RADIOBIOLOGICAL STUDIES OF RADIOIMMUNOTHERAPY AND EXTERNAL-BEAM RADIOTHERAPY IN-VITRO AND IN-VIVO IN HUMAN RENAL-CELL CARCINOMA XENOGRAFTS

BACKGROUND. Previous studies suggest that the radiobiologic characteri stics of in vitro survival curves are important determinants of the re sponse of tumors to both conventional radiotherapy and radioimmunother apy (RIT). The purpose of this study was to elucidate the relationship between in vitro radiation survival curve parameters and the relative sensitivity of tumor to RIT, exponentially decreasing low dose rate ( ED LDR) irradiation and conventional high dose rate (HDR) fractionated external beam radiotherapy. METHODS. Two human renal cell carcinoma c ell lines, Caki-1 and A498, were used in vitro and nude mouse xenograf t studies. HDR external beam gamma irradiation (dose rate, 430 centigr ay [cGy]/minute) and ED LDR irradiation (initial dose rate, 22-25 cGy/ hour) were performed with a cesium-137 (Cs-137) gamma irradiator. RIT was carried out with yttrium-90 (Y-90)-labeled monoclonal antibody NR- LU-10, and the absorbed radiation doses were calculated by medical int ernal radiation dose methodology. A clonogenic assay was used to gener ate radiation survival curves, and a computer FIT program was used to calculate the radiobiologic parameters. The antitumor efficacy of the different treatments was compared in vivo using a tumor regrowth delay assay in these two tumor xenograft models. RESULTS. The radiation sur vival curves showed that the Caki-1 cell line was more sensitive to bo th HDR and ED LDR irradiation than A498 in vitro. The Caki-1 cell line , compared with A498, had a larger alpha (0.39 vs. 0.15 Gy following H DR and 0.32 vs. 0.21 Gy following ED LDR) and alpha-to-beta ratio (6.9 2 vs. 2.60 Gy for HDR and 40.0 vs. 19.2 Gy for ED LDR), a smaller n nu mber (5.13 vs. 23 for HDR and 1.16 vs. 3.53 for ED LDR), a lower quasi -threshold dose (Dq) (1.60 vs. 3.15 Gy for HDR and 0.35 vs. 1.76 Gy fo r ED LDR), and a lower surviving fraction at 2 Gy (SF2) (0.37 vs. 0.60 for HDR and 0.51 vs. 0.61 for ED LDR), suggesting that Caki-1, compar ed with A498, had a steep initial slope and a small shoulder. The fina l slope represented by the beta value and D-0 dose (the dose (Gy) requ ired to reduce the fraction of surviving cells of 37% of its previous value in the exponential region of the survival curves) did not vary s ignificantly between these two cell lines at either HDR or ED LDR irra diation. Tumor volume doubling times were 4.0 +/- 1.5 days for Caki-1 and 4.2 +/- 1.8 days for A498 tumor xenografts. One hundred mu Ci/50 m u g of Y-90-labeled, isotype-matched irrelevant monoclonal antibody CC OO16-3 produced a tumor growth delay time (TGD) of 2.1 days in Caki-1 tumors but had no effect on A498 tumors (P < 0.05). RIT with 100 mu Ci of Y-90-NR-LU-10 resulted in a TGD of 4.8 days for Caki-1 tumors, whe reas 100 mu Ci and 150 mu Ci of Y-90-NR-LU-10 produced a TGD of 1.9 an d 2.7 days for A498 tumors, respectively. Estimated absorbed doses wer e 21.9 Gy in Caki-1 tumors treated with 100 mu Ci of Y-90-NR-LU-10 and 14.5 Gy and 21.8 Gy in A498 tumors treated with 100 mu Ci and 150 mu Ci of Y-90-NR-LU-10, respectively. The weighted normal tissue absorbed doses were 7.4 Gy for Caki-1 tumor-bearing miceand 9.0 Gy for A498 tu mor-bearing mice (P > 0.05). To compare the responses of Caki-1 and A4 98 xenografts to RIT with external beam ED LDR and HDR irradiation, tu mor-bearing mice were treated with equivalent doses (20-22 Gy) of 1) R IT with Y-90-NR-LU-10 (100 mu Ci for Caki-1 and 150 mu Ci for A498), 2 ) continuous ED LDR Cs-137 irradiation with a initial dose rate of 22 cGy/hour, or 3) HDR X-irradiation (2 Gy x 10 fractions in 2 weeks). Th e TGDs produced by RIT, ED LDR, and HDR were 5.3, 9.7, and 8.3 days fo r Caki-1 and 2.7, 5.1, and 5.8 days for A498. The relative efficacy of RIT in these xenograft models correlated well with the radiobiologic parameters (i.e., the size of the initial slope and shoulder) of in vi tro survival curves following HDR and ED LDR irradiation in these cell lines. CONCLUSIONS. The results demonstrated that Caki-1 tumors were more sensitive than A498 tumors to RIT, HDR, and ED LDR irradiation in vitro and in vivo in nude mice. The results suggest that the radiosen sitivity of human tumor cells to low dose rate RIT correlated with the size of the initial slope and the shoulder of in vitro survival curve s. Tumors with radiation survival curves characterized by a steep init ial slope (i.e., a large alpha and alpha/beta ratio) and/or a small sh oulder (i.e., a small n number and a low D-q dose) tend to be relative ly more sensitive to RIT than tumors with a flat initial slope and/or a large shoulder. The SF2 also correlated well with the radiosensitivi ty of these tumor cell lines. The final slope of the survival curves r epresented by the beta value and D-0 was not a useful indicator of the relative radiosensitivity of these two cell lines. These studies sugg est that some of the radiobiologic parameters (the size of the initial slope and the shoulder) of in vitro radiation survival curves may hel p to predict the sensitivity of tumors to RIT. (C) 1997 American Cance r Society.