Radioimmunoimaging with longer-lived positron-emitting radionuclides: potentials and challenges

Abstract

Radioimmunoimaging and therapy has been an area of interest for several decades. Steady progress has been made towards clinical translation of radiolabeled monoclonal antibodies for diagnosis and treatment of diseases.

Tremendous advances have been made in imaging technologies such as positron emission tomography (PET). However, these advances have so far eluded routine translation into clinical radioimmunoimaging applications due to the mismatch between the short half-lives of routinely used positron-emitting radionuclides such as ¹⁸F versus the pharmacokinetics of most intact monoclonal antibodies of interest. The lack of suitable positron-emitting radionuclides that match the pharmacokinetics of intact antibodies has generated interest in exploring the use of longer-lived positron emitters that are more suitable for radioimmunoimaging and dosimetry applications with intact monoclonal antibodies.

In this review, we examine the opportunities and challenges of radioimmunoimaging with select longer-lived positron-emitting radionuclides such as ¹²⁴I, ⁸⁹Zr and ⁸⁶Y with respect to radionuclide production, ease of radiolabeling intact antibodies, imaging characteristics, radiation dosimetry and clinical translation potential.

Introduction and Background

Applications of radiation in medicine have been described now for over 100 years. The use of radiation in medicine branches from many scientific discoveries, most notably the discovery of x-rays in 1895 and its use in surgery (1). Since that time frame, over the next 60 years critical advances in nuclear medicine technology and instrumentation have resulted in methodologies and technologies for the visualization of many of the body’s organs, including liver and spleen scanning, brain tumor localization, and studies of the gastrointestinal tract by the injection of radionuclides (2, 3). The greatest potential use of radiation in medicine is its utility to provide diagnostic information of pathological processes before the outset of structural changes in an organ. In these applications, very small amounts of radioactive material most often labeled or conjugated to “smart” targeting agents such as antibodies, peptides, and small molecules are introduced into the body. These “smart” agents specifically target individual cells instead of just the general tissues or organs, and therefore provide more valuable information about actual pathology. One example of such “smart” targeting agents is the monoclonal antibody (mAb). Currently, 21 monoclonal antibodies (all intact) are approved by the U.S Food and Drugs Administration (U.S.F.D.A) for diagnosis and treatment of various illnesses.

Antibodies in Nuclear Medicine

Currently radiolabeled antibodies are clinically used for numerous applications such as oncology and cardiology (4, 5). Use of radiolabeled antibodies for targeting specific organs in animals have been explored almost half a century ago (6, 7). Radiolabeled antibodies have been used in the clinic for therapeutic and diagnostic purposes now for over 40 years (8, 9). In 1978, Goldenberg and colleagues successfully applied the principles of antibody-antigen binding by visualizing carcinoembryonic antigen (CEA) on tumors of patients with a history of cancer of diverse histopathology by injecting ¹³¹I labeled goat IgG targeting CEA. However, the large scale application of radiolabeled antibodies in the clinic was hindered due to low production yields and concerns of toxic immune reactions after injecting antibodies of animal origin into humans. The introduction of hybridoma technology for monoclonal antibody (mAb) production as developed by Kohler and colleagues along with the evolution of recombinant DNA technology has addressed many of these problems (10). Chimeric and humanized monoclonal antibodies and even completely human antibodies are now the standard. Numerous clinical studies have since been reported describing the use of radiolabeled intact antibodies for diagnosis of cancer using γ-scintigraphy and single photon emission tomography (SPECT) imaging (4, 11–13).

In spite of great successes in pre-clinical animal models, the promise of radioimmunoimaging by γ-scintigraphy has not fully lived up to expectations, mostly due to differences in biodistribution and pharmacokinetic characteristics between animals and humans, and limitations of γ-scintigraphy in terms of intrinsic spatial resolution (14). Radionuclides that decay with γ-energies lower than 100 keV produce too much scatter, while γ-energies over 250 keV are difficult to collimate and therefore pose a challenge for quantitative γ-scintigraphy. To overcome the disadvantages posed by γ-scintigraphy attempts were made to exploit the superiorities of positron emission tomography (PET) for radioimmunoimaging. A single positron decay results in two 511 keV photons being emitted at 180 degrees. Most PET cameras contain a circular array of detectors with coincidence circuits designed to capture 511 keV photons emitted in opposite direction and therefore offer much better resolution and counting efficiency as compared to conventional γ-scintigraphy and SPECT cameras (15). However, a major limitation challenging PET radioimmunoimaging was that the half-life of most of the routinely used PET radionuclides such as ¹⁸F (t_1/2 = 1.8 h) and ¹¹C (t_1/2 = 0.3 h). The half-lives of routinely used PET radionuclides simply did not match well with the biological half-lives (up to several days) and pharmacokinetic parameters of slowly localizing intact antibodies. To overcome the initial blood pool uptake and slow localization of the intact antibody, antibody fragments (Fab) were labeled with ¹⁸F for PET imaging. However, ¹⁸F labeled antibody fragments failed to demonstrate high tumor localization as demonstrated by parent intact antibodies (16, 17). ¹⁸F labeled antibody fragments were cleared from the tumor relatively more quickly than intact antibodies and then rapidly excreted through the urinary route delivering high radiation doses to the kidneys (16, 17). Numerous attempts have been made to modify intact antibodies to increase tumor uptake and residence time, and concomitantly decrease kidney uptake for successful PET with ¹⁸F (18, 19); however, to date results from these strategies have demonstrated low to moderate success. Another drawback of using ¹⁸F as a choice of radionuclide for PET radioimmunoimaging is the cumbersome chemistry involved which requires synthesis of ¹⁸F labeled intermediate precursors for indirect labeling of antibodies with ¹⁸F (17, 20)

As an alternative to the approach of modifying the antibody to match the physical characteristics of ¹⁸F, the use of longer-lived positron emitters has been explored for radioimmunoimaging with intact antibodies (21–23). The half-lives of longer-lived positron emitters such as ¹²⁴I and ⁸⁹Zr (shown in Table 1) have a much closer match with the biological half-life of most intact antibodies. Antibodies labeled with longer-lived positron emitters can be imaged for 2–5 days after injection when the radioactivity in the blood pool has been cleared and target to background ratio is higher. In this report, we have examined the potentials and challenges of selected longer-lived positron-emitting radionuclides for radioimmunoimaging with intact antibodies. The criteria applied for selecting these longer-lived positron emitters (Table 1) included previously published reports with antibodies or clinical studies, a potentially suitable half-life (10 hr–140 h) versus biological half-life, an imageable positron emission, available or potentially available conjugation and radiolabeling chemistry, and reasonable or acceptable production logistics and radiation toxicity.

Table 1.

Characteristics of selected longer-lived positron-emitting radionuclides for radioimmunoimaging. Some of the data presented in the table were obtained from references (4, 159) and http://www.nndc.bnl.gov/nudat2/ (Nuclear Structure and Decay Data Searchable Database, National Nuclear Data Center, Brookhaven National Laboratory, USA) accessed on 03/07/2008.

Radionuclide	Half-life	Production	Daughter (half-life)	β+max in MeV (β+ yields)	γ-energies in MeV (yields)	Mean range (mm)	Intrinsic spatial resolution loss (mm)	Human studies
18F	1.83 h	20Ne(d, α)18F 18O(p, n)18F	18O (stable)	0.63 (97 %)	0.14 (41 %)	0.69	0.7	Yes
124I	100.2 h	124Te(p, n)124I 124Te(d, 2n)124I 125Te(p, 2n)124I	124Te (stable)	2.14 (24 %)	0.60 (61 %)	3.25	2.3	Yes
76Br	16.0 h	76Se(p, n)76Br 75As(3He, 2n)76Br	76Se (stable)	3.98 (56 %)	0.55 (74 %)	5.07	5.3	Yes
52Mn	134.2 h	52Cr(p, n)52Mn 51V(3He, 2n)52Mn 52Cr(3He,t)52Mn	52Cr (stable)	0.58 (29 %)	0.74 (90 %)	5.00	0.6	Yes
55Co	17.5 h	54Fe(d, n)55Co 56Fe(p, 2n)55Co	55Fe ( 2.6 y)	1.50 (76 %)	0.48 (20 %)	5.74	1.6	Yes
66Ga	9.4 h	66Zn(p, n)66Ga 63Cu(4He,n)66Ga	66Zn (stable)	4.15 (56 %)	1.03 (36.9 %)	8.06	5.8	Yes
72As	25.9 h	72Ge(p, n)72As	72Ge (stable)	3.32 (88 %)	0.83 (80 %)	5.01	3.6	Yes
64Cu	12.7 h	64Ni(p, n)64Cu 68Zn(p, n)64Cu 64Ni(d,2 n)64Cu	64Ni,64Zn (stable)	0.66 (18 %)	-	0.70	0.7	Yes
86Y	14.7 h	86Sr(p, n)86Y natRb(3He, 2n)86Y	86Sr (stable)	3.15 (34 %)	1.08 (83 %)	2.46	1.8	Yes
89Zr	78.4 h	89Y(p, n)89Zr 89Y(d, 2n)89Zr	89mY (16 s)	0.90 (23 %)	-	1.18	1.0	Yes

Selected longer-lived positron-emitting radionuclides for radioimmunoimaging

Numerous longer-lived positron-emitting radionuclides can be produced by cyclotrons, accelerators, and reactors. Selected longer-lived positron emitters ranging from halogens to metals are described below with respect to their production, feasibility of use in radiolabeling an intact antibody, biological applications and imaging characteristics.

