Radiolabeled Cyclic RGD Peptides as Integrin α_vβ₃-Targeted Radiotracers: Maximizing Binding Affinity via Bivalency

Abstract

Integrin α_vβ₃ plays a significant role in tumor angiogenesis, and is a receptor for the extracellular matrix proteins with the exposed arginine-glycine-aspartic (RGD) tripeptide sequence. These include vitronectin, fibronectin, fibrinogen, lamin, collagen, Von Willibrand’s factor, osteoponin, and adenovirus particles. Integrin α_vβ₃ is expressed at low levels on epithelial cells and mature endothelial cells, but it is overexpressed on the activated endothelial cells of tumor neovasculature and some tumor cells. The restricted expression of integrin α_vβ₃ during tumor growth, invasion and metastasis present an interesting molecular target for both early detection and treatment of rapidly growing solid tumors. Over the last decade, many radiolabeled linear and cyclic RGD peptide antagonists have been evaluated as the integrin α_vβ₃-targeted radiotracers. Significant progress has been made on their use for imaging integrin α_vβ₃-positive tumors by SPECT or PET. Among the radiotracers evaluated in pre-clinical tumor-bearing models, [¹⁸F]Galacto-RGD (2-[¹⁸F]fluoropropanamide c(RGDfK(SAA); SAA = 7-amino-L-glyero-L-galacto-2,6-anhydro-7-deoxyheptanamide) and [¹⁸F]-AH111585 are currently under clinical investigation for visualization of integrin α_vβ₃ expression in cancer patients. However, their low tumor uptake, high cost and lack of preparative modules for routine radiosynthesis will limit their continued clinical applications. Thus, there is a continuing need for more efficient integrin α_vβ₃-targeted radiotracers that are readily prepared from a kit formulation without further post-labeling purification. This article will focus on different approaches to maximize the targeting capability of cyclic RGD peptides and to improve the radiotracer excretion kinetics from non-cancerous organs. Improvement of tumor uptake and tumor-to-background ratios is important for early detection of integrin α_vβ₃-positive tumors and/or noninvasive monitoring of therapeutic efficacy of antiangiogenic therapy.

Introduction

Cancer is the second leading cause of death worldwide. Although the exact cause of cancer remains unknown, most cancer patients will survive after surgery, radiation therapy, and chemotherapy or a combination thereof if it can be detected at the early stage. Thus, accurate early detection is highly desirable so that appropriate therapy can be given before the primary tumors become widely spread.

Tumors produce many angiogenic factors, which are able to activate endothelial cells in established blood vessels and induce endothelial proliferation, migration, and new vessel formation (angiogenesis) through a series of sequential but partially overlapping steps. Angiogenesis is a requirement for tumor growth and metastasis (1–7). Without the neovasculature to provide oxygen and nutrients, tumors cannot grow beyond 1–2 mm in size. Once vascularized, the tumors begin to grow rapidly. The angiogenic process depends on vascular endothelial cell migration and invasion, is regulated by cell adhesion receptors. Integrins are such a family of proteins that facilitate the cellular adhesion to and the migration on extracellular matrix proteins in the intercellular spaces and basement membranes, and regulate cellular entry and withdraw from the cell cycle (7–10). Integrin α_vβ₃ serves as a receptor for extracellular matrix proteins with exposed arginine-glycine-aspartic (RGD) tripeptide sequence (8–13). These include vitronectin, fibronectin, fibrinogen, lamin, collagen, osteoponin and adenovirus particles. Integrin α_vβ₃ is expressed at low levels on epithelial cells and mature endothelial cells; but it is highly expressed on activated endothelial cells in neovasculature of tumors, including osteosarcomas, neuroblastomas, glioblastomas, melanomas, lung carcinomas and breast cancer (14–16). It has been demonstrated that integrin α_vβ₃ is overexpressed on both the endothelial cells and tumor cells in human breast cancer xenografts (17). The integrin α_vβ₃ expression correlates well with tumor progression and invasiveness of melanoma, glioma, ovarian and breast cancers (10–17). The restricted expression of integrin α_vβ₃ during tumor growth, invasion and metastasis present an interesting molecular target for diagnosis and treatment of the rapidly growing and metastatic tumors (18–32).

Over the last decade, many radiolabeled cyclic RGD peptides have been evaluated as radiotracers for imaging tumors by SPECT or PET (33–68). Several review articles have appeared covering the nuclear medicine applications of radiolabeled cyclic RGD peptide and non-peptide antagonists as radiotracers for diagnosis and radiotherapy of integrin α_vβ₃–positive tumors (17–29). This article is not intended to be an exhaustive review of the current literature on radiolabeled RGD peptides and nuclear medicine applications. Instead, it will focus on different approaches to maximize their targeting capability and to improve the radiotracer excretion kinetics from non-cancerous organs. Improvement of tumor uptake and tumor-to-background (T/B) ratios is important for both diagnostic and therapeutic radiotracers.

Radiotracer Design

Figure 1 shows a schematic illustration of an integrin α_vβ₃-targeted radiotracer. Cyclic RGD peptide serves as the targeting biomolecule to carry radionuclide to the integrin α_vβ₃ overexpressed on tumor cells and activated endothelial cells of tumor neovasculature. A multidentate bifunctional chelator (BFC) is used to attach the metallic radionuclide to the cyclic RGD peptide (69–72), whereas an organic precursor or synthon is often needed for the ¹⁸F-labeling (73). The pharmacokinetic modifying linker (PKM) is used to improve the radiotracer excretion kinetics (69–71). For a new integrin α_vβ₃-targeted radiotracer to be successful, it must show clinical indications for high-incidence tumors (breast, lung, colorectal, prostate and skin cancers). The radiotracer should have high tumor uptake and T/B ratios in a short period of time. To achieve this goal, the radiotracer should have a rapid blood clearance to minimize background radioactivity. Since most high-incidence tumors (breast, lung and colorectal cancers) occur in the torso, renal excretion is necessary to avoid radioactivity accumulation in the gastrointestinal tract. The integrin α_vβ₃-targeted radiotracer should also be able to distinguish between benign and malignant tumors, to follow tumor growth and metastasis, and to predict therapeutic efficacy in integrin α_vβ₃-positive cancer patient. In addition, a kit formulation or preparative module is needed for routine preparation of the radiotracer in high yield and radiochemical purity at low cost.

