Forum Article
`
`pubs.acs.org/IC
`
`Underscoring the Influence of Inorganic Chemistry on Nuclear
`Imaging with Radiometals
`Brian M. Zeglis, Jacob L. Houghton, Michael J. Evans, Nerissa Viola-Villegas, and Jason S. Lewis*
`Department of Radiology and the Program in Molecular Pharmacology and Chemistry, Memorial Sloan-Kettering Cancer Center,
`New York, New York, United States
`
`ABSTRACT: Over the past several decades, radionuclides have matured from largely
`esoteric and experimental
`technologies to indispensible components of medical
`diagnostics. Driving this transition, in part, have been mutually necessary advances in
`biomedical engineering, nuclear medicine, and cancer biology. Somewhat unsung has
`been the seminal role of inorganic chemistry in fostering the development of new
`radiotracers. In this regard, the purpose of this Forum Article is to more visibly highlight
`the significant contributions of inorganic chemistry to nuclear imaging by detailing the development of five metal-based imaging
`agents: 64Cu-ATSM, 68Ga-DOTATOC, 89Zr-transferrin, 99mTc-sestamibi, and 99mTc-colloids. In a concluding section, several
`unmet needs both in and out of the laboratory will be discussed to stimulate conversation between inorganic chemists and the
`imaging community.
`
`■ INTRODUCTION
`
`Over the past 3 decades, nuclear imaging modalities have
`revolutionized clinical medicine, particularly cardiology, neu-
`rology, and oncology.1,2 Indeed, the ability of positron emission
`tomography (PET) and single photon emission computed
`tomography (SPECT) to provide functional and biochemical
`information about tissues to complement the anatomical maps
`provided by other imaging modalities has proven vital in the
`diagnosis and management of disease. The advent of molecular
`imaging has in large part been due to remarkable advances in
`biomedical engineering, medical physics, halogen radiochemistry,
`and cancer biology. Yet the critical role of inorganic chemistry in
`the rise of nuclear imaging has often become lost in the margins.
`In the following pages, we will seek to remedy this oversight. We
`will first discuss the intersection of
`inorganic chemistry,
`radiochemistry, and nuclear imaging in general terms. Then, at
`greater length, we will use five particularly effective or promising
`metal-based imaging agents as case studies both to illustrate the
`fundamental role of inorganic chemistry in the development of
`radiopharmaceuticals and to more visibly celebrate the
`contributions of inorganic chemistry to nuclear imaging.
`Why Use a Metallic Radioisotope? Before we delve any
`deeper into our discussion, we must first answer one simple
`question: “Why use a metallic radioisotope?” This question be-
`comes especially important when considering that PET imaging
`∼
`is largely dominated by a radiohalogen, fluorine-18 (18F, t1/2
`109.8 min). The answer is straightforward: radiometals provide
`flexibility, modularity, and facility unmatched by other imaging
`isotopes.
`First, the wide variety of metallic radionuclides allows for the
`precise tailoring of the physical half-life of the radioisotope to
`the biological half-life of the targeting vector (Figure 1). For
`example, agents with short in vivo residence times can be
`∼ 68 min) or technetium-
`labeled with gallium-68 (68Ga; t1/2
`∼ 6 h), while vectors that require longer
`99m (99mTc; t1/2
`amounts of time to reach their target can be labeled with
`
`∼ 14.7 h),
`∼ 12.7 h), yttrium-86 (86Y; t1/2
`copper-64 (64Cu; t1/2
`∼ 2.8 days), or zirconium-89 (89Zr;
`indium-111 (111In; t1/2
`∼ 3.2 days) (Figure 1 and Tables 1 and 2).3−7
`t1/2
`Second, the simplicity and modularity of using different
`bifunctional chelators and radiometals facilitate the creation of a
`wide variety of imaging agents. For example, with relative ease,
`the same antibody can be conjugated to the chelators
`desferrioxamine (DFO), diethylenetriaminepentaacetic acid
`(DTPA), and 1,4,7,10-tetraazacyclododecane-1,4,7,10-tetraacetic
`acid (DOTA) for labeling with 89Zr for PET imaging, 111In for
`SPECT imaging, or lutetium-177 (177Lu) for radioimmunotherapy.
