVK(Ac)-OH the HPLC conditions were 82/18 A/B at a flow rate of 4.5 mL/min. The retention time and isolated yield were 11.2 min and 78%, respectively. ESI-MS: calculated [M + H]<sup>+</sup> for Ga-DOTA-AmBz-MVK(Ac)-OH C42GaH65N9O13S 1005.4; found 1004.4.

For Ga-DOTA-AmBz-MVK(HTK01166)-OH the HPLC conditions were 74/26 A/B at a flow rate of 4.5 mL/min. The retention time and isolated yield were 8.0 min and 16%, respectively. ESI-MS: calculated [M + H]<sup>+</sup> for Ga-DOTA-AmBz-MVK(HTK01166)-OH C78GaH113N17O23S 1757.7; found 1758.4.

For Ga-DOTA-AmBz-M(O)VK(HTK01166)-OH the HPLC conditions were 74/26 A/B at a flow rate of 4.5 mL/min. The retention time and isolated yield were 10.8 min and 35%, respectively. ESI-MS: calculated [M + 2H]2<sup>+</sup> for Ga-DOTA-AmBz-M(O)VK(HTK01166)-OH C78GaH113N17O24S 887.4; found 887.5.

For Ga-DOTA-AmBz-Met-OH, the HPLC conditions were 83/17 at a flow rate of 4.5 mL/min. The retention time and the isolated yield were 10.5 min and 72%, respectively. ESI-MS: calculated [M + H]<sup>+</sup> for Ga-DOTA-AmBz-Met-OH C29GaH42N6O10S 736.2; found 736.1.

#### *4.5. Synthesis of <sup>68</sup>Ga-Labeled Peptides and Peptidomimetics*

Purification of <sup>68</sup>Ga eluated from <sup>68</sup>Ge/ <sup>68</sup>Ga generator and labeling experiments were performed following our previously published procedures [22,23]. Purified <sup>68</sup>Ga in 0.5 mL of water was added into a 4-mL glass vial preloaded with 0.7 mL of HEPES buffer (2 M, pH 5.0) and 25 µg of the precursor. The radiolabeling reaction was carried out under microwave heating for 1 min. The reaction mixtures were purified using the semipreparative HPLC column and eluted with 84/16 A/B for [ <sup>68</sup>Ga]Ga-DOTA-AmBz-MVK(Ac)-OH, 74/26 A/B for [68Ga]Ga-DOTA-AmBz-MVK(HTK01166)-OH, and 23% CH3CN and 0.1% HCOOH in H2O for [68Ga]Ga-DOTA-AmBz-M(O)VK(HTK01166)-OH at a flow rate of 4.5 mL/min. The retention times for [68Ga]Ga-DOTA-AmBz-MVK(Ac)-OH, [ <sup>68</sup>Ga]Ga-DOTA-AmBz-MVK(HTK01166)-OH and [68Ga]Ga-DOTA-AmBz-M(O)VK(HTK01166)-OH were 24.3, 26.1 and 15.7 min, respectively. The eluate fraction containing the radiolabeled product was collected, diluted with water (50 mL), and passed through a C18 Sep-Pak cartridge that was prewashed with ethanol (10 mL) and water (10 mL). After washing the C18 Sep-Pak cartridge with water (10 mL), the <sup>68</sup>Ga-labeled product was eluted off the cartridge with ethanol (0.4 mL) and diluted with saline for all the in vitro enzyme assays, PET imaging, ex vivo biodistribution and in vivo stability studies.

#### *4.6. In Vitro Enzyme Assay*

The enzymatic recognition of synthesized peptides and peptidomimetics was determined by incubating the <sup>68</sup>Ga-labeled peptides and peptidomimetics with the extracted BBMVs at 37 ◦C for 1 h. For the assay, aliquots (25 µL) of the enzyme solution (1.52 mg/mL) and enzyme buffer (250 mM NaCl, 57.5 mM Tris-base; adjusted to a pH 7.5) were mixed in a 96-well clear bottom plate and incubated at 37 ◦C for 10 min. The mixtures also contained 100 ppm ascorbic acid to prevent oxidation of the Met residue in the tested peptides and peptidomimetics during the assay. The radiolabeled peptide (50 µL, ~3.7 MBq) was then added to the test well. The control well contained phosphoramidon, a potent NEP inhibitor, at a final concentration of 1 mmol/L in addition to the contents of the test well. After 1 h incubation, all reactions were quenched using equal volume of CH3CN and centrifuged at 13,000 rpm for 10 min. The resulting supernatant was collected and analyzed using the analytical HPLC column to identify and quantify radio-metabolite(s). The assay was performed in duplicates.

#### *4.7. Cell Culture*

Human prostate cancer LNCaP cells obtained from ATCC (Manassas, VA, USA) were cultured in RPMI 1640 medium, supplemented with 10% FBS, penicillin (100 U/mL), and streptomycin (100 µg/mL) at 37 ◦C in a Panasonic Healthcare (Tokyo, Japan) MCO-19AIC humidified incubator containing 5% CO2. At 80−90% confluency, cells were washed with sterile phosphate-buffered saline, trypsinized and pooled before they were manually counted using a Bal Supply (Sylvania, OH, USA) 202C counter.

#### *4.8. PET*/*CT Imaging and Ex Vivo Biodistribution in Tumor-Bearing Mice*

All imaging and biodistribution studies were performed using male NOD-*scid* IL2Rgnull (NSG) mice and conducted according to the guidelines established by the Canadian Council on Animal

Care and approved by Animal Ethics Committee of the University of British Columbia. For tumor inoculations, mice were anesthetized by inhalation with 2% isoflurane in oxygen and implanted subcutaneously with 5 × 10<sup>6</sup> LNCaP cells below the left shoulder. Imaging and biodistribution studies were performed only after tumors grew to 5−8 mm in diameter over a period of 5−7 weeks.

For PET/CT imaging studies, ~3−6 MBq of the <sup>68</sup>Ga-labeled tracer was injected through the tail vein. For the blocking study, 2-PMPA (0.2 mg) was co-injected with the tracer. Mice were allowed to recover and roam freely in the cages after injecting the tracer. At 45 min p.i., mice were sedated again and positioned on the scanner. First, a 10 min CT scan was conducted for localization and attenuation correction for reconstruction of PET images, before a 10 min PET image was acquired. Heating pads were used during the entire procedure to keep the mice warm.

For ex vivo biodistribution studies, mice were injected with ~1.5−3 MBq of the <sup>68</sup>Ga-labeled tracer. At 1 h p.i., mice were euthanized, blood was drawn from heart, and organs/tissues of interest were collected, rinsed with PBS, blotted dry, weighed, and counted using an automated gamma counter. The uptake in each organ/tissue was normalized to the injected dose and expressed as the percentage of the injected dose per gram of tissue (%ID/g).

