- Email: [email protected]
Response of mouse tongue epithelium to single doses of bleomycin and radiation
Radiotherapy and Oncology, 27 (1993) 237-244
237
© 1993 Elsevier Scientific Publishers Ireland Ltd. All rights reserved. 0167-8140/93/$06.00 RADION 01159
Response of mouse tongue epithelium to single doses of bleomycin and radiation Wolfgang D6rr, Erwin Hirler and Michaela H6nig GSF-lnstitut .t~r Strahlenbiologie, D-85764 Neuherberg, Germany
(Received 9 October 1992; revision received 14 January 1993; accepted 26 January 1993)
Key words: Mouse tongue mucosa; Bleomycin; Radiation; Interactions
Summary Both bleomycin (BLM) and local X-irradiation (25 kV) induce denudation in the tongue epithelium of the C3H-Neuherberg mouse in a dose-dependent manner. In the present study the effect of BLM alone and of combined single doses of drug and radiation were studied using the incidence of epithelial denudation as the end-point. In 'time-line' experiments, 8 mg/kg BLM were given before or after graded doses of X-rays. BLM treatment required a reduction of the radiation dose (EDs0) from 15 Gy to 5-7 Gy, independent of sequence or time interval. In contrast, the time course of the response was clearly dependent on the treatment interval. Latency decreased when the drug was injected less than 2 h before irradiation with minimum latency observed at 30 min. Isobologram analysis of experiments with varying combinations of X-rays and BLM demonstrated that small drug doses were relatively more effective than larger doses, suggesting an upward concavity of the BLM dose-effect curve in vivo, i.e. a 'negative shoulder' of the curve in the low dose region. In contrast to the response to X-rays alone, which has a constant latent time to ulcer of 10 days, the latency in combined treatment was clearly shortened with increasing drug dose and at high doses eventually approximated the epithelial turnover time of 5 days. The data suggest that BLM both as a single agent and in combination with X-rays reduced the probability of abortive divisions and through this effect shortened the latent time to epithelial denudation.
Introduction Squamous cell carcinoma appear to be highly sensitive to bleomycin (BLM), which for this reason has been introduced in a number of radiochemotherapy protocols for head and neck cancer [2,7,18]. Whenever bleomycin was added to existing treatment protocols, a substantial increase in the acute toxicity to oral mucosal membranes was observed. The effect was most pronounced when radiation and drug were given in close temporal proximity [7,13]. The basic mechanisms underlying the increased toxicity, however, remained unclear. Effects on capacity and kinetics of radiation-induced sublethal damage repair seemed unlikely, but could not be excluded from studies in the mouse lip mucosa [6,24]. In the same model it could be demonstrated that bleomycin was highly effective in suppressing repopulation in con-
tinuously fractionated [25] and split-dose experiments [23]. This effect was attributed to independent cell killing. However, effects on repair of sublethal X-ray damage could not be excluded. As an alternative to the mouse lip system we have established mouse tongue epithelium as a model to investigate radiation effects on oral mucosa [3,4,11]. Local irradiation of the tongue by soft X-rays induces ulcerative lesions of the mucosa in a dose-dependent manner. In the same mouse strain (C3H-Neuherberg), BLM alone provokes epithelial lesions qualitatively similar to radiation-induced ulceration. This marked sensitivity can be exploited for detailed studies on the in vivo effect of combined treatment with drug and radiation. The present paper reports on the effect of BLM alone and the combination of single doses of drug and radia-
Correspondence to: Wolfgang Drrr, GSF-Institut fiir Strahlenbiologie, Neuherberg, Postfach 1129, D-85758 Oberschleissheim, Germany.
238 tion. The influence of sequence and time interval in the application of both agents was investigated in a 'timeline' study. In a second set of experiments the effect of various single-dose combinations of BLM and X-rays were determined using isobologram analysis [15,17]. M a t e r i a l and m e t h o d s
Female mice of the inbred C3H-Neuherberg strain, 10-12 weeks old, were used in all experiments. The animals were housed under specified pathogen free conditions with controlled humidity and temperature. A 12/12 h light-dark cycle was maintained changing at 06:00/18:00 h. Standard pellet diet (Altromin®, Altrogge, Lage, Germany) and water were given ad libitum. Bleomycin (Bleomycinum Mack ®, Mack, Illertissen, Germany) was dissolved in 0.9% NaCI and diluted to concentrations that provided injection volumes (i.p.) of 0.1-0.5 ml per mouse.
Irradiation technique The animals were anaesthetized for irradiation using either hexobarbital sodium (Evipan®, Bayer, Leverkusen, Germany) at a dose of - 150 mg/kg i.p. in the isobologram experiments, or pentobarbital sodium (Nembutal®, WdT, Hannover, Germany) at a dose of - 60 mg/kg i.p. in the time-line experiments. Single radiation doses were applied to a defined area of the inferior tongue surface using a technique recently described in detail [3]. In brief, the anaesthetized animals were placed in the cylindrical bore of a prewarmed (35°C) aluminium block in a supine position with the head supported by a polystyrene wedge, and the tongue was gently pulled out through a small hole in the roof of the block by means of a forceps. The tongue was inflicted cranially and fixed with adhesive tape to expose its ventral surface to a vertical X-ray beam. An aluminium shield (1 mm) with a window of 3 x 3 mm was centrally positioned on the tongue to define the treatment area. The X-rays were generated by a Philips PWll30 machine operated at 25 kV with a beam filter of 0.3 mm AI. When X-ray doses smaller than 4 Gy had to be applied, a shutter (3 mm AI) was used which was opened when the tube voltage was constant at 25 kV. The source-to-surface distance was 15.5 cm. The dose rate at the tongue surface varied between 4.8 and 9.5 Gy/min, resulting in treatment times of 0.6-2 min. An ionization chamber (M23342, PTW Freiburg, Germany) was positioned close to the tongue at identical source distance and connected with a Dosimentor SN4 (PTW Freiburg) for dosimetry in all irradiation sessions.
Scoring Mucosal reactions were scored daily from the occurrence of first clinical symptoms until complete reepithelialization. Following immobilization with either hexobarbital sodium or pentobarbital sodium the tongue was pulled out carefully by means of a forceps and examined under a magnifying lens (3 x ) and a cold light source. After a lag period of 2-7 days, dependent on the experimental design, slight oedema and erythema developed, followed by increased desquamation that progressed to an erosive lesion after sufficient dose levels. Radiation-induced denudation was restricted to the exposed area at the inferior surface. In contrast, high BLM doses yielded ulceration of the whole tongue. The reaction assessed for the dose-response analyses was the incidence of denudation/shallow ulceration within the irradiation test area in order to achieve comparable resuits with both agents. The incidence of this effect was strictly dependent on radiation and/or drug dose. Doseeffect curves were fitted by probit analysis and the EDs0 doses and 95% confidence intervals calculated. At least seven animals were used per dose point.
Experimental design Time-line experiments To study the effect of sequence and time interval between both agents, BLM was applied 24 h, 6 h, 2 h or 30 min before and after irradiation. Furthermore, simultaneous application was tested, i.e. the drug was given immediately (< 30 s) before or after irradiation. The minimum interval between application of drug and anaesthetic was 20 min. The time of irradiation in this set of experiments was kept constant between 09:00 h and 12:00 h. Histological samples of the mobile tongue were taken from two animals each on Days 2, 4 and 6 after 22 mg/kg BLM without irradiation. In addition, samples were taken every second day between Days 2 and 10 after BLM application 6 h and 0.5 h before and 6 h after irradiation with 12 Gy. The specimens were fixed in formalin-alcohol and embedded in methylmethacrylate. Sagittal sections of 3/~m were obtained at several levels and stained with haematoxilin and eosin.
Isobologram study To investigate the effect of combined treatment with varying drug and radiation doses, BLM at nine dose levels ranging from 0 to 16 mg/kg was injected 2 h before irradiation. Graded X-ray doses were applied to generate full dose-response curves of complete denudation for each drug dose level. The time of irradiation was kept constant between 10:00 and 12:00 h in these experiments.
239
Results
Effect of X-rays alone Single dose irradiation with soft X-rays resulted in the first signs of mucositis, i.e. diffuse erythema and oedema, on Day 6-7. The symptoms progressed into marked desquamation, producing a scaly surface within the irradiated area after another 1 or 2 days. The fall-off in dose through the tongue was sufficient to avoid clinical damage on the upper tongue surface. The maximum response presented as a complete denudation or ulcerative lesion in the irradiated area that was usually covered by a pseudomembrane. The individual ulcer duration was 3 4- 1 days with hexobarbital anaesthesia and 4 4- 1 days with pentobarbital anaesthesia used for irradiation. Subsequently, rapid regeneration led to complete healing by Day 16 at the latest. The frequency of animals developing erosions displayed a stringent dose relationship, yielding a sigmoid dose-response curve as shown in Fig. 1A. The EDs0 to irradiation with hexobarbital anaesthesia was 15.0 Gy with 95% confidence limits of 13.5 and 16.3 Gy. Irradiation under pentobarbital anaesthesia resulted in an EDs0 of 13.6 Gy (11.6; 15.1). In contrast to ulcer frequency, the time course of the reaction was independent of the radiation dose including the highest dose level used (22 Gy), as demonstrated in Fig. 1B. The total latency to complete denudation after irradiation was 10.9 4- 1 days, being similar with both anaesthetics used. Maximum prevalence over all dose groups was observed on Day 12. As reported recently [3], only small fluctuations of body weight were observed after local tongue irradiation, independent of radiation dose and ulcer frequency.