Figure 1.

Schematic presentation of the radiotracer design. The targeting biomolecule is a multimeric cyclic RGD peptide. The PKM linker is used to modify radiotracer pharmacokinetics. For the metal-containing radiotracers, a multidentate BFC is often used to attach the metallic radionuclide to the targeting biomolecule. For ¹⁸F-based radiotracers, an organic precursor or synthon is often needed to attach ¹⁸F onto the multimeric cyclic RGD peptide.

Choice of Radionuclide

The choice of radionuclide depends largely on the clinical utility of radiotracer. For planar imaging and SPECT, more than 80% of radiotracers used in nuclear medicine departments are ^99mTc compounds due to optimal nuclear properties of ^99mTc and its easy availability at low cost (69–72). The 6 h half-life is long enough to allow radiopharmacists to carry out radiosynthesis and for physicians to collect clinically useful images. At the same time, it is short enough to permit administration of 20–30 mCi of ^99mTc without imposing a significant radiation dose to the patient. The monochromatic 140 KeV photons are readily collimated to give high quality images with high spatial resolution. The clinically most relevant radionuclides for PET include ¹⁸F, ⁶²Cu, ⁶⁴Cu and ⁶⁸Ga.

¹⁸F is a cyclotron-produced isotope suitable for PET. It has a half-life of 110 min. For several years, ¹⁸F-FDG (FDG = 2-fluoro-2-deoxyglucose) has been widely used as an imaging tool for diagnosis of cancers, brain and cardiovascular diseases. Despite its short half-life, the availability of preparative modules makes ¹⁸F-labeled biomolecules much more accessible to researchers and clinicians in many clinical institutions.

⁶²Cu is a generator-produced radionuclide from the decay of ⁶²Zn. Its 9.7 min half-life allows repeated dosing without imposing a significant radiation burden to the patient. The commercially available ⁶²Zn-⁶²Cu generator has been successfully used in clinical trials (74–77). ⁶⁴Cu is another PET isotope useful for development of target-specific radiotracers. It has a half-life of 12.7 h and a β⁺ emission (18%, E_max = 0.655 MeV). Despite poor nuclear properties, its long half-life makes it feasible to prepare, transport, and deliver the ⁶⁴Cu radiotracer for clinical applications. More importantly, recent breakthroughs in production of ⁶⁴Cu with high specific activity have made it more available to the small research institutions without on-site cyclotron facilities (78). ⁶⁴Cu is a viable alternative to ¹⁸F for research programs that wish to incorporate high sensitivity and high spatial resolution of PET, but cannot afford to maintain the expensive radionuclide production infrastructure. Copper radionuclides and related radiochemistry have been reviewed by Blower et al (79). Nuclear medicine applications of ⁶⁴Cu-labeled monoclonal antibodies and peptides have been reviewed by Anderson et al (75, 80).

⁶⁸Ga is generator-produced PET isotope with a half-life of 68 min. The ⁶⁸Ge-⁶⁸Ga generator can be used for more than a year, allowing PET studies without the on-site cyclotron. If the radiotracer is properly designed, ⁶⁸Ga could become as useful for PET as ^99mTc for SPECT (81, 82). In support of this, the ⁶⁸Ga-labeled somatostatin analogs have been studied extensively for PET imaging of somatostatin-positive tumors in animal models and cancer patients (83–91). Gallium chemistry and related medical applications have been reviewed recently (81, 82, 92).

Bifunctional Chelators

The choice of BFC depends on the radionuclide. ¹⁸F can be incorporated into the cyclic RGD peptide via a covalent bond without the need for BFC. In contrast, BFC is an important part of radiotracers containing a metallic radionuclide (69–72, 92). Among various BFCs (Figure 2), 6-hydazinonicotinic acid (HYNIC) is of great interest due to its high ^99mTc-labeling efficiency (rapid radiolabeling and high radiolabeling yield), the high solution stability of its ^99mTc complexes, and the easy use of different coligands for modification of biodistribution characteristic of ^99mTc-labeled small biomolecules (93). DOTA (1,4,7,10-tetraazacyclododecane-1,4,7,10-tetraacetic acid), NOTA (1,4,7-tritazacyclononane-1,4,7-triacetic acid) and their derivatives (Figure 2) have been used as BFCs for ⁶⁸Ga and ⁶⁴Cu-labeling small biomolecules (84–91). NODAGA is particularly useful for ⁶⁸Ga- and ⁶⁴Cu-labeling due to high hydrophilicity and in vivo stability of its ⁶⁸Ga and ⁶⁴Cu chelates. It has been reported that NOTA derivatives have much higher ⁶⁸Ga and ⁶⁴Cu-labeling efficiency than their DOTA analogs (94–98). The fast and efficient radiolabeling is especially critical for ⁶⁸Ga and ⁶²Cu due to their short half-life (t_1/2 = 68 min for ⁶⁸Ga and 9.7 min for ⁶²Cu).

Figure 2.

BFCs useful for radiolabeling of multimeric cyclic RGD peptides with ^99mTc, ⁶⁸Ga and ⁶⁴Cu. DADA (diamidodithiol), MAMA (monoaminemonoamidedithiol), MADT (diaminedithiol), MAG₂ (2-mecaptoacetylglycylglycyl) and HYNIC are particularly useful for ^99mTc-labeling while DOTA, NOTA and their derivatives are excellent BFCs for chelation of ⁶⁸Ga and ⁶⁴Cu.

