This article is mentioned in:

Contents

Introduction

Hyperthermia

Cisplatin

Gemcitabine

Temozolomide

Halogenated pyrimidines

PARP1 inhibitors

Discussion

Conclusion

Introduction

The treatment of cancer by ionizing radiation is frequently combined with chemotherapy as well as with other agents, in an effort to increase effectiveness. The selection of these combinations has been considerably based on experimental studies with cells in culture and experimental tumours in animals. These studies were designed to obtain insights into the mechanisms of interaction and to derive quantitative information on potential methods of enhancing effectiveness, either through a decrease in cell survival in vitro or an increase in tumour response in vivo.

The results of experiments on combined treatments of cells or tumours are generally expressed in a single sensitisation or enhancement factor (ER), as calculated from dose-effect relationships for the endpoints assessed, following treatment with or without the combined agents. However, such a single sensitisation factor provides only part of the information that can be derived from the complete experimentally assessed dose-effect relationships. Different plating conditions were investigated. Cells were plated prior to irradiation or combined treatment (ppi), plated immediately after irradiation or combined treatment (ip) or plated with a 24-h delay (dp) to establish potentially lethal damage repair (PLDR).

Quantitative information derived from numerous studies on cultured cells, tumours and normal tissues in animals, can be conveniently analysed in terms of mathematical dose-effect relationships based on the linear-quadratic (LQ) model of cell reproductive death as a function of the radiation dose (1–4). The LQ formula for cell reproductive death, the surviving fraction (Sd) of cells exposed to radiation dose (d), compared to the survival of unirradiated cells (S0) is described by an inverse exponential approximation: Sd/S0 = e−(αd+βd2) and contains two parameters, α (Gy−1) and β (Gy−2). The initial slope of cell survival curves and the effectiveness at low doses is determined by α, while β represents the increasing contribution from cumulative damage, presumably due to the interaction of two or more lesions induced by separate ionizing radiations (1). DNA double-strand breaks (DSBs) are generally assumed to be the most relevant lesions.

The α/β ratio represents the dose at which the two terms contribute equally to the total effect. Data from numerous studies on cells, tumours and normal tissues have demonstrated that the values of α/β usually range between 3 and 10 Gy. Since the dose fractions applied in cancer radiotherapy are mostly in the 1.5–2.5 Gy range, it seems clear that the clinical effect of radiotherapy on tumours is largely determined by the linear parameter α.

However, in experimental studies on the enhancement of radiation treatments by chemical or physical agents, the enhancement ratios are typically calculated as the ratio of doses required to obtain equal effects at dose levels for which the effects can be most easily assessed experimentally, namely at doses between 5 and 10 Gy. At these high-fraction doses, the effectiveness is largely determined by the quadratic term βd2 and not by the linear term αd. The common use of a single ER is based on the implicit assumption that any radiosensitiser changes both LQ parameters equally, although this assumption may not hold for all radiosensitising agents. It is therefore of interest to analyse enhancement factors for the linear and quadratic parameters separately, to evaluate their impact at the doses commonly applied in radiotherapy. The parameters α and β were determined from survival curves using SPSS statistical software performing a fit to the data according to the LQ formula by multiple regression analysis.

In this review, radiosensitisation data from our own laboratory are presented and subsequently discussed and compared to the data from the literature. In a number of former studies conducted by our laboratory, radiation sensitisation by a variety of agents in different types of mammalian cells has been investigated and cell survival curves have been analysed using the LQ model (5–11). Radiation enhancement could thus be assessed in separate α- and β-values and in the α/β ratio. The results for various radiosensitising agents are presented in the following sections.

Hyperthermia

Radiosensitisation by hyperthermia

Hyperthermia (HT) refers to heat treatment of cancer cells or tumours by increasing the temperature to a level between 39 and 45°C. It is used in combination with chemo- and/or radiotherapy and it is has been shown to enhance their anticancer effects experimentally and clinically (12–16). A number of in vitro studies on the combination of HT and radiation have demonstrated a synergistic interaction between the two modalities, particularly at higher temperatures (>42°C) (17–19). This interaction possibly results from the inhibition of the repair of radiation-induced DNA damage by HT (20,21). The sequence of combined radiation and HT treatment is important. Optimal sensitisation is achieved when radiation and HT are applied simultaneously or within a short time interval (22). Radiotherapy with concomitant HT is not always feasible in clinical practice. Therefore, in our experiments, HT was also applied sequentially, immediately following radiation treatment.

Despite the clinical goal to realise cytotoxic temperatures as high as 43°C, in practice, tumour temperature distributions are heterogeneous. In large areas of the tumour, temperatures are often <43°C. Nonetheless, satisfactory results have been obtained in locally advanced cervical cancers treated with radiotherapy plus mild HT <43°C (13). Mild temperatures may have more subtle effects than high temperatures, such as tumour reoxygenation (23–26). We recently discovered that mild HT (42°C for 1 h) transiently breaks down the BRCA2 protein (27). In the following sections, the effects of HT for 1 h at 41 or 43°C on the LQ parameters are summarized. Several different cell types have been studied.

Effect of HT on radiosensitivity of SiHa and RKO cells

The SiHa cell line is derived from a human cervical carcinoma. The cells were plated prior to treatment. Mild HT alone (41°C for 1 h) had almost no effect and resulted in a surviving fraction of 0.95±0.2. As can be observed in Fig. 1 and Table I, 1 h at 41.0°C exclusively enhanced the quadratic parameter, β, by a factor of 3.9. The value of the linear parameter, α, was hardly influenced. HT treatment at 43°C for 1 h significantly increased the values of both parameters.

Figure 1 Radiation dose-survival curves of confluent cultures of RKO cells (human colon cancer cells) plated immediately after irradiation (ip, left panel) or 24 h after irradiation (dp, right panel), with or without hyperthermia pre-treatment at 41°C for 1 h. Means with standard errors of at least three experiments are shown.