`In some cases, particularly with the versatile chelators DOTA,
`DTPA, and 1,4,7-triazacyclononane-1,4,7-triacetic acid (NOTA),
`the radiometal may be exchanged without changing the chelator at
`all. Either way, this modularity becomes especially clinically useful
`when an imaging agent labeled with one isotope can be used as a
`companion diagnostic tool for a therapeutic agent bearing another.8
`Third, generally speaking, radiometalation reactions are rapid
`and can be achieved under mild conditions. Purification
`procedures are also quite simple, typically involving cation-
`exchange chromatography or reverse-phase C18 cartridges. It is
`in this area that radiometals likely offer the greatest advantage
`over radiohalogens because probes bearing the latter often
`require multistep syntheses, harsh reaction conditions, and
`complicated purifications.
`Fourth, many radiometalsfor example,
`86Y,
`89Zr, and
`111Inare known to residualize inside cells following the
`uptake of their vector, resulting in increased retention of the
`radioactivity inside the target
`tissue and higher tumor-to-
`background activity ratios than nonresidualizing radiohalogens
`such as 18F, iodine-124 (124I), and bromine-76 (76Br).9
`
`Special Issue: Imaging and Sensing
`
`Received:
`June 25, 2013
`Published: December 6, 2013
`
`© 2013 American Chemical Society
`
`1880
`
`dx.doi.org/10.1021/ic401607z | Inorg. Chem. 2014, 53, 1880−1899
`
`Petitioner GE Healthcare – Ex. 1011, p. 1880
`
`

`

`Inorganic Chemistry
`
`Forum Article
`
`Figure 1. Illustration of the variety of metals with isotopes suitable for nuclear imaging. Elements with isotopes suitable for PET are color-coded
`blue, and elements with isotopes suitable for SPECT are color-coded red. The shading corresponds to half-life, with longer half-lives darker and
`shorter half-lives lighter. Elements with multiple shadings have multiple isotopes suitable for imaging.
`
`Table 1. Physical Properties of Some Common PET Radiometalsa
`
`common
`coordination
`numbers
`4, 5, 6
`
`4, 5, 6
`
`8, 9
`
`8
`
`3+
`
`4+
`
`89Zr
`
`78.4
`
`cyclotron
`
`89Y(p,n)89Zr
`
`395.5(11)
`
`22.74(24)
`
`isotope
`64Cu
`
`half-life/h
`12.7
`
`source
`cyclotron
`
`production
`reaction
`64Ni(p,n)64Cu
`
`68Ga
`
`1.1
`
`generator
`
`68Ge/68Ga
`
`86Y
`
`14.7
`
`cyclotron
`
`86Sr(p,n)86Y
`
`decay mode
`(% branching ratio)
`ε + β+ (61.5)
`β+ (17.6)
`β− (38.5)
`ε + β+ (100)
`β+ (89.1)
`ε + β+ (100)
`β+ (31.9)
`
`Eβ+/keV
`278.2(9)
`
`abundance,
`Iβ+/%
`17.60(22)
`
`Eγ/keV
`(intensity, Iγ/%)
`511.0 (35.2)
`
`relevant
`oxidation states
`1+, 2+
`
`836.02(56)
`
`87.94(12)
`
`511.0 (178.3)
`
`3+
`
`535(7)
`
`11.9(5)
`
`443.1 (16.9)
`511.0 (64)
`627.7 (36.2)
`703.3 (15)
`777.4 (22.4)
`1076.6 (82.5)
`1153.0 (30.5)
`1854.4 (17.2)
`1920.7 (20.8)
`ε + β+ (100)
`511.0 (45.5)
`β+ (22.7)
`909.2 (99.0)
`aUnless otherwise stated, standard deviations are given in parentheses (IT = isomeric transition; ε = electron capture).208
`
`Table 2. Physical Properties of Some Common SPECT Radiometalsa
`
`isotope
`67Ga
`
`half-life/h
`78.2
`
`source
`cyclotron
`
`production
`reaction
`natZn(p,x)67Ga
`68Zn(p,2n)67Ga
`
`decay mode (% branching
`ratio)
`ε (100)
`
`Eγ/keV
`91.265(5)
`93.310(5)
`184.576(10)
`208.950(10)
`300.217(10)
`393.527(10)
`140.511(1)
`
`abundance,
`Iγ/%
`3.11(4)
`38.81(3)
`21.410(10)
`2.460(10)
`16.64(12)
`4.56(24)
`89.06
`
`99mTc
`
`6.0
`
`generator
`
`99Mo/99mTc
`
`111In
`
`67.3
`
`cyclotron
`
`111Cd(p,n)111In
`
`β− (0.0037)
`IT (99.9963)
`ε (100)
`
`90.7(9)
`171.28(3)
`94.1(10)
`245.35(4)
`aUnless otherwise stated, standard deviations are given in parentheses (IT = isomeric transition; ε = electron capture).208
`
`relevant oxidation
`states
`3+
`
`common coordination
`numbers
`4, 5, 6
`
`1− to 7+
`
`3+
`
`4, 5, 6
`
`5, 6, 7, 8
`
`Finally, yet no less critically, metallic radioisotopes present a
`tremendous opportunity to expand the availability of imaging
`agents beyond hospitals with nearby cyclotron facilities because
`
`many radiometals can be produced via portable generator
`systems (e.g., 68Ga and 99mTc) or possess physical half-lives
`long enough such that
`they can be shipped to research
`
`1881
`
`dx.doi.org/10.1021/ic401607z | Inorg. Chem. 2014, 53, 1880−1899
`
`Petitioner GE Healthcare – Ex. 1011, p. 1881
`
`

`

`Inorganic Chemistry
`
`laboratories and hospitals without excessive decay (e.g., 64Cu,
`111In, and 89Zr).