100
•
i
•
i
•
(3."e f
r'"l I--II
/ /
(A)
80
U
This is good evidence that irradiation alone was well tolerated and failed to cause even temporary morbidity. The histological changes following single-dose irradiation of mouse tongue have recently been described in detail [4].
Effects of bleomycin alone Bleomycin alone is able to induce ulcerative lesions in the tongue of the C3H-Neuherberg mouse. The clinical symptoms, i.e. erythema, oedema, increased desquamation and epithelial erosions, were qualitatively similar to the changes induced by radiation. However, in addition to the response of the inferior surface, reactions were observed frequently on the tip and edges of the tongue and partly at the superior surface. To achieve response data comparable to irradiation, scoring was restricted to the site of the irradiation field, i.e. the central portion of the lower tongue surface. The latency to ulceration after BLM, in contrast to the response to X-rays, was clearly dependent on the drug dose. The mean latent times (± SEM) after 14, 16, 18, 20 and 22 mg/kg were 9.5 4- 0.5 days (n = 2), 9.0 4- 0.5 days (n = 4), 8.3 ± 0.3 days (n=6), 7.3 4- 0.4 days ( n = 6 ) and 6.7 4- 0.4 days (n = 7), respectively. The individual ulcer duration (4SD) was 3.4 4- 1.2 days being independent of the drug dose. The frequency of animals developing erosions of the lower tongue surface again yielded a well-defined dose-response curve. The EDs0, calculated by probit analysis, was 17.5 mg/kg (95% confidence limits, 16.2 and 18.8 mg/kg). In contrast to X-rays, some weight loss was observed after BLM with an indication of some dose dependence. Maximum weight loss was 8%. The qualitative histological changes in the tongue epithelium after BLM were similar to those observed after
i
I
I
Ce)
•
.
O
r¢t
I
v"
•
6o
I
¥
I
•
•
,
"1
40 U
,,. 7",: \
°/Z
20
,,f/"
?/J ",
0 8
12 X-roy
16 d o s e [Gy]
2O
8
. 10
Doys
,
. ",
12 ofter
14
.~\r 16
SO
Fig. 1. A, Frequency of animals developing tongue erosions after local single-dose irradiation under pentobarbital (O) or hexobarbital (O) anaesthesia. Dose-response curves were fitted by probit analysis, error bars depict 95% confidence intervals. B, Time course of ulcer prevalence after single-dose irradiation (hexobarbital anaesthesia) with 14 (O), 16 (n), 18 ( 0 ) and 20 Gy (V).
240 TABLE I Results of time-line experiments. Time interval (h)
EDs0 (Gy)
95% Confidence limits (Gy)
Ulcer duration (days)
Latency (days)
-24 -6 -2 -0.5
5.3 6.3 6.7 6.4
2.9; 4.7; 5.2; 4.5;
3,5 3.0 2.3 3.5
lO.O (l.O) 10.9 (0.7) 8.0 (1.O) 5.6 (0.7)
-0 +0
5.6 6.5
2.0; 8.2 3.5; 8.5
4.0 (1.4) 3.6 (1.7)
7.4 (0.7) 7.1 (0.3)
+0.5 +2 +6 +24
6.6 7.0 7.4 6.3
4.4; 5.7; 5.4; 4.4;
3.4 2.7 3.2 3.3
lO.O (l.1) 10.7 (l.1) 10.4 (0.7) 9.8 (1.3)
7.0 7.9 8.3 8.3
8.7 8.3 9.9 8.1
(0.9) (1.2) (0.6) (0.9)
(1.2) (0.9) (1.2) (1.3)
A BLM dose of 8 mg/kg was applied 30 min to 24 h before (negative time interval) or after local tongue irradiation at Time 0. In addition, the drug was injected immediately before (-0) or after (+0) X-ray treatment. EDs0 values and confidence limits were calculated by probit analysis• Numbers in parentheses depict 1 SD.
X-irradiation, although they occurred earlier. Abnormal mitotic figures (chromosome clumping, multipolar spindles) and abnormal nuclei were already found by Day 2 after BLM (22 mg/kg) and culminated on Days 4-6. On Day 4 prominent enlargement of individual cells was observed. As a consequence of abnormal divisions, binucleate cells (10%) and multinucleate cells (6%) reached a maximum on Day 6 in non-denuded regions. Foci of complete denudation were observed by Days 6-8. The erosions were covered by a membrane consisting of cell detritus, keratin and fibrin.
mainder. Cell density within the irradiated area decreased tO 81% (drug before X-rays) and 84% (drug after X-rays) on Day 2, and to 64% and 59% on Day 4. First foci of denudation were found on Day 6, at the time when the average cell count in non-denuded regions had reached 37% and complete denudation was observed on Day 8. Multinucleate keratinocytes were first observed on Day 2 (< 1%) and their number subsequently increased to 6%, 9% and 17% on Days 4, 6 and 8, respectively, to reach a maximum count of 18% on Day 10. While the longer treatment intervals resulted in no change in the latency as compared with radiation
Time-line experiments In these experiments a constant dose of 8 mg/kg BLM was injected 0 to 24 h before or after irradiation or injected immediately (< 30 s) before or after the X-ray treatment. Graded X-ray doses were administered and the resulting dose-response curves were fitted by probit analysis• Table I summarizes the resulting EDs0 values and their 95% confidence limits, average duration of ulcer and the observed mean latent times• The EDs0 values displayed no systematical fluctuations with the variation of the time interval between both agents. The mean EDs0 from all experiments was 6.4 Gy (4.9; 7.9), yielding a dose enhancement factor (DEF = EDs0 radiation alone/EDs0 radiation plus BLM) of 2.13 for 8 mg/kg BLM. Also, the individual ulcer duration was independent of the treatment protocol, giving a mean duration of 3.2 ± 1.2 days. Histological changes after BLM application 6 h before or 6 h after irradiation (12 Gy) were confined to the irradiated area in about 50% of the animals but included also tip and margins of the tongue in the re-
•
.
i
.
.
.
.
i
.
.
.
.
i
.
.
.
.
i
.
.
.
.
i
•
i
-
.
12
10 "0
¢J
.E_
a
¢p o
6
._1
1 B/M b e f o r e [ BLM o f t e r irrodiotlon •
i
-20
.
.
.
.
i
-10
.
.
.
.
i
.
.
.
0
Time intervol
.
i
10
.
.
.
.
20
[hi
Fig. 2. Latent time of tongue ulcer after a single injection of BLM at various intervals before or after local single-dose irradiation. The time of irradiation was kept constant between 10:00 h and 12:00 h.
241 TABLE II Results of the isobologram study. BLM dose (mg/kg)
EDs0 (Gy)
95% Confidence limits (Gy)
Ulcer duration (days)
Latency (days)
0 1 3 5 7 10 12 14 16
15.0 10.1 6.6 6.3 6.2 5.8 4.5 1.5 0.7
13.5; 16.3 7.5; 11.9 5.5; 8.3 3.9; 7.8 4.8; 9.4 4.6; 7.0 3.1; 7.3 -- ; --0.4; 1.5
3.2 3.6 3.6 3.7 4.8 3.7 3.3 3.6 3.3
10.9 (1.0) 10.5 (2.1) 7.5 (1.1) 8.0 (1.4) 6.6 (1.2) 6.4 (1.0) 6.3 (1.1) 6.0 (1.3) 5.1 (1.4)
(1.2) (1.5) (1.1) (1.1) (1.6) (1.1) (1.0) (1.2) (0.8)
Various constant drug doses were combined with graded doses of X-rays. Irradiation was administered 2 h after BLM injection. Confidence limits of EDs0 values were calculated by probit analysis. Numbers in parentheses indicate 1 SD.
alone (10-11 days) a clear shortening was observed when BLM was applied at an interval of 2 h or less prior to radiation. The minimum latency of 5.6 days was found when the drug was applied 30 min before irradiation (Table I and Fig. 2). In accordance with the reduced latent time to ulcer, cell density declined to 50% already on Day 2 and dropped further to 25% on Day 4, when BLM was applied 30 min before X-irradiation. Complete denudation was reached by Day 6. Multinucleate keratinocyte counts were 6.5%, 15% and 18% on Days 2, 4 and 6, respectively.
and 10:00 h, to be followed by graded X-ray doses 2 h later. X-ray doses were chosen to generate full doseresponse curves that allowed the radiation EDs0 required for each BLM dose to be calculated. Table II summarizes the EDs0 values, mean ulcer durations and average latent times of these experiments. The dependence of the EDs0 of X-rays on the BLM dose is graphically presented in Fig. 3A. As the curve demonstrates, the EDs0 values declined in a non-linear manner with increasing BLM dose. The reduction in isoeffective radiation dose was clearly steeper at the lower BLM doses. As BLM per se is able to induce tongue ulcer at doses above 14 mg/kg, the X-ray doseresponse curves above this threshold dose did not include low response frequencies and hence were less well defined, which is expressed in the lacking confidence limits. The DEF (EDs0 radiation alone/ED50 radiation
Combination of varying drug and radiation doses ( isobologram studies) In the isobologram studies nine dose levels of BLM, varying from 0 to 16 mg/kg, were given between 08:00 h
12
15
.-o, o u3 a
t0
E
tlJ 0
I
10
".7,
8
~
e
0 _J
X
i
0
. . . .
|
5
. . . .
|
10
. . . .
i
. . . .
