Influence of tumour volume and cell kinetics on the response of the solid Yoshida sarcoma to hyperthermia (42 degrees C).

The cytokinetic response of the solid Yoshida sarcoma to hyperthermia was examined at two tumour volumes, 1.0-1.5 ml and 3.0-3.5 ml. The tumour, growing on the feet of rats, was heated at 42 degrees C for 1 h by water-bath immersion. The larger tumour grew more slowly than the smaller one (doubling time 144 h vs 36 h) due to a halving in growth fraction from 67.8 to 39.6% and an increase in cell-loss factor from 59 to 75.9%. Cell cycle and phase times were similar at both volumes. The effect of heat on the population kinetics at both volumes was similar but complex, and involved delayed cell death after up to 10 mitoses. Initial cell killing and blockade of cell-cycle progression (0-24 h) was followed by recovery of proliferation due to recruitment of cells from the non-proliferative compartment, cell cycle and phase times remaining unaltered. From 48 h, the proliferation rate declined progressively, and tumours were completely necrotic 7-8 days after heat. The damaging effects of heat were at least as severe in the large tumours with a low labelling index and small growth fraction as in the smaller tumours with a much larger compartment of proliferating cells and shorter doubling time. The results imply that there may be no simple relationship between proliferative status and thermosensitivity of a tumour, and illustrate the difficulty in predicting tumour response to heat on the basis of cytokinetic studies.


volumes.
The effect of heat on the population kinetics at both volumes was similar but complex, and involved delayed cell death after up to 10 mitoses. Initial cell killing and blockade of cell-cycle progression (0-24 h) was followed by recovery of proliferation due to recruitment of cells from the non-proliferative compartment, cell cycle and phase times remaining unaltered. From 48 h, the proliferation rate declined progressively, and tumours were completely necrotic 7-8 days after heat. The damaging effects of heat were at least as severe in the large tumours with a low labelling index and small growth fraction as in the smaller tumours with a much larger compartment of proliferating cells and shorter doubling time.
The results imply that there may be no simple relationship between proliferative status and thermosensitivity of a tumour, and illustrate the difficulty in predicting tumour response to heat on the basis of cytokinetic studies. THE USE of hyperthermia, temperatures above the normal physiological range, for the treatment of cancer is currently receiving renewed interest. A number of different types of cancer cells have been reported to be more heat sensitive than normal cells in vitro, and hyperthermia has been shown to destroy a variety of tumours growing in animals and in man (Wizenberg & Robinson, 1976;Rossi-Fanelli et al., 1977;Streffer, 1978;Dickson, 1978). However, a broader knowledge of the mode of action and optimal means of heat application may be required before the full potential of thermotherapy is realized.
In recent years, the use of both chemotherapy and radiotherapy has been given a more rational basis by applying knowledge of the kinetics of cell populations under normal and therapeutic conditions (Hill & Baserga, 1975;Steel, 1977;Hill, 1978).
Work in vitro has indicated that the effects of hyperthermia may be preferentially phase and cycle specific (Westra & Dewey, 1971;Palzer & Heidelberger, 1973b;Bhuyan et al., 1977) and the synergism between hyperthermia and X-irradiation in the killing of malignant cells both in vivo and in vitro appears to be at least partially explicable in terms of cytokinetic factors (Thrall et al., 1976). Inadequate heating of the Yoshida rat sarcoma (in terms of tumour temperature and/or heating time relative to tumour volume) increased metastasis and suggested an influence of population kinetics on the outcome of hyperthermia in vivo RESPONSE OF YOSHIDA TUMOIJR TO 42 C (Dickson & Ellis, 1974. Preliminary studies on the curative heating of the Yoshida sarcoma in vivo indicated cytokinetic changes in the cell population, with the implication of recruitment of nonproliferating cells into cycle prior to tumour destruction (Dickson & Calderwood, 1976). In the present study, the sensitivity of the Yoshida foot sarcoma in vivo to l h at 42TC has been investigated at 2 tumour volumes, with special reference to the different cell population-kinetic parameters and cure rates of the tumnours.

