- Email: [email protected]
Towards prediction and modulation of treatment response
Radiotherapy and Oncology 50 (1999) 1±11
Breur Award Lecture
Towards prediction and modulation of treatment response Harry Bartelink a,*, Adrian Begg b, Jose Coco Martin b, Mariska van Dijk b, Laura van 't Veer b, Paul van der Vaart b, Marcel Verheij a b
a Department of Radiotherapy, The Netherlands Cancer Institute, Plesmanlaan 121, 1066 CX Amsterdam, The Netherlands Department of Experimental Therapy, The Netherlands Cancer Institute, Plesmanlaan 121, 1066 CX Amsterdam, The Netherlands
Received 24 August 1998; accepted 14 October 1998
Abstract The purpose of this paper is to evaluate new predictive assays and their potential to modulate treatment response. Their impact is presented in the context of three EORTC clinical trials in head and neck, lung and breast cancer, showing an improvement in survival by accelerated fractionation, concomitant use of cisplatin and radiotherapy and adjuvant hormonal treatment, respectively. Assays have been developed to predict the response to treatment by measuring tumor characteristics, such as the growth potential by the labeling index after i.v. injection of IdUrd, the extent of radiation-induced stable and unstable chromosome aberrations and the induction of apoptosis. These assays could guide us in the adaptation of the individual radiation doses and fractionation schedules. The measurement of the effect of cisplatin on DNA has become feasible with the development of antibodies against DNA adducts. In a recently completed phase II dose escalation trial with concomitant radiotherapy and daily cisplatin in lung cancer, we found that patients with high DNA adduct levels measured in the buccal mucosa, had a much better survival rate than patients with a low or undetectable amount of cisplatin DNA adducts. A better understanding of the signal transduction pathways involved in radiation-induced apoptosis may help to design studies aimed at modulating the apoptotic response. We and others have recently shown that alkylphospholipids, which inhibit mitogenic signaling, induce apoptosis in a variety of tumor cell lines. In combination with ionizing radiation, these compounds cause an enhancement of apoptotic cell kill. This type of signaling-based intervention study may form the basis for new therapeutic strategies. Pretreatment levels of apoptosis may be helpful in predicting treatment outcome, although the data so far show inconsistent results. The importance of evaluating other tumorbiological parameters, including cell kinetics should be stressed. Based on assays predicting reliably the response to hormonal therapy, a more appropriate choice can be made for therapeutic intervention with hormonal therapy and for selecting the appropriate adjuvant therapy in breast cancer patients. The development of a functional estrogen receptor assay (ER-FASAY), based on a yeast growth-assay, provides a way of estimating abnormal function of the receptor in tumors with a positive estrogen receptor score as measured by a classical immuno-histochemistry assay. This yeast assay can also detect different DNA mutations of the estrogen receptor existing in an individual tumor specimen. q 1999 Elsevier Science Ireland Ltd. All rights reserved. Keywords: Translational research; Treatment tailoring; Biological assays; Apoptosis; Cisplatin DNA adducts; Estrogen receptor
1. Introduction A new generation of molecular assays and developments in more conventional diagnostic tests have now become available, providing us with much more knowledge of a tumor's characteristics and expected response to therapy. This research has generated new predictive assays that will hopefully allow us to select the most suitable treatment regimen for individual patients. Using these new developments, extra therapeutic gain can be obtained, as illustrated by the results of three EORTC trials which have demon* Corresponding author.
strated a signi®cant bene®t in local control and survival by: (a) accelerated fractionation in head and neck cancer [22] (b) concomitant use of cisplatin and irradiation in inoperable lung cancer [40] (c) adjuvant hormonal treatment in breast cancer patients [4]. These trials have stimulated us to pursue directions in translational research aimed at designing and testing predictive assays for individualized therapy, and to investigate the modulation of treatment outcome by in¯uencing speci®c signal transduction pathways leading to cell death. Insight into the mechanisms responsible for response to treatment, measured by these assays, may guide us in the application of new treatment schedules and the introduction of new drugs in cancer treatment. The avoidance of exposure of patients to ineffective
0167-8140/99/$ - see front matter q 1999 Elsevier Science Ireland Ltd. All rights reserved. PII: S 0167-814 0(99)00009-2
2
H. Bartelink et al. / Radiotherapy and Oncology 50 (1999) 1±11
Fig. 1. (a) Measurements of growth rate of lung metastases and volume reduction of one single dose of X-rays (8 Gy) [2]. (b) Schematic effect of single dose irradiation on lung metastases (Y extrapolated effective residual volume, assuming exponential growth. z theoretical reduction in proliferating cells assuming rapid regrowth). Reproduced by permission of The International Journal of Radiation, Oncology, Biology and Physics.
and unnecessary treatment and its associated toxicity, will also be a positive spin-off of such predictive assays. The purpose of this paper is ®rstly to present some new predictive assays being employed at The Netherlands Cancer Institute and to analyze their possible clinical impact which could be obtained by optimizing presently existing knowledge derived from the above mentioned clinical trials. Secondly, in order to design new ways of modulating treatment outcome, we will attempt to exploit the mechanisms responsible for apoptosis induction after irradiation and estrogen binding for hormonal treatment in breast cancer. 2. Growth rate and radiation sensitivity The kinetics of tumor growth was one of the main interests of the late Professor Klaas Breur, who stimulated us to perform growth rate measurements in patients with lung metastases. For example, we were able to assess tumor volume doubling times in patients with lung metastases,
Fig. 2. Local control (a) and tumor speci®c survival (b) in patients receiving accelerated fractionation (70 Gy/5 weeks) versus conventional fractionation (70 Gy /7 weeks) [22].
and demonstrate a relationship between a large volume reduction in rapidly proliferating tumors and a much smaller reduction in slowly growing tumors; although the growth delay was much longer in these slowly growing tumors (Fig. 1a) [2]. In this way, it was possible to predict a long lasting palliative effect of a single dose half body irradiation in patients with slowly progressing, metastasized breast cancer. These measurements in lung metastases also revealed that rapid repopulation might occur during conventionally fractionated radiotherapy schedules, suggesting that reduction in overall treatment time could lead to improved local control (Fig. 1b). This knowledge, together with observations that an increased local failure rate in head and neck cancer and cervix tumors, occurs as the overall treatment time increases for the same total dose [5,19,33], has lead to the design of new fractionation schedules aiming at reducing the overall treatment time. In a randomized phase III trial by the EORTC Radiotherapy Cooperative Group, the overall treatment time was reduced to 5 weeks by giving three fractions of 1.6 Gy in the 5 week schedule, with rest
H. Bartelink et al. / Radiotherapy and Oncology 50 (1999) 1±11
Fig. 3. Measurement of the labeling index (LI) and s-phase time (Ts) to calculate the potential doubling time (Tpot) with ¯ow cytometry measuring the DNA content per cell and number of cells labeled with anti-IdUrd antibodies (A.C. Begg, The Netherlands Cancer Institute, Amsterdam).
periods in weeks 2 and 3. This was compared with a conventional schedule of 70 Gy with 2 Gy per fraction in 7 weeks. Horiot et al. [22] showed that a reduction of the overall treatment time, indeed resulted in improved local control, with a signi®cant improvement in tumor-speci®c survival (Fig. 2). These results were con®rmed by the CHART trial, where lung cancer patients received a much shorter treatment schedule of 12 consecutive days, compared with a conventional 5 days a week fractionation schedule of 6 weeks [38]. Despite a signi®cant dose reduction, the shorter CHART schedule produced equivalent results in head and neck cancer, indicating the importance of repopulation. Data from Toronto [15], Houston [1] and Denmark [34] showed similar trends.
Fig. 4. Labeling Index (LI) as prognostic factor for conventional fractionation results of 476 patients, divided into four categories: , 5%; 5±10%; 10±15% and . 15% LI (A.C. Begg, The Netherlands Cancer Institute, Amsterdam) [8].
3
These randomized clinical trials established the principle that reduction of overall treatment time is of bene®t to head and neck and lung cancer patients. It can easily be theorized that such a gain will be maximal in patients with rapidly proliferating tumors, while patients with slowly proliferating tumors will pro®t much more from hyperfractionation, as shown in another EORTC trial [23]. In order to select patients for the appropriate fractionation regimen, Begg et al. [6,7] developed a method to assess the potential doubling time (Tpot) of tumors in vivo (Fig. 3). This assay was used on biopsy material from patients with head and neck cancer, randomized in the above-mentioned EORTC trial to test the advantage of accelerated fractionation. A single i.v. injection of IdUrd was given 6 h before taking a biopsy. All samples were sent to our institute where ¯ow cytometry was performed. Initially, promising and signi®cant results were observed [7], but with longer follow-up and a larger sample, it became obvious that Tpot was not a strong or reliable predictive factor that could be used in the clinic. A recent analysis of 476 patients treated with conventional fractionation, in which Tpot measurements were performed in several European centers, showed indeed that Tpot measurements were not informative. It was apparent that the labeling index was a stronger indicator of treatment outcome than Tpot: patients with a high labeling index did worse on conventional fractionation and may therefore bene®t from accelerated fractionation (Fig. 4) [8]. However, in a multivariate analysis including patient and tumor factors, LI lost its statistical signi®cance (P 0:15). It is clear that one single biological parameter will never reliably predict treatment outcome after radiotherapy. Other factors such as intrinsic radiosensitivity and extent of hypoxia need to be taken into account, together with many other tumor characteristics such as TNM stage, volume, histology, etc. Intrinsic radiosensitivity has been shown in cervix cancer patients to correlate strongly with outcome [51], although more rapid assays are needed for its assessment than colony formation. One promising way to estimate intrinsic radiosensitivity is to measure radiationinduced chromosome aberrations. These chromosome aberrations can be visualized rapidly with ¯uorescence in situ hybridization (FISH) in combination with whole-chromosome speci®c DNA probes (chromosome paints). We have now established a good correlation (r 0:90) between cell kill and the number of chromosome aberrations induced by radiation in 12 out of 13 human tumor cell lines tested, including head and neck, ovarian and cervix tumor cell lines (J. Coco Martin, personal communication). In order to use the chromosome aberration approach on fresh tumor tissue from patients, however, several problems need to be solved. Firstly, suf®cient metaphases must be obtained, necessitating short-term culture of the tumor cells, and secondly, a discrimination needs to be made between tumor and normal cells (e.g. ®broblasts). In an attempt to get round these problems, a premature chromosome condensation method (PCC) using phosphatase inhi-
4
H. Bartelink et al. / Radiotherapy and Oncology 50 (1999) 1±11
Fig. 6. (a) Local control and (b) survival in inoperable lung cancer patients treated with radiotherapy alone, weekly cisplatin, or daily cisplatin [40]. Reproduced by permission of New England Journal of Medicine.