#### *4.9. Quantification of Radio-Metabolites in Blood and Urine*

Male NSG mice were injected with 3–17 MBq of the <sup>68</sup>Ga-labeled peptide. For blood profiling, mice were anesthetized with 2% isoflurane in O<sup>2</sup> and euthanized by CO<sup>2</sup> inhalation at 5 min p.i. Blood draw was then performed by cardiac puncture and blood was collected in an eppendorf tube with equal volume of CH3CN. Each tube was then centrifuged at RT for 10–15 min and the resulting supernatant was collected and analyzed using the analytical HPLC column to identify and quantify radio-metabolite(s) in blood. For the purpose of urine profiling, urine was collected after euthanizing the mice at 15 min p.i. The urine samples were also collected and analyzed using the analytical HPLC column to identify and quantify radio-metabolite(s) in urine.

#### *4.10. Statistical Analysis*

Statistical analyses were performed by Student's *t*-test using the Microsoft (Redmond, WA, USA) Excel software. The unpaired, two-tailed test was used to compare tissue uptake and tumor-tobackground (muscle, blood and kidney) contrast ratios of [68Ga]Ga-DOTA-AmBz-MVK(HTK01166)-OH and [68Ga]Ga-DOTA-AmBz-M(O)VK(HTK01166)-OH. The unpaired, one-tailed test was used to compare kidney uptake of [68Ga]Ga-DOTA-AmBz-MVK(HTK01166)-OH (or [68Ga]Ga-DOTA-AmBz-M(O)VK(HTK01166)-OH) with that of the previously reported [68Ga]Ga-HTK01166 [15]. The difference was considered statistically significant when the *p* value was <0.05.

**Supplementary Materials:** The following are available online at http://www.mdpi.com/1420-3049/25/17/3854/s1, Table S1: HPLC conditions and retention times for the purification of DOTA-conjugated peptides and peptidomimetics using the semipreparative column (Luna C18, 5 µm particle size, 100 Å pore size, 250 × 10 mm), Table S2: HPLC conditions and retention times for the purification of nonradioactive Ga-complexed DOTA-conjugated peptides and peptidomimetics using the semipreparative column (Luna C18, 5 µm particle size, 100 Å pore size, 250 × 10 mm), and Table S3: HPLC conditions and retention times for the purification/QC of 68Ga-labeled DOTA-conjugated peptides and peptidomimetics using the semipreparative column—Luna C18, 5 µm particle size, 100 Å pore size, 250 × 10 mm; the analytical (QC) column—Luna C18, 5 µm particle size, 100 Å pore size, 250 × 4.6 mm.

**Author Contributions:** Conceptualization, K.-S.L.; methodology, S.B., Z.Z. and K.-S.L.; formal analysis, S.B., Z.Z. and K.-S.L.; investigation, S.B., Z.Z., H.-T.K., J.R., C.Z., Á.R. and H.M.; resources, F.B. and K.-S.L.; data curation, S.B., Z.Z., and C.Z.; writing—original draft preparation, S.B., Z.Z., and K.-S.L.; writing—review and editing, H.-T.K., J.R., C.Z., Z.Z., Á.R., H.M., and F.B.; supervision, K.-S.L. and F.B.; project administration, H.M.; funding acquisition, K.-S.L. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was funded by the Canadian Institutes of Health Research (grant number PJT-162243), Pancreas Centre BC, BC Cancer Foundation and VGH Foundation.

**Acknowledgments:** The authors would like to thank Nadine Colpo for her technical assistance.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


**Sample Availability:** Samples of all the compounds are available from the authors.

© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

*Article*

## **Radiochemical Synthesis and Evaluation of Novel Radioconjugates of Neurokinin 1 Receptor Antagonist Aprepitant Dedicated for NK1R-Positive Tumors**

**Paweł K. Halik 1,\* , Piotr F. J. Lipi ´nski <sup>2</sup> , Joanna Matali ´nska <sup>2</sup> , Przemysław Ko ´zmi ´nski <sup>1</sup> , Aleksandra Misicka <sup>2</sup> and Ewa Gniazdowska <sup>1</sup>**


Academic Editor: Krishan Kumar

Received: 31 July 2020; Accepted: 15 August 2020; Published: 18 August 2020

**Abstract:** Aprepitant, a lipophilic and small molecular representative of neurokinin 1 receptor antagonists, is known for its anti-proliferative activity on numerous cancer cell lines that are sensitive to Substance P mitogen action. In the presented research, we developed two novel structural modifications of aprepitant to create aprepitant conjugates with different radionuclide chelators. All of them were radiolabeled with <sup>68</sup>Ga and <sup>177</sup>Lu radionuclides and evaluated in terms of their lipophilicity and stability in human serum. Furthermore, fully stable conjugates were examined in molecular modelling with a human neurokinin 1 receptor structure and in a competitive radioligand binding assay using rat brain homogenates in comparison to the aprepitant molecule. This initial research is in the conceptual stage to give potential theranostic-like radiopharmaceutical pairs for the imaging and therapy of neurokinin 1 receptor-overexpressing cancers.

**Keywords:** aprepitant; radiopharmaceuticals; neurokinin 1 receptor antagonist; radionuclide chelators

#### **1. Introduction**

The knowledge of a suitable molecular target and its specificity for a given pathology is a necessary condition in a targeted radionuclide therapy approach. Many malignant tumors possess an infiltrating character with no defined margins or spread out metastases around the whole body. Only the selective binding of a radiopharmaceutical to a molecular target allows for the reliable imaging or safe ablation of cancer lesions with minimal side effects.

Neurokinin 1 receptor (NK1R; tachykinin 1 receptor) is a well-known G protein-coupled receptor for neuropeptide Substance P (SP) and a promising system for an anticancer therapeutic molecular target [1,2]. The activation of the NK1R by its endogenous ligand creates significant proliferative impulses for tumor cells promoting growth and development, including angiogenesis and metastasis. At the same time, the frequent formation of SP-NK1R complexes stimulate the cellular up-regulation of NK1R on tumor cell surfaces [3], thus providing an even greater cell sensitivity for the mitogen action of SP. On the other hand, the blockage of SP action by using antagonists of NK1R on SP-sensitive tumor cells can selectively induce an anti-tumor effect through the mechanism of cell apoptosis [4,5].

