Short-term injection of antiapoptotic cytokine combinations soon after lethal γ-irradiation promotes survival

The pancytopenia that occurs after irradiation is mainly the consequence of early death of hematopoietic stem and progenitor cells (HSPCs), and its severity and duration depend on the level of bone marrow (BM) cell loss.1 Numerous studies have shown the efficacy of hematopoietic growth factors (HGFs) in stimulating neutrophil and platelet recovery in animals and humans after myeloablation.2 Granulocyte-colony-stimulating factor (G-CSF), granulocyte macrophage-colony-stimulating factor (GM-CSF), interleukin-3 (IL-3), IL-11, and thrombopoietin (TPO) are among the most efficient cytokines.3 These molecules have primarily been evaluated as monotherapy, but the simultaneous or sequential administration of 2 complementary cytokines has resulted in better patterns of hematologic recovery.4-7 As a clinical application, victims of the criticality nuclear accident in Tokai-mura, Japan, recently were given G-CSF, megakaryocyte growth and development factor (MGDF), and erythropoietin in combination.8 In most preclinical trials, HGF administration starts the day after myeloablative injury. The duration of treatment covers at least 15 days until neutrophil or platelet recovery. Prolonged administration stimulates HSPCs to cycle, proliferate, and differentiate, but it may also induce adverse effects such as proinflammatory and immunogenic activity.9 This led us to consider short-term treatment schedules. Thus far, few cytokines have been tested as early and acute therapy. IL-1 seemed to be a pivotal hematopoietic cytokine in the therapy for radiation injury before adverse effects were established.10 G-CSF injected into mice 2 hours after irradiation as a single injection has been shown to induce a good level of protection against 30-day mortality.11 Unfortunately, this factor was unsuccessful in restoring thrombopoiesis in myeloablated nonhuman primates. Despite a notable level of platelet recovery observed in sublethally or lethally irradiated mice, TPO did not prevent a profound neutropenic nadir after lethal irradiation, which may be involved in the observed lethality.12 

Apoptosis has been shown to play a major role in HSPC death soon after irradiation in vitro13 and in vivo.14,15 If a subset of highly radioresistant stem cells recapitulates long-term hematopoiesis16,17 (a notion supported by the eventual autologous reconstitution in accidentally irradiated victims despite bone marrow allotransplantation), the preservation of residual cytokine-sensitive HSPCs from radiation-induced (RI) apoptosis would be critical for myeloablated host survival.15 In particular, the mitigation of morbidity and mortality caused by the initial neutropenia and thrombocytopenia requires the preservation and stimulation of less immature HSPCs. In vitro studies dealing with RI apoptosis of CD34⁺ cells have shown that early management is necessary to maintain cell clonogenicity. In fact, cytokines active in the early stages of hematopoiesis preserve CD34⁺ cells, provided they are added in combination soon after irradiation.13 Thus far, we found the best antiapoptotic combination to be stem cell factor (SCF) + FMS-like tyrosine kinase 3 ligand (FLT-3 ligand) + TPO + IL-3 (4F). In addition, stromal derived factor-1 (SDF-1) chemokine has been reported to be an antiapoptotic factor toward HSPC in vitro18,19and to participate in host defenses against DNA damage.20All these data prompted us to use cytokine combinations as an early treatment of RI injury.

The purpose of our study was to assess the capacity of antiapoptotic cytokines to rescue mice from radiation when cytokines are given as an early treatment to counteract cell death, especially potentially lethal lesions, as a combination to ensure multilineage protection, and as a short-term treatment to avert adverse effects. In this study, we report the therapeutic benefit for lethally irradiated mice of the administration at 2 hours and 24 hours after irradiation of multiple cytokine combinations involving SCF, FLT-3 ligand, TPO, IL-3, and SDF-1 in comparison with TPO or SDF-1 alone. The highest 30-day survival rates were obtained with 4F and 4F + SDF-1 (5F). Long-term follow-up revealed delayed death, and treated irradiated mice displayed a reduced long-term hematopoietic potential compared with normal mice. However, approximately half the protection provided by the cytokines at 30 days was still evident 300 days after total body irradiation (TBI). To dissect the mechanism of action of 4F, we used a nonhuman primate model and showed that 4F administration 2 hours after TBI prevented BM CD34⁺ cells from apoptosis and that 4F did not stimulate CD34⁺ cells to cycle within the first 24 hours after in vitro irradiation.

Housing and experiments were approved by the French Army Ethics Committee in accordance with European rules and regulations.

Six-week-old male B6D2F1 mice (Charles River, Les Oncins, France) were quarantined for 7 days before irradiation at the Centre de Transfusion Sanguine des Armées (Clamart, France). Mice were housed in groups of 6 per cage in rooms equipped with a reverse-filtered air barrier and received commercial food and acidified water ad libitum.