Fig. 5. Premature chromosome condensation in (a): unirradiated and (b): irradiated cells derived from a human tumor. Note the fragments in the irradiated cells. (J. Coco Martin, The Netherlands Cancer Institute, Amsterdam).
bitors was tested, using a protocol modi®ed for use on human tumor cells. As such, PCCs could be induced in all the human tumor cell lines tested to date, with up to 70% PCCs achieved [14]. With this technique, interphase cells are forced to condense their chromosomes prematurely by enhancing the hyperphosphorylation of histones H1 and H3, leading to a chemically-induced PCC. The condensed chromosomes can then be probed by standard FISH techniques, thereby allowing the scoring of chromosome aberrations after irradiation in terms of an increase in ¯uorescent domains (Fig. 5). In the ®rst two human tumor cell lines tested, the rate of chromosome aberration induction after
irradiation, was greater in the radiosensitive than in the radioresistant cell line [14]. This now provides the opportunity to score induced chromosome aberrations in fresh human tumor biopsies without the need for extensive cell culturing. This could lead to rapid prediction of radiosensitivity of tumors in individual patients, thereby allowing selection of appropriate radiation dose and schedule, or, for highly radioresistant tumors, providing indications for referral to surgical or other treatments. 2.1. Conclusion For patients with head and neck tumors, accelerated fractionation resulted in improved local control and tumor speci®c survival. Labeling index appeared to be a better parameter to predict local control than potential tumor doubling time for conventional fractionation schedules in these patients. However, labeling index is still a relatively weak predictor for selecting patients for an accelerated or hyperfractionated treatment regimen. Accumulating clinical data suggest that overall treatment time, and thus most probably repopulation, is an important outcome determinant, and that better predictors of repopulation are still needed.
H. Bartelink et al. / Radiotherapy and Oncology 50 (1999) 1±11
Predictors of other radiobiological parameters, such as intrinsic radiosensitivity will also be necessary for more accurate prediction of treatment outcome. Measuring the extent of radiation-induced chromosome aberrations by FISH with whole-chromosome painting may provide such a predictor. 3. Cisplatin-DNA adducts Previous research by ourselves and others into the interaction between cytostatic drugs and radiation, suggested that cisplatin might increase the radiation effect in tumors more than in normal tissues, provided this cytostatic agent is given daily and concomitantly with radiation [3,28]. This led us to successfully complete a phase III study in patients with inoperable lung cancer where, in a three-arm design, radiotherapy was given alone or in combination with a weekly single i.v. dose of 30 mg/m 2, or a daily dose of 6 mg/m 2 cisplatin [40]. The ®nal outcome (Fig. 6) showed that the best results in terms of improved local control and survival, were observed with daily cisplatin, followed by the weekly dose of cisplatin, con®rming the preclinical data. Furthermore, it was recently shown in large phase III trials by Calais et al. [12], Jeremic et al. [25] and Rose et al. [36] that concomitant radiotherapy and cisplatin or carboplatin leads to better local control and survival in lung and cervix cancer patients. These results are consistent with a thorough meta-analysis for head and neck cancer patients receiving neo-adjuvant, concomitant, adjuvant or no chemotherapy in conjunction with standard treatment for their primary tumor [11]. Indeed, this group showed that for patients with head and neck cancer, radiotherapy and concomitant chemotherapy leads to a better locoregional control rate and survival, while no gain was achieved with the other combinations. One possible mechanism for this enhanced effect of the concomitant use of cisplatin and radiation is the formation of extra double strand breaks [9], since, both for cisplatin and for radiation, DNA is the main target leading to cell kill. In order to study the mechanism of interaction and to predict which patients could potentially bene®t from this combined approach, we developed an assay to measure the amount of cisplatin-DNA adducts formed in individual cells using antibodies against these adducts [46]. The amount of cisplatin-DNA adducts in tissues have been scored semi-quantitatively by immunohistochemistry. In laboratory studies, it appeared that the number of platinumDNA adducts is an important determinant of drug-induced cell kill, such that lower numbers of platinum-DNA adducts are associated with resistance at equimolar plasma platinum concentrations. For patients receiving chemotherapy for testicular tumors, we were able to monitor the amount of cisplatin-DNA adducts in buccal cells obtained simply from a scrape of the inner cheek mucosa. Peak staining was observed a few hours after injection with a gradual reduction over the following days, until the next dose of cisplatin was given [10]. In this study, there was a direct correlation
5
between the amount of DNA adduct formation and the treatment response. During the same period we attempted to improve the delivery of both cisplatin and radiation for patients with inoperable lung cancer, resulting in an escalation of the doses of both modalities and a reduction of the overall treatment time [41]. In this patient group, we also measured the amount of cisplatin-DNA adducts in the buccal cells, one week after ®ve cisplatin doses of 6 mg/ m 2. It appeared that the amount of cisplatin-DNA adducts scored in this way was highly predictive for treatment outcome; a better survival was reached in the patient group with high levels of cisplatin-DNA adducts (P. Van der Vaart, pers. commun.). This correlation remained signi®cant after correcting for other prognostic factors in a multivariate analysis, providing us with a potential indicator for selecting patients for combined modality treatment with concomitant administration of cisplatin and radiation. One explanation for these results, is that adduct formation in buccal cells represents integrated systemic exposure of cisplatin. In other words, less adducts in buccal cells implies less tumor exposure. Adduct levels are also likely to re¯ect processes affecting drug uptake, intracellular drug detoxi®cation (dependent on glutathione and glutathione-S-transferase levels), and DNA repair leading to removal of adducts. These will operate independently of exposure level. The association between buccal adduct levels and tumor response could, therefore, also imply that the sum of these cellular factors in tumor cells are related in a systematic way, through genetic factors, to those in buccal cells of the same patient. Without data on tumor cells, it is not possible to assign relative importance to these two possibilities. 3.1. Conclusion In patients with inoperable lung, head and neck or cervix cancer, radiotherapy and concomitant daily cisplatin or carboplatin has resulted in improved local control and survival, possibly by the formation of more DNA double strand breaks. An immunological method was designed to measure the amount of cisplatin-DNA adducts in vivo. The ®rst preliminary results showed that a higher amount of cisplatinDNA adducts in buccal cells related to a signi®cantly improved life expectancy in patients undergoing this combined modality treatment. Patients with a high uptake will bene®t from this combined modality treatment, while patients with a low amount of adduct formation may be selected for an other regimen. 4. Apoptosis For many years, the main cause of cell death after radiation was thought to result from direct damage to the nuclear DNA. However, during the last decade, it has been shown that ionizing radiation also targets the plasma membrane, where it may initiate multiple signal transduction pathways
6
H. Bartelink et al. / Radiotherapy and Oncology 50 (1999) 1±11
Fig. 7. Signaling events in radiation induced apoptosis. For details, see Verheij et al. [49].
leading to different biological responses, including programmed cell death [20,26] (Fig. 7). One of the apoptotic signaling systems is activated by hydrolysis of the membrane phospholipid sphingomyelin by the enzyme sphingomyelinase. This results in the generation of the lipid messenger ceramide, a potent inducer of apoptosis in various cell types [21,27]. We have recently demonstrated that ceramide is an upstream activator of a cytoplasmatic stress-activated protein kinase (SAPK) signaling cascade [48,49]. This ceramide-initiated, SAPK-mediated pathway appeared to be essential in transmitting membrane-derived death signals in certain cell types, since interference of the pathway by dominant-negative constructs blocked ceramide- and radiation-induced apoptosis [13,39,43,48,52].
Although the pathophysiological signi®cance of ceramide-induced apoptosis remains to be determined, several observations suggest that this signaling system plays an important role in radiation-induced cell death. For example, synthetic ceramide analogs were found to mimic radiation in inducing an apoptotic response [20,24,35,48]. Furthermore, cells that lack acidic sphingomyelinase activity due to a genetic defect (Niemann±Pick disease) do not generate ceramide and are resistant to radiation-induced apoptosis [37]. It has been shown in vitro that even after a clinically relevant radiation dose of 2 Gy, ceramide is generated and apoptosis occurs [20,48]. In this respect, fractionation studies on murine ovarian tumor cells indicate that following each dose, a subpopulation of tumor cells reemerges and
H. Bartelink et al. / Radiotherapy and Oncology 50 (1999) 1±11
Fig. 8. The alkylphospholipid HePC enhances radiation-induced apoptosis in murine T-lymphoma cells (M. Verheij, The Netherlands Cancer Institute, Amsterdam).
