In this study, we applied a bone marrow culture system established in the 3-D bioreactors to the investigation of radiation-induced MN-RET, a marker of genotoxicity as a result of DNA damage by clastogens or aneugens. Though the traditional Dexter-style bone marrow cultures in flasks or dishes support the growth of early erythroid progenitors BFU-E and CFU-E, and the suspension cultures show potentials in generating erythrocytes [18
], long-term effective generation of reticulocytes up to 8 weeks has not been previously reported. Our 3-D Long-Term Bone Marrow Culture (3-D LTBMC) is an improvement of Dexter’s culture (2-D culture) that is established in the flat surface of flasks/dishes. The 3-D LTBMC established in the bioreactor is packed with porous microspheres to provide the three dimensional matrix to mimic in vivo bone architectural marrow framework. The cultures are supplemented with hydrocortisone, recombinant human erythropoietin, transferrin, and charcoal-treated FBS to enrich lineage-specific proliferation and differentiation of red blood cells. Such combination of nutrients and cytokines appears to mimic the essential chemical microenvironment for erythropoiesis in bone marrow in vivo.
Our optimized culture condition in the 3-D bioreactor has enriched terminal erythropoiesis for up to 8 weeks, while still maintaining the multiple lineage differentiation. The terminal erythropoiesis reached a peak between week 2 and 3, following by a rebound or staying relatively constant until week 6~8. The expression of % RBC and % RET did vary after 28 days. The exact cause of this observation is not clear. One of the putative mechanisms is that the performance of the culture after Day 28 depends mainly on how many hematopoietic primitive progenitors have survived after 28 days. This process is prone to the random disturbance. Of note that in the traditional 2-D Dexter’s cultures, the cultures have to be recharged with fresh bone marrow cells after 3 weeks to maintain the longevity of culture, because of the heavy loss of hematopoietic primitive progenitors in the first 3 weeks [30
]. In our 3-D culture, “recharge” is not necessary for maintain the culture beyond Day 28. The cell population has varied due to random loss of progenitor cells from weekly sampling for data analyses. Nonetheless, the common feature we wanted to show with our data from the 3-D culture was that the reticulocytes were generated in the significant level in all cultures for up to 6~8 weeks and the activity didn’t seem to die at the end of the cultures.
While comparatively little work in MN-RET has been accomplished with radiation research, we have recently published the work on gamma radiation-induced MN-RET in C57Bl/6 mouse after total body irradiation in vivo [31
]. Because the generation of MN-RET requires the dynamic process of erythroid progenitor and precursor cell proliferation, differentiation, and terminal maturation, almost all MN-RET investigations were conducted in the rodent in vivo testing systems with the exception of a more recent publication using a bone marrow suspension culture system [25
]. Data from this suspension bone marrow culture supports the validation in testing several drugs but did not test radiation effect. In contrast to this reported two-stage suspension culture system that has supported the generation of reticulocytes in the last stage for 2 days, our 3-D LTBMC supports the continuous erythropoiesis and the generation of reticulocytes and mature erythrocytes for as long as 8 weeks. Our observation suggests that the primitive progenitors for erythropoiesis actively proliferated and differentiated in the 3-D LTBMC, and the entire erythroid development chain was expressed in yielding the mature erythrocytes. Therefore, this culture model provides a potentially physiological testing environment, making it possible to investigate bone marrow genotoxic effect of radiation on erythroid lineage in vitro.