`Production and Purification of Radiometals. The first
`step in the synthesis of a radiometal-based imaging agent is
`production of
`the radiometal
`itself. Radiometals can be
`produced via three distinct
`routes: decay of
`longer-lived
`radionuclides in a generator, nuclear bombardment reactions
`in a cyclotron, or nuclear bombardment reactions in a nuclear
`reactor (see Tables 1 and 2). 68Ga, for example, is formed via
`electron capture decay of its parent radionuclide, germanium-
`68 (68Ge), and thus can be produced using a compact, cost-
`effective, and convenient 68Ge/68Ga generator system. 64Cu,
`in contrast, can be produced either on a nuclear reactor
`[via the 63Cu(n,γ)64Cu or 64Zn(n,p)64Cu reaction] or,
`far
`more commonly, by use of a biomedical cyclotron via the
`64Ni(p,n)64Cu reaction.10 As an aside, it is important to note
`that each of these isotopes emits radiation other than the posi-
`trons and photons useful for imaging. Some of these emissions,
`such as the variety of high-energy photons from 86Y and the
`909 keV photon from 89Zr, require special consideration with
`regard to handling, shielding, and dosimetry.11
`Yet the process does not end with the creation of the desired
`radiometal. The radiometal must be purified from its parent
`isotope and other byproducts of the nuclear reaction and
`isolated in a useful form prior to its incorporation into an
`imaging agent. Here lies the first point of intersection between
`inorganic chemistry and radiochemistry.
`86Y, for example, is most often produced via the 86Sr(p,n)86Y
`reaction by the proton bombardment of [86Sr]-enriched SrCO3
`or SrO targets on a cyclotron. A variety of different techniques
`have been employed to separate the 86Y3+ cation from the
`target and byproducts, including cation-exchange chromatog-
`raphy, cation-exchange chromatography followed by coprecipi-
`tation with LaIII or FeIII, and chromatography using Sr-selective
`resins.12,13 Recently, a particularly effective and economical
`isolating 86Y using electrolysis has been
`method for
`developed.14 After irradiation of a [86Sr]-enriched SrO target
`coated onto a platinum disk, the entire target is dissolved in
`nitric acid with NH4NO3 as an electrolyte. This solution is then
`placed in an electrochemical cell
`in which two successive
`rounds of electrolysis are employed to separate 86Y from
`residual Sr via electrodeposition on a platinum-wire electrode.
`This 86Y-coated platinum wire electrode can then be removed
`from the cell and washed with EtOH and HNO3. This solution
`can then be evaporated and reconstituted in 0.1 M HCl to yield
`86Y3+ in very high specific activity and radionuclidic purity.
`Importantly, this method also allows for the efficient recycling
`of the expensive, isotopically enriched 86Sr target material.
`In another example, 89Zr is produced via the 89Y(p,n)89Zr
`reaction by proton bombardment of a solid 89Y target on a
`cyclotron.15,16 In order to produce an aqueous 89Zr4+ species
`suitable for radiolabeling reactions, the solid target is first
`dissolved with 6 M HCl. Yet this process produces aqueous
`89Zr4+ and 89Y3+ species that must be separated. To this end, the
`HCl solution is run through a hydroxamate resin that has high
`affinity for 89Zr4+ and very low affinity for 89Y3+, thus completely
`sequestering the 89Zr4+ cations while allowing the 89Y3+ cations
`89Zr4+ is
`removed from the
`to pass
`through. Finally,
`hydroxamate resin using an eluent of oxalic acid, producing a
`purified solution of 89Zr4+ that can be employed in radiolabeling
`reactions.