15
BLM c l o s e [ m g / k g ]
20
0
5 BLM d o s e
10
15
[mg/kg]
Fig. 3. A, X-ray single-dose resulting in erosions in 50% of the animals as a function of the BLM dose applied 2 h before irradiation. Included is the result of BLM given alone (O). B, Latency to ulceration as a function of the drug dose applied 2 h before local irradiation. X-rays alone (11-20 Gy) result in a constant latency of 10.9 -,- 1 days. At high drug doses, the latency approximates the turnover time (TT) of the tissue.
242 plus BLM) varied between 1.49 (1 mg/kg) and 21.4 (16 mg/kg). The individual ulcer persistence was independent of drug dose. However, in contrast to the picture after radiation alone, the latency to ulceration decreased consistently with increasing BLM dose (Fig. 3B). Drug doses higher than 14 mg/kg resulted in latent times of 5-6 days compared with about 11 days after irradiation alone. The upward concave shape of the curve in Fig. 3B illustrates that low drug doses were again relatively more effective than higher doses for this end-point also. Discussion Clinical experience shows that BLM substantially enhances the acute mucosal response when added to existing radiochemotherapy protocols [7,15]. This increase in toxicity may be based either on independent cell kill or on direct interaction of both agents. In vitro studies demonstrated an increase in radiosensitivity of the target cell population by partial cell cycle synchronization by the drug [8], or an increase in cellular radiosensitivity, mainly by impairment of repair of sublethal damage [20,22]. However, in vivo studies into the functional response are required to investigate the effect of the combined treatment at the tissue level. The acute radiation response of oral epithelium is caused by the sterilization of (stem) cells in the germinal layer. Their proliferative failure translates into a lack of cellular supply to the functional layers and, in the presence of ongoing differentiation and sloughing off of cells or keratin at the surface, results in progressing tissue hypoplasia and eventual denudation. For experimental studies into the problem of acute radiation-induced mucositis the mouse lip mucosa assay was introduced [12,18]. Possible effects of BLM on the capacity and kinetics of repair of sublethal damage have been investigated with this model [6,23]. In combination with single-dose irradiation, BLM increased the mucosal response when administered 4 days or less before the X-ray treatment, with a maximum effect at 2 h. In addition, the response in these experiments occurred significantly earlier than the response to X-rays alone. When continuous drug infusion (7 days) was combined with X-rays of increasing fraction number, the DEF progressively increased. The experimental resuits could plausibly be explained by reduced repair of sublethal damage in addition to independent cytotoxicity. Repopulation in mouse lip mucosa, tested in the interval between split doses was clearly suppressed by BLM [25]. As an alternative animal system for radiobiological studies into the problem of acute oral mucositis the mouse tongue model was established [3,4,5,11]. The response to radiation in this assay is quantified by the frequency of animals that develop complete mucosal
denudation in a defined test area. This end-point obviously depends on stem-cell depletion below a critical level and hence allows extrapolation to cellular responses. In addition, the time course of the reaction gives interesting clues to population kinetics. Assuming both a total proliferative block and a normal differentiation rate after irradiation, the transit time of cells, i.e. the postmitotic life-span of the cells or the time of migration from the basal to the most superficial cell layer, should determine the latent time [9]. In steady state tissues the transit time equals the turnover time of the nucleated cells. Yet, the normal cell turnover time in mouse tongue mucosa has been measured to be about 5 days [4] and thus is considerably shorter than the latent time to complete denudation. The discrepancy can be explained by the fact that abortive cell divisions take place for a longer period, eventually resulting in mostly abnormal (bi- and multinucleate) cells. No influence of the X-ray dose on latency was observed in the present study, suggesting that the division probability and functional performance of abortive daughters is similar over the dose range used. Bleomycin per se induced erosions in tongue epithelium of the C3H-Neuherberg strain. Histologically and in their clinical appearance the mucosal ulcerations were similar to those observed after irradiation; however, in contrast to the response to radiation alone, the latent time to ulceration was dependent on the drug dose. Even small BLM doses significantly shortened the time to ulceration as compared with the latency seen after single-dose irradiation. After high doses of BLM, the time to complete denudation approximated the tissue turnover time of 5 days. The kinetics of developing damage was studied after a BLM dose of 22 mg/kg. This dose (ED95) should induce a cell kill comparable with that obtained by a single radiation dose of 20 Gy (ED95), which has been studied previously [4]. Abnormal mitotic figures after a high dose of BLM culminated already on Days 4-6 as compared with Days 6-8 after the high radiation dose [4]. In consequence, a maximum number of multinucleate keratinocytes was found on Day 6 ( - 16%) after BLM and on Days 7-9 (-40%) after X-rays. Between Days 2 and 4, the efficacy of abortive divisions was similar after both agents, resulting in about 16-20% of multinucleate cells. Their number subsequently further increased after X-rays, but clearly decreased after BLM. A plausible explanation of the discrepancies in latent time and time pattern of histological damage observed between BLM and single-dose irradiation may be that BLM decreases the average number of abortive divisions of doomed cells and leads to earlier occurrence of abnormal mitoses. This should result in a shortening of the latency to complete denudation, but in qualitative changes similar to those after irradiation. No systematical influence of sequence or time interval
243
on the sensitivity of mouse tongue epithelium to combined treatment was evident in the present study. In contrast, enhancement of drug toxicity has been described in the mouse lip model for BLM injection 2 h before irradiation as well as for a 7 days continuous infusion of BLM prior to irradiation [6]. Maximum effects of BLM were also found in rodent skin when the drug was given 15 min to 1 h prior to irradiation [10,26], or 1-6 h before X-rays in intestinal crypts [14,27]. In a mouse tumour model, large fluctuations in tumour growth delay with varying time interval and sequence were observed when BLM was given in intervals from 48 h before to 48 h after irradiation [1]. However, the growth delays in these experiments at most time intervals were equal or less than for X-rays alone. While in our studies the time interval between drug and radiation had no effect on the incidence of ulcer, a sharp drop in the latent time was observed when BLM was administered either at intervals shorter than 2 h before or immediately after irradiation. Injection 0.5 h before irradiation resulted in a latent time of 5.6 days, approximating the normal turnover time of the epithelium [4]. Shortening of the latent time of the response has also been reported for mouse lip mucosa [6] and for mouse skin [261. The histological studies illustrate that the most marked decline in cell density is observed when irradiation is administered 30 min post injection. In addition, the number of multinucleate cells, serving as a measure of effective though abnormal abortive divisions, increases 2 days earlier than with irradiation given within 6 h before or after drug injection. This suggests that a short time interval between BLM and X-rays, similar to high BLM doses alone, decreases the overall number of functional abortive divisions. Thus, the total amount of production of cellular material is decreased after the short time interval, leading to a clear shortening of the latent time. Assuming simple additivity of the cytotoxic effects of both agents, the isobologram curve can be used to derive a tentative cell survival curve to BLM. This is facilitated by the known general shape of the X-ray survival curve (Fig. 4), based on linear quadratic analysis of split-dose experiments in mouse tongue which yielded an or/3 value of 11.6 Gy [5]. The full effect (1.0) in Fig. 4 is the amount of cell kill required to induce ulcer in 50% of the tongues, corresponding to 15 Gy X-rays alone or 17.5 mg/kg BLM alone. In combined treatment studies, the full effect is composed of the known partial effect of Xrays plus the partial effect of BLM. For example, the proportion of the full effect of a given X-ray dose of 6.6 Gy is 0.3. The corresponding BLM dose of 3 mg/kg (Table II) thus has to induce seven-tenth of the full damage (1.0) necessary to result in 50% ulcer. The effect of BLM resulting from the calculations for all dose pairs is shown in Fig. 4, illustrating that the the-
BLM dose [mg/kg] 0
5
I
r--1
20
,
,
.