Table I Values of the linear-quadratic parameters α and β, α/β ratio and enhancement factors from cells treated with ionizing radiation only and following combined radiation and hyperthermia (HT) treatment.

Table I

Values of the linear-quadratic parameters α and β, α/β ratio and enhancement factors from cells treated with ionizing radiation only and following combined radiation and hyperthermia (HT) treatment.

Cells	Treatment °C (h)	α (Gy−1) control	β (Gy−2) control	α/β	α-EF	β-EF
SiHa ppi	Sham	0.33±0.06	0.02±0.01	13.8±6.2
	HT 41 (1)	0.31±0.05	0.09±0.02	3.3±0.7	0.9±0.2	3.9±1.6
	HT 43 (1)	0.76±0.04	0.09±0.01	8.7±0.8	1.4±0.1	2.7±0.5
RKO ip	Sham	0.55±0.09	0.02±0.01	27.5±14.1
	HT 41 (1)	0.93±0.09a	0.05±0.02	18.6±7.7	1.7±0.3	2.5±1.6
RKO dp	Sham	0.47±0.09	0.01±0.01	47.0±47.6b
	HT 41 (1)	0.83±0.08a	0.07±0.02	11.9±3.6	1.8±0.4	7.0±7.3
SW-1573 ip	Sham	0.21±0.02	0.06±0.02	3.5±1.2
	HT 41 (1)	0.06±0.02	0.11±0.03	0.6±0.2	0.3±0.1	1.8±0.8
	HT 43 (1)	0.49±0.04a	0.12±0.03	4.1±1.1	2.3±0.3	2.0±0.8
SW-1573 dp	Sham	0.09±0.02	0.06±0.02	1.5±1.6
	HT 41 (1)	0.05±0.02	0.08±0.02	0.6±0.6	0.6±0.3	1.3±0.6
	HT 43 (1)	0.40±0.04a	0.11±0.03	3.6±1.1	4.4±1.1	1.8±0.8

{ label (or @symbol) needed for fn[@id='tfn1-ijo-42-05-1501'] } Sham, control (radiation only); ppi, plating prior to irradiation; ip, immediately plated; dp, delayed plated.

a Significant from sham P<0.05.

b The α/β has a large variation due to the high uncertainty of the β-value. EF, enhancement factor.

The RKO cell line, derived from a human colon cancer, is relatively sensitive to HT treatment. HT alone for 1 h at 43°C decreased the relative survival to <0.01 and combination with radiation doses in excess of 5 Gy always resulted in the complete absence of colony formation. Mild HT alone (41°C for 1 h) had little effect and resulted in a surviving fraction of 0.8±0.4 in immediately plated (ip) and of 0.9±0.1 in delayed plated (dp) cells. When the cells were heated to 41°C for 1 h immediately prior to irradiation, a significant (P<0.001) enhancement of cellular radiosensitivity was observed in both ip (Fig. 1, left panel) and dp (Fig. 1, right panel) cells (25,28).

The effects of HT on the LQ parameters are summarized in Table I. The value of α increased by a factor of 1.7 to 1.8, while the value of β increased by a factor as high as 2.5 to 7.0. One must bear in mind that the quadratic component βd2 in this cell line is quite small and small absolute changes can result in large relative changes of the numerical values of β.

Effect of HT on radiosensitivity of SW-1573 cells

SW-1573 cells are derived from a human lung tumour and are much less sensitive to HT than RKO cells. Studies have been carried out to evaluate whether pre-treatment with HT at 41 or 43°C can enhance the radiosensitivity of SW-1573 cells (25). HT at 41°C for 1 h without irradiation did not result in a further decrease of the surviving fraction for ip and dp cells, compared to irradiation alone. One-hour HT at 43°C decreased survival to 0.5±0.1 for ip and to 0.4±0.2 for dp cells. Pre-treatment of cells at 41°C for 1 h did not affect cellular radiosensitivity of either ip or dp cells (Fig. 2, left panel). However, 1-h treatment at 43°C resulted in a significant radiation enhancement in both ip and dp cells (p<0.001; Fig. 2, right panel). In Table I, the values of the LQ parameters for radiation alone and for combined treatments are summarised. HT for 1 h at 41°C resulted in an increase of the β-value by a factor of 1.3 to 1.8, while the α-value was decreased. HT treatment for 1 h at 43°C resulted in an increase of the α-value by a factor of 2.3 to 4.4, while the β-value was increased by a factor of 1.8 to 2.0.

Figure 2 Radiation dose-survival curves of confluent cultures of SW-1573 cells (human lung tumour cells) plated immediately after irradiation (ip) or 24 h after irradiation (dp), with or without hyperthermia pre-treatment at 41°C (left panel) or at 43°C (right panel) for 1 h. Means with standard errors of at least three experiments are shown.

Cisplatin

Cisplatin is a widely used anticancer drug that is often combined with radiotherapy (29). Cisplatin-based chemoradiotherapy has become standard treatment for, among others, locally advanced cervical carcinoma (30) and locally advanced non-small-cell lung cancer (NSCLC) (31). There have been numerous studies on the radiation-sensitising effect of cisplatin; however, the results vary from a clear cisplatin-induced radio-sensitisation (24,25,32–34) to a merely additive effect on cell survival (35). Cisplatin and radiation share a common cellular target, DNA (36).

Cisplatin causes DNA damage by inducing the formation of inter- and intrastrand adducts (37). The cisplatin-DNA adducts may cause cell cycle arrest, inhibition of DNA replication and transcription and eventually, apoptosis (38). Repair inhibition of DNA has also been implicated (39). The most important repair pathways reported to be involved in cisplatin-induced DNA damage repair are nucleotide excision repair (NER) and/or homologous recombination (HR) (40,41). An additional route for the repair of cisplatin-DNA interstrand adducts is the post-replication/translation repair pathway which assists the cell in tolerating or bypassing the lesion (42).