`
`Forum Article
`
`Aqueous Coordination Chemistry of Some Common
`Radiometals. Prior to our discussion of metal-based imaging
`agents, a brief discussion of
`the underlying aqueous
`coordination chemistry of the radiometals is in order. For
`more detail, the reader can consult other excellent and more
`exhaustive reviews, chief among them a 2010 Chemical Reviews
`article from Wadas et al.3−5,11,17−21
`To begin, four isotopes of copper have been used for PET
`imaging: 60Cu (t1/2 = 0.4 h; β+ yield = 93%; Eβ+ = 3.9 and
`3.0 MeV), 61Cu (t1/2 = 3.32 h; β+ yield = 62%; Eβ+ = 1.2 and
`1.15 MeV), 62Cu (t1/2 = 0.16 h; β+ yield = 98%; Eβ+ = 2.19 MeV),
`and, most notably, 64Cu (t1/2 = 12.7 h; β+ yield = 19%; Eβ+ =
`0.656 MeV).22 Of course, the chemistry of each is identical. CuII is
`the most biologically relevant oxidation state of the metal. Because
`of its electronic structure, the 3d9 cation typically forms square-
`planar four-coordinate, square-pyramidal or trigonal-bipyramidal
`five-coordinate, or octahedral six-coordinate complexes.21,23
`However, coordinatively saturating six-coordinate ligands have
`generally proven the chelators with the best in vivo perform-
`ance.3,24 Cu2+ is neither a particularly hard nor soft cation, so an
`effective chelator will almost always feature a mixture of uncharged
`nitrogen donors along with anionic oxygen or sulfur donors in
`order to neutralize the 2+ charge of the cation. While DOTA has
`been used as a chelator for Cu2+, the Cu-DOTA complex has been
`shown to be unstable in vivo, often producing elevated levels of
`radiocopper uptake in the liver as a result of demetalation.
`Alternatively, other macrocyclic ligands with smaller or cross-
`bridged cavities,
`such as NOTA (N3O3) or 4,11-bis-
`(carboxymethyl)-1,4,8,11-tetrazabicyclo[6.6.2]hexadecane-4,11-di-
`acetic acid (CB-TE2A; N4O2), have been shown to be excellent
`chelators of the radiometal.25−27 More recently, neutral N6
`macrocyclic chelators based on sarcophagine scaffolds have been
`shown to be extremely adept at chelating the cation.28,29 The
`in vivo reduction of copper from CuII to CuI is possible under
`some circumstances. In most cases, this reduction is undesirable,
`and macrocyclic complexes of CuII generally have reduction
`potentials far below the threshold for in vivo reduction. However,
`in some situations, as we shall see in the Cu-ATSM case study,
`the reduction of CuII to CuI is a critical step in the biological
`mechanism of the tracer.
`Moving on, the only stable oxidation state of gallium in an
`aqueous environment is 3+. The amphoteric nature of Ga3+
`allows for reactions in acidic and alkaline solutions. At pH > 3,
`insoluble Ga(OH)3 precipitates out of aqueous solutions, but
`−] at pH > 7.4.30
`this species redissolves to soluble [Ga(OH)4
`However, on the radiochemical scale, the formation of insoluble
`Ga(OH)3 has been shown to be inconsequential if the overall
`radiometal concentration is kept below ∼2.5 × 10−6 M.30−32
`The Ga3+ cation is smaller and harder than the Cu2+ cation and
`thus typically binds ligands containing multiple anionic oxygen
`donors.33,34 While some tetrahedral
`four-coordinate and
`square-pyramidal five-coordinate complexes are known, octahe-
`dral six-coordinate complexes are far more common. A variety
`of acyclic and macrocyclic chelators have been used with Ga3+,
`with N,N′-ethylenedi-L-cysteine (EC; N2S2O2), N,N′-bis(2-
`hydroxybenzyl)ethylenediamine-N,N′-diacetic acid (HBED;
`N2O4), NOTA (N3O3), 1,4,7-trismercaptoethyl-1,4,7-triaza-
`cyclononane (TACN-TM; N3S3), and DFO (O6) forming par-
`ticularly stable complexes.35−38 As we will discuss later, while
`DOTA has been employed quite often with 68Ga3+, the chelator
`does not form a particularly stable complex with the cation.39 In-
`deed, in this regard, Ga3+ provides an excellent example of the
`importance of the cavity size of macrocyclic chelators. While
`
`1882
`
`dx.doi.org/10.1021/ic401607z | Inorg. Chem. 2014, 53, 1880−1899