PKM Linkers

In general, high lipophilicity often leads to more hepatobiliary excretion and/or high protein binding, which will results in longer blood retention of radioactivity. Hepatobiliary excretion is detrimental for improvement of T/B ratio. Thus, an important aspect of radiotracer development is to improve T/B ratios by modifying pharmacokinetics of radiolabeled cyclic RGD peptides. For example, the negatively charged small peptide sequences or amino acids have been proposed as PKM linkers to reduce renal uptake and kidney retention of radiolabeled small biomolecules (70, 71, 92). The di(cysteic acid) linker has successfully been used to improve the blood clearance and minimize the liver and kidney activity of radiolabeled nonpeptide integrin α_vβ₃ receptor antagonists (99–102). The Asp₃ and Ser₃ tripeptide sequences were also used to modify excretion kinetics of the ^99mTc-labeled cyclic RGD peptide (39). Harris et al reported the use of a PEG₄ (15-amino-4,7,10,13-tetraoxapentadecanoic acid) linker to improve the tumor uptake and T/B ratios of the ^99mTc-labeled nonpeptide integrin α_vβ₃ receptor antagonists (99–102). Kessler et al reported the use of HEG (hexaethylene glycolic acid) as the PKM linker for the ¹⁸F-labeled cyclic RGDfE dimers and tetramers (40–42). Using the HEG linker also increases the distance between the cyclic RGD motifs. Chen et al also found that the introduction of the PEG (polyethylene glycolic acid) linker can improve the tumor uptake and excretion kinetics of ¹²⁵I- and ¹⁸F-labeled c(RGDyK) and ⁶⁴Cu-labeled E[c(RGDyK)]₂ (48, 49, 54, 103).

Targeting Biomolecules

Figure 3 shows several examples of cyclic RGD peptides that have high affinity and selectivity for integrin α_vβ₃. Among the radiotracers evaluated in pre-clinical tumor-bearing models, [¹⁸F]Galacto-RGD (Figure 3: 2-[¹⁸F]fluoropropanamide c(RGDfK(SAA); SAA = 7-amino-L-glyero-L-galacto-2,6-anhydro-7-deoxyheptanamide) and [¹⁸F]-AH111585, the core sequence of which was originally discovered from a phage display library (as ACDRGDCFCG), are currently under clinical investigation for visualization of integrin α_vβ₃ expression in cancer patients (104–109). The results from imaging studies in cancer patients show that there is sufficient integrin α_vβ₃ expression for PET imaging, and the accumulation of radiolabeled RGD peptide radiotracer correlates well with the integrin α_vβ₃ expression levels in cancer patients (108, 109). However, their relatively low tumor uptake, the high cost and lack of preparative modules for routine radiosynthesis will limit their continued clinical applications. Several steps of manual radiosynthesis and post-labeling purification often cause significant radiation exposure to radiopharmacists. Of course, this problem can be solved by developing preparative modules for routine radiosynthesis, which will definitely lead to added cost for the radiotracer. Thus, this is a continuing need for more efficient integrin α_vβ₃-specific radiotracers that are readily prepared from a kit formulation without further post-labeling chromatographic purification.

Figure 3.

Chemdraw structures of cyclic RGD peptides useful as targeting biomolecules for the integrin α_vβ₃-targeted radiotracers. [¹⁸F]Galacto-RGD (2-[¹⁸F]fluoropropanamide c(RGDfK(SAA); SAA = 7-amino-L-glyero-L-galacto-2,6-anhydro-7-deoxyheptanamide) is currently under clinical investigation for visualization of integrin α_vβ₃ expression in cancer patients

Maximizing Binding Affinity via Multimerization

Cyclization to Improve Binding Affinity and Selectivity

Many cyclic RGD peptides have been proposed as the integrin α_vβ₃ antagonists for treatment of cancer. It was found that incorporation of the RGD sequence into a cyclic pentapeptide framework (Figure 3) increases the binding affinity and selectivity for integrin α_vβ₃ over glycoprotein IIb/IIIa (110–113). After extensive structure-activity evaluations, it was concluded that the amino acid in position 5 has no significant impact on integrin α_vβ₃ binding affinity. The valine (V) residue in c(RGDfV) can be replaced by lysine (K) or glutamic acid (E) to afford c(RGDfK) and c(RGDfE), respectively, without changing the integrin α_vβ₃ binding affinity.

Cyclic RGD Dimers

To improve integrin α_vβ₃ binding affinity, multimeric RGD peptides, such as E[c(RGDfK)]₂, have been used to develop integrin α_vβ₃-targeted radiotracers. For example, Rajopadhye et al were the first to use cyclic RGD dimers, such as E[c(RGDfK)]₂ (Figure 4), to develop diagnostic (^99mTc) and therapeutic (⁹⁰Y and ⁶⁴Cu) radiotracers (56, 57, 114, 115). Recently, Chen et al (50,51) reported the ⁶⁴Cu and ¹⁸F-labeled E[c(RGDyK)]₂ (Figure 4) as PET radiotracers. Poethko et al found that the RGDfE dimer [c(RGDfE)HEG]₂-K-Dpr-[¹⁸F]FBOA (Figure 4) had much better targeting capability as evidenced by its higher integrin α_vβ₃ binding affinity and tumor uptake as compared to its monomeric analog c(RGDfE)HEG-Dpr-[¹⁸F]FBOA (40–42).

Figure 4.

Cyclic RGD peptide dimers (E[c(RGDfK)]₂ and E[c(RGDyK)]₂) and [c(RGDfE)HEG]₂-K-Dpr-[¹⁸F]FBOA.

Cyclic RGD Tetramers and Octamers

Several groups have used the multimer concept to prepare cyclic RGD tetramers and octamers. For example, Boturyn et al reported a series of cyclic RGDfK tetramers (116), and found that increasing the peptide multiplicity significantly enhanced the integrin α_vβ₃ binding affinity and internalization. Kessler et al reported a cyclic RGDfE tetramer (Figure 5) that had better integrin α_vβ₃ binding affinity than its monomeric and dimeric analogs (40–42). Liu et al reported the use a cyclic RGDfK tetramer E[E[c(RGDfK)]₂]₂ (Figure 5) for development of integrin α_vβ₃-targeted SPECT and PET radiotracers (52, 64–68). Chen et al recently used ⁶⁴Cu and ¹⁸F-labeled cyclic RGD tetramer E[E[c(RGDxK)]₂]₂ and octamer E[E[E[c(RGDyK)]₂]₂]₂ for tumor imaging by PET (52, 117). Although the results from both in vitro assays and ex-vivo biodistribution studies have demonstrated that radiolabeled (^99mTc, ¹⁸F and ⁶⁴Cu) RGD tetramers and octamer have much better tumor targeting capability (higher integrin α_vβ₃ binding affinity and better radiotracer tumor uptake) than their dimeric analogs, it remains unclear if the cyclic RGD motifs in E[E[c(RGDxK)]₂]₂ (x = f and y) and E[E[E[c(RGDyK)]₂]₂]₂ are really multivalent in binding to integrin α_vβ₃. As the peptide multiplicity increases, the uptake of radiolabeled multimeric RGD peptides in the kidneys, liver, lungs and spleen are also significantly increased. In addition, the cost for E[E[c(RGDfK)]₂]₂ and E[E[E[c(RGDyK)]₂]₂]₂ is prohibitively high for development of integrin α_vβ₃-targeted radiotracers in the future. Thus, an alternate approach is needed to improve integrin α_vβ₃-targeting capability and minimize the radiotracer accumulation in non-cancerous organs.