undergoes apoptotic cell death [30]. If indeed, low doses of radiation cause a recruitment of cells into an apoptosissusceptible subpopulation, fractionated radiotherapy will produce an accumulation of cell kill as compared with single dose irradiation, which is proportional to the number of fractions, and may exceed the number of apoptotic cells induced by a (high) single dose [29,30]. Even if apoptosis induces a relatively small reduction in clonogenic survival, multiple fractions will result in an additional, signi®cant reduction in survival [16]. These ®ndings illustrate that further knowledge and exploitation of pathways involved in apoptosis may lead to new strategies to increase cell death, and consequently tumor cure, after irradiation. Since in each cellular system, both pro-apoptotic and anti-apoptotic signals contribute to the ultimate biological response, the balance between these con¯icting signals may be a potential target for pharmacological intervention. To test this concept, we are currently evaluating the effect of a class of synthetic membranepermeable alkylphospholipids (or ether-phospholipids) on radiation-induced apoptosis. Unlike the classical chemotherapeutic drugs which target the DNA, these alkylphospholipids accumulate preferentially in the plasma membrane of tumor cells where they interfere with several mitogenic signal transduction pathways. As such, these compounds inhibit the anti-apoptotic mitogen-activated protein kinase (MAPK) pathway and, as a result, could favor the SAPK pathway leading to apoptosis. In our clinic, the alkylphospholipid Miltefosine was found to be successful for topical treatment of cutaneous metastases in patients with breast cancer [45]. We have studied the effect of alkylphospholipids and radiation on apoptosis induction in various tumor cells lines in vitro. Whereas treatment with the alkylphospholipids HePC and ET-18-OCH3 or with radiation, caused a modest apoptotic response, the combined treatment with alkylphospholipids and radiation resulted in a more than additive effect, in terms of apoptotic death (Fig. 8). The underlying mechanisms of these ®ndings are currently under investigation. Recently, a great deal of attention has been given to the
7
use of pre-therapeutic apoptotic index (AI) as a prognostic factor. Several animal studies have shown that the AI might indeed predict for tissue responsiveness to radiation [31,44]. A number of clinical studies have now been carried out relating tumor AI to treatment outcome. The emerging picture from the literature, however, is far from uniform. The majority of these studies established a correlation between a high AI and good prognosis, but several other analyses demonstrated the contrary [50]. These opposite ®ndings might be explained, in part, by differences in tumor characteristics among these studies, such as type, stage and grade. Also, variations in other tumor biological parameters including hypoxia, necrosis, mitotic index, p53 and Bcl-2/BAX status, may contribute to this controversy, and should therefore be taken into account as well. Interestingly, in vitro radiation-induced apoptosis was recently found to be useful as a predictive assay of early response to low dose in vivo irradiation in patients with non-Hodgkin's lymphoma [17]. Perhaps this approach is also applicable in other tumor types. 4.1. Conclusion Multiple signal transduction pathways have been identi®ed and implicated in the induction of apoptosis following exposure to ionizing radiation. Several lines of evidence suggest that in each cellular system, both anti- and proapoptotic signaling pathways operate. It has been postulated that the balance between these opposing signals may determine the biological outcome: survival or cell death. This concept provides a potential target for pharmacological modulation which would ultimately lead to a favorable shift in the therapeutic ratio. The clinical data on pretreatment levels of apoptosis for prediction of tumor radioresponsiveness are, as yet, inconclusive and clearly indicate the need for more prospective studies, taking into account additional tumor biological characteristics. 5. Functional estrogen receptor assay A response to hormonal therapy is seen in 60% of the patients with metastasized breast cancer, who have a positive estrogen receptor (ER) assessed with immuno-histochemistry. Response is observed in only 10% of patients with a negative receptor. This means that this classical test is a good, but certainly not fully reliable assay in predicting treatment outcome in patients with breast cancer. Recently obtained knowledge on the effectiveness of adjuvant hormonal therapy in breast cancer patients, illustrates the need for a much more reliable assay in assigning the appropriate adjuvant therapy, as shown for example in the `Scottish' trial [42]. This randomized trial showed that premenopausal patients with a positive receptor, bene®ted from adjuvant hormonal therapy and did not bene®t from adjuvant chemotherapy, while patients with a negative hormonal
8
H. Bartelink et al. / Radiotherapy and Oncology 50 (1999) 1±11
Fig. 9. The trial results changed signi®cantly during the follow-up period. First both at interim analysis and at trial closure the P-values were 0.01 and 0.004 in favor of adjuvant chemotherapy. However, with longer follow-up, this signi®cant advantage disappeared. Adjuvant hormonotherapy did not show a signi®cant bene®t at the beginning. However, with long-term follow-up the gain in survival became signi®cant. Reproduced by permission of W.B. Saunders Company, Journal of Clinical Oncology.
receptor bene®ted from adjuvant chemotherapy. Similar observations were made by Naja et al. [32]. An even more striking effect of adjuvant hormonal therapy on survival was seen in the EORTC trial, in patients with locally advanced breast cancer. Short-term results showed initially a highly signi®cant effect of adjuvant chemotherapy. This bene®cial effect disappeared, however, 2 years after trial closure. Recently, more than 8 years after trial closure, we demonstrated a signi®cant and impressive gain in survival with adjuvant hormonal therapy, while adjuvant chemotherapy did not signi®cantly improve life expectancy (Fig. 9) [4]. The recently performed meta-analysis by the ECBTCG [18], con®rms the value of adjuvant hormonal therapy in breast cancer patients even for premenopausal patients with a positive ER and PR receptor. The above-mentioned failure rate for adequately predicting the response to hormonal therapy, may represent an altered functional activity of the estrogen receptor due to mutations or variations at the messenger RNA level. In order to improve the predictive value of the estrogen receptor assay, a completely new assay was designed in our institute [47]. This so-called ER-FASAY (Estrogen Receptor Functional Analysis of Separated Alleles in Yeast) measures the functional response of the estrogen binding in yeast (Fig. 10a). With this assay, the type and frequency of the variants of the ER in a tumor can be determined and compared with the normal receptor. Messenger RNA is isolated from breast cancer cell lines, tumor biopsies and recently, also from ®ne needle aspirates. The RNA is reverse transcribed and PCR is performed using primers that amplify the ER open reading frame (PCR product is 2004 base pairs). The PCR products are then cloned into a yeast expression vector in vivo by homologous recombination (Fig. 10b). The growth of the
Fig. 10. (a,b) in one ®gure ER- FASAY, a schematic diagram of the functional assay in yeast for the estrogen receptor (M. van Dijk, The Netherlands Cancer Institute, Amsterdam).
H. Bartelink et al. / Radiotherapy and Oncology 50 (1999) 1±11
9
Fig. 11. ER-FASAY of one tumor biopsy. Forty-eight yeast colonies each containing one estrogen receptor molecule are tested for their functional activity. Normal receptor (wild type) is only active in the presence of estrogen (e.g. (B,C), row 1, spot 3) indicated by arrow, wild type. A DNA-binding (DBD) mutant is non-functional in the presence or absence of estrogen (e.g. (B,C), row 1, spots 1 and 2, indicated by arrow DBD mutant). A hormone binding (HBD) mutant is functional in the presence but also in the absence of estrogen (e.g. (B,C) row 1, spots 4, 5 and 6). E 2, estrogen; 2 leu, yeast plate lacking leucine; 2 his, yeast plate lacking histidine. (A) Non-reporter selective 2 leu: not selective for ER activity (only assays the presence of estrogen receptor. (B) Reporter selective 2 leu 2 his 1 E 2: selective for ER activity dependent on estrogen. (C) Reporter selective 2 leu 2 his 2 E 2: selective for ER activity independent of estrogen (M. van Dijk, The Netherlands Cancer Institute, Amsterdam).
yeast Saccharomyces cerevisiae is made, dependent on the functional integrity of the estrogen receptor by using an ERdependent promoter (containing an estrogen response element (ERE)) linked to a gene involved the synthesis of histidine. This allows discrimination between normal and variant receptors and at the same time determines their hormone-dependence (Fig. 11). The average relative expression level of ER in 22 primary breast tumors showed that 67% of ER mRNA in a tumor was wild-type (range 12± 100%), 30% were DNA binding mutants (range 0±64%) and 3% were hormone binding mutants (range 0±40%). All but one were splicing variants, and almost every breast tumor contained more than one ER variant. Almost all exhibited a positive estrogen receptor assessed with the immunohistochemistry assay. At present, predictive reliability of the ER-FASAY is undergoing prospective clinical testing in patients with inoperable or metastasized breast cancer. In the laboratory, the actual mechanism of the estrogen binding to its receptor is being investigated in order to allow therapeutic intervention. 5.1. Conclusion The long-term results of the EORTC trial revealing the improvement of survival by adjuvant hormone therapy for patients with locally advanced breast cancer, demonstrated the potential for selective administration of an adjuvant regime. The functional activity of the estrogen receptor can be measured with the yeast ER-FASAY. This may lead to a better prediction of treatment outcome for hormo-
nal therapy in patients with advanced disease. The ERFASAY may play an important role in designing adjuvant treatment in individual patients, especially as it appeared recently that adjuvant hormonal therapy is very effective in breast cancer patients with a positive estrogen receptor, assessed with an immunohistochemistry assay. 6. Summary The development and testing of the potential of predictive assays was inspired by the positive outcome, in terms of improved local control and survival of three phase III EORTC clinical trials in head and neck cancer, inoperable lung cancer and locally advanced breast cancer. These assays were aimed at predicting treatment outcome of radiotherapy alone, conventional or accelerated, or a combination of radiotherapy with systemic cisplatin or endocrine treatment in lung and breast cancer, respectively. The results obtained from this translational research have shown the potential for predicting treatment outcome by assays measuring proliferation, intrinsic radiosensitivity, the amount of drug-DNA adducts and the function of the estrogen receptor. This offers unique possibilities to modulate treatment response by selecting the appropriate treatment regime for individual patients. Moreover, knowledge of pathways leading to cell death can be exploited to intervene in speci®c signaling pathways, such as in the stimulation of apoptosis Translational research in the areas mentioned is now showing the way towards a real improvement in local
10
H. Bartelink et al. / Radiotherapy and Oncology 50 (1999) 1±11
control and survival in cancer, through individual tailoring of the treatment regime.
Acknowledgements This work was supported in part by grants from The Dutch Cancer Society: NKI 97-1435, NKI 96-1265, Schumacher Kramer Stichting.
