In order to select a charged particle emitting radionuclide to be used in clinical radioimmunotherapy studies, several candidate radionuclides were evaluated for ease of labeling, affect on immunoreactivity, retention within the cell, and radiobiologic evidence of cytocidal effect. The physical properties and labeling chemistry of ¹³¹Iodine [¹³¹I], ⁹⁰Yttrium [⁹⁰Y], and ¹⁷⁷Lutetium [¹⁷⁷Lu] are summarized in Table 1. ¹³¹Iodine, of course, has a long history of effective use in the treatment of thyroid carcinoma and as a radio label in other radioimmunotherapy protocols. Radioiodine labeling techniques are well-known and it is possible to synthesize stable iodinated compounds that withstand enzymatic cleavage of the iodine if the iodine is inserted into double bonds of other than a tyrosine ring. High-specific activity ¹³¹I is readily available and relatively inexpensive. During the 1990s, ⁹⁰Y had come to be considered the preferred radionuclide for therapy because of the β energy greater than that of ¹³¹I with consequently longer range in tissue. The short half-life of ⁹⁰Y is also a potential advantage because of the greater radiobiologic effect of any given radiation dose delivered over a shorter interval (known as the dose rate effect). In the late 1990s, ¹⁷⁷Lu became available (courtesy of the Missouri University Research Reactor (MURR)]. ¹⁷⁷Lu, like ⁹⁰Y, is a radiometal; hence if the radioimmunoglobulin is internalized, it is likely to remain intracellular. Although the chemistry of ¹⁷⁷Lu and ⁹⁰Y is somewhat similar, the physical properties (physical half-life, β energy and accompanying γ emission) of ¹⁷⁷Lu are more similar to ¹³¹I than to ⁹⁰Y. Several methods to prepare chemically stable radiometal-labeled immunoglobulins have been developed. It was decided to evaluate the stability and utility of DOTA (1,4,7,10-tetra aza cyclo dodecane-1,4,7,10-tetra acetic acid) as a complex to covalently bind to the immunoglobulin and serve as a chelating moiety for the metallic radionuclides. The DOTA moiety has a high affinity for both ¹⁷⁷Lu and ⁹⁰Y.

During preliminary studies using ¹¹¹In DOTA-J591 as a surrogate for the radiometals ⁹⁰Y and ¹⁷⁷Lu, the affinity of murine J591 to LnCaP cells and the stability of the bound radiolabeled immunoglobulins were demonstrated in intact animals. Although J591 binds to an external epitope of PSMA, the epitope-antibody complex is internalized and likely digested in the intracellular environment. Nevertheless, ¹¹¹In DOTA-J591 demonstrated high cellular retention of 90–95 % of the radioactivity with a washout half-life greater than 500 h. By contrast, although the ¹³¹I labeled J591 bound effectively to PSMA-expressing LnCaP cells in suspension, the ¹³¹I label was observed to wash out of the cell as the iodine had likely labeled tyrosine moieties within the immunoglobulin. Hence, subsequent evaluation was limited to the radiometals ¹⁷⁷Lu and ⁹⁰Y (Smith-Jones et al. 2000). Significant anti-tumor responses were observed in LnCaP xenograft tumor bearing nude mice and a dose response relationship was observed (Fig. 5). Higher cumulative doses of either ⁹⁰Y or ¹⁷⁷Lu were possible using fractionated dosing (multiple submaximal tolerated dose (MTD)) rather than a single MTD dose. Median survival of the animals improved threefold for fractionated ⁹⁰Y-muJ591 therapy versus control (150 vs. 52 days). With fractionated doses of ¹⁷⁷Lu-muJ591, >80 % of the mice were cured (Smith-Jones et al. 2003).

Following the identification of a series of urea-amino acid based molecules that inhibited the glutamate carboxypeptidase enzymatic activity of PSMA, Babbich, Pomper, and co-workers evaluated several of these molecules that were essentially lysine-urea-lysine or glutamate-urea-lysine analogs (Hillier et al. 2009). One of these molecules, identified as MIP 1095 had a Ki of 0.24 + 0.14 nmol/L indicating high binding affinity. When labeled with ¹²³I, the affinity for human prostate cancer LnCaP cells was 0.81 + 0.39 nmol/L. Similarly, high affinity for PSMA was demonstrated for other similar molecules. This binding was also specific for PSMA as there was no binding to prostate cancer cells that did not express PSMA. This high affinity, selective targeting, was subsequently demonstrated in nude mice bearing LnCaP xenografts (Fig. 3).

Following “humanization” of the murine antibody by removal of a substantial portion of the murine immunoglobulin and substituting nonspecific human gamma globulin for most of the molecule but retaining the immuno-recognition portion, initial clinical studies were performed to determine the safety and effect on biodistribution of incremental amounts of the humanized antibody using ¹¹¹In DOTA-J591. Repetitive dosing was well tolerated up to 500 mg without the development of a human anti-humanized (de-immunized) antibody (HAHA) response or evidence of dose limiting toxicity (DLT); the maximum tolerated dose (MTD) was not reached (Bander et al. 2000). Monoclonal antibody targeting to normal tissues was not observed but significant uptake of the radiolabel was seen in the liver, typical of radiometal labeled immunoglobulin biodistribution patterns (Fig. 6). The liver activity diminishes with increasing dose of antibody but a considerable fraction of the injected dose appears in the liver. Increasing the amount of antibody infused results in longer plasma clearance times (Bander et al. 2001; Vallabhajosula et al. 2005a). These data demonstrated that relatively low amounts of total immunoglobulin, in the order of 20 mg, were sufficient to minimize hepatic uptake and maximize plasma circulation time. Subsequently, a total of 20 mg of the J591 antibody in all clinical studies; that is after labeling at a specified specific activity demonstrated not to adversely influence immuno reactivity, the individual patient doses were supplemented with unlabeled humanized J591, so that the preparation to be administered contained a total of 20 mg/m² of immunoglobulin.