MATERIALS AND METHODS
The Yoshida sarconma is an undifferentiated tumour that arose in a rat after feeding with 0-aminoazotoluol and painting the skin with potassium arsenite (Stewart et al., 1959). The tumour has since b)een maintained in ascitic (Yoshida, 1971) and, as used in the present study, solid forms (Dickson & Suzangar, 1974) by serial implantation in outbred albino rats.
For the present work, the tunour was maintained by serial passage in the thigh muscles of outbred Wistar rats weighing approx. 200 g (Dickson & Suzangar, 1974) and for experimental work the tumour was grown from a s.c. inoculum of 100 mg tumour homogenate in the dorsum of the left hind foot. In this location, the tumour grew as a well-defined mass, the volume of which was best approximated by the formula for an oblate sphere (V= 1/6 7r a2b, wrhere a and b are the major and minor axes respectively). Volume was calculated from caliper measureinents in the antero-posterior and vertical planes, allowance being made for the thickness of the normal tissues of the animal's foot.
Hyperthermia.-For heat treatment of tumours, rats were anaesthetized with i.p. pentobarbitone sodium (Sagatal: May & Baker Ltd, Dagenham). Sagatal was given at a dose of 24 mg/kg body weight (0.1 ml of a 12 mg/ml solution of Sagatal in 0-90o NaCl/50 g rat weight). Temperature monitoring probes were then placed in the tumour and hyperthermia was applied by water-bath immersion. Temperature was measured to +0-1°C at 10 min intervals by means of a Light multiprobe 12-channel direct-reading electric thermometer with a scale range of 36-460C. The instrument had a fast response time of 4 s, recorded temperature with an accuracy of + 0 05C, and was unaffected by changes in ambient temperature.
For intra-tumour and intra-abdominal temperature measurement, the thermistor probes were 5 cm long, needle type 1H, 0-88 mm in diameter, recording temperature only at the needle tip. Polythene-covered probes were used for rectal and water-bath temperature measurement. The intra-tumour sensor was inserted -1 0 cm into the foot tumour, along the line of the limb which acted as a splint, and the probe was immobilised by a non-restricting tape bandage round the leg. Each tumour had an indwelling thermistor during heating. It was established previously that the presence of a temperature probe in tumours (volume 1-10 ml) did not significantly alter the biological behaviour or response to heat of the Yoshida tumour (Dickson & Ellis, 1974;. Central body or "gcore" temperature was monitored by the intra-abdominal needle introduced to a depth of 2-5 cm below the liver in a right paramedian position. Core temperature was also measured by a rectal probe inserted 3 cm into the anus. Taping the rectal probe to the base of the tail prevented dislodgement of the sensor during heating. Before use, temperature sensors were calibrated against a mercury-in-glass thermometer of the National Standards Laboratory, Hemel Hempstead, Herts., England. The heating bath consisted of a perspex tank (33 x 33 x 15 cm) containing 10 1 water heated by a Circotherm Il constant-temperature unit with a 700-watt coil heater and circulating pump with an output of 12 1/min.
At an ambient temperature of 25°C, this unit maintained the bath temperature constant to + 005°C. The rat was placed on a perspex platform resting over the bath. and the tumour-bearing foot was immersed in the water through a 10cm diameter, padded opening. The foot was supported in the bath at a depth that permitted complete submersion of the tumour. Immediately after heat therapy, each rat was given 190 ml of 4%o dextrose in 0.18% NaCl to replace fluid loss.
The animal was then wrapped in a blanket and placed under an infra-red heater for 10-15 min; this helped to control the return of body temperature to normal without an overswing to subnormal temperature. which can occur rapidly in rats follow,Ning hyperthermia, (Dickson 1977 given i.p. to the rats in the following doses: animals to be killed for the percentage labelled mitoses (PLM) and lh (flash) labelling were given a single injection of 2 ,uCi/g body weight in 1 ml 0.9% NaCl; animals for repeated tumour-labelling experiments received 0 5 ,uCi/g [3H]-TdR every 6 h and a further 0-5 ,uCi/g 1 h before killing. Autoradiographs of 4,um paraffin-embedded hemisections of the tumours were prepared using the dipping technique (Baserga & Malamud, 1969) with Ilford K-2 emulsion (Ilford Ltd., Ilford, Essex). After 14 days exposure, the autoradiographs were processed with D-19 developer and F-5 fixer (Rogers, 1967) and stained with haematoxylin and eosin. For the PLM curves, 500 anaphases and metaphases were counted "blind", and for the flash and repeated labelling indices (% labelled cells) a total of 2,000 cells was scored.
