¹¹Boron Delivery Agents for Boron Proton-capture Enhanced Proton Therapy

Boron Proton Capture

Clean energy production research has focused on nuclear fusion reactions potentially obtained from high-intensity-pulsed laser technology. Boron–proton fusion was one of these studied reactions as it is highly energetic (8.7 MeV), is free from high-energy neutrons and generates three alpha particles. The initial reaction between ¹¹B and a proton results in an excited carbon atom (¹²C) which splits into an alpha particle (3.76 MeV) and a beryllium (⁸Be) atom, which subsequently divides into two alpha particles (2.74 MeV). The highest cross section for this reaction occurs with protons with energies around 600-700 keV.

This has led to the idea of exploiting all the advantages of BPC during cancer proton therapy. The first publication on BPC in radiation therapy used Monte Carlo simulations and showed an enhanced high dose in the areas where the boron uptake coincided with the Bragg peak (5). There are a few other studies which used Monte Carlo simulations, but which produced highly divergent results on the potentially achievable BPC dose, ranging from 0.026% to 0.3% based on using different proton energies and ¹¹B concentrations (5-8). The first cell culture experiments were performed at the medical cyclotron in Catania and showed that the BPC procedure increased the biological effectivity of proton irradiation (9). There was a significant enhancement effect on both the cell survival test based on a colony-forming assay and chromosome aberration analysis due the presence of naturally-occurring ¹¹B present in sodium borocaptate (Na₂B₁₂H₁₀-SH; BSH). Such a promising approach requires further investigation.

Therapy using BPC, in contrast to BNCT, does not require such a high differential boron uptake in the target region because proton irradiation is highly spatially selective. This ensures that the majority of BPC will be within the planned region, even in the case of homogeneous boron distribution in a larger volume around the target region (10). To that end, the proton energy distribution must be calculated with high accuracy and should be tailored to the volume of interest where the BPC enhancement is required. A high and long-lasting tumor concentration of ¹¹B and favorable intracellular micro-distribution, preferably close to radiosensitive cell elements (nucleus, mitochondria, membrane) are needed for successful BPCEPT. The vast amount of research and experience with boron compounds used in BNCT is an ideal starting point for seeking optimal boron compounds for BPCEPT research.

Boron Compounds Used in BNCT

BNCT uses the high cross section (3840 barns) of ¹⁰B to capture thermal neutrons producing two alpha (α, ⁴He²⁺) particles and a ⁷Li particle – ¹⁰B + nth → [¹¹B]*→ α + ⁷Li + 2.3 MeV. The fission products deliver energy in tissue to a depth of ≈9 μm and ≈5 μm, respectively, which is highly biologically effective. The therapeutic potential of the boron neutron capture reaction was first recognized by Locher in 1936 (11) and clinical studies started in the early 1950s. Twenty years after the first unsuccessful human applications (11, 12), intensive experimental and careful clinical investigations have led to the acceptance of BNCT, confirming its efficacy in a clinical setting for brain, thyroid, skin and recurrent head and neck tumors (13-16). BNCT is a highly complex approach as a radiotherapy modality and is considerably different from conventional external beam therapy. In particular, BNCT facilities cannot produce clean monoenergetic neutron radiation and thus the beam is polluted with incident photons and neutrons with a wide spectrum of energies. Thermal neutrons, present in the free beam and higher energy neutrons (epithermal, fast neutrons) are thermalized at the treatment depth and these have different probabilities for capture reactions with chemical elements within the human body. Further contributions to the total dose arise from neutron–hydrogen and neutron–nitrogen reactions. The sources for the total absorbed dose include recoil protons from the scattering of incident fast neutrons by hydrogen, protons emitted by the ¹⁴N (n, p) ¹⁴C neutron capture reaction, and incident photons (gamma radiation) and photons generated in the ¹H (n,γ) ²H reaction (17). Photons generated in the irradiated volume will also deposit an additional radiation dose not only in the desired volume, but also to the surrounding tissues. The biological distribution of ¹⁰B also influences the BNCT-related dose to tissue and consequently the total dose distribution is contributed by several dose components. Clinically useful BNCT requires a high ¹⁰B concentration in the target volume with respect to neighboring healthy tissue and a sufficiently high thermal neutron fluence. The most important requirements for a BNCT delivery agent are low systemic toxicity; selective uptake into the tumor cells; rapid clearance from the blood and the tumor surrounding normal tissues; long-lasting, high intratumoral concentration (20 μg/g tumor)/>100 ppm) and a favorable intracellular distribution, preferably in the cell nucleus.