Irradiation causes repairable (potentially lethal) and non-repairable (lethal) DNA lesions, which are independently induced. The ultimate effect of the repairable lesions depends on the competing processes of repair and misrepair. The PLDR is reflected by the difference in survival between ip and dp cells. The inhibition of PLDR has been implicated in cisplatin-induced radiation sensitisation (25). More specifically, cisplatin-induced radiation sensitisation is caused by the inhibition of the non-homologous end joining (NHEJ) pathway and recombination repair (38,40,43).

In this section, the radiation sensitisation of cisplatin on the SW-1573 lung tumour cell line and the SiHa cervical tumour cell line is quantified by changes in the LQ parameters (25). Survival curves for SW-1573 lung tumour cells following radiation alone and radiation combined with cisplatin (1 μM for 1 h) are presented in Fig. 3. Cisplatin was added to the cultures immediately prior to irradiation. The survival curves were obtained directly (ip) and 24 h after (dp) treatment to determine PLDR. A slight, but statistically significant effect of cisplatin on radiosensitivity was only observed in the dp cells (P=0.02). This was also described by an increase in the α- and β-values (Table II). An increase in the α-value by a factor of 2.5 was achieved in the dp cells by cisplatin treatment, whereas an increase in the β-value by a factor of 1.2 was observed under both plating conditions. The effects on the LQ parameters of different plating conditions, 1-h incubation with 1 or 5 μM cisplatin and continuous incubation with cisplatin during the complete duration of the clonogenic assay, are also presented in Table II. The radiosensitizing effects are more evident in the SiHa cervical tumour cell line with 1 μM continous cisplatin incubation compared to the SW-1573 lung tumour cell line.

Figure 3 Radiation dose-survival curves of confluent cultures of SW-1573 cells plated immediately after irradiation, (ip, left panel) or 24 h after irradiation, (dp, right panel), with or without 1 μM cisplatin treatment for 1 h. Means with standard errors of at least three experiments are shown.

Table II Values of the linear-quadratic parameters α and β, α/β ratio and enhancement factors from SW-1573 and SiHa cells treated with ionizing radiation only and following combined radiation and cisplatin (1 μM for 1 h; 1 μM continuously; 5 μM continuously) treatment.

Table II

Values of the linear-quadratic parameters α and β, α/β ratio and enhancement factors from SW-1573 and SiHa cells treated with ionizing radiation only and following combined radiation and cisplatin (1 μM for 1 h; 1 μM continuously; 5 μM continuously) treatment.

Cells	Treatment	α (Gy−1) control	β (Gy−2) control	α/β	α-EF	β-EF
SW-1573 ip	Sham	0.21±0.09	0.061±0.016	3.4±1.7
	1 μM cisplatin (1 h)	0.21±0.08	0.072±0.018	2.9±1.3	1.0±0.6	1.2±0.4
SW-1573 dp	Sham	0.10±0.09	0.063±0.016	1.6±1.5
	1 μM cisplatin (1 h)	0.25±0.09a	0.077±0.017	3.3±1.4	2.5±2.4	1.2±0.4
SW-1573 ppi	Sham	0.37±0.12	0.014±0.034	26.4±64.8b
	1 μM cisplatin (cont)	0.41±0.08	0.019±0.025	21.6±28.7b	1.1±0.4	1.4±3.8
	5 μM cisplatin (cont)	0.58±0.20a	0.030±0.008a	19.3±8.4	1.6±0.7	2.1±5.2
SiHa ppi	Sham	0.41±0.04	0.01±0.01	41.0±41.2b
	1 μM cisplatin (cont)	0.81±0.12a	0.02±0.02	40.5±41.0b	2.0±0.4	2.0±2.8

{ label (or @symbol) needed for fn[@id='tfn4-ijo-42-05-1501'] } Sham, radiation only; ip, immediately plated; dp, delayed plated; ppi, plated prior to irradiation.

a Significant from sham P<0.05. α/β values show that in SW-1573 cells the quadratic term is affected more than the linear term, while in SiHa cells only the linear term is significantly increased.

b The α/β has a large variation due to the high uncertainty of the β-value. EF, enhancement factor; cont, continuously.

Gemcitabine

Gemcitabine (dFdC, difluorodeoxycytidine) is a deoxycytidine analogue with clinical activity in NSCLC and pancreatic cancer (44–47). It requires phosphorylation to its active metabolites, gemcitabine-diphosphate (dF-dCDP) and gemcitabine-triphosphate (dF-dCTP), with the initial phosphorylation by deoxycytidine kinase (dCK) being the rate-limiting step (48,49). dF-dCTP inhibits ribonucleotide reductase, the enzyme regulating the production of deoxy-nucleotides, which are necessary for DNA synthesis and repair (50). Deoxynucleotide depletion leads to the increased incorporation of dF-dCTP into DNA, thereby blocking DNA synthesis (masked chain termination). Following the incorporation of dF-dCTP into DNA, an increase in the number of DNA single-strand breaks (SSBs), chromosome breaks and micronuclei has been observed (51).

In vitro and in vivo studies have demonstrated that gemcitabine is a potent radiosensitiser (39,49,52–59). However, in a previous study on NSCLC patients, radiotherapy with concurrent gemcitabine resulted in unacceptable pulmonary toxicity, due to the large amount of radiation delivered to the lungs (60). Phase I trials have demonstrated that radiotherapy combined with gemcitabine at lower doses is feasible without severe pulmonary toxicity (45,61). Its unique mechanism of action, lack of overlapping toxicity and favourable toxicity profile make gemcitabine an ideal candidate for combination therapy (45). There are numerous ongoing randomized studies in which radiotherapy is combined with gemcitabine.

Gemcitabine radiosensitisation has been investigated in gemcitabine-sensitive and -resistant human lung tumour cells, SWp and SWg, respectively, as well as in gemcitabine-sensitive and -resistant human ovarian tumour cells, A2780 and AG6000, respectively (62–64). Gemcitabine was administered 24 h prior to radiation treatment (64). The SWp cell line is similar to the SW-1573 cell line described above. It is termed SWp to distinguish it from SWg, the gemcitabine-resistant counterpart which was developed by van Bree et al(64). Lung tumour cells exhibit different sensitivities to radiation alone as compared to ovarian cancer cells (62–64).