Iodine-124

Iodine (atomic radius of 1.3 Å and electronegativity of 2.66) belongs to Group VII of the periodic table. ¹²⁴I is a positron-emitting radionuclide with a complex decay scheme associated with numerous high energy γ-emissions (0.6 MeV, 61 % abundance) and high-energy positron emissions (2.14 MeV, 24 % abundance) as shown in Table 1. ¹²⁴I (t_1/2= 100.2 h) decays to a stable 124Te daughter. The relatively longer half-life makes ¹²⁴I conducive for labeling intact antibodies for PET imaging over several days after administration. Additionally, the vast experience of radiolabeling proteins with the other Iodine radionuclides, i.e. ¹³¹I, ¹²⁵I, and ¹²³I, promotes a significant level of confidence since existing chemistry and protocols are directly applicable for use with ¹²⁴I.

Production of ¹²⁴I

¹²⁴I can be produced using a conventional medical cyclotron using the ¹²⁴Te(p,n)¹²⁴I and ¹²⁴Te(d,2n)¹²⁴I reactions [Table 2]. Typically, highly enriched tellurium dioxide is used as the target and ¹²⁴I is isolated using dry distillation with minimal impurities (24–26). The yield (<0.1mCi/μAh) from the ¹²⁴Te(p,n)¹²⁴I reaction is low at 15 MeV (27). Alternatively, ¹²⁴I can be produced through a ¹²⁵Te(p,2n)¹²⁴I reaction with about four times higher yields than the commonly used ¹²⁴Te(p,n)¹²⁴I and ¹²⁴Te(d,2n)¹²⁴I reactions (24, 27). However, the production of ¹²⁴I through the ¹²⁵Te(p,2n)¹²⁴I reaction requires a medium-high energy cyclotron and impurities are relatively higher than the ¹²⁴Te(p,n)¹²⁴I reaction.

Table 2.

Commonly used production routes of selected longed-lived positron-emitting radioimmunoimaging and corresponding yields and radionuclidic impurities (adapted from reference (119)).

Radionuclide	Commonly used production route	Energy range (MeV)	Calculated yield (MBq/μAh)	Radionuclidic impurity
124I	124Te(p, n)124I 124Te(d, 2n)124I	12 → 8 14 → 10	16 17.5	125I (0.1 %) 125I (1.7 %)
76Br	76Se(p, n)76Br 75As(3He, 2n)76Br	15 → 8 18 → 10	360 11	- 77Br (1.6 %)
52Mn	52Cr(p, n)52Mn 52Cr(3He, t)52Mn	20 → 10 36→ 10	0.4 5.6	54Mn (0.5 %) 54Mn (0.8 %)
55Co	54Fe(d, n)55Co 58Ni(p, n)55Co	10 → 5 15 → 7	32 14	56Co, 57Co (0.1 %) 57Co (0.5 %)
66Ga	66Zn(p, n)66Ga	13 → 8	433	-
72As	72Ge(p, n)72As	18 → 8	93	71As (10 %)
64Cu	64Ni(p,n)64Cu 68Zn(p, n)64Cu	12 → 9 35 → 20	240 100	61Cu (0.4 %) 67Cu (25 %)
86Y	86Sr(p, n)86Y	14 → 10	400	87mY (3 %)
89Zr	89Y(p, n)89Zr	12 → 6	43	88Zr (0.1 %)

Radiolabeling of antibodies with ¹²⁴I

Radio-iodinated antibodies have been used for in vivo applications for more than 60 years (6, 7, 28, 29). In the early 1960’s, Hunter and co-workers described a method (Chloramine-T method) to prepare high specific-activity radio-iodinated antibodies using p-toluene sulfonochloramide [Figure 1(a)] (30, 31). Briefly, the Chloramine-T method is an oxidative method which involves exposure of the substrate to Chloramine-T in the presence of NaI for a short time, producing high specific activity proteins labeled with carrier-free radioiodine. However, the major disadvantage of Choramine-T method is the risk of oxidation of thiol groups and protein denaturation due to the presence of high concentrations of strong oxidizing agent which can compromise the intended biological use of the antibody. As an alternative, an enzymatic method using lactoperoxidase as a catalyst was developed for iodination of antibodies (32, 33). Lactoperoxidase catalyzes the oxidation of iodide using hydrogen peroxide as the enzyme substrate and is a milder oxidative agent than Chloramine-T. Due to the continuing concerns of protein denaturation and loss of biological activity by oxidizing agents, a newer technique (Bolton-Hunter method [Figure 1(b)]) was later developed using iodinated 3-(4-hydroxyphenyl)propionic acid N-hydroxysuccinimide ester, which reacts with free amino groups in the protein molecule to attach the radioiodine-labeled groups by amide bonds (34, 35). The Bolton-Hunter method is a non-oxidative technique which is less harsh on the proteins than alternative methods. Similarly, to avoid the interaction of oxidizing agent with the protein, an alternative solid-phase oxidation using 1,3,4,6-tetrachloro-3 alpha,6 alpha-biphenyl glycoluril [Figure 1(c)] was developed and is commonly called as the Iodogen method (36). All of the above methods suffered from the drawback of in vivo deiodination in presence of enzymes. To overcome in vivo deiodination, another method was developed using N-succinimidyl 4-iodobenzoate (PIB) [Figure 1(d)] to specifically label antibodies and proteins (37). This method produced a more stable radio-iodinated antibody as compared to methods like Bolton-Hunter as PIB does not contain an o-phenol moiety which protects PIB derivatives from the deiodinases that can act to delabel o-iodophenols that are formed with the other radio-iodination methods. As a result of increased in vivo stability, the uptake in the thyroid was dramatically reduced when a comparison was made between the same antibody radioiodinated using the PIB and Chloramine-T methods (37). Many monoclonal antibodies are internalized via either clathrin dependent or independent pathways. Antibodies rapidly internalized (within 2–4 hours) via the clathrin-dependent endocytosis pathway are catabolized within lysosomes. Iodotyrosine is known to rapidly exit from the lysosome and the cell after catabolism and as result target to background ratios are poor (38–40). In order to overcome the issue of catabolism of conventional radio-iodinated antibodies, dilactitol-tyramine (DLT) and radioiodinated diethylenetriaminepentaacetic acid-appended peptides have successfully been used to residualize the radio-iodinated antibody within the cells (41–44). As a result of decreased catabolism, the tumor uptake of cell surface binding radio-iodinated antibodies was significantly higher than antibodies radiolabeled with Chloramine-T method (41, 42).

Figure 1.

General scheme and reagents of radio-halogenation of intact antibodies using commonly used methods (a) Chloramine-T method, (b) Bolton-hunter method, (c) Iodogen method and (d) N-succinimidyl 4-iodobenzoate based method.

Detailed methodologies for radio-iodination of proteins can be found in a review article published in Bioconjugate Chemistry (45). Currently, most ¹²⁴I labeled antibodies are prepared using commercial iodination kits based on some of the above described techniques.

Biological studies with ¹²⁴I labeled antibodies

In the early 1990’s, murine monoclonal antibody H17E2 recognizing placental alkaline phosphatase (PLAP) was radiolabeled with ¹²⁴I using the Iodogen method for targeting PLAP on HEp2 human tumor xenografts (21). The ¹²⁴I labeled H17E2 localized in the tumor for at least 7 days demonstrating the feasibility of using monoclonal antibodies labeled with ¹²⁴I for tumor localization studies (21). The utility of antibodies labeled with ¹²⁴I for PET imaging, imaging feasibility, and quantification studies were studied by Pentlow and co-workers (22). In this study, when compared to ¹⁸F, the spatial resolution was only slightly degraded while the linearity was the same (22). This technique was translated to in vivo application by measurements of human neuroblastoma tumors in rats which had been injected with ¹²⁴I labeled 3F8 antibody demonstrating that quantitative PET imaging of ¹²⁴I labeled antibodies was possible in biological systems. Use of ¹²⁴I in PET radioimmunoimaging was further demonstrated by c-erb B2 quantification and visualization in tumor xenografts for up to 160 hours using ¹²⁴I labeled rat monoclonal antibody (ICR12) recognizing the external domain of the human c-erb B2 proto-oncogene product (23). With advances in imaging instrumentation and technology, whole-animal PET studies were performed for non-invasive measurements of tumor vascular endothelial growth factor (VEGF) in animal model using ¹²⁴I-SHPP-VG76e (46). Similarly, ¹²⁴I labeled engineered antibodies have been evaluated for in vivo imaging (47, 48).