Antagonists of NK1R are a very diverse and numerous group of compounds, though clinical applications have only been found for four compounds. They are applied to the prevention of nausea and vomiting induced by chemotherapy or surgical complications [2,6]. One of the best known and

‐

widely studied compounds in this group is aprepitant (APT; Figure 1)—a lipophilic and low molecular weight morpholine derivative with a high and selective affinity for NK1R. APT possesses anti-tumor activity, as has been determined in many cancer cell lines [5,7–12]. Moreover, the phenomenon of the synergism of the anti-tumor activity of NK1R antagonists with an inhibitory effect on the cancer cell growth of other agents has been confirmed [4]. It has been shown in vitro that the application of microtubule destabilizing agents in combination with antagonists of NK1R possess synergism in apoptotic effect in human glioblastoma, bladder, cervical and breast cancer cells [13]. More remarkable cytotoxic synergism has been proven in a combination of aprepitant and ritonavir (an antiretroviral agent) in the human glioblastoma GAMG cell line [14]. The application of these two drugs with temozolomide, an alkylating chemotherapeutic used clinically to treat glioblastoma, gives an even stronger synergistic effect. ‐ ‐ ‐ ‐

‐

‐ ‐

**Aprepitant (APT)**

**Figure 1.** Structure of aprepitant with its key elements marked.

‐ ‐ ‐ ‐ What is most relevant is that APT is a fairly safe drug with a known pharmacological profile, with tolerability similar to placebo- and dose- related action. This could be shown by the fact that aprepitant's half-maximal inhibitory concentration (IC50) value determined for the human embryonic kidney (HEK) 293 cell line (a low expression NK1R control) is higher than the aprepitant IC<sup>100</sup> values determined for numerous tumor cell lines overexpressing NK1R [15]. For the reasons described above, APT's structure is an interesting scaffold for creating conjugates for carrying radionuclides to NK1R-positive tumors.

By looking at aprepitant in terms of molecular structure, it can be seen that the compound (Figure 1) consists of a morpholine core decorated by three 'arms,' which are:


‐ ‐ In the course of extensive structure-activity studies on NK1R antagonists [16–18], it has been established that the first two features (in particular: the distance and mutual positioning of two aromatic rings) are critical for high affinity and, therefore, for NK1R antagonism. On the other hand, the third element, triazolinone ring, can be, at least in some cases, safely modified without a significant loss of affinity [18]. This was exploited in attempts to improve the solubility of aprepitant derivatives, resulting in the derivative L-760,735.

‐ ‐

That this site tolerates some modifications is now well-understood in terms of protein-ligand interactions. A recently reported X-ray structure of an NK1R-aprepitant complex [19] revealed that the triazolinone ring is located relatively close to the extracellular end of the receptor binding pocket, where it participates in hydrogen bonding to E193 and W184. However, E193A mutation has virtually no effect on aprepitant's affinity, thus suggesting that the interactions in this area are of less importance

to high affinity binding. Therefore, it seemed the most rational that a convenient site for functionalizing the APT structure is at this very ring. Nevertheless, the performed functionalization of the APT molecule required confirmation that the obtained conjugate still had a sufficiently high affinity for the receptor.

Based on that knowledge, we focused our efforts on the syntheses and in vitro evaluation of newly designed radioconjugates of aprepitant with gallium-68 or lutetium-177 radionuclides. For this purpose, we have proposed two functionalization routes of the APT molecule, followed by conjugation of different macrocyclic chelators DOTA, Bn-DOTA, and Bn-DOTAGA, as well as acrylic chelator DTPA dedicated to <sup>68</sup>Ga and <sup>177</sup>Lu. For conjugates showing full stability in human serum, molecular modelling studies for human NK1R and preliminary in vitro examination were performed. These reported findings indicate new perspectives of aprepitant applications in the form of selective theranostic-like concept radiopharmaceuticals for NK1R-positive tumors.

#### **2. Results and Discussion**

#### *2.1. Syntheses of Aprepitant-Based Radioconjugates*

#### 2.1.1. Syntheses of Aprepitant Derivatives

The first stage of synthesis concerned the modification of the APT structure in order to introduce a primary amine group. This was realized according to synthetic pathways presented below by using one of selected alkyl linkers (Scheme 1) or acetamide linkers (Scheme 2) so as to receive APT-alkylamine (**2A**–**C**) or APT-acetamide derivatives (**4D**,**E**).

**Scheme 1.** Synthetic route of aprepitant derivatives with aminoalkyl linkers; where n = {2; 3; 4}.

**Scheme 2.** Synthetic route of aprepitant derivatives with acetamide linkers; where k = {0; 2}.

#### 2.1.2. Syntheses of Aprepitant Conjugates

The coupling reactions of APT-ethylamine, **2A**, with different bifunctional chelating agents, were as follows: DOTA-NHS ester, *p*-SCN-Bn-DOTA, *p*-SCN-Bn-DOTAGA, or DTPA dianhydride, as presented in Scheme 3. The use of different chelators allowed for the evaluation of the effect of the chelating moiety on the physicochemical properties of later radioconjugates. Based on the stability results obtained for these radioconjugates (presented in a section below), all other obtained APT derivatives (**2B**, **2C**, **4D**, and **4E**) were only conjugated with selected macrocyclic chelator DOTA. The application of different linkers allowed for the evaluation of their influence on the physicochemical properties of later radioconjugates.

**Scheme 3.** Synthetic routes of conjugations of selected chelators to aprepitant-ethylamine **2A**.

#### 2.1.3. Preparation of Radioconjugates

All APT conjugates with DOTA, Bn-DOTA, and Bn-DOTAGA were radiolabeled with <sup>68</sup>Ga and <sup>177</sup>Lu, while APT conjugates with DTPA were only radiolabeled with <sup>68</sup>Ga. Synthesized radioconjugates were purified using the solid phase extraction (SPE) method before HPLC identification (Figures 2 and 3) and further analyses.

‐ ‐ ‐ ‐ ‐ **Figure 2.** Radiochromatograms of aprepitant (APT)-ethylamine **2A** conjugates with DOTA, Bn-DOTA, Bn-DOTAGA or DTPA radiolabeled with gallium-68 (**upper** two) or with lutetium-177 (**bottom**). ‐ ‐ ‐

‐ ‐ **Figure 3.** Radiochromatograms of DOTA conjugates with all APT derivatives radiolabeled with gallium-68 (**upper** two) or with lutetium-177 (**bottom** two).

‐ ‐ ‐ ‐ ‐ **‐** ‐ ‐ ‐ ‐ **‐** ‐ ‐ ‐ ‐ **‐ ‐ ‐ ‐** ‐ ‐ ‐ **‐** ‐ ‐ ‐ ‐ **‐** ‐ ‐ ‐ ‐ **‐ ‐ ‐ ‐** ‐ As a result of the performed radiosyntheses, all radioconjugates were successfully obtained, except for [68Ga]Ga-DTPA-(Et-APT)<sup>2</sup> (**[ <sup>68</sup>Ga]Ga-9A**), which proved to be immediately unstable. Moreover, in the radiochromatogram of [177Lu]Lu-DOTA-Bn-Et-APT (**[ <sup>177</sup>Lu]Lu-6A**) one can see a small additional signal (about 19.3 min) that is recognized as an early by-product of an interaction with solvent (EtOH) from the purification process. To verify the identity of all synthesized [68Ga]Ga-radioconjugates in a non-carrier added scale, the non-radioactive stable gallium reference compounds (**Ga-5A**–**Ga-9A** and **Ga-5A**–**Ga-5E**) were synthesized and characterized by mass spectrometry. The retention time values of the [68Ga]Ga-radioconjugates and stable references presented below (Tables 1 and 2)