[15]
[16] [17] [18]
References [1] Ang KK, Trotti A, Garden AS, et al. Overall time factor in postoperative radiation: results of a prospective randomized trial (abstract 108). Radiother. Oncol. 1996;40(Suppl.1):S30. [2] Bartelink H, Batterman J, Hart G. Half body irradiation. Int. J. Radiat. Biol. Phys. 1980;6:87±90. [3] Bartelink HB, Kallman RF, Rapacchietta D, Hart GAM. Therapeutic enhancement in mice by clinically relevant dose and fractionation schedules of CDDP and irradiation. Radiother. Oncol. 1986;6:61±74. [4] Bartelink H, Rubens RD, Van der Schueren E, Sylvester R. Hormonal therapy prolongs survival in irradiated locally advanced breast cancer: a European Organization for Research and Treatment of Cancer randomized phase III trial. J. Clin. Oncol. 1997;15:207±215. [5] Barton MB, Keane TJ, Gadalla T. The effect of treatment time on tumor control in the radical radiotherapy of laryngeal cancer. Radiother. Oncol. 1992;23:137±143. [6] Begg AC, McNally NJ, Shrieve DC, Karcher H. A method to measure the duration of DNA synthesis and the potential doubling time from a single sample. Cytometry 1985;6:620±626. [7] Begg AC, Ho¯and K, van Glabbeke M, Bartelink H, Horiot JC. Predictive value of potential doubling time for radiotherapy of head and neck tumour patients: results from the EORTC cooperative trial 22851. Semin. Radiat. Oncol. 1992;1:22±25. [8] Begg AC, Haustermans K, Hart AAM, et al. The value of pretreatment cell kinetic parameters as predictors for radiotherapy outcome in head and neck cancer: a multicenter analysis. Radiat. Oncol 1998; in press. [9] Begg AC. Cisplatin and radiation: interaction probabilities and therapeutic possibilities. Int. J. Radiat. Oncol. Biol. Phys. 1990;19:1183± 1189. [10] Blommaert FA, Michael C, Terheggen PMAB, et al. Drug induced DNA modi®cation in buccal cells of cancer patients receiving carboplatin and cisplatin combination chemotherapy, as determined by an immunocytochemical method: inter-individual variation and correlation with disease response. Cancer Res. 1993;53:5669±5675. [11] Bourhis J, Pignon JP, Designe M, et al. Meta-analysis of chemotherapy in head and neck cancer (MACH-NC): (1) loco-regional treatment vs. same treatment 1 chemotherapy (CT). Proceedings 34th ASCO, May 16±19, Los Angeles, USA (abstract 1487). J. Clin. Oncol. 1998;17:386a. [12] Calais G, Alfonsi M, Bardet E, et al. Randomized study comparing radiation alone (RT) versus RT with concomitant chemotherapy (CT) in stages III and IV oropharynx carcinoma (ARCORO). Preliminary results of the 94±01 study from the French Group of Radiation Oncology for head and neck cancer (GORTEC). Proceedings 34th ASCO, May 16±19, Los Angeles, USA (abstract 1484). J. Clin. Oncol. 1968;17:385a. [13] Chen Y-R, Wang X, Templeton D, Davis RJ, Tan T-H. The role of cJun N-terminal kinase (JNK) in apoptosis induced by ultraviolet C and radiation. J. Biol. Chem. 1996;271:31929±31936. [14] Coco Martin JM, Begg AC. Detection of radiation-induced chromosome aberrations using ¯uorescence in situ hybridization in drug-
[19] [20]
[22]
[23]
[24]
[25]
[26] [27] [28]
[29] [30] [31] [32]
[33] [34]
induced premature chromosome condensations of tumor cell lines with different radiosensitivities. Int. J. Radiat. Biol. 1997;71:265±73. Cummings BJ, Keane TJ, Pintillie M, et al. A prospective randomized trial of hyperfractionated versus conventional once daily radiation for advanced squamous cell carcinomas of the larynx and pharynx (abstract 107). Radiother. Oncol. 1996;40(Suppl. 1):S30. Dewey WC, Ling CC, Meyn RE. Radiation-induced apoptosis: relevance to radiotherapy. Int. J. Radiat. Oncol. Biol. Phys. 1995;33:781± 796. Dubray B, Breton C, Delic J, et al. In vitro radiation-induced apoptosis and early response to low-dose radiotherapy in non-Hodgkin's lymphomas. Radiother. Oncol. 1998;46:185±191. Early Breast Cancer Trialists' Collaborative Group, 1998, Tamoxfen for early breast cancer: an overview of the randomized trials (9114), Lancet, 351:1451-1467. Fyles A, Keane TJ, Barton M, Simm J. The effect of overall treatment duration in the local control of cervix cancer. Radiother. Oncol. 1992;25:273±279. Haimovitz-Friedman A, Kan C-C, Ehleiter D, et al. Ionizing radiation acts on cellular membranes to generate ceramide and initiate apoptosis. J. Exp. Med. 1994;180:525±535. Hannun YA. The sphingomyelin cycle and the second messenger function of ceramide. J. Biol. Chem 1994;269:3125±3128. Horiot JC, Bontemps P, Van den Bogaert W, et al. Accelerated fractionation (AF) compared to conventional fractionation (CF) improves loco-regional control in the radiotherapy of advanced head and neck cancers: results of the EORTC 22851 randomized trial. Radiother. Oncol. 1997;44:111±121. Horiot JC, Le Fur R, Nguyen T, et al. Hyperfractionation versus conventional fractionation in oropharyngeal carcinoma: ®nal analysis of a randomized trial of the EORTC cooperative group of radiotherapy. Radiat Oncol. 1992;25:231±241. Jarvis WD, Kolesnick RN, Fornari FA, Traylor RS, Gewirtz DA, Grant S. Induction of apoptotic DNA damage and cell death by activation of the sphingomyelin pathway. Proc. Natl. Acad. Sci. USA 1994;91:73±77. Jeremic B, Shibamoto Y, Acimovic L, Milisavljevic S. Initial versus delayed accelerated hyperfractionated radiation therapy and concurrent chemotherapy in limited small-cell lung cancer: a randomized study. J. Clin. Oncol. 1997;15:893±900. Kasid U, Suy S, Dent P, Ray S, Whiteside TL, Sturgill TW. Activation of Raf by ionizing radiation. Nature 1996;382:813±816. Kolesnick R. Sphingomyelin and derivatives as cellular signals. Prog. Lipid Res. 1991;30:1±38. Lelieveld P, Scoles MA, Brown JM, Kallman RF. The effect of treatment in fractionated schedules with the combination of X-irradiation and six cytotoxic drug on the RIF-1 tumor and normal mouse skin. Int. J. Radiat. Oncol. Biol. Phys. 1985;11:111±121. Ling CC, Chen CH, Li WX. Apoptosis induced at different dose rates: implication for the shoulder region of cell survival curves. Radiother. Oncol. 1994;32:129±136. Meyn RE, Stephens LC, Hunter NR, Ang KK, Milas L. Reemergence of apoptotic cells between fractionated doses in irradiated murine tumors. Int. J. Radiat. Oncol. Biol. Phys. 1994;30:619±624. Meyn RE, Stephens LC, Ang KK, Hunter N, Milas L, Peters LJ. Heterogeneity in apoptosis development in irradiated murine tumors. Int. J. Radiat. Biol. 1993;64:583±591. Naja A, Roche H, Mihura J, et al. Castration and tamoxifen versus chemotherapy (FAC) for premenopausal, node and receptors positive breast cancer patients: a randomized trial with a 7 years median follow up (abstract 87). Radiother. Oncol. 1997;43(Suppl. 2):S22. Overgaard J, Hjelm-Hansen M, Vendelbo Johansen L, Andersen AP. Comparison of conventional and split course radiotherapy as primary treatment in carcinoma of the larynx. Acta Oncol. 1988;27:147±152. Overgaard J, Sand Hansen H, Overgaard M, et al. Importance of overall treatment time for the outcome of radiotherapy in head and neck carcinoma. Experience from the Danish head and heck cancer
H. Bartelink et al. / Radiotherapy and Oncology 50 (1999) 1±11
[35]
[36]
[37] [38]
[39] [40] [41]
[42]
study. In: Kogelnik HD, Sedlmayer F, et al., editors. Progress in radiooncology. Proceedings of the 6th International Meeting, Salzburg, Austria, May 13±17, Bologna: Monduzzi Editore, 1998. pp. 743. Quintans J, Kilkus J, McShan CL, Gottschalk AR, Dawson G. Ceramide mediates the apoptotic response of WEHI 231 cells to antiimmunoglobulin, corticosteroids and irradiation. Biochem. Biophys. Res. Commun. 1994;202:710±714. Rose P, Bundy B, Thigpen J, et al. Signi®cant preliminary results of a phase III randomized study of concomitant chemotherapy with weekly cisplatin vs. hydroxyurea, 5-Fluorouracil infusion and bolus cisplatin vs. hydroxyurea in advanced cervical cancer: a gynecologic oncology group (GOG) study. Proceedings 34th ASCO, May 16±19, Los Angeles, USA (abstract 1391). J. Clin. Oncol. 1998;17:360a. Santana P, PenÄa LA, Haimovitz-Friedman A, et al. Acid sphingomyelinase-de®cient human lymphoblasts and mice are defective in radiation-induced apoptosis. Cell 1996;86:189±99. Saunders M, Dische S, Barrett A, Harvey A, Gibson D, Parmar M. Continuous hyperfractionated accelerated radiotherapy (CHART) versus conventional radiotherapy in non-small-cell lung cancer: a randomized multicentre trial. Lancet 1997;350:161±165. Sawai H, Okazaki T, Yamamoto H, et al. Requirement of AP-1 for ceramide-induced apoptosis in human leukemia HL-60 cells. J. Biol. Chem. 1995;270:27326±27331. Schaake-Koning C, Van den Bogaert W, Dalesio O, et al. Effects of concomitant cisplatin and radiotherapy on inoperable non-small-cell lung cancer. N. Engl. J. Med. 1992;326:524±530. Schuster-Uitterhoeve ALJ, Van de Vaart PJM, Schaake-Koning CCE, et al. Feasibility of escalating daily doses of cisplatin in combination with accelerated radiotherapy in non-small cell lung cancer. Eur. J. Cancer 1996;32A:1314±1319. Scottish Cancer Trials Breast Group and ICRF Breast Unit, Guy's Hospital, London, 1993, Adjuvant ovarian ablation versus CMF chemotherapy in premenopausal women with pathological stage II breast carcinoma: the Scottish trial, Lancet, 341:1293-1298.
11
[43] Shirakabe K, Yamaguchi K, Shibuya H, et al. TAK1 mediates the ceramide signaling to stress-activated protein kinase/c-Jun N-terminal kinase. J. Biol. Chem. 1997;272:8141±8144. [44] Stephens LC, Ang KK, Schultheiss TE, Milas L, Meyn RE. Apoptosis in irradiated murine tumors. Radiat. Res 1991;127:308±316. [45] Ten Bokkel Huinink WW, Schornagel JH, Hilton A, Somers R, Bartelink H. Topical application of miltefosine against skin metastases of breast cancer. Proceedings of ASCO, San Diego, USA, May 17±19 (abstract 39). J. Clin. Oncol. 1992;11:53. [46] Terheggen PMAB, Emondt JY, Floot BGJ, et al. The correlation between cell killing by cis-diamminedichloroplatinum(II) in six mammalian cell lines and binding of a cis-diamminedichloroplatinum(II)-DNA antiserum. Cancer Res. 1990;50:3556±3561. [47] Van Dijk MAJ, Floore AN, Kloppenborg KIM, Van 't Veer LJ. A functional assay in yeast for human estrogen receptor displays wildtype and variant estrogen receptor messenger RNAs present in breast carcinoma. Cancer Res. 1997;57:3478±3485. [48] Verheij M, Bose R, Lin XH. Requirement for ceramide-initiated SAPK/JNK signalling in stress-induced apoptosis. Nature 1996;380:75±79. [49] Verheij M, Ruiter GA, Zerp SF, et al. The role of the stress-activated protein kinase (SAPK/JNK) signaling pathway in radiation-induced apoptosis. Radiother. Oncol. 1998;47:225±232. [50] Verheij M, van Blitterswijk WJ, Bartelink H. Radiation-induced apoptosis: the ceramide-SAPK signaling pathway and clinical aspects. Acta Oncol. 1998; in press. [51] West CM, Davidson SE, Roberts SA, Hunter RD. The independence of intrinsic radiosensitivity as a prognostic factor for patient response to radiotherapy of carcinoma of the cervix. Br. J. Cancer 1997;76:1184±1190. [52] Zanke BW, Boudreau K, Rubie E, et al. The stress-activated protein kinase pathway mediates cell death following injury induced by cisplatinum, UV irradiation or heat. Curr. Biol 1996;6:606±613.