Table III summarizes the LQ parameters of the different cell lines obtained following analyses of the radiation dose-survival curves for irradiation alone and following combined irradiation and gemcitabine. SWp and SWg cells were almost equally sensitive to ionizing radiation alone with respect to the low-dose region described by the α-value. A slight increase in survival was observed in the SWg cells under the high-radiation dose region (>4 Gy), which was reflected by a slightly lower β-value (0.040±0.006 vs. 0.055±0.008). The A2780 human ovarian carcinoma cell line and its gemcitabine-resistant variant, AG6000, were equally sensitive to ionizing radiation. The surviving fractions of the different cell lines following incubation with gemcitabine alone were as follows: SWp cells: 10 nM, 0.52±0.06; SWg cells: 10 μM, 0.95±0.03; 100 μM, 0.24±0.11; A2780 cells: 2 nM, 0.82±0.08; 10 nM, 0.21±0.08; AG6000 cells: 20 μM, 0.62±0.07; 50 μM, 0.22±0.04.

Table III Values of the linear-quadratic parameters α and β, α/β ratio and enhancement factors from cells treated with ionizing radiation only and gemcitabine-sensitised radiation dose-survival curves of gemcitabine-sensitive (SWp and A2780) and gemcitabine-resistant (SWg and AG6000) cells.

Table III

Values of the linear-quadratic parameters α and β, α/β ratio and enhancement factors from cells treated with ionizing radiation only and gemcitabine-sensitised radiation dose-survival curves of gemcitabine-sensitive (SWp and A2780) and gemcitabine-resistant (SWg and AG6000) cells.

Cells	Treatment	α (Gy−1) control	β (Gy−2) control	α/β	α-EF	β-EF
SWp	Sham	0.10±0.03	0.055±0.008	1.8±0.6
	10 nM gemcitabine	0.30±0.06a	0.053±0.007	5.7±1.4	3.0±2.8	0.96±0.2
SWg	Sham	0.09±0.02	0.040±0.006	2.3±0.6
	100 μM gemcitabine	0.09±0.03	0.090±0.041b	1.0±0.6	1.0±0.5	2.3±1.1
A2780	Sham	0.80±0.10	Na
	10 nM gemcitabine	1.10±0.15a	Na		1.4±0.3
AG6000	Sham	0.83±0.13	Na
	50 μM gemcitabine	1.11±0.20b	Na		1.3±0.3

{ label (or @symbol) needed for fn[@id='tfn7-ijo-42-05-1501'] } Significant difference with

a P<0.01;

b P<0.05; Na, not applicable. EF, enhancement factor.

As depicted in Fig. 4 and Table III, radiosensitisation is observed in both gemcitabine-sensitive and gemcitabine-resistant cells. However, much higher gemcitabine doses were required for the radiation sensitisation of gemcitabine-resistant cells to result in similar cytotoxicity. Both gemcitabine-sensitive cell lines (SWp and A2780) were sensitised by incubation with 10 nM of gemcitabine for 24 h prior to irradiation, while the SWg and AG6000 cell lines were not radiosensitised by this dose of gemcitabine. Radiosensitisation of the two gemcitabine-sensitive cell lines was reflected by an increase in the α-values by a factor of 3 and 1.4, respectively, whereas the β-values were not significantly altered. Higher concentrations of gemcitabine (50 and 100 nM) were required to sensitise the gemcitabine-resistant AG6000 and SWg cells to irradiation. For the SWg cells, the radiosensitisation was reflected by an increase in the β-value by a factor of 2.25 in, whereas in the AG6000 cells, only the α-value was increased by a factor of 1.3.

Figure 4 Radiation sensitisation following a 24-h incubation with different concentrations of gemcitabine in gemcitabine-sensitive SWp and -resistant SWg lung tumour cells and in gemcitabine-sensitive A2780 and -resistant AG6000 ovarian cancer cells. Surviving fractions are corrected for gemcitabine toxicity alone (for values see text). Cells are plated immediately after irradiation. Means with SEM of at least three separate experiments are shown.

Temozolomide

The combination of fractionated radiotherapy with temozolomide (TMZ) has significantly improved the survival of patients with newly diagnosed glioblastoma multiforme (GBM) (65,66). The combination of radiotherapy and TMZ has become standard therapy for GBM patients. The benefits of TMZ are most prominent for tumours with a methylated O6-methylguanine-DNA methyltransferase (MGMT) promoter: methylation of the MGMT promoter has been associated with a longer overall survival of GBM patients treated with radiotherapy and TMZ, compared to radiotherapy alone (67,68). Van Nifterik et al(10) demonstrated a relatively lower cell survival in methylated GBM cell lines following treatment with radiotherapy and TMZ, which suggests an interaction between TMZ and irradiation.

TMZ is a chemotherapeutic prodrug that transforms under physiological conditions into its active unstable methylating metabolite, 5-(3-methyl-1-triazeno)imidazole-4-carboxamide (MTIC). Methylation of DNA by MTIC results in the formation of O6-methylguanine adducts. These adducts are considered to be responsible for the cytotoxic effects of TMZ (69,70). O6-methylguanine adducts can result in failure of the mismatch repair system, leading to DNA double-strand breakage and eventually, cell death (71,72).

O6-methylguanine-DNA methyltransferase is a cytoprotective DNA repair protein that can remove the methyl group from the O6 position of guanine. Therefore, presence of this repair protein may undo, in part, the cytotoxic effect of alkylating agents, hence resulting in tumour resistance to TMZ (73,74). Hypermethylation of the CpG islands in the promoter region of the MGMT gene has been found to be associated with transcriptional silencing (74,75) and, subsequently, with a good clinical response to alkylating agents in glioma patients (76,77).