First-generation boron compounds. The first clinical trials for BNCT were started in the 1950s in the United States by Sweet, Brownell and Farr (18-20) using simple boron agents including sodium tetraborate (borax) (20), sodium pentaborate, p-carboxyphenylboronic acid, or sodium decahydrodecaborate (Na₂B₁₀H₁₀) (21). The failure of patient treatment (22, 23) due to serious treatment sequalae caused both by low penetration of the thermal neutrons and poor tumor/tissue selectivity of the boron delivery agents led to the termination of BNCT in the US at that time. Some decades later, after intensive radiobiological experiments, early clinical research was performed using second-generation boron compounds.

More recent renewed attention has been given to the first simple compounds such as boric acid (BA, H₃BO₃). This small weakly acidic compound is often used as a reference compound for testing other boronated target molecules as it has an accumulation efficiency in cells of 1 (24). This is due to passive diffusion (confirmed by a study in frogs) (25). BA also has no biological activity and has an elimination half-life of 21 h (26).

It has been shown that in GS-9L, U-343MGa, HeLa, V79 and B16 cells, BA distribution is even (27-29). There is, however, favorable BA accumulation in bone and this has potential for use for BNCT of osteosarcoma (30). There is also a US patent for the use of BA in liver BNCT treatment (US20130090513). A dose of about 5-60 mg of ¹⁰B stable isotope per kilogram of patient body weight is used and then the patient is irradiated 10-120 min after the injection when measured blood concentrations of ¹⁰B stable isotope are greater than 20-35 μg /ml. BA is the boron delivery agent of choice for liver BNCT, as the other tested second-generation boron compounds show a preference for accumulation in the pancreas.

Second-generation boron compounds. The neurosurgeon Hatanaka, upon his return to Japan from Sweet's United States research group, continued BNCT treatments in an intraoperative setting (31, 32). In conjunction with Nakagawa, he exposed the surgical cavity surface after brain tumor removal to thermal neutrons in conjunction with BSH, which has more selective accumulation in tumor cells. BSH uptake is a passive process and shows transient binding, with a biological half-life of the order of hours. This agent was the first Food and Drug Administration (FDA)-approved BNCT drug for human application (33). The most relevant human evaluation on the biodistribution of BSH in tumor and normal tissues was published regarding the European Organisation for Research and Treatment of Cancer (EORTC) phase I trial ‘Postoperative treatment of glioblastoma with BNCT at the Petten Irradiation Facility’ (protocol 11961). Twelve hours after boron administration at 100 mg/kg or 22.9 mg/kg body weight, the tumor/blood concentration ratios were 0.6±0.2 and 0.9±0.2, respectively (Table I). The highest boron uptake was detected in the dura and there was an exceptionally low uptake in bone, cerebrospinal fluid and in the brain. BSH is safe for human applications at a dose of 100 mg BSH/kg infused at 1 mg/kg/min for 4 consecutive days (34).

Another BSH biodistribution and pharmacokinetic study looked at 20 patients with high-grade gliomas and confirmed the high interpatient variability of tumor boron concentration and boron concentration in different regions of the tumor (35).

Other representative results from the literature based on the application of second-generation BNCT compounds are summarized in Table I. The BSH dose is typically in the range of 23 200 mg BSH/kg body weight. The use of BSH as a BNCT agent was studied in different types of animal tumor models, such as oral squamous cell carcinoma (36, 37), pigmented melanoma, sarcoma (24) and pancreatic adenocarcinoma in mouse and rat models (38). BSH has already been applied in human clinical trials, such as for patients with glioblastoma (34), with squamous cell carcinoma of the head and neck (37), and in oral mucosal tissue (39).

Favorable ¹⁰B uptake for BNCT treatment would require a tumor/normal tissue ratio greater than 2, and a tumor/blood ratio greater than 0.6 for BSH (40). Table I shows that the measured tumor/normal tissue ratio is frequently below 2 for BSH. The tumor/blood ratio, however, is usually above 0.6 for BSH under the studied conditions.