,
,
10
I
,
o
=
=
I
15 .
.
.
.
I
•
•
\
• o
,
\
X-rays
".T
,, =/p=11.8 oy
40
ID 0
6O
~
8O
100
II
"'" .....
BLM
I
.
.
0
.
.
I
5
.
.
.
.
I
.
.
.
10
.
I
15
X-ray dose [Gy] Fig. 4. Quasi-survival curves to X-rays and BLM, plotting proportion of full effect (50% responders) versus dose. The X-ray curve was established by the linear-quadratic model using an a/~ value of 11.6 Gy [5]. The dose-response curve to BLM was subsequently generated by assuming pure additivity of drug and radiation effects. Thus, to obtain the partial effect of a given BLM dose in combined treatment (according to Table II), the partial effect of the complementary X-ray EDs0 was subtracted from 1.0. Similarly, error bars were obtained by using 95% confidence limits of the X-ray EDs0 doses.
oretical cell survival curve to BLM in vivo has an even more pronounced upward concavity than the combined dose-effect curve. Similarly shaped dose-effect curves have been derived from in vitro experiments using BLM alone [19,21,29] or in combination with X-rays [16]. As a consequence, repeated injections of BLM at low doses may cause repetition of the steep segment of the cell survival curve and thus confer increased mucosal toxicity. Such an enhanced efficacy by fractionating a total BLM dose has indeed been observed in independent experiments on mouse tongue mucosa (Drrr et al., unpublished data). In summary, two interesting effects of bleomycin on oral mucosa in mice could be demonstrated in the present study. The dose- and time-dependent shortening of the latency to ulceration is probably caused by a decrease in the amount and efficacy of abortive cell divisions. The relatively higher effect of small drug doses suggests an upward concavity of the in vivo doseresponse curve to the drug, which deserves further experimental investigation because of its obvious clinical implications.
Acknowledgements This study was in part supported by the Deutsche Forschungsgemeinschaft, grants number Ku576/2-1
244
and Ku576/2-2. All experiments were carried out according to the valid legislation with the permission of the Regierung von Oberbayern. The authors are grateful to Dr. J. Kummermehr for critical reading of and helpful comments on the manuscript. References 1
2
3
4
5
6
7
8
9
10
11 12
13
Begg, A. C., Kane, L. J., Fu, K. K. and Philips, T.L. The effects of combined bleomycin and irradiation on a solid murine tumor. Int. J. Radiat. Oncol., Biol. Phys. 5: 1503-1507, 1979. Cachin, Y., Jortay, A., Sancho, H., Eschwege, F., Madelian, M. Desaulty, A. and Gerard, P. Preliminary results of a randomized E.O.R.T.C. study comparing radiotherapy and concommirant bleomycin to radiotherapy alone in epidermoid carcinomas of the oropharynx. Eur. J. Cancer 13: 1389-1395, 1977. D6rr, W. and Kummermehr, J. Accelerated repopulation of mouse tongue epithelium during fractionated irradiations or following single doses. Radiother. Oncol. 17: 249-259, 1990. D6rr, W. and Kummermehr, J. Proliferation kinetics of mouse tongue epithelium under normal conditions and following single dose irradiation. Virchows Arch. B 60: 287-294, 1991. D6rr, W., Breitner, A. and Kummermehr, J. Capacity and kinetics of SLD repair in mouse tongue epithelium. Radiother. Oncol. 27: 36-45, 1993. Feng, Y,, Vanuytsel, L., Landuyt, W., Ang, K. K. and Van der Schueren, E. The combined effect of bleomycin and irradiation on mouse lip mucosa. I. Influence of timing, sequence and mode of drug administration with single dose irradiation. Radiother. Oncol. 6: 143-151, 1986. Fu, K. K., Philips, T. L., Silverberg, I. J., Jacobs, C., Goffinet, D. R., Chun, C., Friedman, M. A., Kohler, M., McWirter, K. and Carter, S.K. Combined radiotherapy and chemotherapy with bleomycin and methotrexate for advanced inoperable head and neck. J. Clin. Oncol. 5: 1410-1418, 1987. Kimler, B. F. The effect of bleomycin and irradiation on G 2 progression. Int. J. Radiat. Oncol., Biol. Phys. 5: 1523-1526, 1979. Michalowski, A. Effects of radiation on normal tissues: hypothetical mechanisms and limitations of in situ assays of clonogenicity. Radiat. Environ. Biophys. 19: 157-172, 1981. Molin, J., Sogaard, P. E. and Overgaard, J. Experimental studies on the radiation-modifying effect of bleomycin in malignant and normal mouse tissue in vivo. Cancer Treat. Rep. 65: 583-589, 1981. Moses, R. and Kummermehr, J. Radiation response of mouse tongue mucosa. Br. J. Cancer 53 (Suppl. VII): 12-15, 1986. Parkins, C. S., Fowler, J. F. and Yu, S. A murine model of lip epidermal/mucosal reactions to X-irradiation. Radiother. Oncol. 1: 159-165, 1983. Peters, L. J., Harrison, M, L., Dimery, I. W., Fields, R., Goepfert, H. and Oswald, M.J. Acute and late toxicity associated with sequential bleomycin-containing chemotherapy regimens and radiation therapy in the treatment of carcinoma of the nasopharynx. Int. J. Radiat. Oncol., Biol. Phys. 14: 623-633, 1988.
14 Philips, T. L., Ross, G. Y., Goldstein, L. S. and Begg, A. C. The interaction of radiation and bleomycin in intestinal crypt cells. Int. J. Radiat. Oncol., Biol. Phys. 5: 1509-1512, 1979. 15 Redpath, J. L. Mechanisms in combination therapy: isobologram analysis and sequencing. Int. J. Radiat. Biol. 38: 355-356, 1980. 16 Sakamoto, K. The effect of bleomycin and its combined effect with radiation on cultured Chinese hamster cells V79. Eur. J. Cancer 14: 309-313, 1978. 17 Steel, G. G. and Peckham, M . J . Exploitable mechanisms in combined radiotherapy-chemotherapy: the concept of additivity. Int. J. Radiat. Oncol., Biol. Phys. 5: 585-591, 1979. 18 Stell, P. M. and Rawson, N. S.B. Adjuvant chemotherapy in head and neck cancer. Br. J. Cancer 61: 779-787, 1990. 19 Takabe, Y., Katsumata, T., Watanabe, M. and Terasima, T. Lethal effect of bleomycin on cultured mouse L cells: comparison between fractionated and continuous treatment. Gann 63: 645-646, 1972. 20 Takabe, Y., Miyamoto, T., Watanabe, M. and Terasima, T. Synergism of X-rays and bleomycin on Ehrlich ascites tumour cells. Br. J. Cancer 36: 391-395, 1977. 21 Terasima, T., Takabe, Y., Katsumata, T., Watanabe, M. and Umezawa, H. Effect of bleomycin on mammalian cell survival. J. Natl. Cancer Inst. 49: 1093-1100, 1972. 22 Terasima, T., Takabe, Y. and Yasukawa, M. Combined effect of X-ray and bleomycin on cultured mammalian cells. Gann 66: 701-703, 1975. 23 Vanuytsel, L., Feng, Y., Landuyt, W., Leer, J. W. and Van der Schueren, E. The combined effect of bleomycin and irradiation on mouse lip mucosa. 2. Influence on the accumulation and repair of sublethal damage during fractionated irradiation. Radiother. Oncol. 6: 267-273, 1986. 24 Vanuytsel, L. Feng, Y., Landuyt, W., Thames Jr., H. D. and Van der Schueren, E. Effect of continuous infusion of bleomycin on the time for completion of repair of sublethal radiation damage in mouse lip mucosa during fractionated irradiation. Radiother. Oncol. 12: 333-337, 1988 25 Vanuytsel, L., Feng, Y., Landuyt, W. and Van der Schueren, E. Effect of continuous infusion of bleomycin on repopulation in mouse lip mucosa during fractionated irradiation. NCI Monogr. 6: 55-58, 1988. 26 Von der Maase, H. Effect of cancer chemotherapeutic drugs on the radiation-induced skin reactions in mouse feet. Br. J. Radiol. 57: 697-707, 1984. 27 Von der Maase, H. Experimental studies on interactions of radiation and cancer chemotherapeutic drugs in normal tissues and a solid tumour. Radiother. Oncol. 7: 47-68, 1986. 28 Xu, F. -X., Van der Schueren, E. and Ang, K.K. Acute reactions of the mouse lip mucosa of mice to fractionated irradiations. Radiother. Oncol. i: 369-374, 1984. 29 Zylicz, Z., Hofs, H. P., Wagener, T., Van Rennes, H., Wessels, J. M. C., Van den Block, L. A. G. M. and Ottenheim, H. C. J. Modulation of the in vitro cytotoxicity of seven anticancer drugs by protein synthesis inhibition using sparsomycin. Anticancer Res. 9: 1835-1840, 1989.