Figure 5.

A cyclic RGDfK peptide tetramer, E{E[c(RGDfK)]₂}₂ and a ¹⁸F-labeled cyclic RGDfE tetramer, {[c(RGDfE)HEG]₂K}₂-K-Dpr-[¹⁸F]FBOA.

Maximizing Binding Affinity via Bivalency

Factors Influencing Binding Affinity

The success of E[c(RGDfK)]₂ as targeting biomolecules is very intriguing. Given the short distance (6 bonds excluding side-arms of K-residues) between two cyclic RGD motifs in E[c(RGDfK)]₂, it is unlikely that they would bind to two adjacent integrin α_vβ₃ sites simultaneously. However, the binding of one RGD motif to integrin α_vβ₃ will significantly increase the “local concentration” of second RGD motif in the vicinity of integrin α_vβ₃ sites (Figure 6A). The “locally enhanced RGD concentration” may explain the higher tumor uptake of radiolabeled (^99mTc, ¹¹¹In, ⁹⁰Y, ¹⁸F and ⁶⁴Cu) cyclic RGD dimers as compared to their monomeric analogs (60–65). To further improve the integrin α_vβ₃-targeting capability of cyclic RGD dimers, the distance between two cyclic RGD motifs must be increased so that they are to achieve simultaneous integrin α_vβ₃ binding (Figure 6B). The combination of “bivalency” and the “enriched local RGD concentration” is expected to result in higher integrin α_vβ₃ targeting capability of cyclic RGD dimers and better tumor uptake with longer tumor retention time for their corresponding radiotracers.

Figure 6.

Schematic illustration of interactions between cyclic RGD peptide dimers and integrin α_vβ₃. A: The distance between two RGD motifs is not long enough for simultaneous integrin α_vβ₃ binding. However, the RGD concentration is “locally enriched” in the vicinity of neighboring integrin α_vβ₃ once the first RGD motif is bound. B: The distance between two RGD motifs is long due to the presence of two linkers (L). As a result, the cyclic RGD dimer is able to bind integrin α_vβ₃ in a “bivalent” fashion. In both cases, the end-result would be higher integrin α_vβ₃ binding affinity for the multimeric cyclic RGD peptides.

Improve Integrin α_vβ₃ Binding Affinity by Increasing Distance between Cyclic RGD Motifs

To demonstrate this concept, Shi et al recently reported a series of cyclic RGD peptide dimers (Figure 7) containing triglycine (G₃) and PEG₄ linkers, which are used to increase the distance between two cyclic RGD motifs from 6 bonds in E[c(RGDfK)]₂ to 24 bonds in 3G₃-dimer and 38 bonds in 3PEG₄-dimer (118, 119). The integrin α_vβ₃ binding affinities (Table 1) against ¹²⁵I-echistatin bound to U87MG human glioma cells follow the order of HYNIC-tetramer (IC₅₀ = 7 ± 2 nM) > HYNIC-2PEG₄-dimer (IC₅₀ = 52 ± 7 nM) ~ HYNIC-3PEG₄-dimer (IC₅₀ = 60 ± 4 nM) ~ HYNIC-3G₃-dimer (IC₅₀ = 61 ± 2 nM) > HYNIC-PEG₄-dimer (IC₅₀ = 84 ± 7 nM) ~ HYNIC-dimer (IC₅₀ = 112 ± 21 nM) ≫ HYNIC-G₃-monomer (IC₅₀ = 358 ± 8 nM) > HYNIC-PEG₄-monomer (IC₅₀ = 452 ± 11 nM). These data indicate that the G₃ and PEG₄ linkers between two RGD motifs are responsible for the improved integrin α_vβ₃ affinity of HYNIC-3G₃-dimer and HYNIC-3PEG₄-dimer as compared to that of HYNIC-PEG₄-dimer. The higher binding affinity of HYNIC-tetramer is probably due to its two extra cyclic RGD motifs.

Figure 7.

New cyclic RGD peptide dimers useful for the development of integrin α_vβ₃–targeted radiotracers. The PEG₄ and G₃ linkers are used to increase the distance between two RGD motifs and to improve radiotracer excretion kinetics from non-cancerous organs.

Table 1.

Integrin α_vβ₃ binding data for HYNIC-conjugated cyclic RGD peptides against ¹²⁵I-echistatin bound to the integrin α_vβ₃–positive U87MG human glioma cells.

Compound	Radiotracer	IC50 (nM)
HYNIC-G3-monomer	[99mTc(HYNIC-G3-monomer)(tricine)(TPPTS)]	358 ± 8
HYNIC-PEG4-monomer	[99mTc(HYNIC-PEG4-monomer)(tricine)(TPPTS)]	452 ± 11
HYNIC-dimer	[99mTc(HYNIC-dimer)(tricine)(TPPTS)]	112 ± 21
HYNIC-PEG4-dimer	[99mTc(HYNIC-PEG4-dimer)(tricine)(TPPTS)]	84 ± 7
HYNIC-2G3-dimer	[99mTc(HYNIC-2G3-dimer)(tricine)(TPPTS)]	60 ± 4
HYNIC-2PEG4-dimer	[99mTc(HYNIC-2PEG4-dimer)(tricine)(TPPTS)]	52 ± 7
HYNIC-3G3-dimer	[99mTc(HYNIC-3G3-dimer)(tricine)(TPPTS)]	61 ± 2
HYNIC-3PEG4-dimer	[99mTc(HYNIC-3PEG4-dimer)(tricine)(TPPTS)]	62 ± 5
HYNIC-tetramer	[99mTc(HYNIC-dimer)(tricine)(TPPTS)]	7 ± 2
DOTA-dimer	64Cu(DOTA-dimer)/111In(DOTA-dimer)	102 ± 5
DOTA-3G3-dimer	64Cu(DOTA-3G3-dimer)/111In(DOTA-3G3-dimer)	74 ± 3
DOTA-3PEG4-dimer	64Cu(DOTA-3PEG4-dimer)/111In(DOTA-3PEG4-dimer)	62 ± 6
DOTA-tetramer	64Cu(DOTA-tetramer)/111In(DOTA-tetramer)	10 ± 2
NOTA-dimer	68Ga(NOTA-dimer)	100 ± 3
NOTA-2G3-dimer	68Ga(NOTA-2G3-dimer)	66 ± 4
NOTA-3PEG4-dimer	68Ga(NOTA-2PEG4-dimer)	54 ± 2

