The Effects of Gamma and Proton Radiation Exposure on Hematopoietic Cell Counts in the Ferret Model

Female descented ferrets (approximately 12-16 weeks of age) were obtained from Marshall Bioresources (North Rose, NY) and given an acclimation period of 7 days at the Loma Linda University Medical Center (LLUMC). Animals were group-housed and provided access to food and water ad libitum with a 12 hour light-dark cycle. The animals were maintained under standard husbandry conditions and all procedures for the animal care and treatment were approved by the Institutional Animal Care and Use Committees (IACUC) of the University of Pennsylvania and the LLUMC. Ferrets were randomly assigned to treatment groups, with each group consisting of 6–12 animals.

γ-Ray radiation was chosen as the reference radiation for the determination of relative biological effectiveness (RBE) values and was delivered using a ⁶⁰Co source (Eldorado Model ’G’ machine, Atomic Energy of Canada Ltd., Commercial Products Division, Ottawa, Canada) at the LLUMC. For radiation exposures at a high dose rate (HDR) of 0.5 Gy/minute, the source to target distance was 150 cm, with a usable radiation field of 40 x 40 cm² and a field flatness of 5.4% in the horizontal direction and 3.6% in the vertical direction. No additional material was placed between the source and target to modify the dose rate. For radiation exposure at a low-dose rate of 0.5 Gy/hour, the source to target distance was 195 cm with a useable field size of 40 x 40 cm² and a field flatness of 4.0% in the horizontal direction and 2.3% in the vertical direction. To achieve the low dose rate (LDR), an additional 12.7 cm of steel plating was placed between the source and the target to attenuate the flux of the beam. Depth dose measurements were made using a calibrated PTW Markus ionization chamber for comparison with proton irradiation.

For the proton radiation exposures, experiments were performed in the LLUMC horizontal clinical beam-line using an incident beam of 155-MeV. The incident protons were scattered into a uniform field using the clinical 2-stage scattering system and modulated in depth using an 11-cm clinical modulator wheel. At the exit of the beam-line, the beam was degraded to the required energy using a pre-determined thickness of polystyrene. The proton experiments used two different apertures to target distances and two different polystyrene thicknesses to achieve the desired beam sizes and dose rates of 0.5 Gy/minute (HDR) or 0.5 Gy/hour (LDR), respectively. For the LDR proton experiments, the animal cages were placed 122 cm downstream of isocenter. The useable beam at this distance was measured as 22 x 22 cm² while achieving a flatness of ±10%. This field size allowed only one ferret in the radiation chamber (length, width, and height was approximately 24 x 16 x 9 cm) to be irradiated. For the HDR proton experiments, the animal cages were placed at isocenter. The usable beam at this distance was experimentally verified and measured as 19 x 19 cm², again allowing only one ferret in the chamber to be irradiated. The upstream polystyrene degrader was tuned to achieve a fully modulated proton beam of 110-MeV at the inside of the irradiation chamber. Depth dose profiles were measured for the optimized polystyrene degrader thickness using Gafchromic film, type MD-55, and verified with a PTW Markus ionization chamber.

To complete a more efficient proton radiation of multiple ferrets at the LDR, an alternate scattering system was developed and verified in the LLUMC Proton Research Room. The incident protons were scattered using a 2-stage scattering system to a useable radiation field of 50 cm diameter with flatness and depth dose profiles that were comparable to the clinical system described above. This system allowed for 12 ferrets to be irradiated at any one time.

For the γ-ray and the proton radiation experiments, the animals were placed in Plexiglass radiation chambers measuring approximately 24 x 16 x 9 cm in length, width and height, respectively. The custom-made radiation chambers contained adequate holes for proper air circulation and animals were provided with NapaNectar hydrating gel (SE Lab Group Inc., Napa, CA). For the LDR experiments, the chambers were slightly modified to contain a watering system on the outside of each chamber that did not compromise the radiation dose delivered to each animal, so these animals were not provided NapaNectar gel. The LDR chambers also contained a thermistor which enabled real-time monitoring of temperature within the radiation chamber during the long LDR exposures. The average recorded temperature of an empty chamber prior to radiation was 70°C, and the average recorded temperature during/after the HDR and LDR radiation exposures (with an animal in the chamber) was 75°C; thus, these results indicated that the temperature during the containment remained adequate. The animals were not anesthetized during irradiation to total doses ranging from 0.75 Gy to 2.0 Gy and were able to express normal postural movements throughout. For the HDR experiments, although the actual radiation exposures were only minutes long (as opposed to hours in the case of the LDR experiments), the animals were restrained in the radiation chamber for the same amount of time as the LDR animals, corresponding to the appropriate dose.

Animals were anesthetized by isoflurane inhalation for blood collection. Blood samples were collected from the external jugular vein and immediately placed in tubes containing ethylenediaminetetraacetic acid (EDTA). Blood draws were performed on each animal at least 2 days prior to radiation exposure (Pre) and again at 3 hours (3 h) and 48 hours (48 h) post-irradiation. A complete blood count (CBC) with differential was performed using a Bayer Advia 120 Hematology Analyzer within 24 hours of blood collection (Antech Diagnostics, Irvine, CA).