Numerous applications of ¹²⁴I in humans for PET imaging have since been reported (49–51). PET with anti-VEGF ¹²⁴I-HuMV833 was conducted on twenty patients with progressive solid tumors with moderate success (52). A recently published clinical study successfully used a carbonic anhydrase-IX targeted ¹²⁴I-cG250 antibody to accurately identify clear-cell renal carcinoma and to predict the aggressiveness of the disease, information that could be valuable in treatment decision making (53).

Challenges

¹²⁴I is available from numerous academic centers and is also available commercially through sources such as IBA Molecular, but tends to be expensive. Current concerns remain high radiation dose and degraded image quality from high-energy γ-emissions, loss of spatial resolution as a result of high-energy positron [Table 1]. The image quality is further degraded due to decrease in tumor/background ratio as a result of intracellular catabolism of ¹²⁴I. Typically radioiodide localizes in the thyroid delivering high radiation doses and therefore potassium iodide solution has to be infused before protein labeled with ¹²⁴I infusion to block thyroid uptake.

Bromine-76

Bromine (atomic radius of 1.1 Å and electronegativity of 2.96) also belongs to the Group VII of the periodic table. ⁷⁶Br has a fairly high abundance (56%) of high-energy β⁺ emission (3.9 MeV) and decays to a stable ⁷⁶Se daughter as shown in Table 1. The half-life of 16.2 h also makes ⁷⁶Br one of the most attractive radionuclides for PET imaging using intact antibodies that exhibit slow clearance. The positron abundance of ⁷⁶Br is almost double that of ¹²⁴I [Table 1] and the C-Br bond is stronger than the C-I bond in organic compounds therefore making it less susceptible to in vivo dehalogenation, thereby potentially minimizing concerns regarding background activity and associated toxicity (54). Due to the smaller atomic radius and greater electronegativity, ⁷⁶Br has greater polarity than ¹²⁴I. The increased polarity of ⁷⁶Br can alter biological properties and pharmacokinetics of the biomolecule which can either be unfavorable or favorable depending upon the objective and purpose of the study (55, 56).

Production of ⁷⁶Br

⁷⁶Br can be prepared using a low-intermediate energy medical cyclotron. Typically, ⁷⁶Br is produced through the ⁷⁶Se(p,n)⁷⁶Br reaction in which a highly enriched ⁷⁶Se target is bombarded with 15–16 MeV protons (57). Alternatively, ⁷⁶Br can also be prepared through the ⁷⁵As(³He,2n)⁷⁶Br and ^natBr(p, xn)⁷⁶Kr→⁷⁶Br reactions using high energy (over 50 MeV) protons (58, 59). Purification and recovery of ⁷⁶Br is performed by dry distillation methods. ⁷⁶Br produced through alternative routes contains other radionuclide impurities such as ⁷⁷Br and ⁷⁵Br which may result in lower yields due to separation and purification. The calculated yield for ⁷⁶Br produced through the ⁷⁶Se(p,n)⁷⁶Br reaction is almost 32 times greater than ⁷⁶Br produced through the ⁷⁵As(³He,2n)⁷⁶Br reaction [Table 2]. Another disadvantage of ⁷⁶Br produced through the ⁷⁵As(³He,2n)⁷⁶Br reaction is the ⁷⁷Br ( t_1/2=57.03 h) impurity.

Radiolabeling of antibodies with ⁷⁶Br

Numerous antibodies have been brominated using techniques and approaches similar to iodination such as Chloramine-T, enzymatic, N-succinimidyl para-(tri-methylstannyl)benzoate (60–62). Bromination of antibodies by oxidative methods can be more difficult than iodination due to the lower oxidation potential of bromide. Therefore, more powerful oxidizing agents are used for sufficient yield of bromination, potentially proving deleterious to proteins. Recently, a method was described for indirect radiobromination of antibodies to prevent protein denaturation in presence of strong oxidizing agents (63). In this method, indirect labeling was performed using the p-isothiocyanatobenzene derivative of the [⁷⁶Br]undecahydro-bromo-7,8-dicarba-nido-undecaborate(1−) (⁷⁶Br-NBI) as a precursor molecule. ⁷⁶Br-NBI was prepared using the Chloramine-T method and thereafter the coupling was performed without intermediate purification in an “one-pot” reaction (63). This technique resulted in improved intracellular retention of the radiolabeled antibody and decreased the loss of biological activity caused due to direct bromination with oxidizing agents. Similar indirect radio-bromination efforts have been reported using 4-hydroxyphenylethylmaleimide for site-specific bromination of a cysteine containing variant of the anti-HER2 affibody (64).

Biological studies with ⁷⁶Br labeled antibodies

Proteins have been brominated and evaluated in animals now for over 30 years (60). In the past decade, anti-CEA MAb 38S1 was radio-brominated and evaluated in animals (65, 66). A comparative study using ⁷⁶Br-38S1, ¹⁸FDG and ¹¹C-Met demonstrated the superiority of ⁷⁶Br-38S1 in successful identification and imaging of liver metastases in nude rats bearing human colon carcinoma (67). In recent years, more studies have been reported with ⁷⁶Br labeled intact antibodies and engineered antibodies for the purpose of PET imaging (56, 63, 64, 68). In a recent study, an ED-B fibronectin-binding human antibody derivative (L19-SIP) was labeled with ⁷⁶Br via an enzymatic approach using bromoperoxidase resulting in both radiolabeling yields and immunoreactivity of over 80% (56). This radio-brominated immunoconjugate (standardized uptake value of over 2.4) enabled detailed small-animal PET imaging of tumor neovasculature in F9 tumor-bearing animals (56). However, long retention of radioactivity in blood and very slow renal excretion were also observed, and as a result, the background activity in non-target organs was higher than that reported for the ¹²⁵I-labeled L19-SIP, illustrating the differences between iodinated and brominated antibody (56).

Clinically, small molecules labeled with ⁷⁶Br have been successfully used for in vivo quantitative PET neuroimaging (69, 70); however, few if any radio-brominated antibodies have been successfully used in the clinic.

Challenges

Besides the loss of intrinsic spatial resolution of about 5.3 mm due to high-energy positron (3.98 MeV) [Table 1], one of the major problems with ⁷⁶Br is the availability and large scale production. Enriched ⁷⁶Se is relatively expensive due to low natural abundance (1 mg/~US $6) although target material recovery and recycling could obviate that aspect. This may be one of the contributing factors hindering commercialization of large quantities of ⁷⁶Br, although smaller quantities of ⁷⁶Br are available on request through MDS Nordion and Isotrace Technology. The β⁺ energy of ⁷⁶Br is relatively higher than conventional PET radionuclides which may pose concerns of radiation doses for diagnostic purposes; however the same could be viewed as a potential advantage if actually used for therapeutic applications.

Manganese-52

Manganese (atomic radius of 1.8 Å and electronegativity of 1.55) belongs to Group VIIB of the periodic table. ⁵²Mn (t_1/2=134.2 h) decays to a stable 52Cr daughter. The half-life and suitable imaging characteristics as a result of optimum positron energy of 0.58 MeV [Table 1] makes ⁵²Mn one of the most attractive, under-utilized longer-lived positron emitters for PET imaging of intact antibodies.

Production

⁵²Mn can be produced using the ⁵²Cr(p,n)⁵²Mn reaction by irradiating a high purity Cr target with 16–17 MeV protons or by ⁵¹V(³He,2n)⁵²Mn using a high-energy cyclotron (71–73).

Radiolabeling of antibodies with ⁵²Mn

Although ⁵²Mn seems to be a suitable candidate of radioimmunoimaging, relatively little effort has been made with it to label antibodies. Chelating agents such as dipyridoxyl diphosphate (74), macrocyclic Schiff-base ligands (73), cyclopentadienide (74) and porphyrin derivatives (75) have been suggested for Mn in different oxidation states. However, generation of stable ⁵²Mn(II) labeled radioimmunoconjugates in aqueous conditions remains unclear due to complex redox chemistry.

To date, no peer-reviewed reports have been published describing the production or use of ⁵²Mn labeled antibodies.

Biological studies with ⁵²Mn

Radio-manganese has mostly been used as a cationic perfusion tracer and as a tracer to study the involvement of manganese in degenerative neurological diseases (71, 72, 75).

Challenges

⁵²Mn is one of the positron-emitters that have been understudied in nuclear medicine. Therefore, production, chelation and conjugation chemistry, and biological effects of ⁵²Mn all need to be extensively studied prior to embarking on radioimmunoimaging application studies.

Cobalt-55

Cobalt belongs to the Group VIIIB of the periodic table. Cobalt has an atomic radius of 1.7 Å and electronegativity of 1.88. ⁵⁵Co decays to a longer-lived ⁵⁵Fe daughter (t_1/2 = 2.6 y) mainly by positron emission of 1.5 MeV (76% abundant) with a half-life of 17.5 h [Table 1].