It is very important to note that the IC₅₀ value depends largely on the radioligand (¹²⁵I-c(RGDyK) vs. ¹²⁵I-echistatin) and tumor cell lines (U87MG vs. MDA-MB-435) used in the competitive displacement assay. Caution should be taken when comparing the IC₅₀ values of cyclic RGD peptides with those reported in the literature. Whenever possible, a “control compound”, such as c(RGDfK) or c(RGDyK) should be used in each experiment.

Ternary ligand complexes [^99mTc(HYNIC-3PEG₄-dimer)(tricine)(TPPTS)] (^99mTc-3PEG₄-dimer) and [^99mTc(HYNIC-3G₃-dimer)(tricine)(TPPTS)] (^99mTc-3G₃-dimer) have been evaluated in the athymic nude mice bearing U87MG glioma and MDA-MB-435 breast tumor xenografts (118, 119). For comparison purposes, [^99mTc(HYNIC-PEG₄-dimer)(tricine)(TPPTS)] (^99mTc-PEG₄-dimer) and [^99mTc(HYNIC-tetramer)(tricine)(TPPTS)] (^99mTc-tetramer) were also evaluated in the same tumor-bearing animal model (118, 119). As expected, the breast tumor uptake of ^99mTc-3PEG₄-dimer and ^99mTc-3G₃-dimer was comparable to that of ^99mTc-tetramer (Figure 8A), and was >2x higher than that of ^99mTc-PEG₄-dimer (118). These data suggest that 3PEG₄-dimer, 3G₃-dimer and tetramer are most likely “bivalent” (Figure 6B) whereas PEG₄-dimer is monodentate (Figure 6A). If PEG₄-dimer were bivalent, HYNIC-PEG₄-dimer would have shared similar integrin α_vβ₃ binding affinity with HYNIC-3PEG₄-dimer and HYNIC-3G₃-dimer while ^99mTc-PEG₄-dimer would have had the tumor uptake comparable to that of ^99mTc-3PEG₄-dimer and ^99mTc-3G₃-dimer. In addition, ^99mTc-3PEG₄-dimer and ^99mTc-3G₃-dimer had the kidney and liver uptake that was half of that for ^99mTc-tetramer, probably because 3PEG₄-dimer and 3G₃-dimer have only two RGD motifs. Therefore, ^99mTc-3PEG₄-dimer and ^99mTc-3G₃-dimer have significant advantages over ^99mTc-tetramer with respect to T/B ratios (118, 119).

Figure 8.

Top: direct comparison of the tumor uptake for ^99mTc-PEG₄-dimer, ^99mTc-3G₃-dimer, ^99mTc-3PEG₄-dimer and ^99mTc-tetramer in athymic nude mice bearing MDA-MB-435 breast cancer xenografts (A). Bottom: the relationship between the tumor size and tumor uptake (B and C) of ^99mTc-3PEG₄-dimer at 120 min p.i. in athymic nude mice bearing U87MG glioma xenografts. The linear relationship suggests that ^99mTc-3PEG₄-dimer is for monitoring the tumor growth or shrinkage during anti-angiogenic therapy.

Multimeric ≠ Multivalent

It is critical to note that multimeric RGD peptides are not necessarily multivalent. There are two factors (bivalency and enhanced local RGD concentration) contributing to high integrin α_vβ₃ binding affinity of multimeric RGD peptides. The concentration factor exists in all multimeric RGD peptides regardless of the linkers. The key for bivalency is the distance between two cyclic RGD motifs. For example, this distance in 3PEG₄-dimer (38 bonds) and 3G₃-dimer (26 bonds) is long enough for them to achieve bivalency, which leads to the higher integrin α_vβ₃ binding affinity of HYNIC-3PEG₄-dimer and HYNIC-3G₃-dimer than that of HYNIC-PEG₄-dimer, and the much higher breast tumor uptake (Figure 8A) for ^99mTc-3PEG₄-dimer and ^99mTc-3G₃-dimer as compared to that of ^99mTc-PEG₄-dimer. In contrast, the concentration factor is responsible for the better binding affinity of HYNIC-tetramer than that of HYNIC-3PEG₄-dimer and HYNIC-3G₃-dimer. The fact that the breast tumor uptake of ^99mTc-3PEG₄-dimer and ^99mTc-3G₃-dimer is comparable to that of ^99mTc-tetramer suggests that the contribution from the “concentration factor” might not be as significant as that from the “bivalency factor”. In addition, the ability of a multimeric RGD peptide to achieve bivalency also depends on the integrin α_vβ₃ density. If the tumor integrin α_vβ₃ density is high, the distance between two neighboring integrin α_vβ₃ sites will be short, which makes it easier for the multimeric RGD peptide to achieve bivalency. If the integrin α_vβ₃ density is very low, the distance between two neighboring integrin α_vβ₃ sites will be long, and it might be more difficult for the same multimeric RGD peptide to achieve simultaneous integrin α_vβ₃ binding.