Breur Award Lecture
Towards prediction and modulation of treatment response Harry Bartelink a,*, Adrian Begg b, Jose Coco Martin b, Mariska van Dijk b, Laura van 't Veer b, Paul van der Vaart b, Marcel Verheij a b
a Department of Radiotherapy, The Netherlands Cancer Institute, Plesmanlaan 121, 1066 CX Amsterdam, The Netherlands Department of Experimental Therapy, The Netherlands Cancer Institute, Plesmanlaan 121, 1066 CX Amsterdam, The Netherlands
Received 24 August 1998; accepted 14 October 1998
Abstract The purpose of this paper is to evaluate new predictive assays and their potential to modulate treatment response. Their impact is presented in the context of three EORTC clinical trials in head and neck, lung and breast cancer, showing an improvement in survival by accelerated fractionation, concomitant use of cisplatin and radiotherapy and adjuvant hormonal treatment, respectively. Assays have been developed to predict the response to treatment by measuring tumor characteristics, such as the growth potential by the labeling index after i.v. injection of IdUrd, the extent of radiation-induced stable and unstable chromosome aberrations and the induction of apoptosis. These assays could guide us in the adaptation of the individual radiation doses and fractionation schedules. The measurement of the effect of cisplatin on DNA has become feasible with the development of antibodies against DNA adducts. In a recently completed phase II dose escalation trial with concomitant radiotherapy and daily cisplatin in lung cancer, we found that patients with high DNA adduct levels measured in the buccal mucosa, had a much better survival rate than patients with a low or undetectable amount of cisplatin DNA adducts. A better understanding of the signal transduction pathways involved in radiation-induced apoptosis may help to design studies aimed at modulating the apoptotic response. We and others have recently shown that alkylphospholipids, which inhibit mitogenic signaling, induce apoptosis in a variety of tumor cell lines. In combination with ionizing radiation, these compounds cause an enhancement of apoptotic cell kill. This type of signaling-based intervention study may form the basis for new therapeutic strategies. Pretreatment levels of apoptosis may be helpful in predicting treatment outcome, although the data so far show inconsistent results. The importance of evaluating other tumorbiological parameters, including cell kinetics should be stressed. Based on assays predicting reliably the response to hormonal therapy, a more appropriate choice can be made for therapeutic intervention with hormonal therapy and for selecting the appropriate adjuvant therapy in breast cancer patients. The development of a functional estrogen receptor assay (ER-FASAY), based on a yeast growth-assay, provides a way of estimating abnormal function of the receptor in tumors with a positive estrogen receptor score as measured by a classical immuno-histochemistry assay. This yeast assay can also detect different DNA mutations of the estrogen receptor existing in an individual tumor specimen. q 1999 Elsevier Science Ireland Ltd. All rights reserved. Keywords: Translational research; Treatment tailoring; Biological assays; Apoptosis; Cisplatin DNA adducts; Estrogen receptor
1. Introduction A new generation of molecular assays and developments in more conventional diagnostic tests have now become available, providing us with much more knowledge of a tumor's characteristics and expected response to therapy. This research has generated new predictive assays that will hopefully allow us to select the most suitable treatment regimen for individual patients. Using these new developments, extra therapeutic gain can be obtained, as illustrated by the results of three EORTC trials which have demon* Corresponding author.
strated a signi®cant bene®t in local control and survival by: (a) accelerated fractionation in head and neck cancer [22] (b) concomitant use of cisplatin and irradiation in inoperable lung cancer [40] (c) adjuvant hormonal treatment in breast cancer patients [4]. These trials have stimulated us to pursue directions in translational research aimed at designing and testing predictive assays for individualized therapy, and to investigate the modulation of treatment outcome by in¯uencing speci®c signal transduction pathways leading to cell death. Insight into the mechanisms responsible for response to treatment, measured by these assays, may guide us in the application of new treatment schedules and the introduction of new drugs in cancer treatment. The avoidance of exposure of patients to ineffective
0167-8140/99/$ - see front matter q 1999 Elsevier Science Ireland Ltd. All rights reserved. PII: S 0167-814 0(99)00009-2
2
H. Bartelink et al. / Radiotherapy and Oncology 50 (1999) 1±11
Fig. 1. (a) Measurements of growth rate of lung metastases and volume reduction of one single dose of X-rays (8 Gy) [2]. (b) Schematic effect of single dose irradiation on lung metastases (Y extrapolated effective residual volume, assuming exponential growth. z theoretical reduction in proliferating cells assuming rapid regrowth). Reproduced by permission of The International Journal of Radiation, Oncology, Biology and Physics.
and unnecessary treatment and its associated toxicity, will also be a positive spin-off of such predictive assays. The purpose of this paper is ®rstly to present some new predictive assays being employed at The Netherlands Cancer Institute and to analyze their possible clinical impact which could be obtained by optimizing presently existing knowledge derived from the above mentioned clinical trials. Secondly, in order to design new ways of modulating treatment outcome, we will attempt to exploit the mechanisms responsible for apoptosis induction after irradiation and estrogen binding for hormonal treatment in breast cancer. 2. Growth rate and radiation sensitivity The kinetics of tumor growth was one of the main interests of the late Professor Klaas Breur, who stimulated us to perform growth rate measurements in patients with lung metastases. For example, we were able to assess tumor volume doubling times in patients with lung metastases,
Fig. 2. Local control (a) and tumor speci®c survival (b) in patients receiving accelerated fractionation (70 Gy/5 weeks) versus conventional fractionation (70 Gy /7 weeks) [22].
and demonstrate a relationship between a large volume reduction in rapidly proliferating tumors and a much smaller reduction in slowly growing tumors; although the growth delay was much longer in these slowly growing tumors (Fig. 1a) [2]. In this way, it was possible to predict a long lasting palliative effect of a single dose half body irradiation in patients with slowly progressing, metastasized breast cancer. These measurements in lung metastases also revealed that rapid repopulation might occur during conventionally fractionated radiotherapy schedules, suggesting that reduction in overall treatment time could lead to improved local control (Fig. 1b). This knowledge, together with observations that an increased local failure rate in head and neck cancer and cervix tumors, occurs as the overall treatment time increases for the same total dose [5,19,33], has lead to the design of new fractionation schedules aiming at reducing the overall treatment time. In a randomized phase III trial by the EORTC Radiotherapy Cooperative Group, the overall treatment time was reduced to 5 weeks by giving three fractions of 1.6 Gy in the 5 week schedule, with rest
H. Bartelink et al. / Radiotherapy and Oncology 50 (1999) 1±11
Fig. 3. Measurement of the labeling index (LI) and s-phase time (Ts) to calculate the potential doubling time (Tpot) with ¯ow cytometry measuring the DNA content per cell and number of cells labeled with anti-IdUrd antibodies (A.C. Begg, The Netherlands Cancer Institute, Amsterdam).
periods in weeks 2 and 3. This was compared with a conventional schedule of 70 Gy with 2 Gy per fraction in 7 weeks. Horiot et al. [22] showed that a reduction of the overall treatment time, indeed resulted in improved local control, with a signi®cant improvement in tumor-speci®c survival (Fig. 2). These results were con®rmed by the CHART trial, where lung cancer patients received a much shorter treatment schedule of 12 consecutive days, compared with a conventional 5 days a week fractionation schedule of 6 weeks [38]. Despite a signi®cant dose reduction, the shorter CHART schedule produced equivalent results in head and neck cancer, indicating the importance of repopulation. Data from Toronto [15], Houston [1] and Denmark [34] showed similar trends.
Fig. 4. Labeling Index (LI) as prognostic factor for conventional fractionation results of 476 patients, divided into four categories: , 5%; 5±10%; 10±15% and . 15% LI (A.C. Begg, The Netherlands Cancer Institute, Amsterdam) [8].
3
These randomized clinical trials established the principle that reduction of overall treatment time is of bene®t to head and neck and lung cancer patients. It can easily be theorized that such a gain will be maximal in patients with rapidly proliferating tumors, while patients with slowly proliferating tumors will pro®t much more from hyperfractionation, as shown in another EORTC trial [23]. In order to select patients for the appropriate fractionation regimen, Begg et al. [6,7] developed a method to assess the potential doubling time (Tpot) of tumors in vivo (Fig. 3). This assay was used on biopsy material from patients with head and neck cancer, randomized in the above-mentioned EORTC trial to test the advantage of accelerated fractionation. A single i.v. injection of IdUrd was given 6 h before taking a biopsy. All samples were sent to our institute where ¯ow cytometry was performed. Initially, promising and signi®cant results were observed [7], but with longer follow-up and a larger sample, it became obvious that Tpot was not a strong or reliable predictive factor that could be used in the clinic. A recent analysis of 476 patients treated with conventional fractionation, in which Tpot measurements were performed in several European centers, showed indeed that Tpot measurements were not informative. It was apparent that the labeling index was a stronger indicator of treatment outcome than Tpot: patients with a high labeling index did worse on conventional fractionation and may therefore bene®t from accelerated fractionation (Fig. 4) [8]. However, in a multivariate analysis including patient and tumor factors, LI lost its statistical signi®cance (P 0:15). It is clear that one single biological parameter will never reliably predict treatment outcome after radiotherapy. Other factors such as intrinsic radiosensitivity and extent of hypoxia need to be taken into account, together with many other tumor characteristics such as TNM stage, volume, histology, etc. Intrinsic radiosensitivity has been shown in cervix cancer patients to correlate strongly with outcome [51], although more rapid assays are needed for its assessment than colony formation. One promising way to estimate intrinsic radiosensitivity is to measure radiationinduced chromosome aberrations. These chromosome aberrations can be visualized rapidly with ¯uorescence in situ hybridization (FISH) in combination with whole-chromosome speci®c DNA probes (chromosome paints). We have now established a good correlation (r 0:90) between cell kill and the number of chromosome aberrations induced by radiation in 12 out of 13 human tumor cell lines tested, including head and neck, ovarian and cervix tumor cell lines (J. Coco Martin, personal communication). In order to use the chromosome aberration approach on fresh tumor tissue from patients, however, several problems need to be solved. Firstly, suf®cient metaphases must be obtained, necessitating short-term culture of the tumor cells, and secondly, a discrimination needs to be made between tumor and normal cells (e.g. ®broblasts). In an attempt to get round these problems, a premature chromosome condensation method (PCC) using phosphatase inhi-
4
H. Bartelink et al. / Radiotherapy and Oncology 50 (1999) 1±11
Fig. 6. (a) Local control and (b) survival in inoperable lung cancer patients treated with radiotherapy alone, weekly cisplatin, or daily cisplatin [40]. Reproduced by permission of New England Journal of Medicine.
