This invention relates to the use in radiotherapy of gadolinium-157, or other isotopes of elements having a high thermal neutron capture cross-section, which produce Auger electron radiation in response to irradiation with thermal neutrons. More particularly, the invention is concerned with the use of such isotopes as radiosensitisers.

Radiosensitisers are substances which when present

10 during irradiation, enhance the cytotoxic effects of radiation. For example, the hypoxic radiosensitiser Misonidazole, enhances the cytotoxic effect of X- and γ- radiation. Although studied for many years, the interaction between radiation and radiosensitiser is l ^* -- complex and difficult to predict. Moreover, as both the radiosensitiser and the radiation are cytotoxic per se_ their use in therapy is limited.

A recent development has been boron neutron capture therapy (BNCT), which involves the use of an isotope which becomes cytotoxic upon activation by a beam of radiation.

In BNCT, the radiosensitiser is B, which comprises about 20% of the naturally occurring element. B has a high capture cross section for thermal neutrons. Capture of the neutrons by B results in fission, yielding an

7 α particle and a Li atom with a combined energy of

2.4 MeV:-

l D°B + ϊVn ^• > - Z 6lά + A tile + 2.4MeV

These particles are highly damaging (they have a high LET - Linear, Energy Transfer), with ranges of the order of one or two cell diameters. Thus if a tumour can ]-, _β preferentially "loaded-up" with B, irradiation with thermal neutrons can effect tumour-specific cell-kill. The damage to normal tissues in the radiation field is largely dependant upon the B concentration in the tumours and consequently the size of the neutron dose required for cell kill.

The potential application of neutron capture to cancer therapy was recognised soon after the discovery of neutrons by Chadwick in 1930, but research and clinical activity in BNCT has been largely confined to a few groups in the US and Japan. In view of the central importance of specific concentration of B in tumours, the clinical focus of BNCT has been on brain tumours, where breakdown of the blood-brain barrier can be exploited, and particularly glioblastoma multiform, due to its low proclivity to metastasize. Over 100 patients have been treated by Hatanaka in Japan, and for 40 patients with grade III-IV gliomas, the 5- and 10- year survival rates

were 20 and 10% respectively. The results for the superficial subgroup (tumours seated less than 6 cm from the cerebral surface) were even better; 58% and 29%. The Japanese studies indicated that the B compound of choice is mercapto undecahydrodecaborate, 5 a ₂ B ₁₂ H _llS H.

The potential of BNCT is not restricted to brain tumours, and several approaches to selective incorporation of various types of B compounds are being explored.

~- ~ In particular, B-phenylalanine, a melanin precursor, and B chloropro azine, a melanin-binder, have been used in a number of preclinical studies with melanoma both in vitro and in vivo. Professor Mishima in Japan has recently treated three melanoma patients by BNCT, using ⁵ B-boronophenyalanine.

In contrast to the ^" hypoxic cell sensitisers which are toxic per se, BNCT is much closer to the ideal; the B compounds are relatively innocuous and the synergism ⁰ between the B and thermal neutrons is dramatic. Moreover the mechanism of synergism by BNCT is straight-forward; the absolute concentration of B in the tumour and the relative amounts in tumour, blood and other surrounding tissue, are the main determinants of the 5 radiotherapy outcome.

Despite its apparent advantages, BNCT has not yet attracted wide support. The main reasons for this would appear to be:-

thermal neutrons at sufficient fluxes for BNCT are only available from nuclear reactors.

penetration of thermal neutrons is limited and there are technical problems in producing epithermal neutrons of sufficient flux,

the requirement for B concentrations of 10-30 ppm can be difficult to achieve without compromising tumour selectivity; this necessitates a larger neutron flux for cell kill than is desirable to minimize damage to surrounding tissue.

Present-day clinical radiotherapy operates on quite narrow margins in terms of radiation effects on tumour tissue compared with normal tissue. Improvement by a modest factor of say 2 to 3, which is the case for hypoxic cell sensitisers, can have considerable impact on the therapeutic result. While it may be argued that, the degree of specific tumour localisation (of B) required for BNCT does not need to be dramatic to be useful and that the localisation factors common place in diagnostic nuclear medicine may be sufficient, there is obviously room for improvement. The present invention is broadly directed to this end and arises from our investigations into the use of Auger radiation in radiotherapy.