`
`Petitioner GE Healthcare – Ex. 1011, p. 1882
`
`

`

`Inorganic Chemistry
`
`NOTA binds the cation exceptionally tightly (log K = 30.1; pM =
`26.4),
`the larger cavities of
`its cousins DOTA (log K =
`21.3; pM = 15.2) and triethylenetetramine (TETA) (log K =
`19.7; pM = 14.1) make for far less stable complexes.40,41
`Not surprisingly, the chelation chemistry of indium is similar
`indium’s only stable
`to that of gallium. Like its congener,
`aqueous oxidation state is 3+.34,42 However, the In3+ cation is
`larger, has a higher pKa, and exhibits faster water exchange rates
`than its Ga3+ counterpart.34 As a result, In3+ is more tolerant of
`ligands bearing softer thiolate donors and can adopt higher
`coordination numbers than its group 13 neighbor. In part be-
`cause of this flexibility, In3+ has been shown to form complexes
`with a variety of different coordination numbers and geo-
`metries. These include a five-coordinate trigonal-bipyramidal
`complex with tris(2-mercaptobenzyl)amine (NS3 + an
`exogenous ligand), a six-coordinate distorted octahedral com-
`plex with EC (N2S2O2), a six-coordinate distorted octahedral
`complex with NOTA (N3O3), a seven-coordinate pentagonal-
`bipyramidal complex with ethylenediaminetetraacetic acid (EDTA;
`N2O4 + an exogenous ligand), and an eight-coordinate square-
`antiprismatic complex with DTPA (N4O4).36,43−48 In practice,
`however, the vast majority of 111In-labeled bioconjugates have
`employed bifunctional derivatives of DTPA or DOTA.34,42,49−52
`The biologically relevant oxidation state of yttrium is also 3+.
`However, the Y3+ cation is much larger than either Ga3+ or In3+,
`allowing it to form complexes with coordination numbers up to
`8 or 9. Despite its large size, the Y3+ cation is considered to be a
`hard Lewis acid, and thus ligands with multiple anionic oxygen
`donors are usually employed for its chelation. When a ligand
`offers fewer than eight donors, exogenous ligands fill
`the
`cation’s coordination sphere, as in its eight-coordinate distorted
`dodecahedral complex with EDTA (N2O4 + two H2O ligands)
`and nine-coordinate monocapped square-antiprismatic complex
`with 1,4,7-tris(carbamoylmethyl)-1,4,7-triazacyclononane
`(N3O3 + two H2O ligands).43,53−55 Not surprisingly, however,
`it has been shown that
`ligands capable of coordinatively
`saturating the metal form more stable complexes. As a result,
`the two chelators most often used in 86Y-labeled radiopharma-
`ceuticals both offer eight donors: DOTA forms an eight-
`coordinate square-antiprismatic complex with a Kd of ∼22,
`while DTPA forms an eight-coordinate monocapped square-
`antiprismatic complex with a Kd of ∼24.49,50,52,56−60
`As a group IV metal, zirconium exists predominantly in the
`4+ oxidation state in aqueous solution. The aqueous chemistry
`of the Zr(H2O)x species can be quite complex, with both
`speciation between various mononuclear and polynuclear states
`and solubility highly dependent on the pH.61−63 With regard to
`chelation chemistry, however, things simplify somewhat. The
`cation is relatively large, and its high charge makes it a very hard
`Lewis acid. As a result, Zr4+ displays a very strong preference for
`ligands offering anionic oxygen donors in high coordination
`numbers. For example, Zr4+ has been shown to make
`octadentate, dodecahedral complexes with the well-known
`chelators DTPA (N3O5), EDTA (N2O4 + two H2O ligands),
`and DOTA (N4O4).64,65 Interestingly, however, while the
`thermodynamic stability constants for both Zr-EDTA (∼29)
`and Zr-DTPA (∼36) have been shown to be quite high, the
`poor kinetic stability of these complexes has rendered them
`unsuitable for use in vivo.53,58,66 Instead, the vast majority of, if
`not all, published 89Zr-labeled radiotracers have employed DFO
`as the chelator.67−70 DFO is an acyclic siderophore-derived
`molecule that binds 89Zr4+ using three hydroxamate groups,
`thus providing three neutral and three anionic oxygen ligands.