In mice, reticulocytes generated in the bone marrow enter the peripheral blood after residing in the bone marrow for 10 hours, and fully mature to erythrocytes in peripheral blood after approximately 15 hours [29
]. The peak time for MN-RET induction after irradiation in the mouse in vivo model has been previously reported. Published work on the kinetics of MN-RET induction in mice irradiated in vivo shows that MN-RET induction is affected by the radiation doses and the types of radiation. In general the peak of MN-RET expression in the bone marrow is between 24 to 28 hours in the bone marrow in vivo [29
], and it is between 30 to 44 hours in the peripheral blood [29
]. In our 3-D LTBMC, the peak time for radiation-induced MN-RET expression occurs at approximately 24 to 32 hours post-irradiation, which falls between what is observed in the bone marrow and the peripheral blood compartment in mice irradiated in vivo. This timeframe seems appropriate because reticulocytes generated in the 3-D LTBMC are retained in the culture at all times without a relevant peripheral blood compartment. Thus the peak time of MN-RET expression in the 3-D LTBMC (24–32 hours) is expected to occur sometime between the peak time in bone marrow (24–28 hours) and the peripheral blood (30–44 hours) in the in vivo model after total body irradiation of mice. The percentage of RETs in the 3-D LTBMC first declined in a small level at 12 hours, and then big drop took place after 24 hours post irradiation (). The similar phenomena has also been observed in bone marrow in vivo [32
The difference between our 3-D culture model and the in vivo mouse model in the %RET was the intermediate rebound at 24 hours post radiation in our culture model. The putative mechanism behind this was that the cell cycle of most erythroblasts in our 3-D culture might have been synchronized by the daily media change. Thus, the development of erythroblasts was in the pattern of waves. At 0~12 hour post radiation, in the non-synchronized erythroid development, the erythroblasts at the last mitosis that were beyond S phase at the time of radiation would enucleate. In our 3-D culture, the major portion of these erythroblasts would have been synchronized to the previous or next developmental wave so that there were fewer enucleation events, leading to the greater decrease of %RET than that in the non-synchronized erythroid development. Radiation would have caused the delay of the cell cycle and thus increased the interval from the S phase to enucleation in the last mitosis of erythroid development from 9~12 to 24~27 hours [36
]. At 24 hours post radiation, the %RET would rebound to some extent when the delayed enucleation of synchronized erythroblasts was finished, probably before the next enucleation wave in the control culture without irradiation. We have planned to test this hypothesis in the future. The comparable kinetics in the 3-D LTBMC culture and in mice after in vivo irradiation suggests that the time-course for cell cycle progression and enucleation in the culture are similar between the 3-D LTBMC and bone marrow in vivo
Similar to our finding of MN-RET induction after irradiation, the in vivo MN-RET assay reported in the animal models also showed a sensitivity to gamma radiation. It has been reported that the MN-RET assay is able to detect 6~10 cGy X-ray [37
] and 2.5~3 cGy gamma-ray irradiation [34
]. In our 3-D LTBMC system, radiation doses as low as 12.5 cGy (0.125 Gy) gamma-radiation induced significantly higher MN-RET than 0 Gy (p=0.01, two-side paired t-test). Thus, the sensitivity of the 3-D LTBMC was similar to that of in vivo mice model, although we did not test doses lower than 12.5 cGy.
In the 3-D LTBMC system, we observed a near linear increase of MN-RET from 0 Gy to 1 Gy after radiation exposure, with a maximum reached at 1.5–2.0 Gy. This was similar to our previous report of in vivo MN-RET induction in the peripheral blood of C57Bl/6 mice after total body irradiation in vivo [31
]. The previously reported in vivo investigation also showed a lack of increase of MN-RET beyond 1.5 Gy, suggesting that competing radiation cytotoxic effect may be working on the bone marrow precursors, i.e., erythroblasts. By depleting the erythroblasts in the bone marrow through radiation-induced apoptosis, it would be difficult to induce more MN-RET at doses higher than 1.5 Gy in mice. Our finding of the lack of further induction of MN-RET beyond 1.5 – 2.0 Gy in the 3-D LTBMC was similar to our published in vivo data, lending further support that 3-D LTBMC may serve as an in vitro model for MN-RET assay for genotoxicity testing after irradiation.
It appeared that the background of MN-RET is affected by reagents used in the culture. Among three experiments for MN-RET kinetics (), the background percentage of MN-RET in the control cultures of sham irradiation (0 Gy) was at an average of 1.3% (range 0.9 – 1.5%), which was not significantly different (p=0.41; Friedman’s non-parametric 2-way ANOVA), in the two experiments using the same lots of reagents (open diamonds and triangles). It was, however, significantly higher at 2.6% (range 2.4 – 2.7%) (P = 0.003; Friedman’s non-parametric 2-way ANOVA) in the third experiment conducted with different lots of reagents (open squares). The MN-RET expression has long been known to be sensitive to the stimulation by various chemicals [9
]. Thus the higher level of background MN-RET in the 3-D LTBMC may be affected by the sensitivity to the chemicals in different lots of culture reagents. In , the higher background in the third experiment (squares) might be contributing to the higher levels of MN-RET induction below 1 Gy of radiation and the earlier saturation in 1.5 Gy, vs. 2 Gy observed in other two experiments. Our finding supports using similar batches of reagents with adequate internal controls for all experiments using 3-D LTBMC to minimize the influence from chemicals in the culture media.
In summary, a 3-D LTBMC system capable of continuously generating reticulocytes from the progenitor and precursor cells of erythropoiesis for mouse bone marrow culture has been successfully applied to the research in radiation genotoxicity testing, yielding a MN-RET increase in a dose-dependent manner. Our study also reveals the similarities of the dose responses of MN-RET induction after radiation exposure between the mouse in vivo model [31
] and the 3-D LTBMC. The peak time of MN-RET expression in 3-D LTBMC appears to fall between the peak time in the bone marrow in vivo and the peripheral blood of mice after total body irradiation, likely due to the differences in the kinetics of reticulocytes by the physical presence of the bone marrow compartment of experimental animals. Our 3-D LTBMC condition offers the potential for the research of erythropoiesis and bone marrow radiation injury in an in vitro model established in the bone marrow bioreactor.