Auger electrons are named after their discoverer, Pierre Auger, who detected low energy electrons associated with the photoelectric effect, in the early 1920's. In the photoelectric effect, an X-ray photon interacts with an atom in such a way that an inner shell electron is ejected. The ejected photoelectron leaves a vacant electron orbital which is filled by an adjacent electron. The differences in energy between the donor and acceptor orbitals can be emitted from the atom as an X-ray, or as

an electron of characteristic energy, which is termed a Coster- ronig or Auger electron, depending on whether the orbitals involved are from the same or different electron shells, respectively. X-ray fluorescence and electron emission continues until all vacancies are in the outer valence shell(s). The resulting "shower" of electrons have characteristically low energies and short ranges, and constitute a source of high LET damage focussed in the vicinity of the ionised atom.

10 In this specification the term "Auger electrons" includes both Auger and Coster-Kronig electrons.

The basic requirement for Auger electron emission is an inner shell vacancy so it is not restricted to l ⁵ photoelectric ionisation. Radioactive isotopes that decay by electron capture and/or internal conversion may also emit Auger electrons. The most studied of such isotopes is iodine-125 ( 125I) and the molecular radiation damage that occurs upon decay of 125I in the vi.ci.ni.ty of DNA is ^ reasonably well-understood.

It is well known from classic "suicide" experiments, that incorporation of 125I labelled DNA precursors (eg

IUdr) is particularly cytotoxic; only 30-100 I ⁵ decays are required to kill the average cell. Suicide experiments with bacteriophages and bacteria established that each decay of DNA bound 125I results in approximately one double-strand (ds) DNA break. It will be appreciated by those skilled in the art that DNA is the 0 critical radiosensitive target in cells and that the cytotoxic action of ionising radiation is probably largely mediated by the induction of DNA double strand breaks in cell nuclear DNA. Experiments with DNA fragments

containing a single 125I atom in a defined location have shown that the induced ds DNA break is a consequence of a series of single-strand (ss) breaks in each strand, the majority of breaks occurring within 4-5 bp from the decaying atom. Moreover this extent of damage is

⁵ consistent with the calculated energies and ranges of Auger electrons associated with 125I decay. Other studies have suggested that decay of 125I may induce deletions of DNA fragments near the site of decay.

-- ^■ - The decaying I does not need to be covalently associate with DNA to induce a ds DNA break.

125 I-labelled DNA ligands also induce ds DNA breaks and are cytotoxic as shown by studies involving

¹ 1 ² 2 ⁵ 5I I--labelled aminoacridines and 125I-labelled Hoechst

15 33258.

The therapeutic utility of 125I is limited by radio protective considerations. 125I has a long half life, iodo compounds are easily dehalogenated to yield n 125

^*iU I vapour and there is a low permissible body burden of iodine due to the efficient concentration of that element by the thyroid.

We have now devised a method whereby Auger electrons ² ^ can be produced and utilized to disrupt DNA and other critical cellular components, without the disadvantages associated with 125I, which involves the use of isotopes with a high thermal neutron capture cross section in conjunction with neutron radiation.

30

In the prior art, the rationale for using BNCT often begins with a listing of non with large capture cross sections for thermal neutrons, namely 3He, 10B,

149S-m, 235 _τ U _τ , 6 _T Li,, 113C-.,d,, 157G-d-, 135X„e ( .s _a e _a e

Table 1 at the end of this specification) . With the

. π ~ R fϊ exception of B, U and Li, all of these high cross section isotopes undergo n,γ reactions with thermal neutrons. That is, neutron capture (NC) results in the emission of a gamma ray. The characteristics of the γ-rays vary but all are far more penetrating than α-particles. Hence isotopes other than B have been dismissed as uninteresting; there are far easier (more conventional) means of delivering γ-rays to tumours.