‐

‐

overlapped, and the differences between them resulted from the serial connection of UV-Vis and gamma detectors only. ‐

**Table 1.** Retention times (RT) of stable gallium conjugates and [68Ga]Ga-radioconjugates of aprepitant-ethylamine **2A**. **‐ ‐**


**Table 2.** R<sup>T</sup> of stable gallium conjugates and [68Ga]Ga-radioconjugates of all aprepitant derivatives.

‐


#### *2.2. Physiochemical Evaluation of Radioconjugates*

‐

#### 2.2.1. Stability Study

The sine qua non condition of a radionuclide's application in vivo is its radiopharmaceutical stability in biological fluids like serum or cerebrospinal fluid. For this purpose, each isolated and solvent-free radioconjugate was incubated at 37 ◦C in human serum (HS). At specific time points, small samples of radioconjugate mixture were analyzed by the HPLC method for the assessment of the radioconjugate condition. The collected data presented on the charts below (Figure 4) point out that only the DOTA radioconjugates remained stable in the biological fluid; thus, these radioconjugates were selected for further analyses. ‐

‐ ‐ **Figure 4.** Percentage of intact [68Ga]Ga-radioconjugates (**left**) and [177Lu]Lu-radioconjugates (**right**) determined at specific time points during incubation in 37 ◦C human serum.

‐ ‐ ‐ ‐ ‐ ‐ ‐ ‐ We concluded that for the demand of designed aprepitant radioconjugates, the acyclic chelator DTPA showed a poor radionuclide chelating ability during incubation in human serum. DOTA and its analogues presented a satisfactory radionuclide complex stability, however, for the overall stability of the radioconjugate results from the type of the formed chemical bond with the amine terminated aprepitant derivative and the presence of a negative charge on the chelator-metal complex moiety. The amide bond created by the DOTA-NHS ester and uncharged complex in the conjugates remained stable throughout the whole stability study, while the thiourea bonds and negatively charged complexes created by both *p*-SCN-Bn-DOTA and *p*-SCN-Bn-DOTAGA were found to gradually decompose in time. This phenomenon of instability in HS has been observed previously in various radiopharmaceuticals [20].

#### 2.2.2. Lipophilicity Study

Drug distribution in vivo is highly related to both the lipophilicity and charge of a drug. The optimal radiotracer lipophilicity value for blood-brain barrier crossing lies within the range from 2.0 to 3.5 [21]. Non-peptide NK1R antagonists, like aprepitant, are characterized by a high lipophilicity (logD 4.8) [22], while the DOTA chelator is a highly hydrophilic moiety. In seeking to keep in lipophilicity of radioconjugates in a desired range, the choice of a proper linker (primary aprepitant modification) seems essential for distribution and pharmacokinetic aspects.

In the course of the lipophilicity study, each isolated DOTA radioconjugate (determined as fully stable in HS) was examined for distribution in the system of *n*-octanol and a phosphate-buffered saline (PBS) buffer (pH = 7.4) to estimate the lipophilicity of the radiocomplex. The lipophilicity of each radioconjugate (logD), defined as the logarithm of the distribution coefficient (D) is based on the ratio of the radioactivity of the organic phase to the radioactivity of the aqueous phase. The stability of the studied radioconjugate was verified simultaneously during the experiment through the HPLC analysis of the aqueous phase. LogD values of **[ <sup>68</sup>Ga]Ga-5A–[68Ga]Ga-5E** and **[ <sup>177</sup>Lu]Lu-5A–[177Lu]Lu-5E** are listed below in Table 3.


**Table 3.** LogD values of human serum stable radioconjugates determined in *n*-octanol/PBS buffer system.

The APT-alkylamine derivative-based radioconjugates showed similar lipophilicity values that were higher than those of the APT-acetamide derivative-based radioconjugates. The complexes with lutetium were more lipophilic by (on average) 0.6 logD units. However, the logD values for all radioconjugates significantly decreased in comparison to aprepitant, indicating possible divergences in the pharmacokinetic fate of the radioconjugates and the parent drug.

#### *2.3. Binding A*ffi*nity*

An important consideration in the search of conjugate vectors for radionuclides is whether the functionalization of a high affinity ligand would not reduce the binding strength for a desired receptor. For the preliminary addressing of this issue in the case of our conjugates, we measured the affinity of compounds **5A**–**E** (uncomplexed precursors) for the rat neurokinin-1 receptor. The human (hNK1R) and the rat (rNK1R) neurokinin 1 receptors differ in their sequences and pharmacology. It has been established that many (but not all) high affinity NK1R antagonists have a significantly lower affinity for the rat receptor than for that of human origin [23,24]. Still, the results presented below give some tentative insight into the affinity changes caused by the functionalization of the aprepitant structure at the triazolinone ring.

The results of the binding affinity determinations are given in Table 4. The parent compound, aprepitant, was found to exhibit IC<sup>50</sup> = 128.4 nM. This value was roughly consistent with the reported potency of aprepitant in a functional assay. The compound was found to inhibit Substance P-evoked increases in intracellular Ca2<sup>+</sup> mobilization in the cells expressing rNK1R with a pK<sup>B</sup> reading 7.3 [25]. Note that in the assays with cells expressing hNK1R, aprepitant was significantly more potent (pK<sup>B</sup> = 8.7), and the reported binding affinities for the human receptor were of the subnanomolar order (e.g., IC<sup>50</sup> = 0.09 nM [17]).



a IC<sup>50</sup> ± SEM: the half-maximal inhibitory concentration with the standard error of the mean of three independent experiments done in duplicate.

The aprepitant-based conjugates exhibited a diversified range of affinities. The strongest ligand in the set was the compound bearing a propylamine linker, **5B**. It was found to have an IC<sup>50</sup> of 0.69 µM. This value was about five times worse than that of the parent compound. Interestingly, decreasing (**5A**) or increasing (**5C**) the linker length by one methylene unit was associated with much lower affinity of the micromolar order. The shorter **5A** exhibited the lowest binding in the set, with an IC<sup>50</sup> of 6.2 µM. The analogue with the butylamine linker (**5C**) had an IC<sup>50</sup> of 1.8 µM. Similar affinities (IC50~2.5 µM) were found for the conjugates with the acylhydrazine (**5D**) or *N*-aminoethylacetamide (**5E**) linkers.

#### *2.4. Molecular Modelling Study*

In order to get insight into possible interactions between the aprepitant-DOTA conjugates reported herein and the NK1R, the complexes thereof were modelled by molecular docking. The applied procedure consisted in building the appropriate linker-DOTA fragments into the aprepitant structure crystallized with the receptor (Protein Data Bank (PDB) accession code: 6HLO [19]), followed by local search docking executed in AutoDock 4.2.6 [26].