237
© 1993 Elsevier Scientific Publishers Ireland Ltd. All rights reserved. 0167-8140/93/$06.00 RADION 01159
Response of mouse tongue epithelium to single doses of bleomycin and radiation Wolfgang D6rr, Erwin Hirler and Michaela H6nig GSF-lnstitut .t~r Strahlenbiologie, D-85764 Neuherberg, Germany
(Received 9 October 1992; revision received 14 January 1993; accepted 26 January 1993)
Key words: Mouse tongue mucosa; Bleomycin; Radiation; Interactions
Summary Both bleomycin (BLM) and local X-irradiation (25 kV) induce denudation in the tongue epithelium of the C3H-Neuherberg mouse in a dose-dependent manner. In the present study the effect of BLM alone and of combined single doses of drug and radiation were studied using the incidence of epithelial denudation as the end-point. In 'time-line' experiments, 8 mg/kg BLM were given before or after graded doses of X-rays. BLM treatment required a reduction of the radiation dose (EDs0) from 15 Gy to 5-7 Gy, independent of sequence or time interval. In contrast, the time course of the response was clearly dependent on the treatment interval. Latency decreased when the drug was injected less than 2 h before irradiation with minimum latency observed at 30 min. Isobologram analysis of experiments with varying combinations of X-rays and BLM demonstrated that small drug doses were relatively more effective than larger doses, suggesting an upward concavity of the BLM dose-effect curve in vivo, i.e. a 'negative shoulder' of the curve in the low dose region. In contrast to the response to X-rays alone, which has a constant latent time to ulcer of 10 days, the latency in combined treatment was clearly shortened with increasing drug dose and at high doses eventually approximated the epithelial turnover time of 5 days. The data suggest that BLM both as a single agent and in combination with X-rays reduced the probability of abortive divisions and through this effect shortened the latent time to epithelial denudation.
Introduction Squamous cell carcinoma appear to be highly sensitive to bleomycin (BLM), which for this reason has been introduced in a number of radiochemotherapy protocols for head and neck cancer [2,7,18]. Whenever bleomycin was added to existing treatment protocols, a substantial increase in the acute toxicity to oral mucosal membranes was observed. The effect was most pronounced when radiation and drug were given in close temporal proximity [7,13]. The basic mechanisms underlying the increased toxicity, however, remained unclear. Effects on capacity and kinetics of radiation-induced sublethal damage repair seemed unlikely, but could not be excluded from studies in the mouse lip mucosa [6,24]. In the same model it could be demonstrated that bleomycin was highly effective in suppressing repopulation in con-
tinuously fractionated [25] and split-dose experiments [23]. This effect was attributed to independent cell killing. However, effects on repair of sublethal X-ray damage could not be excluded. As an alternative to the mouse lip system we have established mouse tongue epithelium as a model to investigate radiation effects on oral mucosa [3,4,11]. Local irradiation of the tongue by soft X-rays induces ulcerative lesions of the mucosa in a dose-dependent manner. In the same mouse strain (C3H-Neuherberg), BLM alone provokes epithelial lesions qualitatively similar to radiation-induced ulceration. This marked sensitivity can be exploited for detailed studies on the in vivo effect of combined treatment with drug and radiation. The present paper reports on the effect of BLM alone and the combination of single doses of drug and radia-
Correspondence to: Wolfgang Drrr, GSF-Institut fiir Strahlenbiologie, Neuherberg, Postfach 1129, D-85758 Oberschleissheim, Germany.
238 tion. The influence of sequence and time interval in the application of both agents was investigated in a 'timeline' study. In a second set of experiments the effect of various single-dose combinations of BLM and X-rays were determined using isobologram analysis [15,17]. M a t e r i a l and m e t h o d s
Female mice of the inbred C3H-Neuherberg strain, 10-12 weeks old, were used in all experiments. The animals were housed under specified pathogen free conditions with controlled humidity and temperature. A 12/12 h light-dark cycle was maintained changing at 06:00/18:00 h. Standard pellet diet (Altromin®, Altrogge, Lage, Germany) and water were given ad libitum. Bleomycin (Bleomycinum Mack ®, Mack, Illertissen, Germany) was dissolved in 0.9% NaCI and diluted to concentrations that provided injection volumes (i.p.) of 0.1-0.5 ml per mouse.
Irradiation technique The animals were anaesthetized for irradiation using either hexobarbital sodium (Evipan®, Bayer, Leverkusen, Germany) at a dose of - 150 mg/kg i.p. in the isobologram experiments, or pentobarbital sodium (Nembutal®, WdT, Hannover, Germany) at a dose of - 60 mg/kg i.p. in the time-line experiments. Single radiation doses were applied to a defined area of the inferior tongue surface using a technique recently described in detail [3]. In brief, the anaesthetized animals were placed in the cylindrical bore of a prewarmed (35°C) aluminium block in a supine position with the head supported by a polystyrene wedge, and the tongue was gently pulled out through a small hole in the roof of the block by means of a forceps. The tongue was inflicted cranially and fixed with adhesive tape to expose its ventral surface to a vertical X-ray beam. An aluminium shield (1 mm) with a window of 3 x 3 mm was centrally positioned on the tongue to define the treatment area. The X-rays were generated by a Philips PWll30 machine operated at 25 kV with a beam filter of 0.3 mm AI. When X-ray doses smaller than 4 Gy had to be applied, a shutter (3 mm AI) was used which was opened when the tube voltage was constant at 25 kV. The source-to-surface distance was 15.5 cm. The dose rate at the tongue surface varied between 4.8 and 9.5 Gy/min, resulting in treatment times of 0.6-2 min. An ionization chamber (M23342, PTW Freiburg, Germany) was positioned close to the tongue at identical source distance and connected with a Dosimentor SN4 (PTW Freiburg) for dosimetry in all irradiation sessions.
Scoring Mucosal reactions were scored daily from the occurrence of first clinical symptoms until complete reepithelialization. Following immobilization with either hexobarbital sodium or pentobarbital sodium the tongue was pulled out carefully by means of a forceps and examined under a magnifying lens (3 x ) and a cold light source. After a lag period of 2-7 days, dependent on the experimental design, slight oedema and erythema developed, followed by increased desquamation that progressed to an erosive lesion after sufficient dose levels. Radiation-induced denudation was restricted to the exposed area at the inferior surface. In contrast, high BLM doses yielded ulceration of the whole tongue. The reaction assessed for the dose-response analyses was the incidence of denudation/shallow ulceration within the irradiation test area in order to achieve comparable resuits with both agents. The incidence of this effect was strictly dependent on radiation and/or drug dose. Doseeffect curves were fitted by probit analysis and the EDs0 doses and 95% confidence intervals calculated. At least seven animals were used per dose point.
Experimental design Time-line experiments To study the effect of sequence and time interval between both agents, BLM was applied 24 h, 6 h, 2 h or 30 min before and after irradiation. Furthermore, simultaneous application was tested, i.e. the drug was given immediately (< 30 s) before or after irradiation. The minimum interval between application of drug and anaesthetic was 20 min. The time of irradiation in this set of experiments was kept constant between 09:00 h and 12:00 h. Histological samples of the mobile tongue were taken from two animals each on Days 2, 4 and 6 after 22 mg/kg BLM without irradiation. In addition, samples were taken every second day between Days 2 and 10 after BLM application 6 h and 0.5 h before and 6 h after irradiation with 12 Gy. The specimens were fixed in formalin-alcohol and embedded in methylmethacrylate. Sagittal sections of 3/~m were obtained at several levels and stained with haematoxilin and eosin.
Isobologram study To investigate the effect of combined treatment with varying drug and radiation doses, BLM at nine dose levels ranging from 0 to 16 mg/kg was injected 2 h before irradiation. Graded X-ray doses were applied to generate full dose-response curves of complete denudation for each drug dose level. The time of irradiation was kept constant between 10:00 and 12:00 h in these experiments.
239
Results
Effect of X-rays alone Single dose irradiation with soft X-rays resulted in the first signs of mucositis, i.e. diffuse erythema and oedema, on Day 6-7. The symptoms progressed into marked desquamation, producing a scaly surface within the irradiated area after another 1 or 2 days. The fall-off in dose through the tongue was sufficient to avoid clinical damage on the upper tongue surface. The maximum response presented as a complete denudation or ulcerative lesion in the irradiated area that was usually covered by a pseudomembrane. The individual ulcer duration was 3 4- 1 days with hexobarbital anaesthesia and 4 4- 1 days with pentobarbital anaesthesia used for irradiation. Subsequently, rapid regeneration led to complete healing by Day 16 at the latest. The frequency of animals developing erosions displayed a stringent dose relationship, yielding a sigmoid dose-response curve as shown in Fig. 1A. The EDs0 to irradiation with hexobarbital anaesthesia was 15.0 Gy with 95% confidence limits of 13.5 and 16.3 Gy. Irradiation under pentobarbital anaesthesia resulted in an EDs0 of 13.6 Gy (11.6; 15.1). In contrast to ulcer frequency, the time course of the reaction was independent of the radiation dose including the highest dose level used (22 Gy), as demonstrated in Fig. 1B. The total latency to complete denudation after irradiation was 10.9 4- 1 days, being similar with both anaesthetics used. Maximum prevalence over all dose groups was observed on Day 12. As reported recently [3], only small fluctuations of body weight were observed after local tongue irradiation, independent of radiation dose and ulcer frequency.