To determine background counts in autoradiographs, a tumour-bearing animal was injected with [3H]-TdR and the tumour fixed 30 min later. This period is too brief for mitoses to become labelled (Lala, 1971). Therefore, any silver grains over such mitoses in an autoradiograph are due to background radiation. Sections of the tumour were prepared in large numbers and used as standards for processing with each batch of autoradiographs. Background in the majority of autoradiographs averaged less than 1 grain per cell and if this amount of labelling was exceeded the preparations were not used. Cells with 5 or more grains per cell were considered labelled.

PLM curves
The PLM curves of the smaller and larger Yoshida sarcoma are shown in Fig. 2. The TC was similar at both tumour volumes, being 14-1 h in the 1P0-1-5 ml tumour and 13-8 h in the 30-3-5 ml tumour (Table 1)  Kb=birth rate, measured as the rate of entry into mitosis usinig vincristine for mitotic arrest (Aherne et al., 1977). creased with increasing tumour volume from 52 5% in the 1-0-1-5 ml tumour to 28 2% in the 3 0-3 5 ml tumour. The birth rate measured by vincristine blockade (Fig. 3) decreased from 4-7 to 2.1% in the larger tumour (Table II).
The slower growth of the larger tumour compared with the 10-15 ml tumour appeared to be due to a decrease in growth fraction (from 67-8 to 39.6%) and an increase in cell-loss factor from 59 to 75.9%0 (Table II).
Effect of hyperthermia on cell-kinetic parameters Growth curve. The effect of 1 h at 42°C on tumour volume is shown in Fig. 1. Treatment of the tumour at 10-15 ml caused a decrease in tumour volume and complete regression within 14 days. There was a 96% cure rate of the animals. Treatment of the tumour at 3 0-3 5 ml produced a restraint in growth followed by partial regression, although variation in response between tumours was considerable (Fig.  lI). No tumour increased in volume after heat, however. Hyperthermia in rats bearing these larger tumours constituted a hazard to the host, leading to a significantly reduced lifespan of 27-4 + 5*5 days compared to 44-6 + 7-23 days in untreated tumour-bearing controls (P < 0 00 1). At autopsy, animals that died at 27 days showed enhanced spread of tumour locally and to distant sites; the role of heat in this enhanced dissemination has been discussed previously (Dickson, 1976).
The 1-0-1 5 ml tumour The effect of heating at 42°C on the [3H]-TdR LI is illustrated in Fig. 4a. Immediately after hyperthermia, LI was depressed from a control level of 52.5% to 5% labelled cells. Labelling remained inhibited and showed considerable variation between tumours in the period from 0 to 24 h, but recovered to near control levels by 48 h after heating. From -50% at 48 h, the LI declined to 35% at 72 h, 100 at 96 h and zero by 5 days after heat as the tumour regressed.
Changes in tumour-cell kinetics in the 48 h after heat were investigated in more detail by repeated [3H]-TdR labelling ( Fig. 4b). In controls LI increased to a plateau of almost 80% labelled cells after 8 h. After hyperthermia, LI was depressed to 5%0 labelled cells immediately after heat, and then recovered rapidly to a maximum of 60% after 8 h. An equally rapid decline in LI to a nadir of 5oo at 14 h then preceded an erratic recovery to a plateau of 60% labelled cells 36-48 h after hyperthermia. When repeated labelling was carried out from 0 to 14 h after heat and then ceased, the increase in LI from 14 to 48 h was small. When, however, labelling was commenced at 14 h after heat and continued to 48 h, LI recovered to the 60% region by 30 h.
The effect of 1 h at 42°C on entry of tumour cells into mitosis is shown in Fig. 5. Immediately after heat, the rate of entry into mitosis was lower and more variable than in controls (Figs. 5, 3a) Fig. 7b. In untreated controls, the LI increased from 28% at 1 h to a plateau of 55%0 by 18-24 h. Following hyperthermia, the LI varied between 6 and 8% from 0 to 24 h. After 24 h, a slow increase occurred to 32% at 48 h.