According to this procedure, the presence of a linker-DOTA moiety in the aprepitant-based conjugates did not have a major impact on the interactions between the core of the molecule and the receptor. Only a slight repositioning of the morpholine core, 3,5-bis-trifluoromethylphenyl, or *p*-fluorophenyl moieties was observed compared to the 6HLO crystal structure (Figure 5A). Thus, the conjugates were predicted to bind with the 3,5-bis-trifluoromethylphenyl fragment located at the bottom of the ligand-binding pocket and the DOTA moiety closer to the extracellular side of the receptor (Figure 5A).

In the part that was common to all studied derivatives (and the parent aprepitant), the complexes were stabilized by (Figure 5B):


These interactions were identical to those found for the parent aprepitant in 6HLO structure.

On the other hand, the presence of the linker-DOTA fragment was predicted to weaken the contacts that the triazolinone ring of the parent aprepitant had with the receptor in the crystal structure 6HLO [19]. In the optimized complexes for all the conjugates, this ring was displaced compared to the parent structure (Figure 6A,B), so hydrogen bonding to W184 was not possible. On the other hand, a better positioning of this ring for π–π stacking with H197 was predicted for the conjugates.

μ

‐ ‐

**μ** 

μ μ

μ ‐

‐

‐

‐ ‐ ‐

‐

‐ ‐

**‐ ‐ ‐ ‐ ‐ ‐ ‐ ‐ ‐ ‐ ‐ ‐ ‐**

‐

‐

‐

**Figure 5.** Binding mode of the reported conjugates in the neurokinin 1 receptor (NK1R) binding site. (**A**) A generalized view on the binding mode. The receptor is displayed as a yellow surface, with transmembrane helices (TMs) 2 and 6 shown as cylinders. The extracellular loop 2 (ECL2) is shown as a yellow ribbon. The conjugates are represented as colored sticks. (**B**) A view focused on the interactions of the part common to aprepitant and the conjugates. The conjugate shown is compound **5A** (pale blue sticks). Only several residues of the receptor are shown (yellow sticks). ‐ π π

‐ **Figure 6.** (**A**) Interactions of aprepitant's triazolinone ring with W184, E193, and H197 side-chains. (**B**) positioning of the triazolinone ring of the conjugates in the same projection as in (**A**). (**C**–**E**) Relative position of the DOTA moiety in conjugates **5A** (**C**), **5B**–**D** (**D**), and **5E** (**E**). The receptor helices are shown as yellow cylinders.

‐ Regarding the positioning of the linker-DOTA part, in the case of **5A,** this fragment docked closely (Figure 6C) to the extracellular loop 2 (ECL2) and the extracellular terminus of the transmembrane helix 5 (TM5). One of the DOTA's carboxylate oxygens interacted with the side-chain of K190. For the analogues **5B**–**D**, the DOTA moiety was predicted to be located between the extracellular tips of TM5 and TM6 (Figure 6D). Its contacts included residues K194, K190, and P271. In the case of the longest

derivative, **5E**, the docking placed the DOTA moiety close to TM5 and ECL2 (Figure 6E). Here, it could interact with K190 and M181.

Since in the crystal structures 6HLO, 6HLL, and 6HLP [19], several residues by the extracellular end of the receptor were found to adopt different rotamers upon the binding of different ligands, we wanted to see if the flexibility of these residues could affect the docking results. Therefore, the local docking procedure with the enabled flexibility of E193 and H197 was performed. It yielded similar results with only minor adjustments of the side chain rotamers. Its results (in terms of interactions and binding poses) are not discussed herein since they are almost perfectly accounted for by the description of the docking procedure with the rigid receptor.

Regarding the quantitative evaluation (Table 5), AutoDock scoring function predicted that aprepitant would bind with the free energy of −10.43 kcal/mol. For the conjugates, the estimated energy varied between −9.64 kcal/mol (**5D**) and −13.74 kcal/mol (**5E**). The predicted energies did not correlate with the experimental data. This was perhaps due to the problems with estimating the entropic contribution because the conjugates differed with respect to the number of the rotatable bonds.

**Table 5.** Scoring results from molecular docking. The values are the estimated free energy of binding (kcal/mol).


a lowest energy in the best scored cluster; <sup>b</sup> mean energy in the best scored cluster.

Other sources of significant error may have been the way the DOTA moiety was modelled (aimed at mimicking the presence of the cation in a simplified manner) and the fact that the experimentally evaluated conjugates were uncomplexed.

#### **3. Materials and Methods**

Aprepitant (Santa Cruz Biotechnology Inc., Dallas, TX, USA), the DOTA-NHS ester (1,4,7,10-tetraazacyclododecane-1,4,7,10-tetraacetic acid mono-*N*-hydroxysuccinimide ester), *p*-SCN-Bn-DOTAGA (2,2′ ,2′′-(10-(1-carboxy-4-((4-isothiocyanatobenzyl)amino)-4-oxobutyl)-1,4,7,10 tetraaza-cyclododecane-1,4,7-triyl)triacetic acid) (CheMatech, Dijon, France), *p*-SCN-Bn-DOTA (*S*-2-(4-isothiocyanatobenzyl)-1,4,7,10-tetraazacyclododecane tetraacetic acid) (Macrocyclics, Plano, TX, USA), DTPA dianhydride (diethylenetriaminepentaacetic dianhydride), and other substances and solvents (Sigma Aldrich/Merck, Darmstadt, Germany) were commercially available, defined as reagent grade, and applied without further purification. <sup>68</sup>GaCl<sup>3</sup> was eluted from the commercially available <sup>68</sup>Ge/ <sup>68</sup>Ga generator (Eckert & Ziegler, Berlin, Germany). The <sup>177</sup>LuCl<sup>3</sup> solution in 0.04 M HCl was purchased at Radioisotope Centre POLATOM, National Centre for Nuclear Research, Otwock-Swierk, ´ Poland. Sep-Pack® Classic Short C18 Cartridges were purchased from WATERS, Milford, MA, USA. Human serum was isolated and purified at the Centre of Radiobiology and Biological Dosimetry, INCT Warsaw, Poland.

The HPLC conditions and gradient were as follows: a semi-preparative Phenomenex Jupiter Proteo column, 4 µm, 90 Å, 250 × 10 mm, with UV/Vis (220 nm) or/and radio γ-detection at gradient elution: 0–20 min 20 to 80% solvent B; 20–30 min 80% solvent B; 2 mL/min; solvent A: 0.1% (*v*/*v*) trifluoroacetic acid (TFA) in water; and solvent B: 0.1% (*v*/*v*) TFA in acetonitrile.

Mass spectra were measured on a Bruker 3000 Esquire mass spectrometer equipped with electrospray ionization (ESI) (Bruker, Billerica, MA, USA).