100
•
i
•
i
•
(3."e f
r'"l I--II
/ /
(A)
80
U
This is good evidence that irradiation alone was well tolerated and failed to cause even temporary morbidity. The histological changes following single-dose irradiation of mouse tongue have recently been described in detail [4].
Effects of bleomycin alone Bleomycin alone is able to induce ulcerative lesions in the tongue of the C3H-Neuherberg mouse. The clinical symptoms, i.e. erythema, oedema, increased desquamation and epithelial erosions, were qualitatively similar to the changes induced by radiation. However, in addition to the response of the inferior surface, reactions were observed frequently on the tip and edges of the tongue and partly at the superior surface. To achieve response data comparable to irradiation, scoring was restricted to the site of the irradiation field, i.e. the central portion of the lower tongue surface. The latency to ulceration after BLM, in contrast to the response to X-rays, was clearly dependent on the drug dose. The mean latent times (± SEM) after 14, 16, 18, 20 and 22 mg/kg were 9.5 4- 0.5 days (n = 2), 9.0 4- 0.5 days (n = 4), 8.3 ± 0.3 days (n=6), 7.3 4- 0.4 days ( n = 6 ) and 6.7 4- 0.4 days (n = 7), respectively. The individual ulcer duration (4SD) was 3.4 4- 1.2 days being independent of the drug dose. The frequency of animals developing erosions of the lower tongue surface again yielded a well-defined dose-response curve. The EDs0, calculated by probit analysis, was 17.5 mg/kg (95% confidence limits, 16.2 and 18.8 mg/kg). In contrast to X-rays, some weight loss was observed after BLM with an indication of some dose dependence. Maximum weight loss was 8%. The qualitative histological changes in the tongue epithelium after BLM were similar to those observed after
i
I
I
Ce)
•
.
O
r¢t
I
v"
•
6o
I
¥
I
•
•
,
"1
40 U
,,. 7",: \
°/Z
20
,,f/"
?/J ",
0 8
12 X-roy
16 d o s e [Gy]
2O
8
. 10
Doys
,
. ",
12 ofter
14
.~\r 16
SO
Fig. 1. A, Frequency of animals developing tongue erosions after local single-dose irradiation under pentobarbital (O) or hexobarbital (O) anaesthesia. Dose-response curves were fitted by probit analysis, error bars depict 95% confidence intervals. B, Time course of ulcer prevalence after single-dose irradiation (hexobarbital anaesthesia) with 14 (O), 16 (n), 18 ( 0 ) and 20 Gy (V).
240 TABLE I Results of time-line experiments. Time interval (h)
EDs0 (Gy)
95% Confidence limits (Gy)
Ulcer duration (days)
Latency (days)
-24 -6 -2 -0.5
5.3 6.3 6.7 6.4
2.9; 4.7; 5.2; 4.5;
3,5 3.0 2.3 3.5
lO.O (l.O) 10.9 (0.7) 8.0 (1.O) 5.6 (0.7)
-0 +0
5.6 6.5
2.0; 8.2 3.5; 8.5
4.0 (1.4) 3.6 (1.7)
7.4 (0.7) 7.1 (0.3)
+0.5 +2 +6 +24
6.6 7.0 7.4 6.3
4.4; 5.7; 5.4; 4.4;
3.4 2.7 3.2 3.3
lO.O (l.1) 10.7 (l.1) 10.4 (0.7) 9.8 (1.3)
7.0 7.9 8.3 8.3
8.7 8.3 9.9 8.1
(0.9) (1.2) (0.6) (0.9)
(1.2) (0.9) (1.2) (1.3)
A BLM dose of 8 mg/kg was applied 30 min to 24 h before (negative time interval) or after local tongue irradiation at Time 0. In addition, the drug was injected immediately before (-0) or after (+0) X-ray treatment. EDs0 values and confidence limits were calculated by probit analysis• Numbers in parentheses depict 1 SD.
X-irradiation, although they occurred earlier. Abnormal mitotic figures (chromosome clumping, multipolar spindles) and abnormal nuclei were already found by Day 2 after BLM (22 mg/kg) and culminated on Days 4-6. On Day 4 prominent enlargement of individual cells was observed. As a consequence of abnormal divisions, binucleate cells (10%) and multinucleate cells (6%) reached a maximum on Day 6 in non-denuded regions. Foci of complete denudation were observed by Days 6-8. The erosions were covered by a membrane consisting of cell detritus, keratin and fibrin.
mainder. Cell density within the irradiated area decreased tO 81% (drug before X-rays) and 84% (drug after X-rays) on Day 2, and to 64% and 59% on Day 4. First foci of denudation were found on Day 6, at the time when the average cell count in non-denuded regions had reached 37% and complete denudation was observed on Day 8. Multinucleate keratinocytes were first observed on Day 2 (< 1%) and their number subsequently increased to 6%, 9% and 17% on Days 4, 6 and 8, respectively, to reach a maximum count of 18% on Day 10. While the longer treatment intervals resulted in no change in the latency as compared with radiation
Time-line experiments In these experiments a constant dose of 8 mg/kg BLM was injected 0 to 24 h before or after irradiation or injected immediately (< 30 s) before or after the X-ray treatment. Graded X-ray doses were administered and the resulting dose-response curves were fitted by probit analysis• Table I summarizes the resulting EDs0 values and their 95% confidence limits, average duration of ulcer and the observed mean latent times• The EDs0 values displayed no systematical fluctuations with the variation of the time interval between both agents. The mean EDs0 from all experiments was 6.4 Gy (4.9; 7.9), yielding a dose enhancement factor (DEF = EDs0 radiation alone/EDs0 radiation plus BLM) of 2.13 for 8 mg/kg BLM. Also, the individual ulcer duration was independent of the treatment protocol, giving a mean duration of 3.2 ± 1.2 days. Histological changes after BLM application 6 h before or 6 h after irradiation (12 Gy) were confined to the irradiated area in about 50% of the animals but included also tip and margins of the tongue in the re-
•
.
i
.
.
.
.
i
.
.
.
.
i
.
.
.
.
i
.
.
.
.
i
•
i
-
.
12
10 "0
¢J
.E_
a
¢p o
6
._1
1 B/M b e f o r e [ BLM o f t e r irrodiotlon •
i
-20
.
.
.
.
i
-10
.
.
.
.
i
.
.
.
0
Time intervol
.
i
10
.
.
.
.
20
[hi
Fig. 2. Latent time of tongue ulcer after a single injection of BLM at various intervals before or after local single-dose irradiation. The time of irradiation was kept constant between 10:00 h and 12:00 h.
241 TABLE II Results of the isobologram study. BLM dose (mg/kg)
EDs0 (Gy)
95% Confidence limits (Gy)
Ulcer duration (days)
Latency (days)
0 1 3 5 7 10 12 14 16
15.0 10.1 6.6 6.3 6.2 5.8 4.5 1.5 0.7
13.5; 16.3 7.5; 11.9 5.5; 8.3 3.9; 7.8 4.8; 9.4 4.6; 7.0 3.1; 7.3 -- ; --0.4; 1.5
3.2 3.6 3.6 3.7 4.8 3.7 3.3 3.6 3.3
10.9 (1.0) 10.5 (2.1) 7.5 (1.1) 8.0 (1.4) 6.6 (1.2) 6.4 (1.0) 6.3 (1.1) 6.0 (1.3) 5.1 (1.4)
(1.2) (1.5) (1.1) (1.1) (1.6) (1.1) (1.0) (1.2) (0.8)
Various constant drug doses were combined with graded doses of X-rays. Irradiation was administered 2 h after BLM injection. Confidence limits of EDs0 values were calculated by probit analysis. Numbers in parentheses indicate 1 SD.
alone (10-11 days) a clear shortening was observed when BLM was applied at an interval of 2 h or less prior to radiation. The minimum latency of 5.6 days was found when the drug was applied 30 min before irradiation (Table I and Fig. 2). In accordance with the reduced latent time to ulcer, cell density declined to 50% already on Day 2 and dropped further to 25% on Day 4, when BLM was applied 30 min before X-irradiation. Complete denudation was reached by Day 6. Multinucleate keratinocyte counts were 6.5%, 15% and 18% on Days 2, 4 and 6, respectively.
and 10:00 h, to be followed by graded X-ray doses 2 h later. X-ray doses were chosen to generate full doseresponse curves that allowed the radiation EDs0 required for each BLM dose to be calculated. Table II summarizes the EDs0 values, mean ulcer durations and average latent times of these experiments. The dependence of the EDs0 of X-rays on the BLM dose is graphically presented in Fig. 3A. As the curve demonstrates, the EDs0 values declined in a non-linear manner with increasing BLM dose. The reduction in isoeffective radiation dose was clearly steeper at the lower BLM doses. As BLM per se is able to induce tongue ulcer at doses above 14 mg/kg, the X-ray doseresponse curves above this threshold dose did not include low response frequencies and hence were less well defined, which is expressed in the lacking confidence limits. The DEF (EDs0 radiation alone/ED50 radiation
Combination of varying drug and radiation doses ( isobologram studies) In the isobologram studies nine dose levels of BLM, varying from 0 to 16 mg/kg, were given between 08:00 h
12
15
.-o, o u3 a
t0
E
tlJ 0
I
10
".7,
8
~
e
0 _J
X
i
0
. . . .
|
5
. . . .
|
10
. . . .
i
. . . .