Few studies have been published on the radiosensitising potential of TMZ for glioma cell lines using different treatment protocols. In certain cell lines, an enhancement of the radiation effect has been demonstrated, whereas other cell lines have shown no interaction, but merely an additive effect (78–83).

In this review, we present the results of combined TMZ-radiation treatment on three long-term primary TMZ-sensitive glioma cell lines (Table IV). These three cell lines contain a MGMT promoter region that is for the most part, methylated and do not express the MGMT protein (10,84,85).

Table IV Values of the linear-quadratic parameters α and β, α/β ratio and enhancement factors from cells treated with ionizing radiation only and temozolomide-sensitised radiation dose-survival curves of three glioma cell lines AMC-3046, VU-109 and VU-122 with different sensitivities to temozolomide (10).

Table IV

Values of the linear-quadratic parameters α and β, α/β ratio and enhancement factors from cells treated with ionizing radiation only and temozolomide-sensitised radiation dose-survival curves of three glioma cell lines AMC-3046, VU-109 and VU-122 with different sensitivities to temozolomide (10).

Cell line	Treatment	α (Gy−1)	β (Gy−2)	α/β (Gy)	α-EF	β-EF
AMC-3046	Sham	0.014±0.033	0.065±0.007	0.22±0.06
	Temozolomide	0.43±0.025a	0.009±0.005	47.6±26.7	30.7±7.5	0.14±0.8
VU-109	Sham	0.14±0.031	0.037±0.006	3.8±1.0
	Temozolomide	0.19±0.038	0.032±0.008	6.0±1.9	1.4±0.4	0.9±0.3
VU-122	Sham	0.11±0.025	0.063±0.005	1.8±0.4
	Temozolomide	0.21±0.047b	0.067±0.010	3.1±0.8	1.9±0.6	1.1±0.2

{ label (or @symbol) needed for fn[@id='tfn10-ijo-42-05-1501'] } Significant difference with

a P<0.01,

b P<0.05. EF, enhancement factor.

The cells were exposed to isotoxic doses of TMZ for 96 h prior to γ-irradiation. A significant radiosensitising effect (P<0.05) of TMZ was demonstrated in the AMC-3046 glioma cell line (Fig. 5, left panel). The shoulder of the survival curve for irradiated cells disappeared as a result of pre-treatment with TMZ. This was also reflected by TMZ-induced changes in both the α and β parameters of the LQ model (Table IV). No radiosensitisation was observed in the VU-109 glioma cells (P=0.054; Fig. 5, middle panel), as demonstrated by the unaffected α and β parameters (Table IV). The VU-122 glioma cells displayed a small but significant radiosensitising effect of TMZ (P<0.05; Fig. 5, right panel), which was most obvious in the lower-radiation dose range. This difference was reflected by an increase in the α parameter without any change in the β parameter (Table IV).

Figure 5 Radiation dose-survival curves of glioblastoma cells AMC-3046, VU-109 and VU-122 without (closed symbols) or with (open symbols) incubation with temozolomide (TMZ). Cells were incubated with 10 μM (AMC-3046), 3 μM (VU109) and 2.5 μM (VU-122) TMZ for 96 h (daily refreshed) followed by irradiation.

Halogenated pyrimidines

Incorporation of halogenated pyrimidines (HPs), chloro-, bromo- and iodo-deoxyuridine (CldUrd, BrdUrd, IdUrd) into DNA is known to sensitise cells to ionizing radiation (6,8,11,86–93). The induced radiosensitisation increases with the degree of thymidine replacement. The mechanism of radiosensitisation by the HPs has been suggested to be due to an increase in the amount of DNA damage induced by radiation, an influence on repair of sublethal damage (SLD), or an enhanced expression of potentially lethal damage (PLD) (6,94). Since different processes are involved in these phenomena, several mechanisms may contribute to the radiosensitisation.

HPs have been suggested to provide an advantage in radiotherapy as radiosensitisers of cells in rapidly growing tumours, particularly under clinical conditions in which critical normal tissues show limited proliferation, and as a consequence, take up less HP. Labelling depends on the growth fraction, cell loss, cell cycle time and potential doubling time. Of particular importance for sensitisation is the rate at which non-cycling cells are recruited into the proliferative compartment during exposure to HPs and a course of radiotherapy. However, even in rapidly growing tumours, cells may, following proliferative cycles, retreat into a non-proliferative state. This may compromise the degree of radiation sensitisation, since resting cells are less affected by HPs, or are better able to cope with additional damage by PLDR.

In this review, we present the results of radiosensitisation following incubation with 4 μM of IdUrd for 72 h. IdUrd-induced radiosensitisation was observed in all the studied cell lines, SW-1573, RUCII (rat ureteral carcinoma), R1 (rat rhabdomyosarcoma) and V79 (Chinese hamster lung cells), in exponentially growing and in plateau-phase cells. Values of α and β derived by LQ analysis of the survival curves of exponentially growing and plateau-phase cells are summarized in Table V. Fig. 6 depicts the survival curves of SW and V79 cell lines. The plating conditions of the V79 cells, i.e., exponentially growing cells plated prior to or after irradiation (ppi or pai, respectively), and plateau-phase cells, plated immediately or 6–24 h after irradiation (ip or dp, respectively), had no influence on the enhancement factor of the α-value. It is demonstrated that the α-value can be enhanced by a factor of 1.9 to 7.5 and that, in general, low α-values are more enhanced than higher α-values. The value of β is less enhanced and its enhancement factor ranges from 0.7 to 2.4.

Figure 6 Radiation dose-survival curves of exponentially growing cells (left) without IdUrd (open triangles) and after incubation with 4 μM IdUrd (closed triangles) and plateau-phase cells (right) plated immediately after irradiation (dashed lines) and plated 24 h after irradiation (solid lines) without IdUrd (open symbols) and after incubation with 4 μM of IdUrd (closed symbols). Control curve ip: open circles; control curve dp: open squares; idUrd ip: closed circles; idUrd dp: closed squares. Each point represents the mean value of three different experiments ± SEM.

