Owing to their highly specific recognition of cognate antigens, antibodies have long been dream vehicles for targeted therapy and have recently been used to deliver radionuclides, drugs, toxins, and enzymes. Radioimmunotherapy (RIT) is an attractive therapeutic modality that combines the specificity of antibodies to antigens with the toxicity of the radionuclides. Radiolabeled antibodies like Bexxar and Zevalin have been approved by the FDA and are being used successfully to treat hematological malignancies in clinics. Moreover, because of their ability to inhibit the functional activity of the target antigen, some antibodies are also being used as unarmed therapeutics for solid tumors. Examples of these antibodies in clinics are: Avastin, an anti-VEGF monoclonal antibody that is used in the therapy of colorectal cancer, and Herceptin, an anti-Her2 antibody approved for the therapy of Her2-overexpressing metastatic breast and other cancers. For solid tumors, however, radiolabeled antibody-based therapeutics have had limited success, primarily owing to slow uptake, poor penetration, and inadequate retention. Recent research efforts have centered on the strategies to overcome these limitations and optimize tumor uptake of armed antibodies. Several recent articles, including one from Hu et al. in the March 2007 issue of EJNMMI [1], have described the possibility of using cell-penetrating peptides for the intracellular delivery of radiolabeled antibodies to target intracellular epitopes and improve tumor uptake and retention.

Cationic peptide sequences from various proteins have been shown to possess cell-penetrating activity. These are termed cell-penetrating peptides (CPPs) or protein transduction domains, and they have recently gained much attention owing to their ability to transduce a wide variety of cargoes, when co-administered with or linked chemically or genetically to the payload, across the plasma membrane in a cell-, receptor-, and energy-independent manner. Despite the lack of a clear understanding of the mechanism of protein transduction, CPPs have been demonstrated to deliver diverse cargoes such as peptides, proteins, viral particles, liposomes, and nanoparticles both in vitro and in vivo [2, 3]. TAT peptide derived form HIV1-TAT protein and penetratin derived from homeodomain transcription factor Antennapedia are the two most extensively studied CPPs.

The potential of CPPs to improve the uptake, retention, and therapeutic efficacy of radioloabeled antibodies is now being realized. The possible mechanism by which CPPs improve the tumor uptake and retention of antibodies is illustrated in Fig. 1. In their recent article in EJNMMI [1], Hu et al. reported the utility of the cell-penetrating activity of TAT for the transduction of a radiolabeled antibody directed against an intracellular target p21WAF-1/Cip-1. In breast cancer cells, p21WAF-1/Cip-1, which is a cell cycle regulator, is up-regulated by EGF and mediates growth inhibition in response to EGF by inducing G1–S phase cell cycle arrest. TAT peptide resulted in nearly a twofold increase in the accumulation of radiolabeled anti-p21WAF-1/Cip-1 MAb in tumor cells in vitro as compared with unconjugated MAb or MAb conjugated to control peptide, and the proportion of antibody accumulated in the nucleus increased nearly fivefold. Therefore, the total radioactivity accumulated in the nucleus was nearly tenfold more when the antibody was TAT conjugated. Nuclear translocation of the internalized antibody also prevented the deiodination of the internalized antibody, as suggested by the decrease in the proportion of the free iodine in the culture medium. In vitro, from the RIT perspective, TAT resulted in increased nuclear localization of the radioactivity, thus bringing the radioisotopes into close proximity with their therapeutic targets (i.e., DNA). It also improved the intracellular stability of the radioimmunoconjugate by preventing its deiodination, thereby improving its retention. However, these events seemed to be solely attributable to the cell-penetrating activity of TAT, with a minimal contribution from the specific antigen binding. But the in vitro experiments did indicate specific recognition and subsequent inhibition of the target antigen. TAT conjugated anti-p21WAF-1/Cip-1 MAb inhibited the G1–S phase cell cycle arrest induced by EGF while the non-specific IgG had no effect on the EGF-mediated arrest. For the first time, this report demonstrated the functional inhibition of a cell cycle regulator in cancer cells by transduction of target-specific antibody.