### *3.1. Syntheses of Aprepitant Derivatives and Aprepitant-Based Conjugates*

3.1.1. General Procedure of Syntheses of Aprepitant Derivatives with Alkyl Linker, **2A**–**C**

The slight molar excess of the selected *n*-(terminal-bromoalkyl) phthalimide was added into an equimolar mixture of APT and sodium carbonate in dimethylformamide (DMF). The reaction mixture was vigorously stirred in about 50 ◦C for 12–18 h. Then, the triple molar excess of hydrazine was added into the reaction mixture for an additional 3 h. The progress of the reaction was monitored by HPLC. The crude reaction mixture was evaporated, dissolved in the HPLC mobile phase, purified by the HPLC method, and lyophilized. The isolated main product was identified as a mono-substituted APT-alkylamine derivative (**2A**–**C**, ~75% reaction yield) by MS analysis confirmation.

MS: Calculated monoisotopic mass for **APT-Et-NH2, 2A**, C25H26F7N5O3: 577.19; found: 578.27 *m*/*z* [M + H+]

MS: Calculated monoisotopic mass for **APT-Pr-NH2, 2B**, C26H28F7N5O3: 591.21; found: 592.12 *m*/*z* [M + H+]

MS: Calculated monoisotopic mass for **APT-Bu-NH2, 2C**, C27H30F7N5O3: 605.22; found: 606.38 *m*/*z* [M + H+]

3.1.2. General Procedure of Syntheses of Aprepitant Derivatives with Acetamide Linker, **4D** and **4E**

The slight molar excess of ethyl 2-bromoacetate was added into an equimolar mixture of APT and sodium carbonate in DMF. The reaction mixture was vigorously stirred in about 50 ◦C for 24 h. Then, the triple molar excess of hydrazine or ethylenediamine was added into the reaction mixture for an additional 3 h. The progress of the reaction was monitored by HPLC. The crude reaction mixture was evaporated, dissolved in the HPLC mobile phase, purified by the HPLC method, and lyophilized. The isolated main product was identified as a mono-substituted amino-terminated APT-acetamide derivative (**4D** and **4E**, 65–70% reaction yield) by MS analysis confirmation.

MS: Calculated monoisotopic mass for **APT-Ac-HN-NH2, 4D**, C25H25F7N6O4: 606.19; found: 608.07 *m*/*z* [M + H+]

MS: Calculated monoisotopic mass for **APT-Ac-Et-NH2, 4E**, C27H29F7N6O4: 634.21; found: 635.31 *m*/*z* [M + H+]

3.1.3. General Procedure of Syntheses of Aprepitant Conjugates with DOTA, **5A**–**E**

The obtained APT derivative (**2A**–**C**, **4D**, and **4E**) and the DOTA-NHS ester in similar molar ratios were dissolved in DMF purged from oxygen with technical nitrogen and supplemented with a triple molar excess of triethylamine. The reaction mixture was vigorously stirred in about 50 ◦C for 24 h. The progress of the reaction was monitored by HPLC. The crude reaction mixture was evaporated, dissolved in the HPLC mobile phase, purified by the HPLC method, and lyophilized. The isolated main product was identified as a DOTA conjugate with an APT derivative (**5A**–**E**, >90% reaction yield) by MS analysis confirmation.

MS: Calculated monoisotopic mass for **APT-Et-DOTA**, **5A**, C41H52F7N9O10: 963.37; found: 964.27 *m*/*z* [M + H+]

MS: Calculated monoisotopic mass for **APT-Pr-DOTA**, **5B**, C42H54F7N9O10: 977.39; found: 978.42 *m*/*z* [M + H+]

MS: Calculated monoisotopic mass for **APT-Bu-DOTA**, **5C**, C43H56F7N9O10: 991.40; found: 992.41 *m*/*z* [M + H+]

MS: Calculated monoisotopic mass for **APT-Ac-HN-NH-DOTA**, **5D**, C41H51F7N10O11: 992.36; found: 993.17 *m*/*z* [M + H+]

MS: Calculated monoisotopic mass for **APT-Ac-Et-DOTA**, **5E**, C43H55F7N10O11: 1020.39; found: 1021.43 *m*/*z* [M + H+]

3.1.4. Procedure of Syntheses of Aprepitant-Ethylamine Conjugates with *p*-SCN-Bn-DOTA and *p*-SCN-Bn-DOTAGA, **6A** and **7A**

The APT-ethylamine (**2A**) and bifunctional chelating agent in similar molar ratios were dissolved in DMF and supplemented with a 5-fold molar excess of triethylamine. The reaction mixture was vigorously stirred in about 50 ◦C for 24 h. The progress of the reaction was monitored by HPLC. The crude reaction mixture was evaporated, dissolved in the HPLC mobile phase, purified by the HPLC method, and lyophilized. The isolated main product was identified as a DOTA-Bn or DOTAGA-Bn conjugate with an APT derivative (**6A** and **7A**, > 90% reaction yield) by MS analysis confirmation.

MS: Calculated monoisotopic mass for **APT-Et-Bn-DOTA**, **6A**, C49H59F7N10O11S: 1128.40; found: 1129.55 *m*/*z* [M + H+]

MS: Calculated monoisotopic mass for **APT-Et-Bn-DOTAGA**, **7A**, C52H64F7N11O12S: 1199.43; found: 1200.66 *m*/*z* [M + H+]

3.1.5. Procedure of Syntheses of Aprepitant-Ethylamine Conjugates with DTPA Anhydride, **8A**, **9A**

The APT-ethylamine (**2A**) and DTPA anhydride in a 3:2 molar ratio were dissolved in DMF purged from oxygen with technical nitrogen. The reaction mixture was vigorously stirred in room temperature for 2 h. The progress of the reaction was monitored by HPLC. The crude reaction mixture was evaporated, dissolved in the HPLC mobile phase, purified by the HPLC method, and lyophilized. Two isolated main products were identified as DTPA conjugated with one or two molecules of the APT derivative (**8A** and **9A** with ~45% and ~40% reaction yields, respectively) by MS analysis confirmation.

MS: Calculated for monoisotopic mass **APT-Et-DTPA**, **8A**, C39H47F7N8O12: 952.32; found: 953.40 *m*/*z* [M + H+]

MS: Calculated for monoisotopic mass **APT-Et-DTPA-Et-APT**, **9A**, C64H71F14N13O14: 1511.50; found: 1512.64 *m*/*z* [M + H+]

#### *3.2. Preparation of Radioconjugates*

#### 3.2.1. <sup>68</sup>Ga Radiolabeling

The <sup>68</sup>Ga radiolabeling of the DOTA, Bn-DOTA, and Bn-DOTAGA conjugates of APT was performed according to the following procedure: 145 µL of a concentrated solution of [68Ga]GaCl<sup>3</sup> in 0.1 M HCl from the <sup>68</sup>Ge/ <sup>68</sup>Ga generator (4.9 ÷ 7.2 MBq) was added into the solution of 25 nmol of the selected conjugate in 200 µL of a 0.2 M acetate buffer (pH = 4.5) and heated for 5–10 min at 95 ◦C. After this time, each radioconjugate was purified using Sep-Pack® Classic Short C18 Cartridges according to producer recommendations, thereby obtaining an easily vaporized ethanolic solution of each radioconjugate. The effectiveness of the purification was monitored by HPLC. DTPA radioconjugates were obtained via an analogical procedure in room temperature.