`
`Forum Article
`
`To date, neither a solid state nor an NMR structure has been
`determined for Zr-DFO, although density functional theory
`(DFT) calculations suggest that a seven- or eight-coordinate
`complex is formed involving exogenous water molecules in
`addition to the ligand’s six oxygen donors.71
`Finally,
`the chemistry of
`technetium represents a fairly
`significant departure from the radiometals we have discussed so
`far. As a group VIIB metal with a neutral electronic
`configuration of [Kr]4d65s1, the coordination chemistry of
`99mTc is very complex: a large number of oxidation states (1−
`to 7+) and a wide variety of coordination geometries (square-
`pyramidal, octahedral, and heptahedral) are possible.7,18,31,72−74
`This diversity is a double-edged sword:
`it allows
`for
`construction of a range of different 99mTc species, but it also
`gives rise to ample redox chemistry and chemically labile
`species that complicate the design of imaging agents.75
`−,
`Upon elution from the generator as tetrahedral 99mTcO4
`99mTc exists in a 7+ state that is not immediately useful for
`chelation or binding directly to small molecules because of its
`negligible chemical reactivity.18 Indeed, there are very few
`examples of the incorporation of TcVII into imaging agents, with
`99mTc-sulfur-colloid (Tc2S7) standing as
`the only major
`example.31,76 Rather, the vast majority of 99mTc-based imaging
`agents are prepared using 99mTc in a lower oxidation state. As a
`result, a reducing agent or the direct reduction of the metal
`through complexation with hard ligands is necessary in the
`synthesis of these probes.7,31,75
`Not surprisingly, the different oxidation states of technetium
`have different coordination chemistries. TcV is a d2 metal center
`that,
`in aqueous environments,
`typically forms either five-
`coordinate square-pyramidal or
`six-coordinate octahedral
`complexes around a TcVO core or six-coordinate octahedral
`complexes around a TcVO2 core. Ligands featuring donors
`ranging from neutral phosphorus and sulfur atoms to anionic
`oxygen atoms have been employed, although tetradentate
`chelators based on mercaptoacetylglycylglycylglycine, diamine-
`dithiol, or aminoaminedithiol scaffolds have proven most
`common.77,78 Complexes based on technetium(V) nitrido
`cores and the condensation reaction between the TcVO center
`and hydrazinonicotinamide have also been explored as
`alternative TcV coordination strategies.79,80 Unfortunately,
`however, much of the work with TcV cores has ultimately led to
`complexes that are unstable or preparations that are too
`cumbersome for clinical translation. For example, the 99mTcVO
`core is relatively common in radiopharmaceuticals, but these
`complexes are often labile at the trans position or are hydrolytically
`unstable when exposed to physiological environments.31
`Low-spin TcIII d4 complexes have also been studied as
`alternatives to TcV-based constructs. The TcIII center has been
`shown to make both six- and seven-coordinate complexes with
`types.81,82
`ligands
`featuring a variety of different donor
`However, the relatively harsh reducing conditions currently
`employed to form TcIII
`from pertechnetate represent a
`significant obstacle to its routine use.
`Recently, many of the most successful developments have
`centered on 99mTcI, particularly complexes based on the
`kinetically inert, low-spin [99mTc(CO)3]+ d6 core.83,84 Water-
`soluble [99mTc(CO)3(H2O)3]+ can be prepared easily from
`99mTc-pertechnetate under reducing conditions, and the H2O
`ligands are easily exchanged with various types of
`ligands,
`including tris(pyrazoyl)methane derivatives and click-chemistry-
`derived scaffolds.85−90 The lipophilicity of [Tc(CO)3]+ remains
`dx.doi.org/10.1021/ic401607z | Inorg. Chem. 2014, 53, 1880−1899
`
`1883
`
`Petitioner GE Healthcare – Ex. 1011, p. 1883
`
`

`

`Inorganic Chemistry
`
`somewhat of a concern, however. TcI is a relatively soft cation,
`and ligands bearing softer donors
`tend to increase the
`lipophilicity of the complex further. Thus, chelation systems
`must be chosen carefully in order to strike a suitable balance
`between stability and lipophilicity.