Fig. 5. Premature chromosome condensation in (a): unirradiated and (b): irradiated cells derived from a human tumor. Note the fragments in the irradiated cells. (J. Coco Martin, The Netherlands Cancer Institute, Amsterdam).
bitors was tested, using a protocol modi®ed for use on human tumor cells. As such, PCCs could be induced in all the human tumor cell lines tested to date, with up to 70% PCCs achieved [14]. With this technique, interphase cells are forced to condense their chromosomes prematurely by enhancing the hyperphosphorylation of histones H1 and H3, leading to a chemically-induced PCC. The condensed chromosomes can then be probed by standard FISH techniques, thereby allowing the scoring of chromosome aberrations after irradiation in terms of an increase in ¯uorescent domains (Fig. 5). In the ®rst two human tumor cell lines tested, the rate of chromosome aberration induction after
irradiation, was greater in the radiosensitive than in the radioresistant cell line [14]. This now provides the opportunity to score induced chromosome aberrations in fresh human tumor biopsies without the need for extensive cell culturing. This could lead to rapid prediction of radiosensitivity of tumors in individual patients, thereby allowing selection of appropriate radiation dose and schedule, or, for highly radioresistant tumors, providing indications for referral to surgical or other treatments. 2.1. Conclusion For patients with head and neck tumors, accelerated fractionation resulted in improved local control and tumor speci®c survival. Labeling index appeared to be a better parameter to predict local control than potential tumor doubling time for conventional fractionation schedules in these patients. However, labeling index is still a relatively weak predictor for selecting patients for an accelerated or hyperfractionated treatment regimen. Accumulating clinical data suggest that overall treatment time, and thus most probably repopulation, is an important outcome determinant, and that better predictors of repopulation are still needed.
H. Bartelink et al. / Radiotherapy and Oncology 50 (1999) 1±11
Predictors of other radiobiological parameters, such as intrinsic radiosensitivity will also be necessary for more accurate prediction of treatment outcome. Measuring the extent of radiation-induced chromosome aberrations by FISH with whole-chromosome painting may provide such a predictor. 3. Cisplatin-DNA adducts Previous research by ourselves and others into the interaction between cytostatic drugs and radiation, suggested that cisplatin might increase the radiation effect in tumors more than in normal tissues, provided this cytostatic agent is given daily and concomitantly with radiation [3,28]. This led us to successfully complete a phase III study in patients with inoperable lung cancer where, in a three-arm design, radiotherapy was given alone or in combination with a weekly single i.v. dose of 30 mg/m 2, or a daily dose of 6 mg/m 2 cisplatin [40]. The ®nal outcome (Fig. 6) showed that the best results in terms of improved local control and survival, were observed with daily cisplatin, followed by the weekly dose of cisplatin, con®rming the preclinical data. Furthermore, it was recently shown in large phase III trials by Calais et al. [12], Jeremic et al. [25] and Rose et al. [36] that concomitant radiotherapy and cisplatin or carboplatin leads to better local control and survival in lung and cervix cancer patients. These results are consistent with a thorough meta-analysis for head and neck cancer patients receiving neo-adjuvant, concomitant, adjuvant or no chemotherapy in conjunction with standard treatment for their primary tumor [11]. Indeed, this group showed that for patients with head and neck cancer, radiotherapy and concomitant chemotherapy leads to a better locoregional control rate and survival, while no gain was achieved with the other combinations. One possible mechanism for this enhanced effect of the concomitant use of cisplatin and radiation is the formation of extra double strand breaks [9], since, both for cisplatin and for radiation, DNA is the main target leading to cell kill. In order to study the mechanism of interaction and to predict which patients could potentially bene®t from this combined approach, we developed an assay to measure the amount of cisplatin-DNA adducts formed in individual cells using antibodies against these adducts [46]. The amount of cisplatin-DNA adducts in tissues have been scored semi-quantitatively by immunohistochemistry. In laboratory studies, it appeared that the number of platinumDNA adducts is an important determinant of drug-induced cell kill, such that lower numbers of platinum-DNA adducts are associated with resistance at equimolar plasma platinum concentrations. For patients receiving chemotherapy for testicular tumors, we were able to monitor the amount of cisplatin-DNA adducts in buccal cells obtained simply from a scrape of the inner cheek mucosa. Peak staining was observed a few hours after injection with a gradual reduction over the following days, until the next dose of cisplatin was given [10]. In this study, there was a direct correlation
5
between the amount of DNA adduct formation and the treatment response. During the same period we attempted to improve the delivery of both cisplatin and radiation for patients with inoperable lung cancer, resulting in an escalation of the doses of both modalities and a reduction of the overall treatment time [41]. In this patient group, we also measured the amount of cisplatin-DNA adducts in the buccal cells, one week after ®ve cisplatin doses of 6 mg/ m 2. It appeared that the amount of cisplatin-DNA adducts scored in this way was highly predictive for treatment outcome; a better survival was reached in the patient group with high levels of cisplatin-DNA adducts (P. Van der Vaart, pers. commun.). This correlation remained signi®cant after correcting for other prognostic factors in a multivariate analysis, providing us with a potential indicator for selecting patients for combined modality treatment with concomitant administration of cisplatin and radiation. One explanation for these results, is that adduct formation in buccal cells represents integrated systemic exposure of cisplatin. In other words, less adducts in buccal cells implies less tumor exposure. Adduct levels are also likely to re¯ect processes affecting drug uptake, intracellular drug detoxi®cation (dependent on glutathione and glutathione-S-transferase levels), and DNA repair leading to removal of adducts. These will operate independently of exposure level. The association between buccal adduct levels and tumor response could, therefore, also imply that the sum of these cellular factors in tumor cells are related in a systematic way, through genetic factors, to those in buccal cells of the same patient. Without data on tumor cells, it is not possible to assign relative importance to these two possibilities. 3.1. Conclusion In patients with inoperable lung, head and neck or cervix cancer, radiotherapy and concomitant daily cisplatin or carboplatin has resulted in improved local control and survival, possibly by the formation of more DNA double strand breaks. An immunological method was designed to measure the amount of cisplatin-DNA adducts in vivo. The ®rst preliminary results showed that a higher amount of cisplatinDNA adducts in buccal cells related to a signi®cantly improved life expectancy in patients undergoing this combined modality treatment. Patients with a high uptake will bene®t from this combined modality treatment, while patients with a low amount of adduct formation may be selected for an other regimen. 4. Apoptosis For many years, the main cause of cell death after radiation was thought to result from direct damage to the nuclear DNA. However, during the last decade, it has been shown that ionizing radiation also targets the plasma membrane, where it may initiate multiple signal transduction pathways
6
H. Bartelink et al. / Radiotherapy and Oncology 50 (1999) 1±11
Fig. 7. Signaling events in radiation induced apoptosis. For details, see Verheij et al. [49].
leading to different biological responses, including programmed cell death [20,26] (Fig. 7). One of the apoptotic signaling systems is activated by hydrolysis of the membrane phospholipid sphingomyelin by the enzyme sphingomyelinase. This results in the generation of the lipid messenger ceramide, a potent inducer of apoptosis in various cell types [21,27]. We have recently demonstrated that ceramide is an upstream activator of a cytoplasmatic stress-activated protein kinase (SAPK) signaling cascade [48,49]. This ceramide-initiated, SAPK-mediated pathway appeared to be essential in transmitting membrane-derived death signals in certain cell types, since interference of the pathway by dominant-negative constructs blocked ceramide- and radiation-induced apoptosis [13,39,43,48,52].
Although the pathophysiological signi®cance of ceramide-induced apoptosis remains to be determined, several observations suggest that this signaling system plays an important role in radiation-induced cell death. For example, synthetic ceramide analogs were found to mimic radiation in inducing an apoptotic response [20,24,35,48]. Furthermore, cells that lack acidic sphingomyelinase activity due to a genetic defect (Niemann±Pick disease) do not generate ceramide and are resistant to radiation-induced apoptosis [37]. It has been shown in vitro that even after a clinically relevant radiation dose of 2 Gy, ceramide is generated and apoptosis occurs [20,48]. In this respect, fractionation studies on murine ovarian tumor cells indicate that following each dose, a subpopulation of tumor cells reemerges and
H. Bartelink et al. / Radiotherapy and Oncology 50 (1999) 1±11
Fig. 8. The alkylphospholipid HePC enhances radiation-induced apoptosis in murine T-lymphoma cells (M. Verheij, The Netherlands Cancer Institute, Amsterdam).