Boron-containing amino acid (boronophenylalanine) (BPA) was synthesized in 1958 by Snyder et al. (41) and was first used in Japan for BNCT of melanoma (42). Clinical trials were once more initiated in 1990s in the United States and used the epithermal neutron beam at the Medical Research Reactor and at Harvard-MIT. A soluble fructose complex, BPA-F (concentration 400 mg/kg/2 h, i.v.) was used as the boron delivery agent. It was postulated to be taken up by active metabolic pathways in cells, resulting in favorable tumor/blood and similar tumor/brain ratio of more than 1.5-2 (43). The uptake of BPA into glioma and melanoma was measured by tissue sampling in several clinical and preclinical trials. Marked variability was detected in a study using BPA-F involving 98 patients with glioma (44) (Table I). A Japanese clinical study investigated the uptake of BPA in squamous cell carcinoma of the head and neck showed a promising ¹⁰B concentration and, for the first time, demonstrated the change of boron concentration in the tumor over time (36). All these investigations demonstrated significantly larger tumor/normal tissue boron concentration ratios when compared to BSH studies, e.g. 4.2-4.7 in oral squamous cell carcinoma (36), 4.5-7.8 in clear cell sarcoma (45), 8.7-37.3 in a lung metastasis model of clear cell sarcoma (46), and 3.2 in B16 pigmented melanoma (24), however with notable intra-individual and intertumoral variability in boron concentration both in humans and in rodent tumor models (Table I). BPA only has one boron atom per molecule but can be applied at higher doses (180-700 mg/kg body weight) compared with BSH (23-200 mg/kg body weight). Morris et al. applied 55 mg/kg BSH (30 mg ¹⁰B/kg) and 700 mg/kg BPA (34 mg ¹⁰B/kg) which resulted in similar boron dosing and yet the BPA treatment resulted in higher accumulated concentration of boron (6.1 versus 21.5 ppm) in oral mucosal tissue (39). High boron concentrations were found in melanocytic malignant tumor (74 mg ¹⁰B/kg wet tumor tissue) and in a lung metastasis model of clear cell sarcoma (59.8 mg 10B/g). It is assumed that BPA is selectively accumulated in cancer cells via L-amino acid transporter (LAT1)-mediated uptake (47, 48). T98G and A172 glioblastoma cells were used in a recent study published by Wada et al. in which the influence of oxygen concentration on cell viability was investigated, as well as the mRNA expression of LAT1 and the uptake of ¹⁰B-BPA (49). It was concluded that hypoxia below 10% O₂ significantly reduced the mRNA expression of LAT1 in both cell lines. This means that reduced uptake of ¹⁰B-BPA in glioblastoma in hypoxic conditions may be due to a lower expression of this transporter protein. Wongthai et al. reported that BPA is transported by amino acid transporter (ATB^0,+), and LAT1 and LAT2 aromatic amino acid transporters (50). Subsequently, ¹⁸F-labeled BPA derivatives were used for positron-emission tomography in order to determine the pharmacokinetic parameters of BPA and to predict the effectiveness of BNCT, as well as to detect post-treatment radiation effects (51-53).

BSH and BPA have been used independently or in combination for decades in clinical studies in Japan, in the US and in Europe (the multinational European trial, in Czech Republic and in Finland) (34, 39, 40) for BNCT of patients with high-grade glioma, melanoma and recurrent head and neck cancer. The safety and feasibility for clinical application of these two boron delivery agents was established on the basis of the numerous preclinical and clinical trials (40), although neither BSH, nor BPA-F proved to be an optimal boron compound for BNCT.

Third-generation boron compounds. The considerable requirements for boron delivery agents have resulted in the need for further developments in order to achieve a more selective, higher intra-tumoral boron concentration and long-term intra-tumoral binding with favorable subcellular localization, i.e. close to the cell nucleus.

A collection of third-generation low- and high-molecular-weight boron delivery agents is summarized in Table II (54). Amongst the low-molecular-weight BNCT compounds, there are boronated derivatives of other natural amino acids, such as aspartic acid, tyrosine, cysteine, methionine and serine, and boron-containing unnatural amino acids, e.g. designated derivatives of 1-aminocyclobutane-1-carboxylic acid and 1-amino-3-boronocyclopentanecarboxylic acid, and boron-containing linear and cyclic peptides (55). These peptides were developed due to their low toxicity, high tissue-penetrating properties (56) and increased water solubility compared to BPA.