Fig. 1
figure 1

Schematic representation of the rationale for using CPPs for optimization of RIT (RE renal elimination, TIS tumor interstitial space, TC tumor cell, BV blood vessel). The radiolabeled antibody is represented with a starred Y, and the size is representative of the concentration of radiolabeled antibody at various places. 1 Immediately after administration, the concentration of radiolabeled Ab is high in the blood. Most of it undergoes RE, while a small amount diffuses into TIS. 2 Diffusion into TIS continues until a dynamic equilibrium is attained between the concentration of radiolabeled antibody in blood and TIS. The MAb localized in the TIS binds to the antigen present on the tumor cell surface as a function of its affinity to target antigen. 3 The concentration of radiolabeled Ab in blood continues to fall due to RE and a point comes when the concentration in the TIS is higher than that in blood. At this point, the antibody localized in TIS leaches out into the blood, resulting in a decreased concentration of radiolabeled Ab in the TIS. Lower TIS concentration might lead to the disassociation of antibody bound to the antigen on the cell surface. 4 If, by the inclusion of CPPs, the surface-bound antibody is internalized into the cells, then the intracellular antibody will be more resistant to clearing by blood than the unbound antibody in the TIS and, thus, will have an increased tumor residence time (retention). Moreover, due to internalization of the surface antibody, the dynamic equilibrium between the cell surface-bound and unbound Ab in the TIS will be shifted towards cell surface binding. As a result, an increase level of antibody will bind to the cell surface in steps 2 and 3, thereby increasing the net tumor uptake

Full size image

A few other reports have also demonstrated the utility of CPPs for intracellular delivery of antibodies and subsequent neutralization of target antigen in vitro. Stein et al. demonstrated that Fab fragments of anti-tetanus toxin antibodies, when conjugated to TAT (37–72), resulted in the neutralization of tetanus toxin inside bovine chromaffin cells [4]. The antibody fragments were conjugated to TAT either by thioether or by disulfide linkage, and only the disulfide conjugates effectively neutralized the tetanus toxin. In comparison to the thioether conjugate, disulfide conjugate also exhibited increased nuclear translocation. Thus the nature of linkage affects the subcellular distribution and functional activity of the antibody. Along these lines, it will be of interest to study the effect of various linkages between TAT and anti-p21WAF-1/Cip-1 MAb on the functional inhibition of the target antigen. Another study exploited the membrane translocating sequence (MTS) derived from cell-penetrating polyreactive antibodies directed against DNA and nuclear ribonucleoprotein that occur naturally in autoimmune disorders. MTS conjugated monoclonal and polyclonal antibodies directed against active caspase-3 inhibited actinomycin-D-induced apoptosis in Jurkat cells while the unmodified antibodies had no effect [5]. Such an approach can have applications in treating disorders where apoptosis is involved in disease progression, such as the neurodegenerative disorders.

Hu et al. also studied the in vivo biodistribution of TAT-conjugated anti-p21WAF-1/Cip-1 MAb in tumor xenograft-bearing mice. TAT-conjugated MAb exhibited increased tumor accumulation at 48 h as compared with the unmodified antibody. The overexpression of the target antigen in response to EGF further improved the tumor accumulation of the radiolabeled antibody. However, it could not be determined whether the amount of antibody accumulated in the tumor was sufficient to inhibit p21 activity significantly. Since the tumor uptake was studied only at a single time point, it could not be determined whether high %ID/g (percentage injected dose/gram) values for 123I-TAT-anti-p21WAF-1/Cip-1in a tumor are a reflection of increased net tumor uptake or improved tumor retention. TAT-modified antibody also showed increased accumulation in other tissues and elevated levels in the blood in comparison to the unmodified Fab. The tumor to normal tissue ratios did not improve following conjugation with TAT. Apparently, the non-specific cell-penetrating activity of TAT prevailed over the antigen targeting ability of the antibody, either because the cellular penetration preceded antigen recognition owing to the intracellular location of the antigen or because of a high ratio of CPP to antibody. Previous reports have demonstrated that this ratio is important and can disrupt the specific antigen binding and hence targeting at high ratios. Anderson et al. studied the uptake of TAT-conjugated Fab fragments of MAb NRLU-10 in antigen-expressing and antigen-negative cells [6]. Although TAT-conjugated Fab exhibited increased binding and internalization, only the conjugates with the 1.1 peptide/Fab retained their specificity toward antigen-positive cells. In contrast, conjugates with 1.6 peptide/Fab exhibited increased binding and internalization in both antigen-positive and antigen-negative cells. Similarly, in vivo results from the co-administration of penetratin and divalent scFv against TAG72 suggested that tumor-specific improved retention was observed only at a low peptide to antibody ratio. At higher peptide to antibody ratios, the cell-penetrating activity of penetratin predominated over the specific antigen binding of antibody fragments, resulting in increased localization and retention in non-target tissues [7]. Thus, to preserve the specific antigen recognition, which is the essence of targeted delivery, it is of the utmost importance to determine the optimal ratio of CPPs to antibody.