undergoes apoptotic cell death [30]. If indeed, low doses of radiation cause a recruitment of cells into an apoptosissusceptible subpopulation, fractionated radiotherapy will produce an accumulation of cell kill as compared with single dose irradiation, which is proportional to the number of fractions, and may exceed the number of apoptotic cells induced by a (high) single dose [29,30]. Even if apoptosis induces a relatively small reduction in clonogenic survival, multiple fractions will result in an additional, signi®cant reduction in survival [16]. These ®ndings illustrate that further knowledge and exploitation of pathways involved in apoptosis may lead to new strategies to increase cell death, and consequently tumor cure, after irradiation. Since in each cellular system, both pro-apoptotic and anti-apoptotic signals contribute to the ultimate biological response, the balance between these con¯icting signals may be a potential target for pharmacological intervention. To test this concept, we are currently evaluating the effect of a class of synthetic membranepermeable alkylphospholipids (or ether-phospholipids) on radiation-induced apoptosis. Unlike the classical chemotherapeutic drugs which target the DNA, these alkylphospholipids accumulate preferentially in the plasma membrane of tumor cells where they interfere with several mitogenic signal transduction pathways. As such, these compounds inhibit the anti-apoptotic mitogen-activated protein kinase (MAPK) pathway and, as a result, could favor the SAPK pathway leading to apoptosis. In our clinic, the alkylphospholipid Miltefosine was found to be successful for topical treatment of cutaneous metastases in patients with breast cancer [45]. We have studied the effect of alkylphospholipids and radiation on apoptosis induction in various tumor cells lines in vitro. Whereas treatment with the alkylphospholipids HePC and ET-18-OCH3 or with radiation, caused a modest apoptotic response, the combined treatment with alkylphospholipids and radiation resulted in a more than additive effect, in terms of apoptotic death (Fig. 8). The underlying mechanisms of these ®ndings are currently under investigation. Recently, a great deal of attention has been given to the
7
use of pre-therapeutic apoptotic index (AI) as a prognostic factor. Several animal studies have shown that the AI might indeed predict for tissue responsiveness to radiation [31,44]. A number of clinical studies have now been carried out relating tumor AI to treatment outcome. The emerging picture from the literature, however, is far from uniform. The majority of these studies established a correlation between a high AI and good prognosis, but several other analyses demonstrated the contrary [50]. These opposite ®ndings might be explained, in part, by differences in tumor characteristics among these studies, such as type, stage and grade. Also, variations in other tumor biological parameters including hypoxia, necrosis, mitotic index, p53 and Bcl-2/BAX status, may contribute to this controversy, and should therefore be taken into account as well. Interestingly, in vitro radiation-induced apoptosis was recently found to be useful as a predictive assay of early response to low dose in vivo irradiation in patients with non-Hodgkin's lymphoma [17]. Perhaps this approach is also applicable in other tumor types. 4.1. Conclusion Multiple signal transduction pathways have been identi®ed and implicated in the induction of apoptosis following exposure to ionizing radiation. Several lines of evidence suggest that in each cellular system, both anti- and proapoptotic signaling pathways operate. It has been postulated that the balance between these opposing signals may determine the biological outcome: survival or cell death. This concept provides a potential target for pharmacological modulation which would ultimately lead to a favorable shift in the therapeutic ratio. The clinical data on pretreatment levels of apoptosis for prediction of tumor radioresponsiveness are, as yet, inconclusive and clearly indicate the need for more prospective studies, taking into account additional tumor biological characteristics. 5. Functional estrogen receptor assay A response to hormonal therapy is seen in 60% of the patients with metastasized breast cancer, who have a positive estrogen receptor (ER) assessed with immuno-histochemistry. Response is observed in only 10% of patients with a negative receptor. This means that this classical test is a good, but certainly not fully reliable assay in predicting treatment outcome in patients with breast cancer. Recently obtained knowledge on the effectiveness of adjuvant hormonal therapy in breast cancer patients, illustrates the need for a much more reliable assay in assigning the appropriate adjuvant therapy, as shown for example in the `Scottish' trial [42]. This randomized trial showed that premenopausal patients with a positive receptor, bene®ted from adjuvant hormonal therapy and did not bene®t from adjuvant chemotherapy, while patients with a negative hormonal
8
H. Bartelink et al. / Radiotherapy and Oncology 50 (1999) 1±11
Fig. 9. The trial results changed signi®cantly during the follow-up period. First both at interim analysis and at trial closure the P-values were 0.01 and 0.004 in favor of adjuvant chemotherapy. However, with longer follow-up, this signi®cant advantage disappeared. Adjuvant hormonotherapy did not show a signi®cant bene®t at the beginning. However, with long-term follow-up the gain in survival became signi®cant. Reproduced by permission of W.B. Saunders Company, Journal of Clinical Oncology.
receptor bene®ted from adjuvant chemotherapy. Similar observations were made by Naja et al. [32]. An even more striking effect of adjuvant hormonal therapy on survival was seen in the EORTC trial, in patients with locally advanced breast cancer. Short-term results showed initially a highly signi®cant effect of adjuvant chemotherapy. This bene®cial effect disappeared, however, 2 years after trial closure. Recently, more than 8 years after trial closure, we demonstrated a signi®cant and impressive gain in survival with adjuvant hormonal therapy, while adjuvant chemotherapy did not signi®cantly improve life expectancy (Fig. 9) [4]. The recently performed meta-analysis by the ECBTCG [18], con®rms the value of adjuvant hormonal therapy in breast cancer patients even for premenopausal patients with a positive ER and PR receptor. The above-mentioned failure rate for adequately predicting the response to hormonal therapy, may represent an altered functional activity of the estrogen receptor due to mutations or variations at the messenger RNA level. In order to improve the predictive value of the estrogen receptor assay, a completely new assay was designed in our institute [47]. This so-called ER-FASAY (Estrogen Receptor Functional Analysis of Separated Alleles in Yeast) measures the functional response of the estrogen binding in yeast (Fig. 10a). With this assay, the type and frequency of the variants of the ER in a tumor can be determined and compared with the normal receptor. Messenger RNA is isolated from breast cancer cell lines, tumor biopsies and recently, also from ®ne needle aspirates. The RNA is reverse transcribed and PCR is performed using primers that amplify the ER open reading frame (PCR product is 2004 base pairs). The PCR products are then cloned into a yeast expression vector in vivo by homologous recombination (Fig. 10b). The growth of the
Fig. 10. (a,b) in one ®gure ER- FASAY, a schematic diagram of the functional assay in yeast for the estrogen receptor (M. van Dijk, The Netherlands Cancer Institute, Amsterdam).
H. Bartelink et al. / Radiotherapy and Oncology 50 (1999) 1±11
9
Fig. 11. ER-FASAY of one tumor biopsy. Forty-eight yeast colonies each containing one estrogen receptor molecule are tested for their functional activity. Normal receptor (wild type) is only active in the presence of estrogen (e.g. (B,C), row 1, spot 3) indicated by arrow, wild type. A DNA-binding (DBD) mutant is non-functional in the presence or absence of estrogen (e.g. (B,C), row 1, spots 1 and 2, indicated by arrow DBD mutant). A hormone binding (HBD) mutant is functional in the presence but also in the absence of estrogen (e.g. (B,C) row 1, spots 4, 5 and 6). E 2, estrogen; 2 leu, yeast plate lacking leucine; 2 his, yeast plate lacking histidine. (A) Non-reporter selective 2 leu: not selective for ER activity (only assays the presence of estrogen receptor. (B) Reporter selective 2 leu 2 his 1 E 2: selective for ER activity dependent on estrogen. (C) Reporter selective 2 leu 2 his 2 E 2: selective for ER activity independent of estrogen (M. van Dijk, The Netherlands Cancer Institute, Amsterdam).
yeast Saccharomyces cerevisiae is made, dependent on the functional integrity of the estrogen receptor by using an ERdependent promoter (containing an estrogen response element (ERE)) linked to a gene involved the synthesis of histidine. This allows discrimination between normal and variant receptors and at the same time determines their hormone-dependence (Fig. 11). The average relative expression level of ER in 22 primary breast tumors showed that 67% of ER mRNA in a tumor was wild-type (range 12± 100%), 30% were DNA binding mutants (range 0±64%) and 3% were hormone binding mutants (range 0±40%). All but one were splicing variants, and almost every breast tumor contained more than one ER variant. Almost all exhibited a positive estrogen receptor assessed with the immunohistochemistry assay. At present, predictive reliability of the ER-FASAY is undergoing prospective clinical testing in patients with inoperable or metastasized breast cancer. In the laboratory, the actual mechanism of the estrogen binding to its receptor is being investigated in order to allow therapeutic intervention. 5.1. Conclusion The long-term results of the EORTC trial revealing the improvement of survival by adjuvant hormone therapy for patients with locally advanced breast cancer, demonstrated the potential for selective administration of an adjuvant regime. The functional activity of the estrogen receptor can be measured with the yeast ER-FASAY. This may lead to a better prediction of treatment outcome for hormo-
nal therapy in patients with advanced disease. The ERFASAY may play an important role in designing adjuvant treatment in individual patients, especially as it appeared recently that adjuvant hormonal therapy is very effective in breast cancer patients with a positive estrogen receptor, assessed with an immunohistochemistry assay. 6. Summary The development and testing of the potential of predictive assays was inspired by the positive outcome, in terms of improved local control and survival of three phase III EORTC clinical trials in head and neck cancer, inoperable lung cancer and locally advanced breast cancer. These assays were aimed at predicting treatment outcome of radiotherapy alone, conventional or accelerated, or a combination of radiotherapy with systemic cisplatin or endocrine treatment in lung and breast cancer, respectively. The results obtained from this translational research have shown the potential for predicting treatment outcome by assays measuring proliferation, intrinsic radiosensitivity, the amount of drug-DNA adducts and the function of the estrogen receptor. This offers unique possibilities to modulate treatment response by selecting the appropriate treatment regime for individual patients. Moreover, knowledge of pathways leading to cell death can be exploited to intervene in speci®c signaling pathways, such as in the stimulation of apoptosis Translational research in the areas mentioned is now showing the way towards a real improvement in local
10
H. Bartelink et al. / Radiotherapy and Oncology 50 (1999) 1±11
control and survival in cancer, through individual tailoring of the treatment regime.
Acknowledgements This work was supported in part by grants from The Dutch Cancer Society: NKI 97-1435, NKI 96-1265, Schumacher Kramer Stichting.