`Regardless of the identity of the metal, synthesis on the
`radiochemical scale has a few critical features that set it apart
`from the macroscale synthesis of “cold” complexes. The limited
`amount of time allowed for synthesis and purification is the
`most obvious difference, because reaction and purification
`conditions must often be designed with the half-life of the
`radionuclide in mind. A less apparent difference is the strikingly
`low absolute concentration of radiometals in most radiolabeling
`reactions. Generally, the concentration of radiometal is at least 3
`(and often more) orders of magnitude lower than that of any
`other reactants in a radiochemical reaction. This contrasts
`dramatically with the excess of metal typically employed in
`macroscale reactions that aim to achieve the best possible
`chemical yield. For this reason, during radiosynthesis reactions,
`any potential contaminants, particularly metals that may compete
`with the radiometal of interest, become a major concern.
`Design and Structure of Radiometal-Based Imaging
`Agents. From a design perspective, radiometalated imaging
`agents can be grouped into three classes:
`small metal
`complexes, chelator-based conjugates, and colloids. Small-
`metal-complex radiotracers are the most structurally straightfor-
`ward class, comprised of
`two essential parts: a central
`radiometal and a set of coordinating ligands. These agents
`represent the purest points of intersection between inorganic
`chemistry and nuclear
`imaging,
`for
`the metal complexes
`themselves are solely responsible for in vivo targeting, uptake,
`and retention. A number of small-metal-complex PET and
`SPECT imaging agents have had a significant impact in the
`including 99mTc-bisphosphonates for bone imaging,
`clinic,
`99mTc-sestamibi for myocardial perfusion imaging, and 64Cu-
`PTSM for blood perfusion imaging.6
`Chelator-based conjugates, on the other hand, have four
`parts: a targeting vector, a radiometal, a chelator, and a linker
`connecting the chelator and targeting vector.4,5,7 The targeting
`vector is typically a biomolecule such as a peptide, protein, or
`antibody. However, synthetic vectors such as nanoparticles and
`liposomes have come into vogue in recent years. The selection
`of a radiometal is governed by both the imaging modality and
`the biological half-life of the targeting vector. The most pop-
`ular radiometals for SPECT imaging are 111In and 99mTc, and
`the most popular radiometals for PET imaging are 68Ga, 64Cu,
`86Y, and 89Zr. However, a variety of other metallic radioisotopes
`including gallium-67 (67Ga), copper-60 (60Cu), titanium-45
`(45Ti), and technetium-94m (94mTc) have also been produced
`and used. Once an imaging modality has been chosen,
`matching the radioactive half-life of the isotope to the biological
`half-life of the biomolecule is critical. For example, 68Ga and
`99mTc would not be ideal choices for labeling antibodies
`because the radionuclides would decay significantly before the
`antibody reaches its optimal concentration at
`the target.
`Conversely, neither 89Zr nor 111In would be the best choice
`for labeling a short peptide because their multiday half-lives
`would far exceed the residence time of the peptidic agent.
`The job of the chelatorinterestingly, from the Greek χηλή
`(che̅le̅) meaning “claw”is simple: form a kinetically inert and
`thermodynamically stable complex with the radiometal in order
`to prevent its inadvertent release in vivo. Radiometal chelators
`
`Forum Article
`
`fall into two structural classes: macrocylic and acyclic chelators.
`While macrocyclic chelators typically offer greater thermody-
`namic stability, acyclic chelators usually have faster rates of
`metal binding.18 Generally, transition-metal chelators offer at
`least four (and usually six or more) coordinating atoms arrayed
`in a configuration that suits the preferred geometry of the metal
`in question. As we have discussed above, different metals prefer
`different chelators, and therefore the choice of chelator is
`dictated by the identity of the radiometal.
`For the linkage between the chelator and targeting vector,
`the only requirements are that the link must be stable under
`physiological conditions and must not significantly compromise
`the binding strength or specificity of the vector. The specific
`chemical nature of the conjugation method is dependent on
`both the type of vector and the availability of bifunctional
`variants of the desired chelator. For vectors with free thiol
`groups, the reaction between a thiol and a maleimide has
`proven a popular route; for vectors with free amine groups, the
`formation of thiourea bonds using isothiocyanates or peptide
`bonds using activated carboxylic acids has been widely
`employed. It is important to remember, however, that the
`conjugation of a chelator to a vector may alter its ability to
`coordinate a given radiometal. For example, conjugating DOTA
`to a peptide using one of its carboxylate arms leaves only a
`three-armed DOTA, more properly termed DO3A,
`for
`chelation of
`the radiometal.