DISCUSSION
The decreased growth rate of the Yoshida sarcoma at the larger volume (Td 144 h at 3-5 ml compared to 36 h at 1.5 ml) resulted from a decreased growth . A RESPONSE OF YOSHIDA TUMOUR TO 420c fraction and an increase in cell-loss factor; the cell-cycle time of the tumour was unaltered (Tables). A similar finding has been reported previously in solid tumours, in which TC appears to be a relatively constant factor and alterations in growth rate are due mainly to changes in GF and 0 (Watson, 1976;Aherne et al., 1977;Steel, 1977). The median Tc of the Yoshida sarcoma remained a stable parameter in cells repopulating the tumour after hyperthermia (Fig. 6); the changes produced by curative heating concerned cell loss and an alteration in the relationship between the P and Q cell compartments (vide infra) rather than cell-generation times.
Alterations in the [3H]-TdR flash LI indicated a complex sequence of events in the tumour cell population after heating (Fig. 4a). The decreased LI immediately after hyperthermia reflected killing of cells in S phase and/or a 2-3 h block in the progression of cells into and through S phase. After release of the block, the partially synchronized cells appeared to enter S phase and cause a maximum in LI at 8 h (Fig. 4b). The decline in LI to a minimum at 14 h must be a consequence of further cell death. Thus, more than 90°0 of the cells proliferating after hyperthermia had been lost by 14 h, the median Tc of the tumour. This would imply that cells damaged at the time of heating had progressed through one cell cycle and died, possibly in mitosis. It seems unlikely that the rapid recovery of labelling from 14 to 36 h was caused by cells in cycle at the time of heat treatment. A more likely explanation is the entry into S phase of a population of cells not proliferating at the time of heat treatment; this would repopulate the tumour from 14 h onwards. Evidence in favour of this hypothesis (Fig. 4b) is that when repeated labelling was terminated 14 h after heating (the time of maximum cell death) there was only a small increase in subsequent labelling. When repeated labelling was commenced at 14 h, LI increased to near control values. These two experiments support the hypothesis that the recovery in proliferation from 14 h after heat was mainly due to recruitment of cells from the Q compartment of the tumour population. A similar effect was noted by Luieke-Huhle & Dertinger (1977) using V79 spheroids in vitro. Cells in the centre of the spheroids had a low proportion of P cells under normal conditions. Heating at 42°C for 4 h caused extensive loss from the population of cells with a high GF at the periphery of the spheroids, and appeared to stimulate the entry of Q cells in the centre of the spheroids into cycle. This caused an increase in the proportion of cells in S phase at the interior of the spheroids for 12-24 h after heating. Preferential destruction of P cells followed by recruitment of Q cells into cycle has been reported after radiation or chemotherapy to solid tumours, and the timing of recovery has been used to plan fractionated therapy regimens (see Steel (1977) for refs.).
The stathmokinetic study detailed in Fig. 5 gave supportive evidence for the inferences drawn from Fig. 4. The blockade of entry into mitosis immediately after heat paralleled the inhibition of entry into S phase at this time. This blocking effect of heat on progression round the cell cycle has previously been noted in vitro by Rao & Engelberg (1965) and Sisken et al. (1965) and more recently by Palzer & Heidelberger (1973a) Kase & Hahn (1975) Lucke-Htihle & Dertinger (1977 and Sapareto et al. (1978). The release of the mitotic block in the Yoshida sarcoma and entry into mitosis at 6 h (Fig. 5) again paralleled the entry of cells into S phase at this time (Fig. 4a). The high base-line mitotic index of 5%0 at 6 h (compared to 2.3% in controls) could indicate cells blocked in mitosis, possibly in the early stages of mitotic death. The decline in mitotic rate to almost 0 cells/h at 12 h and 18 h (Fig. 5) is consistent with destruction of the P cell population from 8-14 h after heating, as inferred from the repeated-labelling study. Recovery of the mitotic rate by 48 h ( [3H]-TdR studies (Figs 4a, b) that proliferation in the Yoshida sarcoma had regained control values by this time after heating. The progressive decline in mitotic rate after 48 h (Fig. 5) to zero at 6 days as the tumour regressed, followed a similar time course to the decline in flash LI (Fig.  4a). By Day 9 the tumour was totally necrotic, and mitoses and [3H]-TdR labelled cells were no longer seen in sections of such tumours. It is not clear to what extent the reduction in LI and mitotic rate from 2-9 days was due to inhibition of cell division and wastage by cell loss or to delayed cell killing due to a cytotoxic effect of heat. The mitotic rate and LI of cells in viable areas of the tumour 72 and 96 h after heat (Figs 4a,5) were considerably less than in controls, so a reduced rate of cell production was at least partially implicated in the destruction of the tumour-cell population 2-9 days after heat.