Both in vitro and in vivo data from the current study indicate the predominance of the cell-penetrating activity of TAT over the targeting ability of the antibody. However, caution is necessary for cargoes like radioimmunoconjugates where the effectors (i.e., the radioisotopes) may be delivered to normal as well as target cancer cells and subject them to similar disruptive mechanisms. In the present study, 123I-TAT-anti-p21WAF-1/Cip-1 exhibited increased uptake in non-target tissues, too. What will be the consequences of the inhibition of p21WAF-1/Cip-1 inhibition in these tissues? In vivo biodistribution experiments also indicate significantly higher radioactivity in the blood following 123I-TAT-anti-p21WAF-1/Cip-1 administration. Future studies should address the key question of whether the increased circulating radioactivity is a reflection of internalization of 123I-TAT-anti-p21WAF-1/Cip-1 in the blood cells. A recent study involving the modification of antibody Fab fragments with another CPP derived from HIV-1 Rev protein was performed in rats. Kameyama et al. demonstrated that the blood and plasma concentration of 125I-Rev-Fab decreased rapidly in the initial phase and was about 3.5–4.3 times less than that of unmodified Fab [8]. By 24 h, nearly 70% of the injected dose of 125I-Rev-Fab was eliminated by degradation into low molecular weight components followed by urinary excretion. The amount of radioactivity associated with the whole blood and plasma was determined separately. The ratio of radioactivity in the plasma to blood decreased from 1.84 for unmodified Fab to 1.5 for Rev-Fab, suggesting that some proportion of the Rev-Fab did associate with blood cells. Other studies have also demonstrated that TAT-conjugated proteins do penetrate into blood cells [9].

In another study, Kameyama et al. compared the biodistribution of Fab fragments modified by three different peptides, TAT, Rev and Antennapedia (Antp) peptide (penetratin) [10]. In cultured HeLa cells, all three peptides improved the cellular uptake of the Fab, with Rev-Fab exhibiting greater uptake than Tat-Fab and Antp-Fab. More importantly, following intravenous administration in rats, there was a considerable difference in distribution and retention of the radiolabeled Fab modified by the three CPPs. At 4 h, the highest concentration of Rev-Fab was observed in the liver, spleen, and adrenal gland, while TAT-Fab was undetectable in the adrenal gland but showed increased accumulation in the liver and spleen. The distribution of Antp-Fab was similar to that of Rev-Fab, with the latter exhibiting greater penetration. TAT-Fab exhibited the earliest elimination, with the levels decreasing to those of unmodified Fab at 24 h. Rev-Fab exhibited high retention in the spleen and renal medulla, while Antp-Fab was retained primarily in the spleen and renal cortex. The study also suggested that TAT-modified Fab might be eliminated through the gastrointestinal tract while Rev- and Antp-modified Fabs follow urinary excretion. Thus, modification with various CPPs results in differential distribution and retention in various organs, and a distinct metabolic fate of the radiolabeled antibodies.

It remains to be seen how the tumor distribution of specific antibodies will be affected by various CPPs and whether changing the ratio of antibody to CPPs will alter the non-target tissue distribution. In this direction, a few other reports have studied the biodistribution of specific radiolabeled antibodies or their derivatives following modification with CPPs. Neisner et al. conjugated scFv (L19), a single-chain Fv directed against the ED-B domain of fibronectin, with TAT and studied its biodistribution in murine F9 teratocarcinoma-bearing mice [11]. The tumor-targeting ability of the scFv (L19) was completely abolished following conjugation with TAT-conjugated scFv, exhibiting decreased tumor uptake and rapid uptake in the liver. In contrast, we recently demonstrated that penetratin is more efficient than TAT in improving the tumor retention of the divalent antibody fragments [7]. Co-administration of penetratin with radiolabeled scFv fragments directed against tumor-associated glycoprotein 72 (TAG72) resulted in the retention of nearly 80% of the peak accumulated tumor dose as compared to 48.5% retention with TAT and 27% with the control treatment by 24 h. While co-administration with penetratin did not result in altered retention in the non-target tissues, TAT resulted in increased uptake in the lungs. However, at higher peptide to scFv ratios, penetratin also resulted in increased non-specific uptake. Although direct evidence for improved therapeutic efficacy is still missing, these studies nevertheless accentuate the need to find the right balance between the specific “tumor targeting” of antibodies and the non-specific “cell-penetrating” activities of the CPPs. For the specific delivery of targeted radiopharmaceuticals, CPPs can be more useful vectors if their cell-penetrating activity is exerted specifically in the tumor tissue. A recent report has demonstrated the feasibility of generating tumor-specific CPPs [12]. The cell-penetrating activity of polyarginine-based cationic peptide was inhibited by fusion with negatively charged peptide. The polyanionic and polycationic domains were separated by a linker sequence which is sensitive to cleavage by proteases that are generally overexpressed in tumor tissues like matrix metalloproteases 2 and 9. In vivo administration of fluorescently labeled activateable CPPs (as they were called) resulted in increased tumor uptake as compared with normal tissues. It will be interesting to evaluate the utility of such tumor-specific CPPs in improving RIT.