[15]
[16] [17] [18]
References [1] Ang KK, Trotti A, Garden AS, et al. Overall time factor in postoperative radiation: results of a prospective randomized trial (abstract 108). Radiother. Oncol. 1996;40(Suppl.1):S30. [2] Bartelink H, Batterman J, Hart G. Half body irradiation. Int. J. Radiat. Biol. Phys. 1980;6:87±90. [3] Bartelink HB, Kallman RF, Rapacchietta D, Hart GAM. Therapeutic enhancement in mice by clinically relevant dose and fractionation schedules of CDDP and irradiation. Radiother. Oncol. 1986;6:61±74. [4] Bartelink H, Rubens RD, Van der Schueren E, Sylvester R. Hormonal therapy prolongs survival in irradiated locally advanced breast cancer: a European Organization for Research and Treatment of Cancer randomized phase III trial. J. Clin. Oncol. 1997;15:207±215. [5] Barton MB, Keane TJ, Gadalla T. The effect of treatment time on tumor control in the radical radiotherapy of laryngeal cancer. Radiother. Oncol. 1992;23:137±143. [6] Begg AC, McNally NJ, Shrieve DC, Karcher H. A method to measure the duration of DNA synthesis and the potential doubling time from a single sample. Cytometry 1985;6:620±626. [7] Begg AC, Ho¯and K, van Glabbeke M, Bartelink H, Horiot JC. Predictive value of potential doubling time for radiotherapy of head and neck tumour patients: results from the EORTC cooperative trial 22851. Semin. Radiat. Oncol. 1992;1:22±25. [8] Begg AC, Haustermans K, Hart AAM, et al. The value of pretreatment cell kinetic parameters as predictors for radiotherapy outcome in head and neck cancer: a multicenter analysis. Radiat. Oncol 1998; in press. [9] Begg AC. Cisplatin and radiation: interaction probabilities and therapeutic possibilities. Int. J. Radiat. Oncol. Biol. Phys. 1990;19:1183± 1189. [10] Blommaert FA, Michael C, Terheggen PMAB, et al. Drug induced DNA modi®cation in buccal cells of cancer patients receiving carboplatin and cisplatin combination chemotherapy, as determined by an immunocytochemical method: inter-individual variation and correlation with disease response. Cancer Res. 1993;53:5669±5675. [11] Bourhis J, Pignon JP, Designe M, et al. Meta-analysis of chemotherapy in head and neck cancer (MACH-NC): (1) loco-regional treatment vs. same treatment 1 chemotherapy (CT). Proceedings 34th ASCO, May 16±19, Los Angeles, USA (abstract 1487). J. Clin. Oncol. 1998;17:386a. [12] Calais G, Alfonsi M, Bardet E, et al. Randomized study comparing radiation alone (RT) versus RT with concomitant chemotherapy (CT) in stages III and IV oropharynx carcinoma (ARCORO). Preliminary results of the 94±01 study from the French Group of Radiation Oncology for head and neck cancer (GORTEC). Proceedings 34th ASCO, May 16±19, Los Angeles, USA (abstract 1484). J. Clin. Oncol. 1968;17:385a. [13] Chen Y-R, Wang X, Templeton D, Davis RJ, Tan T-H. The role of cJun N-terminal kinase (JNK) in apoptosis induced by ultraviolet C and radiation. J. Biol. Chem. 1996;271:31929±31936. [14] Coco Martin JM, Begg AC. Detection of radiation-induced chromosome aberrations using ¯uorescence in situ hybridization in drug-
[19] [20]
[22]
[23]
[24]
[25]
[26] [27] [28]
[29] [30] [31] [32]
[33] [34]
induced premature chromosome condensations of tumor cell lines with different radiosensitivities. Int. J. Radiat. Biol. 1997;71:265±73. Cummings BJ, Keane TJ, Pintillie M, et al. A prospective randomized trial of hyperfractionated versus conventional once daily radiation for advanced squamous cell carcinomas of the larynx and pharynx (abstract 107). Radiother. Oncol. 1996;40(Suppl. 1):S30. Dewey WC, Ling CC, Meyn RE. Radiation-induced apoptosis: relevance to radiotherapy. Int. J. Radiat. Oncol. Biol. Phys. 1995;33:781± 796. Dubray B, Breton C, Delic J, et al. In vitro radiation-induced apoptosis and early response to low-dose radiotherapy in non-Hodgkin's lymphomas. Radiother. Oncol. 1998;46:185±191. Early Breast Cancer Trialists' Collaborative Group, 1998, Tamoxfen for early breast cancer: an overview of the randomized trials (9114), Lancet, 351:1451-1467. Fyles A, Keane TJ, Barton M, Simm J. The effect of overall treatment duration in the local control of cervix cancer. Radiother. Oncol. 1992;25:273±279. Haimovitz-Friedman A, Kan C-C, Ehleiter D, et al. Ionizing radiation acts on cellular membranes to generate ceramide and initiate apoptosis. J. Exp. Med. 1994;180:525±535. Hannun YA. The sphingomyelin cycle and the second messenger function of ceramide. J. Biol. Chem 1994;269:3125±3128. Horiot JC, Bontemps P, Van den Bogaert W, et al. Accelerated fractionation (AF) compared to conventional fractionation (CF) improves loco-regional control in the radiotherapy of advanced head and neck cancers: results of the EORTC 22851 randomized trial. Radiother. Oncol. 1997;44:111±121. Horiot JC, Le Fur R, Nguyen T, et al. Hyperfractionation versus conventional fractionation in oropharyngeal carcinoma: ®nal analysis of a randomized trial of the EORTC cooperative group of radiotherapy. Radiat Oncol. 1992;25:231±241. Jarvis WD, Kolesnick RN, Fornari FA, Traylor RS, Gewirtz DA, Grant S. Induction of apoptotic DNA damage and cell death by activation of the sphingomyelin pathway. Proc. Natl. Acad. Sci. USA 1994;91:73±77. Jeremic B, Shibamoto Y, Acimovic L, Milisavljevic S. Initial versus delayed accelerated hyperfractionated radiation therapy and concurrent chemotherapy in limited small-cell lung cancer: a randomized study. J. Clin. Oncol. 1997;15:893±900. Kasid U, Suy S, Dent P, Ray S, Whiteside TL, Sturgill TW. Activation of Raf by ionizing radiation. Nature 1996;382:813±816. Kolesnick R. Sphingomyelin and derivatives as cellular signals. Prog. Lipid Res. 1991;30:1±38. Lelieveld P, Scoles MA, Brown JM, Kallman RF. The effect of treatment in fractionated schedules with the combination of X-irradiation and six cytotoxic drug on the RIF-1 tumor and normal mouse skin. Int. J. Radiat. Oncol. Biol. Phys. 1985;11:111±121. Ling CC, Chen CH, Li WX. Apoptosis induced at different dose rates: implication for the shoulder region of cell survival curves. Radiother. Oncol. 1994;32:129±136. Meyn RE, Stephens LC, Hunter NR, Ang KK, Milas L. Reemergence of apoptotic cells between fractionated doses in irradiated murine tumors. Int. J. Radiat. Oncol. Biol. Phys. 1994;30:619±624. Meyn RE, Stephens LC, Ang KK, Hunter N, Milas L, Peters LJ. Heterogeneity in apoptosis development in irradiated murine tumors. Int. J. Radiat. Biol. 1993;64:583±591. Naja A, Roche H, Mihura J, et al. Castration and tamoxifen versus chemotherapy (FAC) for premenopausal, node and receptors positive breast cancer patients: a randomized trial with a 7 years median follow up (abstract 87). Radiother. Oncol. 1997;43(Suppl. 2):S22. Overgaard J, Hjelm-Hansen M, Vendelbo Johansen L, Andersen AP. Comparison of conventional and split course radiotherapy as primary treatment in carcinoma of the larynx. Acta Oncol. 1988;27:147±152. Overgaard J, Sand Hansen H, Overgaard M, et al. Importance of overall treatment time for the outcome of radiotherapy in head and neck carcinoma. Experience from the Danish head and heck cancer
H. Bartelink et al. / Radiotherapy and Oncology 50 (1999) 1±11
[35]
[36]
[37] [38]
[39] [40] [41]
[42]
study. In: Kogelnik HD, Sedlmayer F, et al., editors. Progress in radiooncology. Proceedings of the 6th International Meeting, Salzburg, Austria, May 13±17, Bologna: Monduzzi Editore, 1998. pp. 743. Quintans J, Kilkus J, McShan CL, Gottschalk AR, Dawson G. Ceramide mediates the apoptotic response of WEHI 231 cells to antiimmunoglobulin, corticosteroids and irradiation. Biochem. Biophys. Res. Commun. 1994;202:710±714. Rose P, Bundy B, Thigpen J, et al. Signi®cant preliminary results of a phase III randomized study of concomitant chemotherapy with weekly cisplatin vs. hydroxyurea, 5-Fluorouracil infusion and bolus cisplatin vs. hydroxyurea in advanced cervical cancer: a gynecologic oncology group (GOG) study. Proceedings 34th ASCO, May 16±19, Los Angeles, USA (abstract 1391). J. Clin. Oncol. 1998;17:360a. Santana P, PenÄa LA, Haimovitz-Friedman A, et al. Acid sphingomyelinase-de®cient human lymphoblasts and mice are defective in radiation-induced apoptosis. Cell 1996;86:189±99. Saunders M, Dische S, Barrett A, Harvey A, Gibson D, Parmar M. Continuous hyperfractionated accelerated radiotherapy (CHART) versus conventional radiotherapy in non-small-cell lung cancer: a randomized multicentre trial. Lancet 1997;350:161±165. Sawai H, Okazaki T, Yamamoto H, et al. Requirement of AP-1 for ceramide-induced apoptosis in human leukemia HL-60 cells. J. Biol. Chem. 1995;270:27326±27331. Schaake-Koning C, Van den Bogaert W, Dalesio O, et al. Effects of concomitant cisplatin and radiotherapy on inoperable non-small-cell lung cancer. N. Engl. J. Med. 1992;326:524±530. Schuster-Uitterhoeve ALJ, Van de Vaart PJM, Schaake-Koning CCE, et al. Feasibility of escalating daily doses of cisplatin in combination with accelerated radiotherapy in non-small cell lung cancer. Eur. J. Cancer 1996;32A:1314±1319. Scottish Cancer Trials Breast Group and ICRF Breast Unit, Guy's Hospital, London, 1993, Adjuvant ovarian ablation versus CMF chemotherapy in premenopausal women with pathological stage II breast carcinoma: the Scottish trial, Lancet, 341:1293-1298.
11
[43] Shirakabe K, Yamaguchi K, Shibuya H, et al. TAK1 mediates the ceramide signaling to stress-activated protein kinase/c-Jun N-terminal kinase. J. Biol. Chem. 1997;272:8141±8144. [44] Stephens LC, Ang KK, Schultheiss TE, Milas L, Meyn RE. Apoptosis in irradiated murine tumors. Radiat. Res 1991;127:308±316. [45] Ten Bokkel Huinink WW, Schornagel JH, Hilton A, Somers R, Bartelink H. Topical application of miltefosine against skin metastases of breast cancer. Proceedings of ASCO, San Diego, USA, May 17±19 (abstract 39). J. Clin. Oncol. 1992;11:53. [46] Terheggen PMAB, Emondt JY, Floot BGJ, et al. The correlation between cell killing by cis-diamminedichloroplatinum(II) in six mammalian cell lines and binding of a cis-diamminedichloroplatinum(II)-DNA antiserum. Cancer Res. 1990;50:3556±3561. [47] Van Dijk MAJ, Floore AN, Kloppenborg KIM, Van 't Veer LJ. A functional assay in yeast for human estrogen receptor displays wildtype and variant estrogen receptor messenger RNAs present in breast carcinoma. Cancer Res. 1997;57:3478±3485. [48] Verheij M, Bose R, Lin XH. Requirement for ceramide-initiated SAPK/JNK signalling in stress-induced apoptosis. Nature 1996;380:75±79. [49] Verheij M, Ruiter GA, Zerp SF, et al. The role of the stress-activated protein kinase (SAPK/JNK) signaling pathway in radiation-induced apoptosis. Radiother. Oncol. 1998;47:225±232. [50] Verheij M, van Blitterswijk WJ, Bartelink H. Radiation-induced apoptosis: the ceramide-SAPK signaling pathway and clinical aspects. Acta Oncol. 1998; in press. [51] West CM, Davidson SE, Roberts SA, Hunter RD. The independence of intrinsic radiosensitivity as a prognostic factor for patient response to radiotherapy of carcinoma of the cervix. Br. J. Cancer 1997;76:1184±1190. [52] Zanke BW, Boudreau K, Rubie E, et al. The stress-activated protein kinase pathway mediates cell death following injury induced by cisplatinum, UV irradiation or heat. Curr. Biol 1996;6:606±613.