15
BLM c l o s e [ m g / k g ]
20
0
5 BLM d o s e
10
15
[mg/kg]
Fig. 3. A, X-ray single-dose resulting in erosions in 50% of the animals as a function of the BLM dose applied 2 h before irradiation. Included is the result of BLM given alone (O). B, Latency to ulceration as a function of the drug dose applied 2 h before local irradiation. X-rays alone (11-20 Gy) result in a constant latency of 10.9 -,- 1 days. At high drug doses, the latency approximates the turnover time (TT) of the tissue.
242 plus BLM) varied between 1.49 (1 mg/kg) and 21.4 (16 mg/kg). The individual ulcer persistence was independent of drug dose. However, in contrast to the picture after radiation alone, the latency to ulceration decreased consistently with increasing BLM dose (Fig. 3B). Drug doses higher than 14 mg/kg resulted in latent times of 5-6 days compared with about 11 days after irradiation alone. The upward concave shape of the curve in Fig. 3B illustrates that low drug doses were again relatively more effective than higher doses for this end-point also. Discussion Clinical experience shows that BLM substantially enhances the acute mucosal response when added to existing radiochemotherapy protocols [7,15]. This increase in toxicity may be based either on independent cell kill or on direct interaction of both agents. In vitro studies demonstrated an increase in radiosensitivity of the target cell population by partial cell cycle synchronization by the drug [8], or an increase in cellular radiosensitivity, mainly by impairment of repair of sublethal damage [20,22]. However, in vivo studies into the functional response are required to investigate the effect of the combined treatment at the tissue level. The acute radiation response of oral epithelium is caused by the sterilization of (stem) cells in the germinal layer. Their proliferative failure translates into a lack of cellular supply to the functional layers and, in the presence of ongoing differentiation and sloughing off of cells or keratin at the surface, results in progressing tissue hypoplasia and eventual denudation. For experimental studies into the problem of acute radiation-induced mucositis the mouse lip mucosa assay was introduced [12,18]. Possible effects of BLM on the capacity and kinetics of repair of sublethal damage have been investigated with this model [6,23]. In combination with single-dose irradiation, BLM increased the mucosal response when administered 4 days or less before the X-ray treatment, with a maximum effect at 2 h. In addition, the response in these experiments occurred significantly earlier than the response to X-rays alone. When continuous drug infusion (7 days) was combined with X-rays of increasing fraction number, the DEF progressively increased. The experimental resuits could plausibly be explained by reduced repair of sublethal damage in addition to independent cytotoxicity. Repopulation in mouse lip mucosa, tested in the interval between split doses was clearly suppressed by BLM [25]. As an alternative animal system for radiobiological studies into the problem of acute oral mucositis the mouse tongue model was established [3,4,5,11]. The response to radiation in this assay is quantified by the frequency of animals that develop complete mucosal
denudation in a defined test area. This end-point obviously depends on stem-cell depletion below a critical level and hence allows extrapolation to cellular responses. In addition, the time course of the reaction gives interesting clues to population kinetics. Assuming both a total proliferative block and a normal differentiation rate after irradiation, the transit time of cells, i.e. the postmitotic life-span of the cells or the time of migration from the basal to the most superficial cell layer, should determine the latent time [9]. In steady state tissues the transit time equals the turnover time of the nucleated cells. Yet, the normal cell turnover time in mouse tongue mucosa has been measured to be about 5 days [4] and thus is considerably shorter than the latent time to complete denudation. The discrepancy can be explained by the fact that abortive cell divisions take place for a longer period, eventually resulting in mostly abnormal (bi- and multinucleate) cells. No influence of the X-ray dose on latency was observed in the present study, suggesting that the division probability and functional performance of abortive daughters is similar over the dose range used. Bleomycin per se induced erosions in tongue epithelium of the C3H-Neuherberg strain. Histologically and in their clinical appearance the mucosal ulcerations were similar to those observed after irradiation; however, in contrast to the response to radiation alone, the latent time to ulceration was dependent on the drug dose. Even small BLM doses significantly shortened the time to ulceration as compared with the latency seen after single-dose irradiation. After high doses of BLM, the time to complete denudation approximated the tissue turnover time of 5 days. The kinetics of developing damage was studied after a BLM dose of 22 mg/kg. This dose (ED95) should induce a cell kill comparable with that obtained by a single radiation dose of 20 Gy (ED95), which has been studied previously [4]. Abnormal mitotic figures after a high dose of BLM culminated already on Days 4-6 as compared with Days 6-8 after the high radiation dose [4]. In consequence, a maximum number of multinucleate keratinocytes was found on Day 6 ( - 16%) after BLM and on Days 7-9 (-40%) after X-rays. Between Days 2 and 4, the efficacy of abortive divisions was similar after both agents, resulting in about 16-20% of multinucleate cells. Their number subsequently further increased after X-rays, but clearly decreased after BLM. A plausible explanation of the discrepancies in latent time and time pattern of histological damage observed between BLM and single-dose irradiation may be that BLM decreases the average number of abortive divisions of doomed cells and leads to earlier occurrence of abnormal mitoses. This should result in a shortening of the latency to complete denudation, but in qualitative changes similar to those after irradiation. No systematical influence of sequence or time interval
243
on the sensitivity of mouse tongue epithelium to combined treatment was evident in the present study. In contrast, enhancement of drug toxicity has been described in the mouse lip model for BLM injection 2 h before irradiation as well as for a 7 days continuous infusion of BLM prior to irradiation [6]. Maximum effects of BLM were also found in rodent skin when the drug was given 15 min to 1 h prior to irradiation [10,26], or 1-6 h before X-rays in intestinal crypts [14,27]. In a mouse tumour model, large fluctuations in tumour growth delay with varying time interval and sequence were observed when BLM was given in intervals from 48 h before to 48 h after irradiation [1]. However, the growth delays in these experiments at most time intervals were equal or less than for X-rays alone. While in our studies the time interval between drug and radiation had no effect on the incidence of ulcer, a sharp drop in the latent time was observed when BLM was administered either at intervals shorter than 2 h before or immediately after irradiation. Injection 0.5 h before irradiation resulted in a latent time of 5.6 days, approximating the normal turnover time of the epithelium [4]. Shortening of the latent time of the response has also been reported for mouse lip mucosa [6] and for mouse skin [261. The histological studies illustrate that the most marked decline in cell density is observed when irradiation is administered 30 min post injection. In addition, the number of multinucleate cells, serving as a measure of effective though abnormal abortive divisions, increases 2 days earlier than with irradiation given within 6 h before or after drug injection. This suggests that a short time interval between BLM and X-rays, similar to high BLM doses alone, decreases the overall number of functional abortive divisions. Thus, the total amount of production of cellular material is decreased after the short time interval, leading to a clear shortening of the latent time. Assuming simple additivity of the cytotoxic effects of both agents, the isobologram curve can be used to derive a tentative cell survival curve to BLM. This is facilitated by the known general shape of the X-ray survival curve (Fig. 4), based on linear quadratic analysis of split-dose experiments in mouse tongue which yielded an or/3 value of 11.6 Gy [5]. The full effect (1.0) in Fig. 4 is the amount of cell kill required to induce ulcer in 50% of the tongues, corresponding to 15 Gy X-rays alone or 17.5 mg/kg BLM alone. In combined treatment studies, the full effect is composed of the known partial effect of Xrays plus the partial effect of BLM. For example, the proportion of the full effect of a given X-ray dose of 6.6 Gy is 0.3. The corresponding BLM dose of 3 mg/kg (Table II) thus has to induce seven-tenth of the full damage (1.0) necessary to result in 50% ulcer. The effect of BLM resulting from the calculations for all dose pairs is shown in Fig. 4, illustrating that the the-
BLM dose [mg/kg] 0
5
I
r--1
20
,
,
.