The dominance of cell-penetrating activity can still be exploited if the payload is not toxic to the non-target cells; for example, a drug whose molecular target is expressed only in the target cells or the therapeutic moiety is activated specifically in the target cells. In such a situation, the accumulation of the therapeutic moiety in the normal tissues will not have any harmful effect. Several studies have exploited this approach for intracellular delivery of such selective inhibitory agents. Vocero-Akbani et al. delivered HIV-1 protease-activated caspase-3 by generating a fusion protein with a TAT peptide [13]. The caspase-3 was modified such that it was activated only in HIV-1-infected cells owing to the replacement of the endogenous proteolytic sites with HIV-1 protease-sensitive sequences. The TAT–caspase-3 fusion resulted in the specific killing of the HIV-infected cells with active protease. Similarly, the fusion of TAT with tumor suppressor protein p27Kip1 and dominant negative CDK2 resulted in G1 phase cell cycle arrest in hepatocellular carcinoma cells. Additionally, TAT-p27Kip-1 resulted in prolonged survival of mice bearing intraperitoneal tumors. Along similar lines, it will be of interest to investigate whether the transduction of p21WAF-1/Cip-1 in tumor cells triggers cell cycle arrest and induces apoptosis in tumor cells in vitro and in vivo.

Antibody-mediated targeting has been limited to extracellular, cell surface antigens or antigens specifically expressed by tumor vasculature, primarily owing to their accessibility to the circulating antibodies. The first step in tumor targeting is recognition of antigens expressed by the tumor cells, and internalization can be a means to improve retention or inhibit functional activity of the target. However, for intracellular antigens, internalization with CPPs will precede specific antigen recognition and can limit the delivery to the target. The non-specific nature of the CPPs compromises the specificity and efficiency of targeting, as the CPP-conjugated antibodies will also be internalized in non-target cells which do not express the antigen. From a RIT perspective, this is not desirable as it will result in decreased tumor localization and increased accumulation in the non-target tissues, thereby decreasing therapeutic efficacy and increasing toxicity. Still, some intracellular antigens that are overexpressed in tumors can serve as good targets for RIT or targeted drug delivery. In large tumors, intracellular antigens are accessible in dead or dying cells with permeable membranes. The best example of such an intracellular antigen is melanin, which is abundant and accessible in dying melanoma cells. In a recent study, 188Re-labeled MAb directed against melanin resulted in the growth inhibition of large tumors, regression of smaller tumors, and prolonged survival in the animal model [14]. Similarly, MAb 2C5 directed against a nucleoprotein epitope aberrantly expressed on the cell surface of several cancer types is exploited for targeted delivery of liposomes and nanoparticles [15].

For the successful exploitation of CPPs in order to improve RIT, the key issue will be to optimize the ratio of TAT and antibody and find a balance between the antigen-specific targeting ability of the antibody and the non-specific cell-penetrating activity. CPPs differ in terms of the extent of penetration and the tissue distribution of the cargoes. Also, the manner in which CPPs are linked to antibodies affects the biodistribution and intracellular activity and availability of the antibody, as does the format of the antibody itself. Several groups have developed conjugates of various CPPs with intact antibodies, Fab fragments, or scFvs. CPPs have been attached in various ways: chemically, by disulfide thiol linkages, by means of complex formation via biotin, or even as genetically engineered fusion proteins. With so many variables, at present no general rules seem to be emerging to fine tune and exploit the useful properties of CPPs and antibodies for the improvement of RIT. Concerted research efforts will address these issues in the near future, and Trojan peptides will add more to the magic of “magic bullets”, leading to the development of a new generation of radiopharmaceuticals.