`In light of
`this,
`the use of
`bifunctional chelators with pendant conjugation handles, e.g.,
`[S-2-(aminobenzyl)1,4,7-triazacyclononane-1,4,7-triacetic acid
`(p-NH2Bn-NOTA) or N-(2-aminoethyl)-trans-1,2-diaminocy-
`clohexane-N,N′,N″-pentaacetic acid (CHX-A″-DTPA), is often
`preferable.
`The third class of radiometal-based imaging agents, colloids,
`is the oldest of the three, yet it boasts only one prominent
`example: the family of 99mTc-radiocolloids.91−93 Nevertheless,
`99mTc-radiocolloids have had a profound impact on the clinical
`imaging of
`the reticuloendothelial system (RES). Broadly
`speaking, colloids are particles that range in size from 1 nm to 4
`μm. In the body, they are typically removed from circulation via
`phagocytosis, a process especially active in macrophages.
`Consequently, when radiolabeled, they can be used to image
`tissues with high concentrations of macrophages, such as the
`liver, spleen, bone marrow, and lymph nodes. As a result,
`99mTc-radiocolloids have proven especially important in the
`imaging of the lymphatic system in oncology. 99mTc-colloids of
`a wide range of diameters have been created using a variety of
`materials,
`including denatured human albumin, sulfur, anti-
`mony, and stannous phytate. Somewhat surprisingly,
`the
`literature contains very few allusions to the use of other
`imaging agents.94 A more detailed
`radiometals in colloidal
`discussion of the synthesis and application of 99mTc-colloids can
`be found in the last of the five case studies.
`
`■ CASE STUDIES
`In the following pages, our hope is to use five metal-based radio-
`pharmaceuticals64Cu-ATSM, 68Ga-DOTATOC, 89Zr-transferrin,
`99mTc-sestamibi, and 99mTc-colloidas lenses to illustrate the
`fundamental role of inorganic chemistry in the development of both
`well-established and next-generation nuclear imaging agents. Taken
`together, we believe that these vignettes will provide both a sound
`overview of the different ways inorganic chemistry influences
`radiopharmaceuticals and an arena for the celebration of the integral
`contributions of
`inorganic chemistry to nuclear imaging, while
`
`1884
`
`dx.doi.org/10.1021/ic401607z | Inorg. Chem. 2014, 53, 1880−1899
`
`Petitioner GE Healthcare – Ex. 1011, p. 1884
`
`

`

`Inorganic Chemistry
`
`simultaneously pointing out areas in which inorganic chemistry
`could play a role moving forward.
`
`■ CU-ATSM: TARGETING TUMOR PHENOTYPE WITH
`A SMALL METAL COMPLEX
`The first PET imaging agent we will discuss is the hypoxia-
`targeting small-metal-complex copper(II) diacetylbis(N4-methyl-
`thiosemicarbazone), more commonly referred to as Cu-ATSM.95−97
`Structurally, Cu-ATSM is a relatively simple metal complex: a CuII 3d9
`metal center coordinated in a square-planar geometry by two
`nitrogen atoms and two sulfur atoms of a tetradentate
`bis(thiosemicarbazone) ligand (Figure 2A). Cu-ATSM has
`
`Figure 2. (A) Structures of hypoxia-selective Cu-ATSM and
`nonselective Cu-PTSM. (B) Possible mechanistic scheme for the
`uptake and retention of Cu-ATSM in hypoxic cells.
`
`been radiolabeled and studied with all four positron-emitting
`radioisotopes of copper: 60Cu, 61Cu, 62Cu, and 64Cu. Regardless
`of the isotope, however, Cu-ATSM is prepared through the
`simple incubation of CuCl2 and the free ligand H2ATSM and
`purified using a reverse-phase C18 cartridge.
`Background and In Vitro Characterization. As its name
`suggests, the term “hypoxia” describes the pathological condition in
`which a tissue is deprived of normal physiological levels of oxygen.
`Under normal conditions, the mean arterial partial pressure of
`oxygen (pO2) is 70−100 mmHg. In cancerous tissues, however, the
`erratic and disorganized vasculature of the growing tumor often
`in many cases to <10 mmHg
`results in dramatic reductions in pO2
`and occasionally to the point of complete anoxia (0 mmHg)with
`dangerous consequences for the patient.98 Hypoxia is associated not
`only with significant resistance to radiation therapy but also with
`resistance to chemotherapies,
`increased tumor aggressiveness,
`increased metastatic potential, and higher rates of recurrence.99,100
`Given these relationships, the development of nuclear imaging tools
`for the noninvasive delineation of tumor hypoxia in vivo has been
`an incredibly imp