Production of ⁵⁵Co

The ⁵⁴Fe(d,n)⁵⁵Co reaction has been used to obtain ⁵⁵Co in good yields (32 MBq/μAh) and acceptable purity of over 99% using a medical cyclotron [Table 2] (76). Alternatively, ⁵⁵Co can produced through ⁵⁸Ni(p,n)⁵⁵Co on medical cyclotron but yields are less than half that of ⁵⁴Fe(d,n)⁵⁵Co reaction [Table 2]. No-carrier-added ⁵⁵Co can be prepared using the ⁵⁶Fe(p,2n)⁵⁵Co reaction by bombardment with 28 MeV protons using a cyclotron or 40 MeV protons using a linear accelerator (77, 78). Better yields and decreased impurities of ⁵⁶Co are achieved when the reaction is carried out using 28 MeV protons and high-purity target material. Typically, radiochemical separation is performed by solvent extraction techniques (78, 79).

Radiolabeling of antibodies with ⁵⁵Co

In studies performed at Brookhaven National Laboratory, an anti-CEA antibody fragment was labeled with ⁵⁵Co using cyclohexyl EDTA monoanhydride (CDTA-MA), 4-isothiocyanato-trans-1,2,-diaminocyclohexane N,N,N′,N′-tetraacetic acid (4-ICE) and diethylenetriamine pentaacetic acid dianhydride (DTPA-DA) as bifunctional chelating agents (77). The radiolabeling was carried out at pH 5.5 at room temperature with an incubation time ranging from 2–4 hours. All three of the immunoconjugates were obtained in high labeling yields and demonstrated good serum stability after 4 days (77).

Biological studies with ⁵⁵Co labeled antibodies

⁵⁵Co labeled anti-CEA F(ab′)₂ using DTPA-DA and 4-ICE as bifunctional chelating agents was evaluated in athymic mice bearing human colon carcinoma LS174T xenografts as potential PET imaging agent (77). In the same study, to evaluate the feasibility of the use of ⁵⁵Co as a PET tracer, a Jaszczak phantom study was performed on a whole body PET scanner, providing images of satisfactory resolution (77). In vivo biodistribution studies revealed the superiority of 4-ICE as a better ⁵⁵Co chelating agent as compared to DTPA-DA as the tumor uptake of 4-ICE conjugated antibody was 2.8 times with greater than that of DTPA-DA conjugated antibody.

In the past, ⁵⁵Co labeled bleomycin has been evaluated for imaging of tumors in pre-clinical and clinical settings (76, 78, 80). Other applications of ⁵⁵Co include evaluation of renal function (79). However, most applications of ⁵⁵Co in clinic are in the field of neuro-imaging to assess neuronal damage in stroke and traumatic brain injury (81, 82).

Challenges

Since the initial reports of successful use of ⁵⁵Co in radioimmunoimaging in early 1990’s, no peer-reviewed reports appear to have been published in over 15 years demonstrating the use of ⁵⁵Co in radioimmunoimaging. Issues surrounding appropriate chelation chemistry should be addressable extrapolating from the prior reports and use of 1,4,7,10-tetraaminocyclododecane-N,N,N′,N′-tetraacetic acid (DOTA) derivatives might also prove useful in this application. The dissociation kinetics of Co(II)-DOTA complexes have been suggested to be very slow at neutral and slightly acidic pH and therefore may offer better in vivo stability than DTPA complexes (83, 84). Cobalt has been used to determine the mean number of DOTA chelating groups conjugated per antibody using a metal-binding assay (85, 86). Size considerations might also direct one to investigate NOTA. This is an area of research that remains unsettled. Another potential challenge might be the radiation dose from the ⁵⁶Co impurity and the generated presence of the longer-lived ⁵⁵Fe daughter (t_1/2=2.6 y) (87). It is known that uncomplexed cobalt localizes in the bone and liver which could pose issues related to radiation toxicity in case the ⁵⁵Co complex is unstable in vivo (88).

Gallium-66

Gallium belongs to Group III of the periodic table. Gallium has an atomic radius of 1.8 Å and electronegativity of 1.81. ⁶⁶Ga has one of the highest energies of the candidates considered here for PET imaging (E_max=4.15 MeV, 56% abundant). Along with a complex spectrum of γ-rays, ⁶⁶Ga is one of the more difficult positron-emitters for quantitative PET imaging. However, due its favorable half-life of 9.4 h, although at the shorter end of this criterion, use of ⁶⁶Ga has been explored for radioimmunoimaging in the past decade (89, 90).

Production of ⁶⁶Ga

⁶⁶Ga can be produced using a medical cyclotron employing the ⁶⁶Zn(p,n)⁶⁶Ga reaction [Table 2] (91). Alternatively, ⁶⁶Ga can be produced by ⁴He bombardment of natural copper targets by means of the ⁶³Cu (⁴He, n) ⁶⁶Ga reaction (89, 92, 93). However, this method also yields an impurity in the form of ⁶⁷Ga (t_1/2= 3.26 d). Chemical purification of 66Ga from the copper target using different techniques is described in detail in these published reports (89, 92, 93).

Radiolabeling of antibodies with ⁶⁶Ga

The co-ordination chemistry of Ga(III) is well established as it is known that Ga(III) readily binds to the harder oxygen donor chelating agents (94). Initially, GaCl₃ is complexed as an acetate or a citrate and this weak coordination complex is then used to prepare complexes of higher stability (94).

Monoclonal antimyosin antibody, Imciromab was labeled with ⁶⁶Ga in high yield (99%) by transcomplexation from an acetate buffer to the DTPA-conjugated protein (89, 90). One of the major concerns of using ⁶⁶Ga for labeling antibodies is the formation of the insoluble hydrolyzed Ga(OH)₃ species above pH 4.5. The radiolabeling had to be carried out in a pH range between 3–4.5 conditions which may potentially degrade intact antibodies. Suitable chelation chemistry for Ga(III) radionuclides that has been evaluated includes EDTA, DTPA derivatives, and DOTA, the latter of which may be adequately stable in vivo within the context of the half-lives of the positron emitting gallium radionuclides (95). A critical parameter for in vivo stability of the Ga(III)-complex is it’s relatively stability to the Ga(III)-transferrin complex. The reported formation constant for Ga(III)-transferrin complex is log K₁ = 20.3 (96) and the high plasma concentration of transferrin increases the risks of transchelation. The majority of the Ga(III) complexes used in nuclear medicine are thermodynamically stable and kinetically inert to exchange with transferrin in vivo. However, the complex formed with the chelator 1,4,7-triazacyclononane-N,N′,N″-triacetic acid (NOTA) is exceptionally stable and quite probably will emerge as the agent of choice for use with Ga(III) radionuclides (97, 98). Moreover, NOTA is a more suitable chelating agent for Ga(III) labeled antibodies as the Ga(III)-NOTA complex is formed at room temperature and therefore the protein is exposed to less harsh conditions as compared to heating as in the use of DOTA (99).

Biological studies with ⁶⁶Ga labeled antibodies

⁶⁶Ga-DTPA-antimyosin antibody was evaluated in dogs for imaging of myocardial necrosis by PET (89). Most biological applications of ⁶⁶Ga in molecular imaging have been with either peptides or small molecules as the radiolabeling in acidic conditions is more favorable as compared to use with antibodies (100, 101). The use of ⁶⁶Ga in humans dates back to 1954 (102), and in that study ⁶⁶Ga was used to investigate bone disease. As compared to ⁶⁶Ga, short-lived positron emitter ⁶⁸Ga (t_1/2= 1.1 h) obtained from a 68Ge/68Ga generator has gained immense popularity in the recent years due to availability of commercial ⁶⁸Ge/⁶⁸Ga generators and clinical successes with ⁶⁸Ga labeled somatostatin peptide (103, 104). In spite of such a short half-life, excellent results have been obtained from a ⁶⁸Ga labeled trastuzumab fragment in studying the tumor response to the heat shock protein 90 (Hsp90) inhibitor 17-allylamino-17-demethoxygeldanamycin (17-AAG) (105, 106). In these studies, a ⁶⁸Ga labeled Herceptin fragment was successfully used to image the pharmacodynamics of HER2 degradation in response to 17-AAG and was an earlier predictor of tumor response to 17-AAG therapy than ¹⁸F-FDG PET (105, 106). However, ⁶⁸Ga may not be a suitable candidate for imaging with intact antibodies due to their slower pharmacokinetics and tumor localization.

Challenges

Apart from unfavorable antibody radiolabeling procedures, the utility of ⁶⁶Ga in quantitative radioimmunoimaging is severely impaired by high positron energy (4.15 MeV) which results in loss of intrinsic spatial resolution of about 5.8 mm and additional high-energy γ-emission (1.03 MeV) resulting to poor signal to noise ratio [Table 1]. The half-life of 9.4 h may be relatively longer than existing radionuclides such ¹⁸F (t_1/2= 1.83 h), but the half-life is shorter with respect to the blood clearance and tumor localization of an intact antibody. Traditionally, gallium (⁶⁷Ga) is used for bone scans and therefore; ⁶⁶Ga lost from the chelate (if any) will localize in the bones delivering high radiation dose.