The average counts of white blood cells (WBC), neutrophils, lymphocytes, monocytes and eosinophils were determined prior to the radiation exposure as baseline values. The blood cell counts obtained in animals at different time points were divided by the respective baseline values and expressed as fractions of control for statistical analyses. The effects of experimental factors (Radiation Type, Dose Rate, and Dose) and their interactions on the blood cell counts were determined by variance analyses using a general linear model, which was performed using a Minitab statistical software, release 15 (Minitab Inc., State College, PA). The relationship between the dose and blood cell count was determined by fitting the data to a linear quadratic model, y = exp (-αD-βD²), where y is blood cell count expressed as fraction of pre-irradiation baseline value, D stands for radiation dose (Gray or Gy), and α and β are the fitted coefficients for the linear and quadratic components of the linear quadratic model. The linear quadratic curve fitting was performed using SigmaPlot graphics software (SPSS Inc., Chicago, IL). RBE values were calculated by first determining the dose of γ-ray radiation needed to produce the same blood cell counts as proton radiation, and secondly dividing the derived γ-ray radiation dose by the proton radiation dose. The RBE values were plotted against proton radiation dose and analyzed by a non-linear regression analysis to show the relationship between RBE value and proton radiation dose. The non-linear regression analysis showing the relationship between the RBE value and proton radiation dose and the associated 95% confidence interval (CI) was also performed using Minitab statistical software (release 15).

The effect of confinement in the irradiation chamber was determined in a group of 5 animals that were placed in the irradiation chamber for up to 7 hours without radiation exposure. Using one-way ANOVA, no statistically significant differences are reported for each blood cell type (WBC, neutrophils, lymphocytes, monocytes, eosinophils, and platelets) when comparing the absolute cell count at each time point to the other time points (data not shown). Thus, the confinement up to 7 hours did not affect blood cell counts.

The effects of the experimental factors, which consisted of Radiation Type, Dose, and Dose Rate, were determined by variance analyses using the general linear model. The data obtained at 3 and 48 hours after irradiation were analyzed separately to compare the effects observed at these two time points. The results showed that the WBC count decreased within 3 hours after irradiation and the magnitude of the decrease observed at 48 hours after irradiation was approximately three times the decrease observed at 3 hours after irradiation (Figure 1). At 3 hours after irradiation, Dose was a significant influencing factor of the WBC count. The Dose Rate was also a significant influencing factor of the WBC count, which resulted in larger decreased counts after irradiation at HDR than at LDR. Radiation Type (γ-rays vs. protons) had almost no effect on the WBC count. At 48 hours after irradiation, the Dose was the only significant influencing factor of the WBC count, which decreased in a dose-dependent manner. As shown in the dose response curves in Figure 2, a significant and dose-dependent decrease in WBC count was observed in animals at 3 hours after irradiation with HDR γ-rays (Figure 2A), HDR protons (Figure 2A). and LDR γ-rays (Figure 2B), but not with LDR protons (Figure 2B). The coefficient for the linear component of the dose response curve slope was 0.266 and 0.333 for the animals irradiated with HDR γ-rays and protons, which were approximately twice or more of that for the animals irradiated with LDR γ-rays and protons. The coefficient for the quadratic component of the dose response curve slopes was negligible (≤ 3.469 x 10^-18) except for the animals irradiated with HDR γ-rays. The effective dose (ED) ED₁₀, ED₅₀, and ED₉₀ for the HDR γ-ray radiation were bracketed by the 95% confidence intervals of the ED₁₀, ED50, and ED₉₀ for the HDR proton radiation (Table 1), indicating that all three ED values were not significantly different between the HDR γ-ray and HDR proton radiation. The ED₁₀, ED₅₀, and ED₉₀ values were not compared between the LDR γ-ray and proton radiation due to the absence of a significant dose response for the LDR proton irradiated animals. The fitted RBE values were 2.15 and 4.02 for HDR and LDR protons at 0.75 Gy, respectively (Table 2), which were significantly above 1.00. The RBE decreased with the increase of the proton radiation dose and the fitted RBE values for HDR and LDR protons at 1 and 2 Gy were not significantly different from 1.00, which was within the 95% confidence intervals for the respective RBE values.

Figure 1.

Figure 2.