Relationship between Tumor Size and Radiotracer Tumor Uptake

The ability to quantify the integrin α_vβ₃ in vivo provides opportunities to select the patients more appropriately for anti-angiogenic treatment and to monitor the therapeutic efficacy of integrin α_vβ₃-positive tumors (120, 121). The %ID tumor uptake reflects the total integrin α_vβ₃ level while the %ID/g tumor uptake reflects the integrin α_vβ₃ density. Figure 8B shows the relationship between tumor size and tumor uptake (%ID and %ID/g) of ^99mTc-3PEG₄-dimer. There was a linear relationship between tumor size and %ID tumor uptake with R² = 0.9164 (Figure 8B), suggesting that ^99mTc-3PEG₄-dimer might be useful for monitoring tumor growth during anti-angiogenic therapy (118). If the tumor uptake is expressed as %ID/g (Figure 8C), it seems that ^99mTc-3PEG₄-dimer has a narrow window to achieve an optimal tumor uptake. When tumor size is small (<0.05 g), there is little angiogenesis with low blood flow, and ^99mTc-3PEG₄-dimer has low %ID/g tumor uptake. When the tumor size is 0.1 g–0.25 g, the microvessel density and integrin α_vβ₃ density are high. The %ID/g tumor uptake of ^99mTc-3PEG₄-dimer is ~10 %ID/g (Figure 8C). As tumors grow, the total integrin α_vβ₃ level is higher, and the %ID tumor uptake increases (Figure 8B). However, the microvessel density decreases due to maturity of blood vessels. The integrin α_vβ₃ density also decreases due to larger interstitial space (122). In addition, parts of the tumor may become necrotic, leading to the lower integrin α_vβ₃ density. As a result, larger tumors have lower %ID/g tumor uptake than smaller ones (Figure 8C).

Radiotracer Tumor Uptake and Tumor Cell Integrin α_vβ₃ Expression

Figure 9 compares the tumor uptake of ^99mTc-3PEG₄-dimer and ^99mTc-3G₃-dimer in athymic nude mice bearing U87MG glioma and HT29 colon cancer xenografts, and the integrin α_vβ₃ expression levels on U87MG glioma and HT29 cells. Both ^99mTc-3PEG₄-dimer and ^99mTc-3G₃-dimer have significantly higher uptake (%ID/g) in the glioma than HT29 tumors (Figure 9: top), which is supported by the presence of higher level of integrin α_vβ₃ expression on U87MG glioma cells than those on HT29 cells (Figure 9: bottom). These data clearly show that the integrin α_vβ₃ level on tumor cells (U87MG > HT29) plays a significant role in the radiotracer tumor uptake. More fluorescent staining studies are needed to better quantify the contributions from the integrin α_vβ₃ expressed on the tumor cells and neovasculature in the tumor tissue.

Figure 9.

Top: direct comparison of the tumor uptake for ^99mTc-3PEG₄-dimer and ^99mTc-3G₃-dimer in athymic nude mice bearing U87MG glioma and HT29 colon cancer xenografts. Bottom: confocal microscope images of U87MG glioma (left) and HT29 (right) tumor cells. The blue color indicates the presence cell nuclei (the presence of 4′,6′-diamidino-2-phenylindole). The red color indicates the presence of integrin α_vβ₃ due to the binding of LM609 and TRITC-coupled goat-anti-mouse IgG. These data show that the integrin α_vβ₃ level on tumor cells (U87MG > HT29) plays a significant role in the radiotracer tumor uptake.

⁶⁴Cu(DOTA-3PEG₄-dimer) and ⁶⁴Cu(DOTA-3G₃-dimer)

Recently, Shi et al (123) reported two cyclic RGD dimer conjugates: DOTA-3PEG₄-dimer and DOTA-3G₃-dimer (Figure 10). It was found that the integrin α_vβ₃ binding affinity (Table 1) follow the order of DOTA-tetramer (IC₅₀ = 10 ± 2 nM) > DOTA-3G₃-dimer (IC₅₀ = 74 ± 3 nM) ~ DOTA-3PEG₄-dimer (IC₅₀ = 62 ± 6, nM) > DOTA-dimer (IC₅₀ = 102 ± 5 nM) against ¹²⁵I-echistatin bound to U87MG glioma cells. Once again, bivalency is likely responsible for the higher integrin α_vβ₃ binding affinity of DOTA-3PEG₄-dimer and DOTA-3G₃-dimer than that of DOTA-dimer. This conclusion is completely consistent with the higher tumor uptake of ⁶⁴Cu(DOTA-3PEG₄-dimer) and ⁶⁴Cu(DOTA-3G₃-dimer) than that of ⁶⁴Cu(DOTA-dimer) (123). In contrast, the concentration factor is most likely responsible for higher integrin α_vβ₃ binding affinity of DOTA-tetramer than that of DOTA-3PEG₄-dimer and DOTA-3G₃-dimer, and the higher initial tumor uptake of ⁶⁴Cu(DOTA-tetramer) as compared to that of ⁶⁴Cu(DOTA-3PEG₄-dimer) and ⁶⁴Cu(DOTA-3G₃-dimer) (132). However, the uptake of ⁶⁴Cu(DOTA-3PEG₄-dimer) in the liver and kidneys was significantly lower than that reported for ⁶⁴Cu(DOTA-tetramer) (57), due to the presence of four R-residues in E[E[c(RGDfK)]₂]₂ as compared to only two R-residues in ⁶⁴Cu(DOTA-3PEG₄-dimer) and ⁶⁴Cu(DOTA-3G₃-dimer) (132).

Figure 10.

Structures of DOTA-3G₃-dimer and DOTA-3PEG₄-dimer.