Adult male cynomolgus monkeys (Macaca fascicularis) weighing 8 kg each were housed at the Centre de Recherches du Service de Santé des Armées (CRSSA)–accredited animal facility in individual cages in holding rooms equipped with a reverse-filtered air barrier (10 air changes per hour, 100% fresh air) and full-spectrum light (8:00 am to 8:00 pm) and conditioned to 23°C with a relative humidity of 60%. Animals were fed commercial primate chow supplemented with fresh fruit and were provided tap water ad libitum.

Cells were irradiated at 2.5 Gy using a cobalt Co 60 γ source (dose rate, 100 cGy/min). This dose appears relevant because it is greater than the D₀ (dose required to reduce cell survival to 37% of its initial value) value, estimated to be approximately 1 Gy for HSPC.21 

Mice placed in ventilated polycarbonate canisters underwent TBI with a single 8 Gy dose using a cesium Cs 137 gamma source (dose rate, 10 Gy/min). Dosimetry was performed using an ionization chamber and thermoluminescent aluminum oxide dosimeters. Survival curves previously performed established that 8 Gy constitutes the 90% ± 5% lethal dose at 30 days.

Anesthetized monkeys (n = 6) placed in a restraining chair were exposed to global unilateral front irradiation with a ⁶⁰Co γ source at a total midline tissue dose of 5 Gy (measured in air at the anterior iliac crest level) with a dose rate of 20 cGy/min. Dosimetry was performed using ionizing chambers and thermoluminescent dosimeters.

Recombinant murine SCF (S), FLT-3 ligand (F), TPO (T), IL-3 (3), and SDF-1α were used (R&D Systems, Abingdon, United Kingdom) for the mouse study. Recombinant human (rh) TPO, rhSCF, rhFLT-3 ligand, and rhIL-3 (R&D Systems) were used for the monkey study.

For the in vitro study, baboon-mobilized peripheral blood (PB) CD34⁺ cells were collected using leukapheresis as previously described.22 

Within 30 minutes of irradiation or sham irradiation, cells were washed and resuspended (1 × 10⁵ cells/mL) in serum-free medium H00 (RTM-Mabio, Tourcoing, France). Unirradiated and irradiated cells were then incubated with the 4F cytokine combination (FLT-3 ligand + TPO + SCF + IL-3; 50 ng/mL each) in 6-well plates at 37°C in a humidified incubator with 5% CO₂ in air. Cells were analyzed after 24 hours and 3 days of incubation relative to fresh control cells.

DNA and nuclear protein Ki67 (only present during the G₁and S/G₂/M phases of the cell cycle) cell content were evaluated using flow cytometry.23 Briefly, after CD34⁺ cell labeling with antihuman CD34, clone 566, monoclonal antibody (mAb) conjugated to phycoerythrin (PE) (BD Biosciences, Le Pont de Claix, France), and cell fixation in phosphate-buffered saline (PBS)–formaldehyde 1%, cells were permeabilized in PBS-Triton 0.1% (30 minutes on ice). After two washes, cells were incubated with anti-Ki-67 MIB-1 mAb conjugated to fluorescein isothiocyanate (FITC; Beckman Coulter, Villepinte, France) for 1 hour at 4°C. Finally cells were washed and resuspended in PBS 7-aminoactinomycin (7 AAD, 0.5μg/mL; Sigma, St Louis, MO). Flow cytometric analysis was performed within 1hour, and G₀, G₁, S/G₂/M cell contents were determined (FCS Express; De novo Software, Toronto, ON, Canada).

Current methodologies used for evaluating in vitro cell death are inefficient for in vivo evaluation because apoptotic cells are progressively eliminated by phagocytes. It is, therefore, necessary to identify a pertinent early marker. We chose to measure the mitochondrial transmembrane potential (MTP) at the CD34⁺cell level by flow cytometry using 3, 3′ dihexyloxacarbocianine iodide (DiOC_6; Interchim, Montluçon, France).24MTP is known to be increased during the early phase of apoptosis,25 before the final step of depolarization. Humeral BM was harvested after TBI, and mononuclear cells were isolated by cell density separation and were DiOC₆/CD34 double-labeled in 2 steps. First, cells were incubated with 25 nM/L DiOC₆ for 15 minutes at 37°C in PBS. Cells were then put on ice to prevent the leak of DiOC₆ and were then stained with anti-CD34 mAb for 1 hour. Nucleated cells were selected for analysis using 0.5 μg/mL LDS751 (Interchim) addition. DiOC6 uptake was expressed in terms of mean fluorescence intensity (MFI). For each sample, DiOC6 uptake in the granulocyte population was used as an internal standard. Thus, the ratio of DiOC6 uptake in CD34⁺cells to DiOC6 uptake in granulocytes was evaluated before and 6 hours after TBI. The increase in the ratio accounts for a status of hyperpolarization of the mitochondrial membrane, which is considered to be an early marker of apoptosis.