Arsenic-72

Arsenic belongs to Group III of the periodic table and has an atomic radius of 1.3 Å, and electronegativity of 2.18. ⁷²As (t_1/2 = 25.9 h) decays to a stable 72Ge daughter with 88% high-energy positron emission [Table 1]. The longer half-life and positron abundance makes ⁷²As one of the more attractive potential candidates of PET radioimmunoimaging. ⁷²As can also serve as a potential imaging and dosimetry surrogate for the β⁻-emitting ⁷⁷As (t_1/2 = 38.4 h) which can be used for radioimmunotherapy.

Production of ⁷²As

⁷²As can produced by irradiating an enriched germanium target using the ⁷²Ge(p,n)⁷²As reaction, but the radionuclidic purity is low due to the formation of ⁷¹As (10%) as shown in Table 2 (107). The target material can be recovered by dry distillation for reprocessing. Alternatively, ⁷²As can be produced from ⁷²Se/⁷²As radionuclide generators based on solid phase extraction and distillation in which ⁷²As can be eluted in aqueous solvents with over 40–70 % yield with less than 0.1 % contamination of ⁷²Se (108–110). However, for generator based production of ⁷²As, the production of ⁷²Se is tedious as it requires over 35 MeV bombardment of natural germanium target with helium ions. Another limitation of the ⁷²Se/⁷²As radionuclide generator is the relatively short half-life of ⁷²Se (8.40 d).

Radiolabeling of antibodies by ⁷²As

Radio-arsenic (⁷⁴As and ⁷⁷As) has been used to label the phosphatidylserine targeting chimeric monoclonal antibody Bavituximab (111). Bavituximab was first modified with N-succinimidyl S-acetylthioacetate (SATA) using a commercial kit. Direct labeling with radio-arsenic was performed after deprotection of the sulfhydryl groups using hydroxylamine [Figure 2]. The resultant radioimmunoconjugate demonstrated good stability over 72 hours during incubation in fetal bovine serum as determined by size-exclusion HPLC (111).

Figure 2.

Representative scheme of radiolabeling intact antibodies with radio-arsenic using SATA approach.

Biological studies with ⁷²As labeled antibodies

Bavituximab was labeled with ⁷⁴As (t_1/2 = 427 h, 30 % positron emission) as a surrogate for ⁷²As for a proof in principle study and evaluated for PET imaging in rats bearing Dunning prostate R3327-AT1 tumors (111). Tumors were clearly visualized after 48 hours with injection of 10 MBq of the radioimmunoconjugate with the highest tumor to background ratio being achieved at 72 hours. Localization of the radio-arsenic labeled immunoconjugate was confirmed by autoradiography. As a proof of principle, this study demonstrated successful use of radio-arsenic in PET radioimmunoimaging. However, one might argue that ⁷²As is a better candidate for radioimmunoimaging due to its more appropriate half-life and higher positron abundance. ⁷²As has been used in humans to study environmental toxicity caused by high doses of natural arsenic (112).

Challenges

Production of ⁷²As is limited to a few centers around the world and therefore the issue of a steady supply of clinical grade ⁷²As remains unclear. There have been efforts by the United States Department of Energy to assist in development of ⁷²As generators; however, the availability of a commercial generator seems unlikely in near future. Another drawback of ⁷²As is its association with a relatively high radiation dose which could restrict certain applications particularly when hematologic toxicity is a concern since the uncomplexed arsenic localizes in red blood cells (88).

Copper-64

Copper belongs to the Group IB of the periodic table and has an atomic radius of 1.6 Å, and an electronegativity of 1.90. ⁶⁴Cu has been a radionuclide of considerable interest for PET imaging due to its attractive physical and chemical properties. As compared to ¹⁸F, ⁶⁴Cu has a longer half-life of 12.7 h and decays to stable ⁶⁴Ni and ⁶⁴Zn daughters. Besides the positron emission, ⁶⁴Cu also emits Auger electrons and has a β⁻-emission, and therefore could potentially be used for radioimmunotherapy (113). Imaging characteristics as a result of optimum positron energy (0.66 MeV) and no high-energy γ-emission are one of the biggest factors that make ⁶⁴Cu an attractive candidate for PET imaging [Table 1]. The image quality and spatial resolution are equivalent to ¹⁸F and require no major adjustments in data processing and analysis (114).

Production of ⁶⁴Cu

⁶⁴Cu can be produced in numerous ways. Initially, ⁶⁴Cu was prepared by use of the ⁶⁴Zn(n,p)⁶⁴Cu reaction, but due to low yields and high contamination of ⁶⁷Cu, the ⁶⁴Ni(p,n)⁶⁴Cu reaction was later employed to produce therapeutic quantities of ⁶⁴Cu (115–117). ⁶⁴Cu can also be produced through the ⁶⁸Zn(p,αn)⁶⁴Cu reaction using a high-energy instrument (118–120). ⁶⁸Zn(p,αn)⁶⁴Cu reaction gives a good yield, but the radionuclidic purity is as low as 75% and the final product contains around 25% of ⁶⁷Cu as shown in Table 2 (118, 120). High yield production of good quality ⁶⁴Cu has greatly improved over the past couple of years and this radionuclide is now commercially available from several sources such as Trace Life Science, Isotrace Technologies, MDS Nordion, and IBA Molecular.

Radiolabeling of antibodies with ⁶⁴Cu

Copper can exist in two primary oxidation states, Cu(I) and Cu(II). It has been demonstrated that kinetic inertness is more important than thermodynamic stability of Cu(II) complexes as Cu(II) complexes of EDTA and DTPA rapidly dissociated in human serum and a Cu(II)-albumin complex was formed (121)

To improve the in vivo stability, antibodies have been labeled with ⁶⁴Cu using various bifunctional chelating agents [Figure 3] such as derivatives of 1,4,8,11-tetraazacyclotetradecane (cyclam) (122), 1,4,8,11-tetraazacyclotetradecane-N, N′,N″,N‴-tetraacetic acid (TETA) (123, 124), and DOTA (125). However, all these bifunctional chelates and their derivatives were unsuccessful in preventing metabolic processing of the ⁶⁴Cu complex, presumably a factor of the redox chemistry of Cu(II), and as a result elevated liver uptake was observed. To overcome this in vivo instability of Cu(II) complexes, cross-bridged tetraamine chelating agents are currently being developed and evaluated (126) These cross-bridged chelating agents provided better geometric coordination and were inert in aqueous solutions (126). Most application of these chelating agents has been for labeling peptides (127). To our best knowledge, intact antibodies have not been labeled with ⁶⁴Cu using cross-bridged tetraamine chelating agents as yet due to required harsh radiolabeling conditions utilizing high temperature and acidic pH that are not conducive for labeling heat and acid sensitive antibodies. To overcome these problems, a novel chelating agent SarAr: 1-N-(4-aminobenzyl)-3,6,10,13,16,19-hexaazabicyclo[6.6.6]eicosane-1,8-diamine conjugated to anti-GD2 monoclonal murine antibody 14.G2a and the corresponding chimeric monoclonal antibody ch14.8 were evaluated recently (128). SarAr is based on the macrobicyclic sarcophagine cage structure and was modified to contain the reactive aminobenzyl group for conjugation with the antibody. At pH 5.0and room temperature conditions, the SarAr chelating agent forms a stable complex with Cu(II) within the multiple macrocyclic rings comprising the sarcophagine cage structure. Therefore, radiolabeling of SarAr-mAb was performed at less harsh conditions (128). The resulting radioimmunoconjugate was shown to be stable in presence of liver proteins known to transchelate with ⁶⁴Cu (128). Alternately, attempts have been made to radiolabel antibodies with ⁶⁴Cu using porphyrin based chelating agents (129).

Figure 3.

Selected chelating agents used for radiolabeling intact antibodies with ⁶⁴Cu.

Biological studies with ⁶⁴Cu labeled antibodies

Numerous studies using ⁶⁴Cu labeled antibodies have been reported for PET radioimmunoimaging and radioimmunotherapy. Clinically approved antibodies such as Cetuximab and Abegrin have been labeled with ⁶⁴Cu for PET imaging in pre-clinical models (130–132). ⁶⁴Cu-DOTA-cetuximab was evaluated in mice bearing EGFR positive and EGFR negative tumors for quantitative PET imaging (132). In this study, ⁶⁴Cu-DOTA-cetuximab showed prominent uptake in EGFR-expressing tumors, but low accumulation in EGFR-negative tumors. A good linear correlation between the calculated %ID/g values as measured by PET imaging and the EGFR expression level as measured by western blot was exhibited (132). Subcutaneous xenografts and peritoneal metastases were clearly visualized by PET imaging with an anti-L1-CAM antibody chCE7 labeled with ⁶⁴Cu in mice bearing L1-CAM tumors (133, 134). However, this PET study also revealed that radio-metabolites of the copper immunoconjugate non-specifically localized in the lymph nodes through the reticuloendothelial system which could be critical in imaging lymph node metastasis (134). Recently, novel approaches in terms of chelating agents and linkers have been used to decrease the liver uptake and increase the target/background ratio (128, 135). The evaluation of ⁶⁴Cu-SarAr-14.G2a demonstrated that the use of a better chelating agent could greatly improve the tumor/background ratio as MicroPET imaging confirmed significant uptake of the radioimmunoconjugate in GD-2-positive tumors, with minimal uptake in GD-2-negative tumors and non-target tissues of interest such as the liver (128). Another use of ⁶⁴Cu labeled antibodies was to serve as a PET surrogate for ⁶⁷Cu labeled antibodies used for radioimmunotherapy (13). On comparison with ⁶⁷Cu, it was found that the ⁶⁴Cu labeled antibody alone reduced tumor growth, however, a much higher dose (5 times higher) of ⁶⁴Cu was required to produce effects comparable to ⁶⁷Cu and the therapy with ⁶⁴Cu was largely ineffective in treating larger tumors (136). To date, there has been one clinical study reported that utilized ⁶⁴Cu-TETA-1A3 to evaluate 36 patients with suspected primary or advanced colorectal cancer (137). In this study, all patients had CT scans and PET scans 4 to 36 hours after being injected with 370 MBq of ⁶⁴Cu-TETA-1A3 directed towards lipid antigens present in human colon carcinoma. The positive predictive value of ⁶⁴Cu-TETA-1A3 PET radioimmunoimaging was excellent, ranging from 89% (40/45) to 96% (43/45) and detected 11 new occult tumor sites including 9 small abdominopelvic foci less than 2.0 cm in diameter that were previously not detected by CT or MRI. This study demonstrated that PET radioimmunoimaging with ⁶⁴Cu may have important applications in clinical oncology, particularly for detecting smaller lesions undetected by CT or MRI.