Table V Values of the linear-quadratic parameters α and β, α/β ratio and enhancement factors of several cell lines treated with ionizing radiation only and after sensitisation with iododeoxyuridine (IdUrd) (incubation with 4 μM of IdUrd for 72 h).

Table V

Values of the linear-quadratic parameters α and β, α/β ratio and enhancement factors of several cell lines treated with ionizing radiation only and after sensitisation with iododeoxyuridine (IdUrd) (incubation with 4 μM of IdUrd for 72 h).

Cell line	α (Gy−1) control	β (Gy−2) control	α (Gy−1) IdUrd-sens	β (Gy−2) IdUrd-sens	α/β control	α/β IdUrd-sens	α-EF	β-EF
SW-1573 cells Exp growing ip	0.22±0.01	0.022±0.001	0.83±0.06	Na	10.0±0.6	Na	3.8±0.3	Na
SW-1573 cells Plateau-phase ip	0.17±0.03	0.042±0.004	0.31±0.03	0.047±0.005	4.1±0.8	6.6±1.0	1.8±0.4	1.1±0.2
SW-1573 cells Plateau-phase dp	0.09±0.02	0.046±0.002	0.37±0.04	0.033±0.006	2.0±0.4	11.2±2.4	4.1±1.0	0.7±0.1
RUCII cells Exp growing ppi	0.008±0.007	0.025±0.001	0.06±0.02	0.026±0.001	0.3±0.3	2.3±0.8	7.5±7.0	1.0±0.1
R1 cells Exp growing ppi	0.23±0.01	0.068±0.003	0.44±0.05	0.075±0.016	3.4±0.2	5.9±1.4	1.9±0.3	1.1±0.2
V79 cells Exp growing ip	0.18±0.02	0.017±0.003	0.38±0.04	0.023±0.007	10.6±2.2	16.5±5.3	2.1±0.3	1.4±0.5
V79 cells Exp growing ppi	0.15±0.02	0.013±0.003	0.29±0.03	0.016±0.004	11.5±3.1	18.1±4.9	1.9±0.3	1.2±0.4
V79 cells Plateau-phase ip	0.09±0.03	0.026±0.004	0.17±0.02	0.062±0.005	3.5±1.3	2.7±0.4	1.9±0.7	2.4±0.4
V79 cells Plateau-phase dp	0.07±0.02	0.020±0.002	0.30±0.03	0.024±0.004	3.5±1.1	12.5±2.4	4.3±1.3	1.2±0.2

[i] Means with SEM of at least three separate experiments are shown. ip, immediately plated after irradiation; dp, delayed plated after irradiation; ppi, plated prior to irradiation; Na, not applicable. EF, enhancement factor; IdUrd-sens, sensitisation with iododeoxyuridine; Exp growing, exponentially growing.

The direct comparison between immediate and delayed plating of plateau-phase cells and between plateau-phase and exponentially growing cells shows significant quantitative differences. The data on the LQ parameters presented herein provide various new insights into the interpretation of radio-sensitisation of dp plateau-phase cells. It is demonstrated that in dp HP-sensitised plateau-phase cells PLD is not abolished.

PARP1 inhibitors

The effect of inhibition of poly(ADP-ribose) polymerase-1 (PARP1) by olaparib on the LQ parameters was examined in mouse embryonic fibroblasts (MEFs). PARP1 is an enzyme which is involved in the repair of DNA SSBs. The DNA SSBs induced by ionizing radiation are mostly repaired by the base excision repair (BER) system, whereas the DNA DSBs are repaired by NHEJ or by HR. Inhibiting PARP1 activity reduces the repair of SSBs (95). Apart from its role in BER, PARP1 is further involved in a number of nuclear processes, such as DNA replication, transcription, DSB repair, apoptosis and genome stability (95–97). It was recently hypothesised that cells deficient in BRCA2 or BRCA1 are particularly sensitive to PARP1 inhibition (27,96). SSBs are induced during DNA replication. In the absence of PARP1, these SSBs transform into DSBs. These DSBs are repaired with HR. Therefore, cells deficient in HR (e.g., BRCA1 or BRCA2 tumours) may be sensitive to PARP1 inhibitors. Since PARP1 is involved in numerous DNA repair processes, PARP1 inhibitors may function effectively as radiosensitisers (97). As can be observed in Fig. 7, we achieved a modest sensitisation effect by the PARP1 inhibitor NU-1025 in the MEF cell lines. The increase of the α-value in the repair-deficient cell line was greater than in the repair-proficient cell line, 1.4 vs. 1.2, respectively (Table VI). The radiation dose-survival curves of these MEF cells did not exhibit a shoulder and therefore the quadratic parameter β could not be determined.

Figure 7 Radiation dose-survival curves of mouse embryonic fibroblasts (MEFs). Ligase IV- and RAD54-positive (LigIV+/+, Rad54+/+, left panel) and Ligase IV-and RAD54-deficient (LigIV−/−, Rad54−/−, right panel) MEFs. Open circles, radiation only curves; closed squares, radiation with PARP inhibitor (parpi). Cells were treated with 100 μM of Nu-1025 for 24 h prior to irradiation.

Table VI Values of the linear parameter α and the enhancement factors from repair-proficient and repair-deficient MEF cells.

Table VI

Values of the linear parameter α and the enhancement factors from repair-proficient and repair-deficient MEF cells.

MEF cells	Treatment with PARPi	α (Gy−1)	β (Gy−2)	α-EF
LigIV+/+, Rad54+/+	No	0.28±0.01	Na
LigIV+/+, Rad54+/+	Yes	0.33±0.03	Na	1.2
LigIV−/−, Rad54−/−	No	1.59±0.18	Na
LigIV−/−, Rad54−/−	Yes	2.28±0.42	Na	1.4

[i] The quadratic parameter β could not be determined in these MEF cells. Na, not applicable; EF, enhancement factor. PARPi, PARP inhibitor.