Boronated linear and cyclic peptides have a low toxicity and high tumor selectivity (57). In tumor cells, the ligands for up-regulated receptors, especially epidermal growth factor (EGFR), vascular endothelial growth factor (VEGFR) and somatostatin receptors, are of particular interest (58).

Boron-containing purines, pyrimidines, thymidines, nucleosides, and nucleotides have been investigated as boron delivery agents, especially 3-carboranyl thymidine analogs. For example, the 3-(dihydroxypropyl-carboranyl-pentyl) thymidine derivate designated N5-2OH has shown low toxicity, selective tumor uptake, and high rate of phosphorylation into the corresponding nucleotide (56). Convection-enhanced delivery of N5-2OH was effective in selectively delivering therapeutic amounts of boron to rats bearing intracerebral implants of RG2 glioma (59).

Another approach was to improve the delivery of second-generation agents, using BSH lipiodol emulsion, BSH-transferrin polyethylene glycol (PEG) liposomes, and BPA derivatives with a higher percentage of boron by weight (60). Boron-containing nucleosides and nucleotides, purines, pyrimidines, thymidines (59) and boronated DNA intercalators were designed to achieve high boron concentration in or near to the cell nucleus. Boronated DNA-binding molecules, e.g. alkylating agents, intercalators, minor-groove binders, polyamines, derivatives of acridines, Pt(II) complexes, tri-methoxyindoles, carboranylpolyamines and metal complexes have been investigated and these compounds typically have low tumor selectivity and sometimes high toxicity due to their multiple cationic charges and ability to bind to DNA (60).

Boronated carbohydrates, such as derivates of galactose, glucose, lactose, maltose, mannose and ribose, have also been investigated. These compounds usually display low toxicity and poor tumor uptake, perhaps as a result of their low hydrophobicity and a fast clearance from tissues (60).

A novel idea was to bind boron to porphyrin derivatives, which have been already used for photodynamic therapy, some with FDA approval, in order to establish dual tumor-killing modality (54, 61). Several boron-containing porphyrin derivatives, including chlorin, bacteriochlorin, tetrabenzoporphyrin have been investigated as BNCT agents (62) porphyrins of high-boron content. They are well-known, non toxic, tumor-specific agents, which are exhibited several superior features in comparison to BSH, BPA, such as longer-lasting high tumor accumulation and high tumor/brain, tumor/blood boron concentration ratios. Boronated chlorins and bacterochlorins are promising agents as dual BNCT and photodynamic therapy photosensitizers due to a strong absorption of near-infrared light, which penetrates deeper into human tissues. Boronated derivatives of other amino acids, such as aspartic acid, tyrosine, cysteine, methionine and serine, designated derivatives of 1-aminocyclobutane-1-carboxylic acid and 1-amino-3-boronocyclopentanecarboxylic acid and boron-containing linear and cyclic peptides were developed because of their low toxicity and high tissue-penetrating properties (63).

The radioresistance of hypoxic cells is well known (64), therefore boron compounds with high capacity to accumulate in hypoxic cells would be highly useful for BNCT of large tumors. A boronated 2-nitroimidazole derivative (B-381) was found to have a minimal cytotoxicity and preferentially accumulate in hypoxic cells, and demonstrated longer tumor retention in an in vivo glioma model compared to BPA. It achieved significant tumor/normal tissue ratios as well as tumor/blood ratios but suffered from extremely fast elimination from plasma (64).

Immunoliposomes have been shown to deliver small boronated molecules such as BSH into brain tumors via incorporation into the corresponding lipid bilayers (58). Liposomes have been used to intracellularly transport large amounts of boronated molecules e.g. PEG-based liposomes and particles generally show increased tumor boron uptake (65). High cellular boron concentrations can be achieved by encapsulating low-molecular-weight compounds in tumor-selective unilamellar liposomes. However, in vivo therapeutic efficacy has yet to be demonstrated. These particles avoid problems associated with leakage of encapsulated boronated compounds and were shown to deliver large amounts of boron to tumor-bearing mice with increased survival times following BNCT (54). Other boron-containing nanoparticles have also been proposed and investigated as delivery agents for BNCT (66).