Challenges

Excellent results have been obtained while imaging intact antibodies with ⁶⁴Cu, but ⁶⁴Cu seems to be a more suitable candidate for imaging when applied to use antibody fragments, diabodies, and other engineered targeting vectors which have faster localization and clearance (124, 132). Copper is a constituent of many metalloenzymes related to cellular respiratory oxidation process (88) and therefore, the release and accumulation of ⁶⁴Cu from the most traditional radioimmunoconjugates in the organs such as liver remains a concern in spite of efforts to reduce the non-specific uptake by developing better chelating agents and linkers.

Yttrium-86

Yttrium belongs to Group IIIB of the periodic table. Yttrium has an atomic radius of 2.3 Å and electronegativity of 1.22. The primary purpose for exploring the viability of ⁸⁶Y for PET imaging was to use it as an imaging and dosimetry surrogate for the therapeutic radionuclide ⁹⁰Y which has no imageable emission. ⁸⁶Y is an attractive candidate for studying ⁹⁰Y due its half-life (14. 2 h) which allows imaging over several days.

Production of ⁸⁶Y

⁸⁶Y can be produced on a medical cyclotron by irradiating SrCO₃ or SrO with 2–6 microA of beam current for <4 hours applying the ⁸⁶Sr(p,n)⁸⁶Y reaction. As compared to SrCO₃, SrO is a superior target as a SrO target can withstand at least a 6 microA of beam current, a significant improvement over a maximum of 2 microA on the SrCO₃ target resulting in yields that are almost doubled with minimum contaminants (138). At our own institution, we have developed a simple and efficient semi-automated method to purify ⁸⁶Y which does not compromise the biological properties of the ⁸⁶Y labeled antibody (139, 140).

Radiolabeling of antibodies with ⁸⁶Y

In physiological conditions, the primary oxidation state of yttrium is Y(III). Chelation chemistry is well established for Y(III) radionuclides and several bifunctional derivatives of both DOTA and DTPA [Figure 4] have demonstrated their utility in this arena (95). The majority of the chelation chemistry for yttrium labeled antibodies was carried out with ⁹⁰Y and ⁸⁸Y. In 1976, ⁹⁰Y-DTPA was used for radionuclide therapy of brain tumors in patients (141), and since then cyclic anhydride-DTPA was used to conjugate to antibodies for radiolabeling with ⁹⁰Y (142). However, when ⁹⁰Y-DTPA antibodies were evaluated in vivo, ⁹⁰Y was released from that DTPA and deposited in the bones causing radiation toxicity (143, 144). In order to improve the in vivo stability of the radioimmunoconjugate, backbone substituted DTPA derivatives such as p-isothiocyanatobenzyl- diethylenetriaminepentaacetic acid (SCN-Bz-DTPA) and a methyl derivative of p-isothiocyanatobenzyl-diethylenetriaminepentaacetic acid (SCN-Bz-Mx-DTPA also known as 1B4M-DTPA) (143–145) were evaluated. The 4-isothiocyanatobenzyl group (SCN-Bz) substituted onto the carbon backbone of DTPA for use in linkage to the antibody and the methyl groups were incorporated onto the backbone to sterically hinder the release of radiometal from the chelate. Comparative biological studies in mice demonstrated the superiority of backbone substituted DTPA radioimmunoconjugates over cyclic anhydride DTPA radioimmunoconjugates (143, 144). With the better understanding of the coordination chemistry and geometry of DTPA derivatives, a cyclohexyl derivative of DTPA was developed which demonstrated excellent stability in vitro and in vivo (146, 147). Interestingly, stereochemistry of CHX-DTPA played a major role in in vivo stability of ⁹⁰Y/⁸⁸Y radioimmunoconjugate and subsequent bone uptake (146, 148). The single enantiomeric [(R)-2-amino-3-(4-isothiocyanatophenyl)propyl]-trans-(S,S)-cyclohexane-1,2-diamine-pentaacetic acid (CHX-A″-DTPA) radioimmunoconjugate was found to be more stable as compared to the corresponding diastereomeric pair of CHX-B-DTPA radioimmunoconjugates (148). After 7 days post injection, bone uptake of the ⁸⁸Y labeled CHX-B′-DTPA or CHX-B″-DTPA radioimmunoconjugates was 5 times higher than that of the ⁸⁸Y labeled CHX-A″-DTPA radioimmunoconjugate (148). In order to further improve the stability, DOTA derivatives such as SCN-Bz-DOTA were explored for radiolabeling antibodies with ⁹⁰Y (149, 150). In a comparative study, it was found that DOTA conjugated antibody was better than DTPA conjugated antibody in terms of stability (150). Since then numerous pre-clinical and clinical studies have been published reporting DTPA and DOTA based chelating agents for ⁹⁰Y radioimmunotherapy and thus extrapolating to ⁸⁶Y for PET imaging (151–153).

Figure 4.

Selected chelating agents used for radiolabeling intact antibodies with ⁸⁶Y/⁹⁰Y.

Biological studies with ⁸⁶Y labeled antibodies

⁸⁶Y labeled antibodies were most recently evaluated to perform dosimetry for ⁹⁰Y labeled antibodies used for radioimmunotherapy as they were expected to be superior to ¹¹¹In labeled antibodies in terms of similarities in both chemical and biological behavior (153). Numerous studies have been reported with ⁸⁶Y labeled antibodies including clinically approved trastuzumab (Herceptin) (152, 154) and cetuximab (Erbitux) (154). In a recent study, ⁸⁶Y-CHX-A″-DTPA-trastuzumab was evaluated in mice bearing intraperitoneal SKOV3 ovarian tumors (152). In this study, PET and MRI studies were performed to assess the therapeutic potential of ⁹⁰Y-CHX-A″-DTPA-trastuzumab using ⁸⁶Y-CHX-A″-DTPA-trastuzumab as an imaging surrogate for dosimetry. Intraperitoneal tumors were visualized by PET imaging 2 days after injecting ⁸⁶Y-CHX-A″-DTPA-trastuzumab and the tumor localization was confirmed by MRI studies. ⁸⁶Y-CHX-A″-DTPA-trastuzumab localized to sites of disease with minimal normal organ uptake (152). In the study performed at our institution, we could clearly visualize LS-174T colon carcinoma xenografts in mice after 2 days of injecting ⁸⁶Y-CHX-A″-DTPA trastuzumab (tumor uptake was greater than 30% ID/g) (154). In the same study, the HER1-targeted ⁸⁶Y-CHX-A″-DTPA cetuximab was evaluated in mice bearing LS-174T colon carcinoma, SKOV3 ovarian carcinoma and DU-145 prostate carcinoma (154). All the tumors were clearly visualized at 1 day post-injection with the tumor uptake ranging from 20–30% ID/g.

Clinical use of ⁸⁶Y labeled agent used for dosimetry ⁹⁰Y labeled has been illustrated by the use of ⁸⁶Y-DOTA-Phe¹-Tyr³-octreotide (SMT-487) for dosimetry of ⁹⁰Y-SMT-487 in patients with metastatic carcinoid tumors (155). In this study, the maximum dose that could be delivered to the tumor without exceeding the 23-Gy cut-off dose to kidneys was successfully calculated using PET imaging, and therefore presumably reduced the incidence of potential renal toxicity caused by ⁹⁰Y-SMT-487 therapy.

Challenges

⁸⁶Y is commercially available (Isotrace Technology) in a limited capacity and is also only available at centers capable of producing ⁸⁶Y on a regular basis. The high positron energy of 3.14 MeV (34% abundance) and addition γ-emission of 1.08 MeV (83 % abundance) significantly affects the spatial resolution and image quality respectively [Table 1]. Uncomplexed ⁸⁶Y localizes in the bone and therefore can cause radiation toxicity in the events of “leakage” from the immunoconjugate and thus may be not a suitable candidate to image bone metastasis.