Discussion

Radiosensitisation by a variety of chemotherapeutic agents is in most cases reflected by an increase of the linear or α component of the LQ model, which corresponds to an enhanced direct PLD at low radiation doses (1,4–7,98,99). The β component, which presumably depends on the interaction of repairable SLD, is affected by HT treatment. Furthermore, it appears that radiosensitisation is more pronounced in radioresistant than in radiosensitive cell lines. In addition, it can be concluded that the extent of radiosensitisation also depends on cell cycle stage (plateau or exponentially growing phase) and post-treatment plating conditions.

Hyperthermia

Hyperthermia (HT) is a very potent radiosensitiser, already effective at mild temperatures. HT for 1 h at 41°C without radiation exerted only a slight cytotoxic effect in both heat-sensitive and heat-resistant cell lines. This is in agreement with the general idea of cell kill induction at temperatures ≥42°C for 1 h or more (23). HT at 43°C for 1 h did not have a significant cytotoxic effect in heat-resistant SW-1573 cells. Radiosensitisation by HT at 41°C was observed in SiHa and RKO, but not in SW-1573 cells. The ability of mild HT (40–42°C) to increase radiosensitivity of human tumour cells has been shown to be cell line-dependent (8,26,100–105). In a study by Xu et al, pre-treatment of cells at 41.1°C for 1 h did not induce radiosensitisation, whereas treatment for 2 h or more resulted in radiosensitisation in the HT-resistant but not in the HT-sensitive cell line (106). However, simultaneous treatment of the sensitive cell line with 1-h 41.1°C HT combined with irradiation increased cellular radiosensitivity (107). In vivo radiosensitisation by mild HT is usually attributed to reoxygenation of tumours by an increase in blood flow (108–110). We recently discovered that the BRCA2 protein is transiently inhibited by mild HT (27). Translocation of the Mre11 DSB repair protein from the nucleus to the cytoplasm has also been implicated (106,111). However, disappearance of Mre11 protein foci at the sites of irradiation-induced DNA DSBs was not observed by pre-incubation of cells at 41°C (24,27). A role for mitotic catastrophe, occurring as a result of G2/M checkpoint abrogation, has also been suggested (112). It has been demonstrated that radiosensitisation by HT at 41–43°C correlates with an increased number of chromosomal fragments, but not of colour junctions, 24 h after treatment, compared to radiation alone (101). HT at clinically reachable temperatures mainly enhances the quadratic parameter, β, which represents the frequency of induction-repairable SLD. The fact that HT breaks down the repair protein BRCA2 and in this way influences DNA DSB repair correlates well with the effect on the repairable factor β.

Cisplatin

Cisplatin causes radiosensitisation as measured by clonogenic survival, but only after allowing a PLDR time of 24 h. These results are in agreement with those of Wilkins et al(113), who investigated the effect of cisplatin and radiation on PLDR in confluent cultures of two different brain tumour cell lines (113). Wilkins et al did not observe radiosensitisation by cisplatin in ip cells, whereas a cisplatin-induced radiosensitisation was observed in dp cells 8 h following irradiation (114). Their results indicate that the radiosensitising effect of cisplatin is caused by the inhibition of post-irradiation recovery. The strongest inhibition of PLDR was achieved when cisplatin was administered shortly before or after irradiation (113,114). In our experiments, cells were irradiated with cisplatin present in the medium.

Results from studies using exponentially growing cell cultures vary from a cisplatin-induced radiosensitisation (31–33,115), to a merely additive effect (29,31,116–118). The effect of cisplatin treatment on radiosensitivity may depend on the cell type used. Loprevite et al(31) observed synergism in a squamous lung carcinoma cell line when exposed to cisplatin, whereas an adenocarcinoma of the lung was not sensitised by cisplatin (31). Even cell lines derived from a single biopsy can differ in their response to combined treatment with cisplatin and irradiation (116).

Although dependence on cell cycle phase (119,120), cisplatin incubation time and sequence of treatment modalities have been implicated (29,119,120), there is currently no consensus to account for the varying response of cells to cisplatin and irradiation.

The mechanism of cisplatin-induced radiosensitisation may be due to the inhibition of the DNA repair NHEJ and HR pathways (38,43). The Ku protein complex, which plays an important role in NHEJ, was demonstrated to show a reduced ability to translocate on DNA containing cisplatin-DNA adducts compared with undamaged DNA. This resulted in a decreased interaction between Ku and DNA-dependent protein kinase catalytic subunit (DNA-PKcs) (121). However, the biochemical processes that cisplatin undergoes in the cell are complex and its intracellular fate may be linked to copper transport (122). Therefore, other processes, such as the formation of peroxy complexes inside the cell, may be involved in cisplatin-induced radiosensitisation (123). Bergs et al(34) demonstrated an increase in the induction of apoptosis 24 h after combined treatment as compared to radiation or cisplatin alone. This was confirmed by several other studies (124,125). These apoptotic effects observed by Bergs et al correlated with clonogenic survival (34). Fujita et al(126) also observed an inhibitory effect of the combination of cisplatin and radiation on the survival of lung tumour cells and ascribed this effect to the induction of tumour cell apoptosis (126).

In conclusion, the radiosensitising effect of cisplatin on cell survival was observed in confluent cultures when cells were replated after a 24-hour incubation period during which PLDR was allowed to occur. By contrast, cisplatin did not induce a significant radiosensitisation after immediate plating.