A similar pattern in cytokinetic response was seen in the 3 0-3*5 ml tumour after heating to that found in the 1 0-1-5 ml tumour (Figs 4a, 7a). However, in the 48 h immediately after hyperthermia, there were differences in response between the tumours, indicating that cells in the larger tumour may have suffered more damage. In the 3 0-3 5 ml tumours, labelling did not increase significantly until 48 h after hyperthermia, and there was no peak in LI at 6-8 h as in the 1-0-1*5 ml tumour (Figs 4b, 7b). It is apparent that the larger tumour with a smaller growth fraction (39.6% vs 67.8%, Table II) and a larger fraction of Q cells, was at least as heatsensitive as the smaller tumour. Thus, it would seem that only a fraction of the Q-cell population in the larger tumour was able to enter the cell cycle and repopulate the tumour. The Q-cell compartment of tumours is thought to be heterogeneous (Sarna, 1974;Gelfant, 1977) and may contain cells of differing heat sensitivity. Untreated 3 0-3*5 ml tumours contained large necrotic zones, probably bordered by Q cells remote from the tumour micro-circulation. Such cells would be deficient in 02 and nutrients, conditions which have been shown to sensitize cells to heat in vitro (Gerweck et al., 1974;Bass et al., 1978). The results imply that there is no simple relationship between the proliferative status and the thermosensitivity of tumour-cell populations in vivo.
In both tumour sizes, failure of cells surviving 48 h after heat treatment to maintain the growth of the tumour may be due to 3 mechanisms: (1) Failure to repair sublethal damage followed by delayed, heat-induced cell killing. The finding of proliferating cells in the tumour up to 6 days after curative heating confirms in vitro findings that several mitoses (up to 10 in Yoshida sarcoma) may occur before the expression of lethal hyperthermic damage (Palzer & Heidelberger, 1973a). The role of repair processes in hyperthermic cell damage has been discussed by Bronk (1976).
(2) Preferential eradication of clonogenic cells in the tumour and cell population decline due to cell loss.
(3) The operation of host factors in the destruction of the tumour. It has been demonstrated that regression of the Guerin carcinoma in the rat (Szmigielski & Janiak, 1978) and the VX2 carcinoma in the rabbit (Shah & Dickson, 1978) after local hyperthermia, are accompanied by stimulation of a host anti-tumour immune response. A recent review (Dickson, 1978) indicates that in both inbred and outbred animals (and also in man) immunogenic tumours are more readily cured by heat that nonimmunogenic tumours, and it has been reported that cure of the immunogenic MC7 sarcoma in rats and the non-immunogenic VX2 carcinoma in rabbits may be abrogated by immunosuppression of the host (Shah & Dickson, 1979). Little definitive information is available on the immunogenicity of the Yoshida sarcoma, although immune factors seem to be involved in cure of the tumour by chemotherapy (Fox & Gregory, 1972). In rats with 1.0-15 ml Yoshida sarcomas, metastatic tumour cells are present in the regional lymph nodes. Cure of such animals by heating the primary tumour for 1 h at 42TC, and subsequent resistance of the hosts to tumour inoculation, implies the generation of an anti-tumour response by hyperthermia (Dickson & Ellis, 1976). The two distinct exponential phases in the growth curve of the Yoshida sarcoma (Fig. 1) could be interpreted as the operation of anti-tumour immunity from 10 days after implantation. However, in tumours grown in the thigh muscles or s.c. in the flank, growth retardation did not occur until tumour volumes of 8-10 ml were attained, and the growth pattern was more Gompertz-like, with a smooth decrease in growth from an initial exponential phase. It is believed, therefore, that the volume curve of the tumour reflects the anatomical characteristics and functional restrictions of the foot as a site for growth rather than a host anti-tumour response.