Zirconium-89

Zirconium belongs to Group IVB of the periodic table, has an atomic radius of 2.2 Å, and an electronegativity of 1.33. ⁸⁹Zr (t_1/2= 78.4 h) decays to an ultra short-lived daughter ^89mY (t_1/2= 16 s) with a 23% β⁺ emission and 77% electron capture [Table 1]. Due to its appropriate half-life and imaging characteristics, ⁸⁹Zr labeled antibodies are currently being evaluated in the clinic as PET surrogates for both ⁹⁰Y and ¹⁷⁷Lu therapy studies despite the possibility of using the elementally matched ⁸⁶Y for the former and that the latter possesses an imagable γ-emission (156).

Production of ⁸⁹Zr

⁸⁹Zr can be produced through the ⁸⁹Y(p,n)⁸⁹Zr reaction in which inexpensive natural yttrium foil is irradiated with a 14–14.5 MeV proton beam and the target is separated by ion-exchange chromatography or solvent extraction (26, 157, 158). Alternatively, ⁸⁹Zr can also be produced through the ⁸⁹Y(d,2n)⁸⁹Zr reaction (159). A relatively high yield of 43 MBq/μAh is obtained with the ⁸⁹Y(p,n)⁸⁹Zr reaction with less than 0.2% impurity of ⁸⁸Zr (t_1/2=85 d and daughter ⁸⁸Y; t_1/2=105 d) [Table 2].

Radiolabeling of antibodies with ⁸⁹Zr

Over the past 20 years attempts have been made to label proteins and antibodies with ⁸⁹Zr using DTPA derivatives and porphyrins as chelating agents (160, 161). Proteins and antibodies were labeled with ⁸⁹Zr using desferal (Df) conjugates (162, 163). Briefly, incorporation of Df-groups onto monoclonal antibodies was performed in a two-step procedure. In the first step, maleimide groups were incorporated into the protein with the assistance of succinimidyl-4-(N-maleimidomethyl)cyclohexane-1-carboxylate (SMCC). In the second step, the thioester of Df formed by conjugation to SATA was further converted to a free thiol with hydroxylamine to facilitate reaction with the maleimide groups of the antibody with subsequent radiolabeling with ⁸⁹Zr. The final purified product was obtained through gel filtration (162, 163). This protocol was further modified by formation of N-succinyldesferrioxamine B of Df (N-SucDf) as an intermediate (164). In this method, the hydroxamate groups of N-sucDf were temporarily blocked with formation of the Fe(III) complex. The N-sucDf-Fe complex was then esterified and coupled to antibodies. Thereafter, Fe(III) was removed by reduction to Fe(II) and transchelation to EDTA, mAb-N-sucDf is then radiolabeled with ⁸⁹Zr at pH 7.2-7.4 with a 30 minute incubation at room temperature (164).

Biological studies with ⁸⁹Zr labeled antibodies

Considering the success of ⁹⁰Y-labeled mAbs in clinical radioimmunotherapy (12, 151, 165), ⁸⁶Y-labeled mAbs are currently being developed in pre-clinical models for clinical dosimetry purposes (139, 152, 153). However, the half-life of ⁹⁰Y is nearly 4 times greater than that of ⁸⁶Y and therefore predicting ⁹⁰Y dosimetry of slowly clearing antibody may not be feasible or even appropriate. Due to these reasons, ⁸⁹Zr was explored as one of the possible radionuclides that could be used for dosimetry analysis of ⁹⁰Y-labeled mAbs (166). Clinically used antibodies such as HER1-targeted cetuximab (156), VEGF-targeted bevacizumab (167) and CD-20-targeted ibritumomab tiuxetan have been labeled with ⁸⁹Zr-Df conjugate (168). All the above studies demonstrated the use of ⁸⁹Zr labeled antibodies in quantitative PET imaging; additionally, the tumor localization of the ⁸⁹Zr-mAb was similar to that of the ¹¹¹In/⁹⁰Y-mAb.

Due to the feasibility in production and purification of ⁸⁹Zr labeled mAb, clinical studies were performed with CD44V6-specific chimeric mAb, U36 labeled with ⁸⁹Zr (169). In this study, ⁸⁹Zr-U36 was evaluated in patients with squamous cell carcinoma of the head and neck (HNSCC), who were at high risk of having neck lymph node metastases. All primary tumors (n = 17) as well as lymph node metastases in 18 of 25 positive levels (sensitivity 72%) and in 11 of 15 positive sides (sensitivity 73%) were detected using PET radioimmunoimaging (169). Interpretation of PET was correct in 112 of 121 operated levels (accuracy 93%) and in 19 of 25 operated sides (accuracy 76%). For CT/MRI, sensitivities of 60% and 73% and accuracies of 90% and 80% were found per level and side, respectively. ⁸⁹Zr-immunoPET demonstrated a sensitivity of 85% and accuracy of 95% as compared to sensitivity of 62 % and accuracy of 88% for¹⁸F-FDG-PET (169). Thus, this study demonstrated the utility and advantages of ⁸⁹Zr PET radioimmunoimaging in clinical setting in comparison with CT/MRI and ¹⁸F-FDG (169). Similarly, all known tumor lesions previously identified by ¹⁸F-FDG were successfully imaged by ⁸⁹Zr-Df-Zevalin in a pilot PET imaging study in a patient with Non-Hodgkin’s Lymphoma (168)

Challenges

Uncomplexed ⁸⁹Zr localizes in the bone and therefore can deliver high radiation dose to the bone marrow, therefore the need for stable chelation chemistry remains an important consideration with this element (88). In spite of the recent success of ⁸⁹Zr PET imaging, clinical grade ⁸⁹Zr is not commercially available yet (although, ⁸⁹Zr can be obtained from IBA Molecular on request). Another major disadvantage is the currently used ⁸⁹Zr-mAb production scheme. The currently used conjugation and radiolabeling schemes are often tedious, time-consuming and involve numerous steps that increase the margin for error for the cGMP (current Good Manufacturing Practice) production. In the currently used procedure, the Fe(III) complex is first formed with desferal which is then activated for conjugation through extension of the terminal amine with succinic anhydride followed by the conversion of the formed carboxylate into an active ester. After conjugation of the antibody, the Fe(III) is then reduced and displaced with ⁸⁹Zr (164). Alternate, simpler conjugation and radiolabeling schemes that can be successfully carried out at local radiopharmacies are warranted to bolster theincreased appeal of ⁸⁹Zr PET radioimmunoimaging. One of the suggested applications of ⁸⁹Zr has been to use it as a surrogate for ⁹⁰Y and ¹³¹I labeled antibodies, however, ⁸⁹Zr labeled antibodies, while close to, also do not exactly match the biodistribution and pharmacokinetics of ⁹⁰Y and ¹³¹I labeled antibodies (168, 170).

Opportunities and Challenges

Production

The reliable large-scale availability of the above desired longer-lived positron emitting radionuclides in suitable quality for clinical applications is critical for potential clinical translation. Many of the above described radionuclides such as ⁶⁴Cu, ⁷⁶Br, ⁸⁶Y and ⁷²As can be produced in good yields at low-energy medical cyclotrons (26). Most of the longer-lived positron-emitting radionuclides are currently produced by employing low-energy nuclear reactions such as (p, n) and (d, n) which may require isotopically enriched targets which are usually expensive. Alternatively, a higher-energy dual-particle cyclotron or a multi-particle intermediate-energy cyclotron could be used for production of some these radionuclides employing nuclear reactions such as (p, xn) and (³He, xn). Two of the most important criteria in medical application of positron-emitters are quality and yield, and therefore the target material is critical. It is absolutely essential to produce target material that is economical, may be efficiently recovered for reuse, can withstand high-energy beam fluxes of charged particles, and most importantly, also be easily isolated from the product to give high-purity final product. One such example that meets these criteria has been the production of ⁶⁴Cu. Originally, ⁶⁴Cu was produced through the ⁶⁴Zn(n,p)⁶⁴Cu reaction with resulting in low yields and a high contamination of ⁶⁷Cu, unless highly enriched pure ⁶⁴Zn was used. The low yields compounded with contamination of the β⁻-emitting therapeutic radionuclide ⁶⁷Cu were impediments to routine large-scale clinical applications of ⁶⁴Cu for PET imaging. As an alternative, the ⁶⁴Ni(p,n)⁶⁴Cu reaction was employed to obtain higher yields and better quality ⁶⁴Cu (115, 116). However, the biggest drawback of this method remains the cost of the highly enriched ⁶⁴Ni target material which can be a hindrance in large-scale production of ⁶⁴Cu at multiple centers across the world. With better understanding and characterization of cross-section data, integral yields and chemical processing methods, it is possible to obtain good quality positron-emitters in large yields using a medical cyclotron that can be potentially used for radioimmunoimaging. Most commonly used production routes of selected longer-lived positron emitters for radioimmunoimaging with their corresponding yields and radionuclidic impurities are listed in Table 2.