#### 3.2.2. <sup>177</sup>Lu Radiolabeling

The <sup>177</sup>Lu radiolabeling of the DOTA, Bn-DOTA, and Bn-DOTAGA conjugates of APT was performed according to the following procedure: 2.7 ÷ 5.3 µL of a [177Lu]LuCl<sup>3</sup> n.c.a. solution in 0.04 M HCl (4.6 ÷ 5.2 MBq) was added into the solution of 2.5 nmol of the selected conjugate in 200 µL of a 0.02 M acetate buffer (pH 4.5) and heated for 10 min at 95 ◦C. After this time, each radioconjugate was purified using Sep-Pack® C18 Cartridges according to the producer recommendations, thereby obtaining an easily vaporized ethanolic solution of each radioconjugate. The effectiveness of the purification was monitored by HPLC.

#### 3.2.3. Preparation of Non-Radioactive References

The non-radioactive Ga labelling of the DOTA, Bn-DOTA, and Bn-DOTAGA conjugates of APT was performed according to the following procedure: 145 µL of a concentrated solution of 20 mM

GaCl<sup>3</sup> in 0.1 M HCl was added into the solution of 50 nmol of the selected conjugate in 200 µL of a 0.2 M acetate buffer (pH = 4.5) and heated for 5–10 min at 95 ◦C. After this time, each reaction mixture was purified by the HPLC method, lyophilized, and characterized by mass spectrometry. DTPA conjugates were obtained via an analogical procedure in room temperature.

MS: Calculated for monoisotopic mass **APT-Et-DOTA-Ga**, **5A-Ga**, C41H50F7N9O10Ga: 1030.80 and 1032.28; found: 1030.40 and 1032.40 *m*/*z* [M+]

MS: Calculated for monoisotopic mass **APT-Pr-DOTA-Ga**, **5B-Ga**, C42H52F7N9O10Ga: 1044.30 and 1046.30; found: 1044.38 and 1046.39 *m*/*z* [M+]

MS: Calculated for monoisotopic mass **APT-Bu-DOTA-Ga**, **5C-Ga**, C43H54F7N9O10Ga: 1058.31 and 1060.31; found: 1058.37 and 1060.40 *m*/*z* [M+]

MS: Calculated for monoisotopic mass **APT-Ac-HN-NH-DOTA-Ga**, **5D-Ga**, C41H49F7N10O11Ga: 1059.27 and 1061.27; found: 1059.31 and 1061.40 *m*/*z* [M+]

MS: Calculated for monoisotopic mass **APT-Ac-Et-DOTA-Ga**, **5E-Ga**, C43H53F7N10O11Ga: 1087.30 and 1089.30; found: 1087.37 and 1089.44 *m*/*z* [M+]

MS: Calculated for monoisotopic mass **APT-Et-Bn-DOTA-Ga**, **6A-Ga**, C49H57F7N10O11SGa: 1095.31 and 1097.31; found: 1195.54 and 1197.51 *m*/*z* [M+]

MS: Calculated for monoisotopic mass **APT-Et-Bn-DOTAGA-Ga**, **7A-Ga**, C52H62F7N11O12Sga: 1266.34 and 1268.34; found: 1266.47 and 1268.47 *m*/*z* [M+]

MS: Calculated for monoisotopic mass **APT-Et-DTPA-Ga**, **8A-Ga**, C39H44F7N8O12Ga: 1018.22 and 1020.22; found: 1019.35 and 1021.37 *m*/*z* [M + H+]

MS: Calculated for monoisotopic mass **APT-Et-DTPA-**(**Ga**)**-Et-APT**, **9A-Ga**, C64H68F14N13O14Ga: 1577.40 and 1579.40; found: 1578.64 and 1580.66 *m*/*z* [M + H+]

#### *3.3. Physiochemical Evaluation of Radioconjugates*

#### 3.3.1. Stability Study

All obtained radioconjugates (isolated from the reaction mixtures using the SPE method and being solvent-free) were examined in terms of stability in human serum using HPLC analyses. A solution of each isolated selected radioconjugate in 100 µL of a 0.1M PBS buffer pH 7.40 was added to 900 µL of human serum and incubated at 37 ◦C for 4 h (68Ga radioconjugates) or 14 days (177Lu radioconjugates). At specific time points, 400 µL of the incubated mixture was added into 500 µL of ethanol, vigorously stirred to precipitate serum proteins, and centrifuged (13,500 rpm for 5 min) to separate the supernatant for HPLC analysis.

#### 3.3.2. Lipophilicity Study

The lipophilicity values of the radioconjugates (logD), expressed as the logarithm of its D in the *n*-octanol/PBS (pH 7.40) system, mimicking the physiological conditions (Product Properties Test Guidelines of the Office of Prevention, Pesticides and Toxic Substances 830.7550, 1996), were determined right after the SPE method purification and ethanol evaporation processes. A solution of isolated selected radioconjugate in 500 µL of a 0.1 M PBS buffer at pH 7.40 and 500 µL of *n*-octanol was vigorously stirred and centrifuged (13,500 rpm for 5 min) to separate the immiscible phases. The radioactivities of the aqueous and organic layers were determined using a well-type NaI(Tl) detector. The distribution coefficient was calculated as the ratio of the radioactivity of the radioconjugate in the organic phase to that in the aqueous phase. Each measurement was performed in triplicate and averaged. Simultaneously, the aqueous phases were analyzed by HPLC to check whether the studied radioconjugate remained intact during the experiment.