Recently, novel nanotechniques, such as hollow boron nitride nanospheres with potential for intratumoral boron release from within the sphere, have produced nanoparticles with promising results. Hexagonal BN has a high dispersibility in water which makes this material useful for BNCT (67).

Boron Concentration Measurement and Distribution Imaging

A critical requirement at every stage in the development of boronating agents for both BNCT and BPCEPT is the ability to measure the concentration of and image the spatial distribution of boron in biological samples and in living organisms. A multidisciplinary, international effort has resulted in various methods for measurement and imaging boron in practical ways to enable BNCT dose calculation in human treatment and establishing sophisticated techniques for research on subcellular distribution analysis. Prompt gamma-ray analysis (PGRA) is based on gamma-ray spectroscopy following neutron capture in ¹⁰B, resulting in the emission of 478 keV photons. The photon emission rate is proportional to the reaction rate of the neutron capture reaction and therefore the ¹⁰B concentration. The emission of 715 keV PGR due to BPC reaction is controversial, but the same PGRA techniques can be used if there is an optimal mixture of ¹⁰B and ¹¹B and an appropriate neutron beam. PGRA is fast and simple but requires a relatively large sample and produces an average concentration of the sample. Further application of PGRA for BNCT includes the possibility of in vivo gamma-ray spectroscopy of the patient during treatment. This method, however, needs further improvement for implementation in a clinical routine.

Inductively coupled plasma-atomic emission spectroscopy (ICP-AES) uses plasma to produce excited atoms that emit electromagnetic radiation at a wavelength characteristic of a particular element. ICP-mass spectrometry (MS) is a combined method capable of monitoring isotopic speciation for desired ions. ICP-AES and ICP-MS are rapid, reliable methods for quantification of ¹⁰B concentration in macroscopic samples; they have been used in conjunction with BNCT for many years and can be applied for ¹¹B concentration determination for therapy with BPC. PGRA, ICP-AES and ICP-MS measurements integrate the boron concentration over a relatively large volume, therefore individual boron distribution and subcellular boron accumulation cannot be directly obtained.

Further measuring techniques analyze the spatial distribution of boron in a sample e.g. high-resolution alpha-track autoradiography and neutron capture radiography result in a similar resolution to histopathological imaging. Laser microprobe mass analysis was the first technique to reach a subcellular resolution in BNCT (68). Time-of-flight secondary ion MS, particularly laser secondary neutral MS, can quantitatively map the spatial boron distribution in biological matrices with a sub-ppm detection limit. Electron microscopic methods e.g. electron energy loss spectroscopy, are less sensitive but have a much higher spatial resolution (69). However, the latter methods are extremely time-, labor- and cost-intensive (40).

Another important aspect that has had only minimal investigation is the metabolism of BPA and BSH that are actually used in clinical trials for BNCT. Ion-trap MS in combination with proteomic technologies offers a possibility to investigate metabolites and transport of target molecules. Such information is mandatory for optimizing the application of these drugs on a scientifically solid basis. ¹⁸F-Labeling of BPA and the use of positron-emission tomography to assess the molecule in a patient and the detection of boron compounds by magnetic resonance imaging might be feasible (46, 70, 71). Positron-emission tomography with ¹⁸F-BPA has been used for patient selection in recent clinical trials in Japan (72) and Finland (29). It is essential in BPCEPT to be able to measure the ¹¹B concentration with a high resolution at different time points relevant to proton irradiation in the concerned target volume both for patient selection and for the calculation of dose and biological effects.

Boron Compounds for BPCEPT

Comparing the requirements for boron compounds for BPCEPT and BNCT. The common feature of BNCT and BPC is the binary nature of these high LET radiation modalities requiring high intra-tumoral boron concentration which is delivered by a non-toxic boronated agent. A suitable compound can theoretically use the same chemical for both types of treatment, but each enriched with suitable boron isotope (¹⁰B for BNCT, ¹¹B for BPC). However the optimal boron delivery agent for BPCEPT should consider and take into account the differences between BNCT and BPC (38).