At 48 hours after irradiation, a significant and dose-dependent decrease in WBC count was observed in animals of all groups with the coefficient for the linear component of the dose response curve slope ranging from 0.786 to 1.268 (Figures 2C and 2D). The coefficient for the quadratic component of the dose response curve slopes was again negligible (≤ 1.388 x 10^-17) except for the animals irradiated with HDR γ-rays. The ED₁₀, ED₅₀, and ED₉₀ for the HDR γ-ray radiation were greater than the upper limit of the 95% confidence intervals of the ED₁₀, ED₅₀,, and ED₉₀ for the HDR proton radiation (Table 1), indicating that HDR proton radiation was more effective than HDR γ-rays in reducing the number of WBCs. The ED₁₀, ED₅₀,, and ED₉₀ for the LDR γ-ray radiation were within the 95% confidence intervals of the three corresponding ED values for LDR proton radiation, suggesting that the LDR γ-rays and protons were not significantly different in the effectiveness of reducing the number of WBCs. The RBE values derived from results observed at 48 hours after irradiation also decreased with the increase of the proton radiation dose, but the downward trend was not as prominent as the trend observed at 3 hours after irradiation. The fitted RBE values ranged from 1.60 - 1.19 at 0.75 Gy and 1.04 - 0.77 at 2 Gy for HDR and LDR protons, respectively (Table 2), and were significantly above 1.00 for HDR protons at 0.75 and 1 Gy but significantly below 1.00 for LDR protons at 2 Gy.

Of the three experimental factors evaluated, Radiation Dose was the most significant influencing factor of the neutrophil count, which increased at 3 hours after irradiation (Figure 3A) but decreased at 48 hours after irradiation (Figure 3B), both in a dose-dependent manner. Radiation Dose Rate significantly affected the neutrophil count at 3 hours after irradiation with a larger reduction after irradiation at HDR than at LDR. The Radiation Dose Rate did not significantly affect the neutrophil count at 48 hours after irradiation. The Radiation Type did not significantly affect the neutrophil count at either time points after irradiation. At 48 hours after irradiation with HDR γ-rays, HDR protons, LDR γ-rays, and LDR protons, the coefficient for the linear component of the dose response curve slope was 0.334, 0.582, 0.386, and 0.231, respectively (Figures 4A and 4B). The coefficient for the quadratic component of the dose response curve slopes was small (≤ 0.003) except for the animals irradiated with LDR protons. The ED₁₀, ED₅₀,, and ED₉₀ for the HDR γ-ray radiation were greater than the upper limit of the 95% confidence intervals of the ED₁₀, ED₅₀,, and ED₉₀ for the HDR proton radiation (Table 1), indicating that HDR protons were significantly more effective than HDR γ-ray radiation in reducing neutrophils. The ED₁₀, ED₅₀, and ED₉₀ for the LDR γ-ray radiation were bracketed by the 95% confidence intervals of the ED₁₀, ED₅₀, and ED₉₀ for the LDR proton radiation, suggesting that LDR γ-rays and protons were not significantly different in their effectiveness of killing neutrophils. The fitted RBE values for irradiation with HDR protons at 0.75 and 1 Gy and with LDR protons at 0.75 Gy were 1.88, 2.04, and 2.11, respectively (Table 2), which were significantly above 1.00. The fitted RBE values for irradiation with HDR protons at 2 Gy and with LDR protons at 1 and 2 Gy were 1.55, 0.62, and 0.90, respectively, and were not significantly different from 1.00.

Figure 3.

Figure 4.

Of the three experimental factors evaluated, Radiation Dose was the only significant influencing factor of the lymphocyte count at 3 hours (Figure 5A) and 48 hours (Figure 5B) after irradiation. Neither the Radiation Type nor the Dose Rate significantly affected the lymphocyte count at either time points after irradiation. The dose response curves for the lymphocytes displayed a significant and dose-dependent decrease in lymphocyte count at 3 hours and 48 hours after irradiation with either γ-rays or protons at HDR or LDR (Figure 6). The coefficient for the linear component of the dose response curve slope ranged from 1.698 to 2.148 and the coefficient for the quadratic component of the dose response curve slopes was negligible (≤ 1.180 x 10^-16) for all groups except the animals at 48 hours after irradiation with HDR protons and LDR protons, for which the coefficient for the linear component of the dose response curve slopes were 1.227 and 0.858 whereas the coefficient for the quadratic component of the dose response curve slopes were 1.159 and 0.782, respectively. The ED₁₀, ED₅₀, and ED₉₀ for animals at 3 and 48 hours after irradiation with HDR and LDR γ-rays were within the 95% confidence intervals of the corresponding ED₁₀, ED₅₀, and ED₉₀ for the HDR and LDR proton radiation (Table 1), indicating that γ-rays and protons were not significantly different in their effectiveness of reducing lymphocytes at HDR or LDR. The fitted RBE values for HDR and LDR protons at 0.75 Gy ranged from 0.83 to 1.41 at 3 and 48 hours after irradiation (Table 2). With the increase of proton radiation dose to 2 Gy, the fitted RBE values decreased to a range of 0.67 to 0.84 at 3 and 48 hours after irradiation at HDR or LDR. The RBE values were not significantly different from 1.00; however, the RBE was significantly higher than 1.00 for 0.75 Gy protons at LDR and significantly lower than 1.00 for 2 Gy HDR and LDR protons at 3 hours after irradiation and for 2 Gy LDR protons at 48 hours after irradiation.

Figure 5.

Figure 6.

Monoctye, eosinophil, and platelet counts were statistically insignificant at both the 3 hour and 48 hour time points (data not shown).