¹¹¹In(DOTA-3PEG₄-dimer) and ¹¹¹In(DOTA-3G₃-dimer)

To explore the impact of radiometal chelates, ¹¹¹In(DOTA-3PEG₄-dimer) and ¹¹¹In(DOTA-3G₃-dimer) were evaluated in the same animal model (124). ¹¹¹In(DOTA-3PEG₄-dimer) and ¹¹¹In(DOTA-3G₃-dimer) share the same DOTA chelator, have almost identical lipophilicity (log P = −4.13 ± 0.08 and −4.20 ± 0.21, respectively), and show very similar metabolic stability (124). The tumor uptake of ¹¹¹In(DOTA-3G₃-dimer) is comparable to that of ¹¹¹In(DOTA-3PEG₄-dimer) (Figure 11A). Planar imaging studies showed that they both had very high tumor uptake with a long tumor retention and excellent tumor-to-background contrast (124). The activity accumulation in the chest and abdominal regions almost completely disappeared at 60 min p.i. The combination of the hydrophilic ¹¹¹In(DOTA) chelate with PEG₄/G₃ linkers is responsible for the low liver uptake for ¹¹¹In(DOTA-3PEG₄-dimer) and ¹¹¹In(DOTA-3G₃-dimer), leading to high tumor/liver ratios (Figure 11A). Both ¹¹¹In(DOTA-3PEG₄-dimer) and ¹¹¹In(DOTA-3G₃-dimer) also had low kidney and muscle uptake with very high tumor/kidney and tumor/muscle ratios (124). Their integrin α_vβ₃-specificity has been clearly demonstrated by the blocking experiment, in which ¹¹¹In(DOTA-3PEG₄-dimer) was used as the radiotracer and E[c(RGDfK)]₂ as the blocking agent (124). The RGD-specificity of ¹¹¹In-labeled cyclic RGD dimers was demonstrated by the higher integrin α_vβ₃ affinity of DOTA-3PEG₄-NS (IC₅₀ = 715 ± 45 nM; 3PEG₄-NS = PEG₄-E[PEG₄-(RGKfD)]₂, a scrambled nonsense peptide) than that of DOTA-3PEG₄-dimer (1.3 ± 0.3 nM), and the much better tumor uptake of ¹¹¹In(DOTA-3PEG₄-dimer) (10.06 ± 3.52 %ID/g) than that of ¹¹¹In(DOTA-3PEG₄-NS) (0.30 ± 0.09 %ID/g) in the same animal model. On the basis of the integrin α_vβ₃- and RGD-specificity of the ¹¹¹In-labeled cyclic RGD peptides, it has been suggested that their uptake in several normal organs (e.g. intestine, kidneys, liver, lungs and spleen) might be also integrin α_vβ₃-mediated (124).

Figure 11.

A: comparison of tumor uptake and tumor/liver ratios for ⁶⁴Cu(DOTA-3PEG₄-dimer), ⁶⁴Cu(DOTA-3G₃-dimer), ¹¹¹In(DOTA-3PEG₄-dimer) and ¹¹¹In(DOTA-3G₃-dimer) in athymic nude mice bearing U87MG human glioma xenografts. B: Chem-3D presentation of ⁶⁴Cu(DOTA-monoamide) (6-coordinated) and ¹¹¹In(DOTA-monoamide) (7-coordinated).

¹¹¹In(DOTA-3PEG₄-dimer) and ⁶⁴Cu(DOTA-3PEG₄-dimer) share the same DOTA-conjugate. In spite of their difference in radiometal, the tumor uptake ¹¹¹In(DOTA-3PEG₄-dimer) (10.89 ± 2.55 and 7.65 ± 3.17 %ID/g at 30and 240 min p.i., respectively) was close to that of ⁶⁴Cu(DOTA-3PEG₄-dimer) (8.23 ± 1.97 and 6.43 ± 1.22 %ID/g at 30 and 240 min p.i., respectively). They also share similar uptake in normal organs. For example, the kidney uptake of ¹¹¹In(DOTA-3PEG₄-dimer) was 5.80 ± 0.95 and 2.78 ± 0.20 %ID/g at 30 and 240 min p.i., respectively, and was comparable to that of ⁶⁴Cu(DOTA-3PEG₄-dimer) (6.59 ± 0.93 %ID/g at 30 min p.i. and 2.81 ± 0.36 % ID/g at 240 min p.i.). The liver uptake of ¹¹¹In(DOTA-3PEG₄-dimer) is 2.52 ± 0.57 %ID/g at 30 min and 1.61 ± 0.06 %ID/g at 240 min p.i. while ⁶⁴Cu(DOTA-3PEG₄-dimer) has the liver uptake of 2.80 ± 0.35 %ID/g at 30 min p.i. and 1.87 ± 0.51 %ID/g at 240 min p.i. Similar conclusion can be made by comparing biodistribution properties of ⁶⁴Cu(DOTA-3G₃-dimer) and ¹¹¹In(DOTA-3G₃-dimer). These data suggest that the radiometal chelate (Figure 11B) has minimal impact on radiotracer tumor uptake and excretion kinetics.

⁶⁸Ga(NOTA-2PEG₄-dimer) and ⁶⁸Ga(NOTA-2G₃-dimer)

We also prepared conjugates NOTA-2PEG₄-dimer and NOTA-2G₃-dimer (Figure 12: NOTA = 1,4,7-triaazacyclononane-1,4,7-tetracetic acid) and their ⁶⁸Ga complexes, ⁶⁸Ga(NOTA-2PEG₄-dimer) and ⁶⁸Ga(NOTA-2G₃-dimer) (125). The integrin α_vβ₃ binding affinity (Table 1) of NOTA-dimer (IC₅₀ = 100 ± 3 nM), NOTA-2G₃-dimer (IC₅₀ = 66 ± 4 nM) and NOTA-2PEG₄-dimer (IC₅₀ = 54 ± 2 nM) were very close to those for DOTA-dimer (IC₅₀ = 102 ± 5 nM), DOTA-3G₃-dimer (IC₅₀ = 74 ± 3 nM) and DOTA-3PEG₄-dimer (IC₅₀ = 62 ± 6 nM), respectively. These data further suggest that the G₃ and PEG₄ linkers between cyclic RGD motifs in dimeric RGD peptides make it possible for them to bind integrin α_vβ₃ in a bivalent fashion. The tumor uptake of ⁶⁸Ga(NOTA-2G₃-dimer) and ⁶⁸Ga(NOTA-2PEG₄-dimer) was much higher than that of ⁶⁸Ga(NOTA-dimer) in the same tumor-bearing animal model (125), suggesting that the addition of G₃ and PEG₄ linkers between two cyclic RGD motifs increases the radiotracer tumor uptake. In all cases, the tumors can be clearly visualized as early as 30 min p.i. with excellent contrast (Figure 12). It was also found that the tumor uptake of ⁶⁸Ga(NOTA-2PEG₄-dimer) in MDA-MB-435 breast tumor was significantly lower than that in U87MG glioma. Similar results were also obtained for ¹⁸F-labeled 3PEG₄-dimer (126), which is consistent with that fact that the MDA-MB-435 breast tumors have lower integrin α_vβ₃ expression than U87MG glioma (52–55).