BNCT is based on a reaction in which a thermal neutron particle cannot be generated in a form of mono-particle, monoenergetic beam specifically tailored to the target with high accuracy. The BPC reaction is only considered as an additional high LET dose component to standard proton therapy. The cross section of proton boron fusion reaction is about 0.9 barn with 675 keV resonant energy resulting in the possibility of triggering a BPC reaction with extremely high accuracy in the target even in the case of less selective boron accumulation in the body. The much lower cross section of the proton-boron capture reaction does however require a much higher boron concentration to achieve a relevant contribution to the proton dose. Furthermore, BNCT is delivered in a single session or as 2-4 fractions after intravenous infusion with boronated agent but in the majority of proton irradiation cases, the dose is delivered with conventional fractionation during 30-35 consecutive working days. Therefore one of the most important requirements for ¹¹B delivery agent is a long-lasting high intracellular concentration as well as tolerability, which allows the administration of the boron compound on weekly basis (9).

Strategies for ¹¹B compound use and development for BPCEPT. FDA-approved BSH and BPA are useful starting agents for BPCEPT research. Different means of enhancing bio-availability may result in an increase in intracellular concentration and longer intracellular retention. The large number of third-generation boron compounds developed for BNCT has provided a basis for testing agents for BPCEPT which have exhibited high boron delivery in the vicinity of the cell nucleus, with favorable pharmacokinetic features. One promising approach is the development of boronated nucleic acids characterized by low cytotoxicity. Some types of high boron-loaded DNA oligonucleotide modifications as potential dual-action anticancer agents with antisense/anti-oxidant and proposed BNCT activities might also work for BPCEPT. The important biological features of these metallocarborane-modified DNA oligomers are their low cytotoxicity, increased lipophilicity, formation of stable heteroduplexes with complementary RNA strands and their function as effective antisense oligonucleotides (73). These compounds deliver ¹¹B to the cell nuclei, therefore further increasing the biological effect. Moreover, pro- and antioxidant activities were detected depending on the concentration of the modified oligomer used. The metallocarborane modification serves as an infrared-sensitive nucleic acid label for detection. This approach may be of importance for designing therapeutic nucleic acids both for gene silencing as future biotherapeutics and for studies on boron clusters as nucleic acid-modifying units for BPCEPT (73).

Significantly higher tumor/normal tissue ratio for BSH can be achieved by using a BSH-polymer conjugate [PEG-b-poly(glutamic acid) copolymer]. The tumor/muscle tissue ratio was about 5-6 for PEG-b-poly(glutamic acid) copolymer-SS-BSH in subcutaneous human pancreatic adenocarcinoma tumors in mice (Table I). High absolute boron concentration (90 ppm) can be achieved in tumor when BSH polymer conjugate is applied (38).

Some porphyrin derivatives have remarkably high absolute intratumoral boron concentration and long retention time, the two main prerequisites for efficient BPCEPT. Wu et al. reported the synthesis of three copper (II) tetracarboranylphenylporphyrins containing 40 or 80 boron atoms per molecule (74). EMT-6 tumor-bearing mice were given porphyrins in six i.p. injections (3/day within 8 h) over a period of 2 days using a volume of 0.01 ml/g body weight per injection. One of the porphyrins resulted in about 200 ppm tumor boron concentration after 2 days from the last injection and 175 ppm after 4 days of the last injection.

Boron nitride nanotubes exhibit supermagnetic properties and thus have potential as nanovectors for magnetic drug targeting – the delivery of a drug with the required concentration to the target site under the influence of an external magnetic field (75, 76). The intended drug and a suitable magnetically active carrier component would need to be combined into a pharmacologically stable compound. The intrinsic magnetic properties of coated boron nitride nanotubes and their subsequent binding ability make them highly suitable for use as targeted drug-delivery systems (76). Hollow boron nitride spheres are also promising in cancer therapy as a dual-action substrate. They show excellent water solubility and degradability. Typical particle sizes are around 200 nm and they undergo further degradation, providing enhanced permeability and retention effects, thereby enabling passive targeting to tumor sites. A new therapeutic property of hollow boron nitride spheres was demonstrated with controlled release of boron for prostate cancer treatment. Hollow boron nitride spheres induced apoptosis and inhibited the proliferation of both androgen-sensitive and androgen-independent prostate cancer cells (77).