10 The rationale is then completed with the various reasons in ^* why B is preferable to Li or U.

We have now found that isotopes which undergo an n,γ-type reaction with thermal neutrons and have high

^** - ^** ' neutron capture cross sections are efficient radiosensitisers when irradiated with neutron radiation if the Auger electrons produced by internal conversion are utilized. The process of internal conversion involves decay of a metastable nucleus by emission of a

²⁰ γ-ray from the nucleus, and a proportion of these γ-emissions are converted so that an equivalent quantum of energy is emitted as a electron - a "conversion electron". It is the orbital vacancy left by the conversion electron that results in the Auger emissions, ^D ς as discussed earlier. In the case of 125I, the conversion energy is 35.4 KeV and the efficiency of conversion is 93%. We have investigated whether there might be any conversion of the γ-ray energy in n,γ-type reactions, and have focussed our attention on

Gd, because of its spectacularly high cross-section.

157 Gd has a thermal neutron capture cross section approximately 50 times that of B. This high cross

section ensures that even modest concentrations of

157 Gd, say around 10 ppm, on DNA or other critical cellular components will ensure that 157Gd neutron capture will account for the major part of thermal neutron dose and focus the effects of the Auger electrons produced on these components.

Fast neutrons are also used for cancer radiotherapy in some centres, and when the fast neutrons are absorbed by the tissue they are eventually thermalised, but the fluence of such thermal neutrons is much lower than that provided by thermal or epithermal neutron beams from reactors. Combination of fast neutron therapy and BNCT by inclusion of B compounds only boosts the fast neutron dose by a few to several percent, even a very high B concentration (lOOppm) . However, the higher neutron capture cross-section of 157Gd indicates a potential for using 157Gd-labelled DNA ligands in combination with fast neutron therapy.

Likewise, the use of Californium-252 as a source of neutrons (about 2MeV) for brachytherapy of some cancers, may also be enhanced by using 1.57Gd-labelled DNA ligands.

According to one aspect of the present invention, there is provided a radiosensitiser for use in

157 radiotherapy which comprises gadolιnιum-157 ( Gd) or another isotope of an element which has a high thermal neutron capture ^' cross-section and produces Auger electron radiation as a result of irradiation with thermal neutrons.

According to another aspect of the present invention, there is provided a method for inducing

radiation damage in critical cellular components, such as DNA, which comprises irradiating with Auger electron radiation produced in situ by the capture of thermal neutrons by 157Gd or another isotope of an element which has a high thermal neutron capture cross-section.

Usually it will be preferable for the 157Gd and

(or other isotope) to be attached to, or form part of, a ligand or other reactive molecule capable of interaction with the DNA or other critical cell component, although in some cases, it may be possible for the isotope to directly interact with the DNA or other component.

According to another aspect of the present invention, there is provided a method for tumour radiotherapy which comprises administering to a patient in need of such therapy a DNA ligand containing 157Gd or ^' another isotope of an element which has a high thermal neutron capture cross-section and subjecting the locus of the tumour to a flux of thermal neutrons.

The flux of thermal neutrons is preferably provided by a beam of thermal neutrons from any suitable source, e.g. a nuclear reactor. As indicated above, however, fast or epithermal neutrons, which are thermalised by passage through the tumour or its surrounding tissue, may also be used.

The DNA ligand may be of any suitable known type, e.g. an aminoacridine or a bis-benzimidazole. It may also contain or be conjugated with a metal-chelatmg group.

One example of such a ligand is methidium propyl-EDTA.

Advantageously, the 157Gd or other isotope is attached to a ligand or other reactive molecule which, as well as binding to DNA or some other critical cell component, allows enhanced uptake by endocytosis or other means of the radiosensitiser into cells.

5

In a particularly preferred embodiment the ligand is one which is preferentially accumulated in tumour cells, such as tetracycline.

^~ -- The invention is further described and illustrated by the following non-limiting examples.

Reference will be made to the accompanying drawing which is an autoradiograph of an ethidium bromide agarose I ⁵ electrophoresis gel of various plasmid DNA samples.