Figure 12.

Top: structures of NOTA-2G₃-dimer and NOTA-2PEG₄-dimer. Bottom: microPET images for ⁶⁸Ga(NOTA-2G₃-dimer) and ⁶⁸Ga(NOTA-2PEG₄-dimer). Arrows indicate the presence of tumors.

Conclusions

Over the last several years, many multimeric cyclic RGD peptides have been used to increase the integrin α_vβ₃–targeting capability. It was found that increasing the peptide multiplicity can significantly enhance their integrin α_vβ₃ binding affinity, and improve the radiotracers tumor targeting ability. However, as peptide multiplicity increases, the uptake of radiolabeled multimeric RGD peptides is also significantly increased in normal organs. There is no significant advantage in using radiolabeled tetramers E{E[c(RGDxK)]₂}₂ (x = f and y) over their dimeric analogs E[c(RGDxK)]₂ (x = f and y) with respect to T/B ratios. In addition, the cost for E{E[c(RGDxK)]₂}₂ (x = f and y), E{E[E[c(RGDyK)]₂]₂}₂ is too high for them to be useful for future radiotracer development.

Recent studies on cyclic RGD dimers suggest that two factors (bivalency and enhanced local RGD concentration) contribute to the high integrin α_vβ₃ binding affinity of multimeric cyclic RGD peptides (118, 119, 123–126). The concentration factor exists in all multimeric RGD peptides regardless of the linker length between two cyclic RGD motifs. To achieve bivalency, the distance between two RGD motifs must be long enough for them to bind the neighboring integrin α_vβ₃ sites simultaneously. Among the cyclic RGD peptides, 2PEG₄-dimer/3PEG₄-dimer and 2G₃-dimer/3G₃-dimer show most promising results with respect to the tumor uptake and T/B ratios of their radiotracers (^99mTc, ¹¹¹In, ⁶⁴Cu and ⁶⁸Ga). Thus, 2PEG₄-dimer/3PEG₄-dimer and 2G₃-dimer/3G₃-dimer are better integrin α_vβ₃-targeting biomolecules than the tetramer E{E[c(RGDfK)]₂}₂ with respect to their cost and the T/B ratios of their ^99mTc, ¹¹¹In, ⁶⁴Cu and ⁶⁸Ga radiotracers.

While current research efforts on the integrin α_vβ₃–targeted radiotracers have been focused on new cyclic RGD peptides, the formulation development for routine preparation of radiotracers is often neglected. It must be emphasized that it is common to use post-labeling chromatographic separation for improvement of radiotracer purity and specific activity for research purposes. In clinical settings, however, post-labeling purification is not practical. Regardless of the beauty of the science involved in the discovery of a new radiotracer, its success relies largely on the availability and capability to improve the quality of cancer patient’s life. As a matter of fact, the main challenge for [¹⁸F]Galacto-RGD and [¹⁸F]-AH111585 to assume a wide-spread clinical utility is not their biological performance, but their clinical availability at reasonable cost. In this respect, the integrin α_vβ₃-targeted ^99mTc radiotracers will offer significant advantages over their corresponding ¹⁸F analogs because of the clinical availability of ⁹⁹Mo-^99mTc generators, and the kit formulation for routine preparation of ^99mTc radiotracers at low cost. However, both planar imaging and SPECT suffer a significant drawback with respect to quantification of the radiotracer “absolute” organ uptake, the speed of dynamic imaging, spatial resolution and tissue attenuation.

The successful application of the ⁶⁸Ga-labeled somatostatin analogs for imaging tumors in cancer patients has clearly demonstrated that ⁶⁸Ga is an excellent alternative to ¹⁸F. ⁶⁸Ga is available from an in-house commercially available ⁶⁸Ge/⁶⁸Ga generator (127–129), and its short half-life is best suited for the fast excretion kinetics of many ⁶⁸Ga-labeled small peptides. ⁶⁴Cu is another viable alternative to ¹⁸F. The use of NOTA and its derivatives as BFCs allows the development of the kit formulation for routine preparation of ⁶⁸Ga and ⁶⁴Cu radiotracers with high specific activity. In addition, ⁶⁸Ga(NOTA) and ⁶⁴Cu(NOTA) chelates have very high hydrophilicity, which is extremely important for improving the radiotracer clearance kinetics from non-cancerous organs.

It is very important to note that integrin α_vβ₃ is also overexpressed on the activated endothelial cells during wound healing and post-infarct remodeling, in rheumatoid arthritis, and atherosclerotic plaque (1–4, 130, 131). The integrin α_vβ₃–targeted radiotracers have been proposed for imaging myocardial angiogenesis (132), inflammatory diseases (133), and hindlimb ischemia (134). Recent results showed that the ¹¹¹In-labeled nonpeptide integrin α_vβ₃ antagonist (RP748) was able to image the angiogenesis in the heart with myocardial infarction (132), and the radiotracer uptake in the infarct region was associated with the integrin α_vβ₃ expression level. The results reported by Pichler et al suggest that [¹⁸F]Galacto-RGD might be a powerful tool to distinguish between acute and chronic phases of T-cell mediated immune responses (133). Studies have also demonstrated the value of a ^99mTc-labeled cyclic RGD monomer (NC100692) for imaging the integrin α_vβ₃ in rodent models of hindlimb ischemia (134). These promising results suggest that the newer and more effective integrin α_vβ₃–targeted radiotracers under development for tumor imaging might become valuable non-invasive markers of angiogenesis after ischemic injury, myocardial infarction and inflammation. In addition, the combination of high tumor uptake, long tumor retention with favorable pharmacokinetic of ¹¹¹In(DOTA-3G₃-dimer) and ¹¹¹In(DOTA-3PEG₄-dimer) suggests that their corresponding ⁹⁰Y and ¹⁷⁷Lu analogs, M(DOTA-RGD) (M = ⁹⁰Y and ¹⁷⁷Lu; and RGD = 3G₃-dimer and 3PEG₄-dimer), might be useful for the treatment of integrin α_vβ₃-positive tumors.