,
,
10
I
,
o
=
=
I
15 .
.
.
.
I
•
•
\
• o
,
\
X-rays
".T
,, =/p=11.8 oy
40
ID 0
6O
~
8O
100
II
"'" .....
BLM
I
.
.
0
.
.
I
5
.
.
.
.
I
.
.
.
10
.
I
15
X-ray dose [Gy] Fig. 4. Quasi-survival curves to X-rays and BLM, plotting proportion of full effect (50% responders) versus dose. The X-ray curve was established by the linear-quadratic model using an a/~ value of 11.6 Gy [5]. The dose-response curve to BLM was subsequently generated by assuming pure additivity of drug and radiation effects. Thus, to obtain the partial effect of a given BLM dose in combined treatment (according to Table II), the partial effect of the complementary X-ray EDs0 was subtracted from 1.0. Similarly, error bars were obtained by using 95% confidence limits of the X-ray EDs0 doses.
oretical cell survival curve to BLM in vivo has an even more pronounced upward concavity than the combined dose-effect curve. Similarly shaped dose-effect curves have been derived from in vitro experiments using BLM alone [19,21,29] or in combination with X-rays [16]. As a consequence, repeated injections of BLM at low doses may cause repetition of the steep segment of the cell survival curve and thus confer increased mucosal toxicity. Such an enhanced efficacy by fractionating a total BLM dose has indeed been observed in independent experiments on mouse tongue mucosa (Drrr et al., unpublished data). In summary, two interesting effects of bleomycin on oral mucosa in mice could be demonstrated in the present study. The dose- and time-dependent shortening of the latency to ulceration is probably caused by a decrease in the amount and efficacy of abortive cell divisions. The relatively higher effect of small drug doses suggests an upward concavity of the in vivo doseresponse curve to the drug, which deserves further experimental investigation because of its obvious clinical implications.
Acknowledgements This study was in part supported by the Deutsche Forschungsgemeinschaft, grants number Ku576/2-1
244
and Ku576/2-2. All experiments were carried out according to the valid legislation with the permission of the Regierung von Oberbayern. The authors are grateful to Dr. J. Kummermehr for critical reading of and helpful comments on the manuscript. References 1
2
3
4
5
6
7
8
9
10
11 12
13
Begg, A. C., Kane, L. J., Fu, K. K. and Philips, T.L. The effects of combined bleomycin and irradiation on a solid murine tumor. Int. J. Radiat. Oncol., Biol. Phys. 5: 1503-1507, 1979. Cachin, Y., Jortay, A., Sancho, H., Eschwege, F., Madelian, M. Desaulty, A. and Gerard, P. Preliminary results of a randomized E.O.R.T.C. study comparing radiotherapy and concommirant bleomycin to radiotherapy alone in epidermoid carcinomas of the oropharynx. Eur. J. Cancer 13: 1389-1395, 1977. D6rr, W. and Kummermehr, J. Accelerated repopulation of mouse tongue epithelium during fractionated irradiations or following single doses. Radiother. Oncol. 17: 249-259, 1990. D6rr, W. and Kummermehr, J. Proliferation kinetics of mouse tongue epithelium under normal conditions and following single dose irradiation. Virchows Arch. B 60: 287-294, 1991. D6rr, W., Breitner, A. and Kummermehr, J. Capacity and kinetics of SLD repair in mouse tongue epithelium. Radiother. Oncol. 27: 36-45, 1993. Feng, Y,, Vanuytsel, L., Landuyt, W., Ang, K. K. and Van der Schueren, E. The combined effect of bleomycin and irradiation on mouse lip mucosa. I. Influence of timing, sequence and mode of drug administration with single dose irradiation. Radiother. Oncol. 6: 143-151, 1986. Fu, K. K., Philips, T. L., Silverberg, I. J., Jacobs, C., Goffinet, D. R., Chun, C., Friedman, M. A., Kohler, M., McWirter, K. and Carter, S.K. Combined radiotherapy and chemotherapy with bleomycin and methotrexate for advanced inoperable head and neck. J. Clin. Oncol. 5: 1410-1418, 1987. Kimler, B. F. The effect of bleomycin and irradiation on G 2 progression. Int. J. Radiat. Oncol., Biol. Phys. 5: 1523-1526, 1979. Michalowski, A. Effects of radiation on normal tissues: hypothetical mechanisms and limitations of in situ assays of clonogenicity. Radiat. Environ. Biophys. 19: 157-172, 1981. Molin, J., Sogaard, P. E. and Overgaard, J. Experimental studies on the radiation-modifying effect of bleomycin in malignant and normal mouse tissue in vivo. Cancer Treat. Rep. 65: 583-589, 1981. Moses, R. and Kummermehr, J. Radiation response of mouse tongue mucosa. Br. J. Cancer 53 (Suppl. VII): 12-15, 1986. Parkins, C. S., Fowler, J. F. and Yu, S. A murine model of lip epidermal/mucosal reactions to X-irradiation. Radiother. Oncol. 1: 159-165, 1983. Peters, L. J., Harrison, M, L., Dimery, I. W., Fields, R., Goepfert, H. and Oswald, M.J. Acute and late toxicity associated with sequential bleomycin-containing chemotherapy regimens and radiation therapy in the treatment of carcinoma of the nasopharynx. Int. J. Radiat. Oncol., Biol. Phys. 14: 623-633, 1988.
14 Philips, T. L., Ross, G. Y., Goldstein, L. S. and Begg, A. C. The interaction of radiation and bleomycin in intestinal crypt cells. Int. J. Radiat. Oncol., Biol. Phys. 5: 1509-1512, 1979. 15 Redpath, J. L. Mechanisms in combination therapy: isobologram analysis and sequencing. Int. J. Radiat. Biol. 38: 355-356, 1980. 16 Sakamoto, K. The effect of bleomycin and its combined effect with radiation on cultured Chinese hamster cells V79. Eur. J. Cancer 14: 309-313, 1978. 17 Steel, G. G. and Peckham, M . J . Exploitable mechanisms in combined radiotherapy-chemotherapy: the concept of additivity. Int. J. Radiat. Oncol., Biol. Phys. 5: 585-591, 1979. 18 Stell, P. M. and Rawson, N. S.B. Adjuvant chemotherapy in head and neck cancer. Br. J. Cancer 61: 779-787, 1990. 19 Takabe, Y., Katsumata, T., Watanabe, M. and Terasima, T. Lethal effect of bleomycin on cultured mouse L cells: comparison between fractionated and continuous treatment. Gann 63: 645-646, 1972. 20 Takabe, Y., Miyamoto, T., Watanabe, M. and Terasima, T. Synergism of X-rays and bleomycin on Ehrlich ascites tumour cells. Br. J. Cancer 36: 391-395, 1977. 21 Terasima, T., Takabe, Y., Katsumata, T., Watanabe, M. and Umezawa, H. Effect of bleomycin on mammalian cell survival. J. Natl. Cancer Inst. 49: 1093-1100, 1972. 22 Terasima, T., Takabe, Y. and Yasukawa, M. Combined effect of X-ray and bleomycin on cultured mammalian cells. Gann 66: 701-703, 1975. 23 Vanuytsel, L., Feng, Y., Landuyt, W., Leer, J. W. and Van der Schueren, E. The combined effect of bleomycin and irradiation on mouse lip mucosa. 2. Influence on the accumulation and repair of sublethal damage during fractionated irradiation. Radiother. Oncol. 6: 267-273, 1986. 24 Vanuytsel, L. Feng, Y., Landuyt, W., Thames Jr., H. D. and Van der Schueren, E. Effect of continuous infusion of bleomycin on the time for completion of repair of sublethal radiation damage in mouse lip mucosa during fractionated irradiation. Radiother. Oncol. 12: 333-337, 1988 25 Vanuytsel, L., Feng, Y., Landuyt, W. and Van der Schueren, E. Effect of continuous infusion of bleomycin on repopulation in mouse lip mucosa during fractionated irradiation. NCI Monogr. 6: 55-58, 1988. 26 Von der Maase, H. Effect of cancer chemotherapeutic drugs on the radiation-induced skin reactions in mouse feet. Br. J. Radiol. 57: 697-707, 1984. 27 Von der Maase, H. Experimental studies on interactions of radiation and cancer chemotherapeutic drugs in normal tissues and a solid tumour. Radiother. Oncol. 7: 47-68, 1986. 28 Xu, F. -X., Van der Schueren, E. and Ang, K.K. Acute reactions of the mouse lip mucosa of mice to fractionated irradiations. Radiother. Oncol. i: 369-374, 1984. 29 Zylicz, Z., Hofs, H. P., Wagener, T., Van Rennes, H., Wessels, J. M. C., Van den Block, L. A. G. M. and Ottenheim, H. C. J. Modulation of the in vitro cytotoxicity of seven anticancer drugs by protein synthesis inhibition using sparsomycin. Anticancer Res. 9: 1835-1840, 1989.