In the examples, the assay system used was the same as that previously used to detect induction of ds DNA breaks by 125I-labelled DNA ligands. Circular plasmid

20 DNA is incubated/irradiated and then analysed by agarose gel electrophoresis; induction of ds DNA breaks are reflected as conversion of circular to linear plasmid DNA species. On an electrophoresis gel, unbroken, supercoiled plasmids run faster than the linear species that result ²⁵ from induction of a DNA ds break. Relaxed ss nicked plasmids run more slowly than both. The neutron source used was the MOATA reactor at ANSTO, Lucas Heights, in an especially constructed column which produces a dose rate up to about 6Gy per hour (approx. 25% γ component) at a

³⁰ flux of 10 neutrons cm ^" /sec. This experiment was facilitated by the fact that Gd ³⁺ itself binds to DNA by electrostatic interaction with phosphate groups, but with a modest binding constant.

Example 1

Induction of ds DNA breaks by 157Gd neutron capture

Samples of 5 μg of pBR322 in 100 ml of a buffer containing lOmM each of Hepes and Tris HC1 at pH 7.5

(lanes 1-12 in Figure 1) or 50mM Tris HC1 (lanes 13,14). GdCl„ was added to final concentrations of 0.5 mM

(lanes 5,9) or 2.5 mM (lanes 6,7,8,10,11,12) using either naturally occurring Gd (lanes 5,6,7,8) or enriched 157Gd (lanes 9,10,11,12). Some samples contained 10 mM EDTA prior to irradiation (lanes 2,4,7,11,13,14) and the others had the same amount of EDTA after irradiation. Unirradiated controls (lanes 1,2,8,12) were kept at room temperature while the other samples were irradiated in the neutron column at full reactor power for 5 hrs (lanes 13,14) or 6hrs 40 min (lanes 3-7, 9-11). 50μl of plasmid preparation was analysed on ethidium bromide agarose gels. The results are shown in Fig. 1. Control samples were unirradiated (lanes 1,2, 8 and 12) or irradiated in buffer only (lanes 3, 4 and 13). Lanes 5 and 6 show the effect of inclusion of Gd 3+ to final concentration of 0.5 mM and 2.5 mM respectively; ds DNA breakage is evident. Lanes 7 and 8 also have 2.5 mM Gd 3+ but the sample for the former included excess EDTA which reduced ds DNA break induction. Lanes 9-12 are essentially repeats of lanes 5-8, except that enriched (79.7%) 157Gd was used (compared to the natural 15.7%), with consequent increase in DNA damage. Lanes 13 and 14 show the result of a separate experiment in which one sample (lane 14) contained borate to a final concentration of 30 mM in the presence of EDTA. In this case the ds DNA breakage is presumably due to the action of high LET

particles from boron neutron capture. The result also hhiigghhlliigghhttss tthhee ffaacctt tthhaatt 1 ³ 57GGdd iiss a much more active neutron sensitiser than is boron.

The above results show that irradiation of mixtures of Gd 3+ and pBR322 DNA with about 35 Gy of neutrons produced detectable linear species. Addition of EDTA protected against the induction of double-stranded breaks, indicating that the Gd when DNA-bound during irradiation with thermal neutrons resulted in more DNA damage.

10 Titration experiments in which the input ratio of Gd 3+/DNA was varied revealed that ds break induction increased with the input ratio up to about 10 Gd ⁺ per bp, consistent with the low binding constant. Under the conditions of this experiment, breakage was detected at

-*- ⁵ 0.2 mM Gd. The same dose of neutrons required 30 mM of borate (not enriched for B) to produce a similar level of breakage.

TABLE 1 - THERMAL NEUTRON CAPTURE CROSS-SECTIONS (BARNS)*

Isotopes with high cross sections (% abundance) :

³ He: 5,327 (0.00013%) ⁶ Li: 940 ( 7.42%)

¹⁰ B: 3,837 (19.6%) ¹¹³ Cd: 15,787 (12.3%)

¹⁴⁹ Sm: 41 _r 000 (13.8%) ¹⁵⁷ Gd: 242,000 (15.7%)

²³⁵ U: 693 (0.72%) ¹³⁵ Xe: 2.7 x 10 ⁶

* From "Boron Neutron Capture Therapy for Tumours", Ed. H. Hatanaka, Nishimara Co. (1986) and "CRC Handbook of Chemistry and Physics", 49th Edn., 1968/9.