Imaging characteristics

One of the major factors impacting image quality is the additional γ-emissions and the energy of the positron which can contribute to the background noise and the loss of intrinsic spatial resolution respectively. Some of the longer-lived positron-emitting radionuclides such as ⁷⁶Br, ⁷²As, ¹²⁴I and ⁶⁶Ga have high-energy positron emissions which result in loss of intrinsic spatial resolution and poor signal to noise ratios [Table 1]. In clinical settings, both radiation dosimetry and chemistry as a result of inadequate specific activity and stability may limit the amount of radioactivity injected, thereby further degrading image quality. Loss of spatial resolution could also limit the use of radioimmunoconjugates labeled with some of the described longer-lived positron-emitting radionuclides for the detection of metastasis and smaller lesions which could be critical for their use in assessing or predicting clinical outcomes. Successful attempts have been made to improve the image characteristics of high-energy positron emitters by developing methods to correct for scatter fractions and spurious coincidences (171–173). Image reconstruction algorithms have been developed which include models of the statistical nature of nuclear decays and tomography responses which improve the spatial resolution, and at the same time reduce the image noise (114, 171–173). Further advances in instrumentation and image processing may additionally improve spatial resolution and the ability to perform more accurate quantitative analysis. However, implementing these improvements in the clinic may be complex as new instruments and imaging software will require FDA approval.

Potential therapeutic applications and radiation toxicity

Tumor therapy using antibodies carrying radioactivity has been reported now for over 50 years (174, 175). In 1966, McCradle and co-workers reported the use of ¹³¹I labeled antibody to human fibrinogen as a clinical diagnostic and therapeutic agent (8). In this study, 50 patients with reported neoplasm were imaged and a high percentage of patients showed positive scans for metastases to the brain which was a remarkable achievement during that period when targeted non-invasive radioimmunoimaging was still in its infancy. This study demonstrated the value of molecular imaging with radiolabeled antibodies for the assessment of primary and metastatic lesions (8). In this same study, two previously imaged patients were treated with higher doses of ¹³¹I labeled antibody to human fibrinogen resulting in regression of their lesions (8). Since that time numerous studies have been reported demonstrating the utility of antibodies labeled with radionuclides such as ¹¹¹In, ¹⁷⁷Lu and ⁶⁴Cu which could be used for diagnosis as well as therapy (4, 176). For better clinical acceptance, cost-benefit effectiveness and numerous other reasons including the need of surrogate radionuclide, a radiolabeled antibody that could be used as a diagnostic as well as therapeutic becomes a desirable combination.

One of the major concerns in radioimmunotherapy is the non-uniform absorbed dose distribution in the tumor resulting in less effective therapy in spite of dose escalation (177). One possible solution might be use of longer-lived radionuclides with longer emission ranges which may increase mean absorbed dose to the tumor as they are more likely to penetrate into the tumor before decaying in the body than shorter-lived radionuclides (178). Longer-lived positron-emitting radionuclides can deliver high absorbed doses to the target site as the doses delivered by positrons are as high as negatrons of the same energy. Considering the success of negatron-emitting radioimmunotherapy in pre-clinical and clinical studies, longer-lived positron-emitting radioimmunoconjugates may potentially be used for therapeutic applications. The longer half-life may minimize the effects of non-uniformity up to an extent. However, the choice of radionuclide will be based upon the function of biological and pharmacokinetic properties of the targeting antibody (normal organs uptake and clearance), disease state, physical half-life of the radionuclide, abundance and energy of the positron emission and additional γ-rays. For example, assuming a biological half-life of 100 hours and whole-body uniform distribution in an adult human, based on the values given in the International Commission on Radiological Protection (ICRP) publications, the effective dose from ⁵²Mn (t_1/2= 134.2 h) will be approximately 0.9–1.1 mSv/MBq as compared to 0.02–0.04 mSv/MBq from ⁶⁴Cu (t_1/2= 12.7 h) and 0.2–0.3 mSv/MBq from 76Br (t_1/2= 16.2 h) (159). Another important parameter in terms of therapeutic applications of longer-lived positron emitters is the positron abundance and the dose rate; a dose rate of 2 – 3 cGy/hr has been suggested as being required to overcome the proliferation of most types of tumor cells (179). Therefore, the portion of the total dose that is delivered below this rate is essentially unnecessary and can potentially pose a hazard. Theoretically, high specific activity longer-lived positron-emitting radioimmunoconjugates could be used for therapy; however toxicity and dose to the normal organs may pose a few challenges and will depend on whether the radioimmunoconjugate is delivered systemically or locally.

Economics, regulatory affairs and clinical acceptance

Currently, the use of PET in the clinical setting revolves around ¹⁸F-FDG, which has great clinical acceptance as it can be used to image most metabolically active tumors and lesions irrespective of any expression of cell surface molecular targets such as EGFR and HER2/μ. This characteristics of ¹⁸F-FDG can also be disadvantageous too as it does not provide information regarding the status of cell surface targets which could be essential for patients about to undergo immunotherapy or radioimmunotherapy. Under those circumstances, PET radioimmunoimaging could play a critical role in risk stratification. In the past year, sales of Herceptin (trastuzumab), Avastin (Bevacizumab), Rituxan (Rituximab) and Erbitux (Cetuximab) each all exceeded one billion U.S. dollars (180). Based on previous results, not all patients may respond to immunotherapy due to low-density of target expression or changes in the target levels. In all likelihood, not all patients undergo some form of diagnostic scan to evaluate the status of target in the progression of illness. PET radioimmunoimaging could serve as one of the potential tools for screening patients that may or may not respond to therapy, thereby acting as a cost containment tool. Contrary to that positive aspect, a sizeable financial investment and commitment is required to develop a clinically FDA approved radiopharmaceutical. In order to recover the cost spent in developing the PET radioimmunoconjugate, sales and use volumes have to be significantly high which will be directly related to the clinical acceptance of PET radioimmunoimaging.

The number of radiopharmaceuticals approved by the US FDA in the past decade has been limited due to more scrutiny and a tedious approval process. Radiation toxicity and required pre-approval studies could potentially be a major obstacle for the approval of radioimmunoconjugates labeled with longer-lived positron emitters. Thus, one can question the rationale behind the objective of developing a novel PET radioimmunoconjugate considering the level of required investment and resources versus failure to acquire approval of the agent by the FDA.

Other regulatory barriers include waste disposal of large quantities of longer-lived high energy positron emitters, decontamination, occupational hazards, and the radiation exposure to normal population by out-patients undergoing high-dose longer-lived positron-emitting radionuclide protocols.

Lack of standards for calibration, decay characteristics

A recent study has highlighted the differences in published values of decay data and energies of some of the described radionuclides (119). These differences in values could impact the overall biological and dosimetry results as the same amount of radioactivity quantified from dose calibrator and PET image could vary from institution to institution depending of the values used. Therefore, there is a need to establish standards for calibration of instruments for these longer-lived positron emitters to avoid variances which could result in errors.

Future perspective

It is clear from the selected examples cited in this report that efforts are continuing to being made to develop radioimmunoconjugates labeled with longer-lived positron-emitting radionuclides for PET imaging. The superiority of PET in radioimmunoimaging was demonstrated by a recent clinical study in which a comparison was made between ¹³¹I-G250, ¹²⁴I-cG250 and ¹⁸F-FDG for detection of metastatic renal-cell carcinoma (53, 181). ¹³¹I-cG250 was able to detect 30% of metastatic lesions as compared to 69% for ¹⁸F-FDG (181); however, the results were far better when ¹²⁴I-cG250 was used as the sensitivity of detection of tumor was 94 % (53, 181). When compared to ¹²⁴I-cG250, ¹³¹I-cG250 did not provide the same resolution, contrast, or the potential for quantification as ¹²⁴I-cG250 (53). Considering the advantages of PET radioimmunoimaging, further efforts are warranted to improve PET instrumentation and image processing, increase production of clinical grade high-yield longer-lived positron emitters, and continue to create efficient radiolabeling methods to produce high-yield biologically active stable radioimmunoconjugates in the least amount of time and variability. And, lastly, but perhaps the most critical criterion for successful clinical translation, continued efforts have to be applied towards making wise selections of the most appropriate antibody matched with the most appropriate radionuclide for the desired application within an appropriate patient population. However, a few questions still remain unanswered: how many positron-emitting radionuclides do we really need to evaluate to “fit the bill” and to be able to ask and answer questions related to radioimmunoimaging? What are the alternatives in case of unexpected unreliable availability of the radionuclide? Is it worth the capital investments related to PET imaging if techniques such as biopsies and CT imaging are going to remain the mainstay of cancer diagnosis? Can future generation SPECT systems be as good as currently used PET (182, 183)? If so, then why develop newer positron-emitting products if ¹¹¹In based SPECT can offer precise quantitation in future? And finally, should be there be a common set of criteria and guidelines for choosing the radionuclide?

In conclusion, we believe that radioimmunoimaging with longer-lived positron-emitting radionuclides holds significant promise, but several scientific and non-scientific challenges have to be conquered before realizing the true potentials of PET radioimmunoimaging with intact antibodies.