Gemcitabine

A number of previous studies have demonstrated that gemcitabine is a potent sensitiser to ionizing radiation (49,55,127). Among other proposed mechanisms of action, the effect of gemcitabine on cell cycle distribution may be the most important (55,57). In our study, the gemcitabine-sensitive cell lines SWp and A2780 were sensitised to irradiation following administration of cytotoxic gemcitabine treatments. The radiosensitisation was accompanied by a clear arrest of cells in early S phase, which has been argued to be vital for gemcitabine-induced radiosensitisation (54). Both cell lines showed an increase in the α-value, indicating the efficacy of gemcitabine-induced radiosensitisation in the clinically relevant dose range. The gemcitabine-resistant cells were also sensitised, although with only much higher gemcitabine doses. In the resistant AG6000 ovarian carcinoma cell line, this was demonstrated by an increase in the α-value. By contrast, in the gemcitabine-resistant lung tumour cell line an increase in the β-value was observed, while the α-value was not affected. In both gemcitabine-resistant cell lines the sensitivity to ionizing radiation alone was not altered. It is reported that gemcitabine-resistant tumours are cross-resistant to related drugs like Ara-C (128,129). In the AG6000 and SWg gemcitabine-resistant cell lines, this was indeed the case (64). Moreover, the AG6000 cells were more resistant to cisplatin and taxoids as well (62). However, no altered sensitivity was observed in SWg cells for cisplatin, paclitaxel, methotrexate (MTX) and 5-fluorouracil (5-FU), while AG6000 cells were 2.5-fold more sensitive to MTX (62). These findings indicate that patients previously treated with gemcitabine may still benefit from radiotherapy combined with cisplatin or paclitaxel.

Temozolomide

The potential of TMZ to enhance the radiation response in long-term primary GBM cell lines has been clearly demonstrated (10) and therewith, the rationale for the clinical use of this drug concomitantly with radiotherapy. A distinct increase of the α parameter is shown following treatment with TMZ. The three MGMT promoter methylated cell lines discussed above responded differently to the combination treatment, even though they were treated at similar TMZ sensitivity levels. The combined effect of TMZ and radiation was found to be synergistic and is at least additive.

Halogenated pyrimidines

Radiosensitisation by halogenated pyrimidines (HPs) is mainly due to an increase in the linear parameter α. The quadratic parameter β is rarely influenced. Different radiosensitisation mechanisms induced by HPs have been described. Wang et al(130) suggested that increased DNA damage was the major component of radiosensitisation in exponentially growing cells, while in plateau-phase cells, radiosensitisation occurred through inhibited repair and/or enhanced fixation of PLD (6,90). The increase of the α-values for exponentially growing cells, as presented in our study, indicates an increase in the number of directly lethal events due to the HPs. This is in agreement with the observations of Webb et al(131) and Jones et al(94), which suggested that an important mechanism of radiosensitisation involves an increase of effective DNA DSBs (87,89,130). Miller et al(92,93) suggested that radiation-induced damage in cells which have HPs incorporated into the DNA after low-LET radiation resembles the damage produced by high-LET radiation. In plateau-phase cells plated immediately after irradiation, the increase of α might be due to the same mechanism as involved in exponentially growing cells. In these cells an increase of β was also observed, indicating that accumulation of SLD was a major contributor (2). Due to the immediate plating after irradiation this SLD may be fixated.

The most significant increases in the α value were observed in dp plateau-phase cells. This radiosensitisation may be interpreted as an enhanced fixation of PLD due to immediate DNA damage and/or to damaged DNA repair function in these cells, expressed during the interval before delayed plating. The value of β in these cells returned to values found in cells not containing HPs. This demonstrates that SLD was repaired in HP-containing plateau-phase cells.

PARP1 inhibitors

Since PARP1 has been implicated in several DNA repair processes, PARP1 inhibitors may be good radiosensitisers. Several studies have already demonstrated the radiosensitising effect of PARP1 inhibitors (27,95,132). Löser et al(97) concluded that the effects of PARP1 inhibitors are more pronounced on rapidly dividing and/or DNA repair-deficient cells (95). In our study, at the time of treatment most of the cells in culture were accumulated in G1 phase. Therefore, radiosensitisation effects were modest. However, the increase in the α-value in repair-deficient cells was more pronounced following PARP1 inhibition than in repair-proficient cells.

Conclusion

The increase in the α parameter by the various radiosensitising agents yields promising perspective for clinical practice. The radiation tolerance dose is generally expressed quantitatively in terms of the biologically effective dose (BED), as defined by the LQ model. BED takes total dose, dose per fraction, dose rate and overall treatment time into account. By definition, BED is the total dose required to obtain an equal biological effect E (isoeffect) for a certain endpoint, e.g., few log cell kill (e.g., 10−2 cell survival), a normal tissue effect (e.g., 1% complication rate) or tumour response (e.g., tumour cure rate of 50%) when applying an infinite number (∞) of tiny dose fractions (∼0). At this point, the α parameter is inversely proportional to BED [BED = E/α; (133)]. With increasing α, BED is decreasing, resulting in a lower radiation tolerance dose for a certain isoeffect.

The radiosensitising effect of HT on the LQ parameters seems to be temperature-dependent. HT for 1 h at 41°C increases the β-value while HT for 1 h at 43°C increases both the α- and β-values. Increase of the β-value consequently lowers the α/β ratio, which makes the tumour more sensitive to higher fraction doses. The effects of HT on BED remain to be elucidated.

Radiosensitising agents that selectively sensitise tumour cells and not normal tissue cells will be of therapeutic benefit. This is due to the increase of the α parameter of tumour cells only and, as a consequence, the decrease in tumour response dose proportionally to the relative increase in α. This effect could be further exploited using smaller sized fractions in external beam radiotherapy or by lowering the dose rate in brachytherapy. With lower fraction size or lower dose rate, normal tissue cells with low α/β ratios will tolerate a higher total dose. Tumour tissues with high α/β ratios will not exhibit an increased sparing effect, since an increase in α by a radio-sensitising agent will further increase the α/β ratio, resulting in even less sensitivity to a modification in fraction size. However, since an increase in the number of fractions or a lower dose rate might increase the overall treatment time, tumour cell repopulation rate should be taken into account.

Acknowledgements

We thank Jan Sijbrands for his technical support. The authors are thankful for financial support from several foundations. The Maurits and Anna de Kock and the Nijbakker Morra foundations are acknowledged for sponsoring laboratory equipment. The Dutch Cancer Foundation (grant nos. UVA 2006-3484, UVA 2008-4019 and UVA 2012-5540) and the Stichting Vanderes are acknowledged for financing personnel support.

References