#### *3.4. Binding A*ffi*nity Determination*

The binding affinity of aprepitant and compounds **5A**–**E** for rNK1R was determined in a competitive radioligand binding assay using rat brain homogenates, following a previously described method [27]. In brief, the membrane preparations obtained from rat brains were incubated at 25 ◦C for 60 min in the presence of a selective radioligand [3H]-[Sar<sup>9</sup> ,Met(O2) <sup>11</sup>]-Substance P obtained from PerkinElmer, (Waltham, MA, USA) and the increasing concentrations of the tested compounds (each concentration in duplicate). Non-specific binding was measured in the presence of 10 µM cold Substance P. The assay buffer was composed of 50 mM Tris-HCl (pH 7.4), 5 mM MnCl2, bovine serum albumin (BSA) (0.1 mg/mL), bacitracin (100 µg/mL), bestatin (30 µM), phenylmethylsulfonyl fluoride (30 µg/mL), and captopril (10 µM). The reaction total volume was 1 mL. With the incubation having been terminated, a rapid filtration through GF/B Whatman glass fiber strips was done with a M-24 Cell Harvester (Brandel, Gaithersburg, MD, USA). The filters were pre-soaked overnight with 0.5% polyethyleneimine so that the extent of non-specific binding could be minimized. After the filtration, the strips were dried, the filter discs were placed separately in 24-well plates, and a Betaplate Scint scintillation solution (PerkinElmer, Waltham, MA, USA) was added to each well. Radioactivity was measured with a MicroBeta LS scintillation counter, Trilux (PerkinElmer, Waltham, MA, USA). The data came from three independent experiments done in duplicate. The results are presented as IC<sup>50</sup> with SEM.

#### *3.5. Docking*

In order to obtain the probable structures of the complexes of the neurokinin 1 receptor with the conjugates **5A**–**E**, the following modelling procedure was performed. The aprepitant structure (with neutral charge) in the complex with the receptor (PDB accession code: 6HLO [19]) was expanded by attaching to the triazolinone ring the appropriate linkers and the DOTA moiety. Such initial complexes were subjected to local search docking in AutoDock 4.2.6 [26].

The DOTA geometry was set based on the NOJYIU entry [28] of The Cambridge Structural Database [29]. This structure is a DOTA complex with Lu3<sup>+</sup> (diaqua-lutetium(III)-sodium trihydrate). For the purposes of our modelling, DOTA carboxylate arms were protonated and frozen in the conformation found in the crystal structure of lutetium (III) chelate of DOTA (after removing the Lu3<sup>+</sup> cation, Na<sup>+</sup> cations and waters). The rationale behind this gambit was the fact that the carboxylates would be primarily engaged in the interactions with a cation; therefore, they might have been expected to retain the conformation they had in the solid state structure. This approach could also give a rough approximation of the DOTA's steric influence on the binding of the conjugates despite a lack of properly scaled and validated parameters for modelling and scoring the complexes with the cations of interest.

The used receptor structure was a refined one (as provided by the GPCRdb service [30]) in order to have the mutated residues replaced with the native ones and to supply the side chains missing in the original PDB structure. The structure was pre-processed in AutoDock Tools [26]. The box was set around the experimental position of aprepitant in 6HLO and extended towards the extracellular part of the receptor so as to cover the expected length of the expanded conjugate. The grids were calculated with AutoGrid 4 [26]. We considered two variants of docking with respect to the flexibility of the receptor structure. In the first variant, all receptor residues were rigid. In the second variant, E193 and H197 side-chains were set to be flexible.

The docking procedure was the local search with the following parameters: 500 individuals in population, 500 iterations of the Solis-Wets local search, the *sw*\_*rho* parameter of the local search space set to 20.0, and 1000 local search runs. The structures resulting from the local search were clustered, and the representative models of the lowest scored (on average) cluster were taken for further analysis. For the qualitative assessment of the binding energy, both the lowest and the mean energy of the clusters were collected. The molecular graphics were prepared in PyMol [31].

For comparative and validation purposes, the very same procedure of local docking (with and without the flexibility of the mentioned two residues) was performed for the parent aprepitant.

#### **4. Conclusions**

The presented paper describes the evaluation of aprepitant functionalization in order to provide an application of this NK1R antagonist in nuclear medicine.

Out of the corresponding <sup>68</sup>Ga/ <sup>177</sup>Lu radioconjugates of APT-ethylamine **2A** with DOTA, Bn-DOTA, Bn-DOTAGA, and DTPA, only the DOTA amide conjugates showed satisfactory stability in human serum throughout the whole incubation time. The evaluation of the linker effect on radioconjugate lipophilicity indicated APT-alkylamine derivatives as more promising biovectors with features closer to parent aprepitant. The physicochemical properties of obtained APT-alkylamine-DOTA derivatives labelled with <sup>68</sup>Ga (**[ <sup>68</sup>Ga]Ga-5A–[68Ga]Ga-5C**) can be compared with those of [67Ga]Ga-NOTA-NK1R radioligands based on another NK1R antagonist—L-733,060 [32]. The <sup>67</sup>/68Ga-radioligands based on these two high affinity NK1R antagonists turned out to be very similar, as evidenced by the following parameters:


Regarding the affinity studies of the **5A**–**E** conjugates, on the assumption that the human NK1R affinities for aprepitant derivatives were generally much higher than the rat NK1R affinities and that structure-affinity trends were parallel in both species, all the synthesized compounds might be considered to retain reasonable NK1R affinity compared to their parent. In particular, the analogue **5B** (which only suffered a few-times decrease in affinity compared to APT) seems to be especially interesting for further development. Obtained results suggest that the functionalizing of the aprepitant structure via the triazolinone ring is the right strategy.

Itis alsoworthmentioning that, in general, radiopharmaceuticals based on small non-peptidemolecules (e.g., aprepitant and L-733,060) have many advantages over peptide-based radiopharmaceuticals [2]. They usually have lower molecular weights, higher lipophilicity values, and, hence, different pharmacokinetics; they are stable in vivo, but, more importantly, their radiosyntheses can be carried out at higher temperatures and in a wider pH range. Moreover, according to the literature, radiopharmaceuticals based on non-peptide antagonists interact with a receptor through more binding sites and accumulate better and for a longer time period in cancer cells [33,34]. Even though the further evaluation of aprepitant-based radiopharmaceuticals is still needed, the findings reported herein provide insight on the perspectives of their application in the theranostics paradigm.

#### **5. Patent**

In course of this study, the following national patent application was submitted: No. P430136 "The modified drug substance molecule, method of its production, diagnostic or therapeutic receptor radiopharmaceutical based on this molecule, method of its production and its application".

**Author Contributions:** Conceptualization, P.K.H. and P.F.J.L.; methodology, P.K.H., P.F.J.L., J.M., and P.K.; investigation, P.K.H. and J.M.; writing—original draft preparation, P.K.H. and P.F.J.L.; writing—review and editing, P.K. and E.G.; visualization, P.F.J.L.; supervision, A.M. and E.G.; project administration, A.M. and E.G.; funding acquisition, A.M. and E.G. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was funded by National Science Centre (Poland), grant number 2017/25/B/NZ7/01896.

**Acknowledgments:** The contribution of Paweł Krzysztof Halik has been done in the frame of the National Centre for Research and Development Project No POWR.03.02.00-00-I009/17 (Radiopharmaceuticals for molecularly targeted diagnosis and therapy, RadFarm. Operational Project Knowledge Education Development 2014–2020 co-financed by European Social Fund).

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **Abbreviations**


#### **References**


**Sample Availability:** Samples of the compounds **2A**–**C**, **4D**, **E** are available from the authors.

© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

*Article*