Twenty-four mice were randomly assigned to each group. Seven treatments were evaluated using S, F, T, 3, and SDF-1 alone or in combination: TPO, SDF-1, 3F (S + F + T), 4F (S + F + T + 3), and 5F (4F + SDF-1) were given 2 hours and 24 hours after TBI. Each cytokine was administered at 50 μg/kg for each administration. The last 2 treatments were 4Fsingle (4F given only at 2 hours after TBI, each cytokine given at 50 μg/kg) and 4F10 (4F given at 2 hours and 24 hours after TBI, each cytokine given at 10 μg/kg). Preliminary experiments showed that 4F induced a similarly protective effect by either the intravenous or the intraperitoneal route. Cytokine solutions were prepared extemporaneously with pyrogen-free sterile Dulbecco phosphate-buffered solution containing 0.5% human serum albumin and injected intraperitoneally in a volume of 200 ± 20 μL (according to the animal's weight) at 2 hours and 24 hours after irradiation. Four control groups that received cytokine diluent were evaluated throughout the study; 4F and 5F evaluation was carried out twice.

Six monkeys underwent 5 Gy γ irradiation. Three of them were treated with 4F at 2 hours after TBI (each cytokine was administered at 50 μg/kg intravenously), and 3 control animals received the cytokine diluent. Humeral BM of control and treated animals was harvested before irradiation and 6 hours after TBI for the evaluation of apoptosis at the CD34⁺ cell level.

During the first 30 days, mice were observed daily to record survival. Weight and behavior were monitored. After 30 days, animals were observed 3 times a week. Autopsies were performed in long-term deceased animals. After 360 days, surviving mice were killed by cervical dislocation and were autopsied. Spleens and femurs were collected. Spleens were stained with formalin and weighed. BM was harvested, and total nucleated cell (TNC) counts were assessed, as were differential cell counts using cytospins (Shandon, Pittsburgh, PA) stained with May-Grünwald-Giemsa. BM clonogenicity (granulocyte macrophage-colony-forming unit [GM-CFU] and erythroid burst-forming unit [BFU-e]) was evaluated using a short-term methylcellulose assay (RTM Mabio). Briefly, 15 × 10³/mL femoral BM TNCs were plated in triplicate in 6-well plates with recombinant murine SCF, FLT-3 ligand, IL-3, GM-CSF, and EPO (R&D Systems), and CFU scoring was performed at 10 days of culture.

The mouse study was conducted over 4 experiments, and no significant difference in survival among the 4 control groups was observed (χ² evaluation), allowing for a global statistical comparison. Survival curves were compared using Kaplan-Meier analysis, and statistical significance was determined between groups using χ² analysis with Yates correction. Spleen weights, BM TNC, and clonogenicity were compared using the Mann-Whitney U test. The latter test was also used to analyze the effect of 4F on the cell cycle distribution of baboon BM-irradiated CD34⁺ cells versus unirradiated cells.

The onset of lethality was delayed by 3 days in mice treated with 5F in comparison with the other 2 treatments and with the control group (11.5 days instead of 8 days). Thirty-day survival rates are presented in Figure 1A and Table1. The average survival rate of controls was 8.3%. All treated groups displayed 30-day survival rates significantly higher than those of the control group. Multiple cytokine treatments proved to be more efficient than single therapy. In fact, 5F and 4F afforded the best protection (survival rates, 87.5% and 81.25%, respectively) when given at 50 μg/kg 2 hours and 24 hours after TBI. Neither was statistically different from the other (P = .36). Both treatments were more efficient than TPO (58.3%), 3F (50%), 4F10 (45.8%), and SDF-1 (29.2%), as shown in Table 2, and 5F was better than 4Fsingle (62.5%). The benefit conferred by 2 injections of 4F rather than a single injection was not significant. Thus, the combination of cytokines administered appears crucial for efficacy.

Freshly isolated baboon-mobilized PB CD34⁺ cells are mainly in the G₀/G₁ phases of the cell cycle, with only a reduced subset in the S and G₂M phases (67.8% ± 5.8%, 29.9% ± 4.8%, and 2.4% ± 1.2%, respectively), as shown in Table 3 and Figure 2. Unirradiated (unir) CD34⁺ cells incubated for 24 hours in serum-free medium in the presence of 4F exhibited an important reduction of G₀cell subset and an increase of G₁ and S/G₂/M cell subsets (40.2% ± 12.6%, 39.5% ± 6%, 19.8% ± 7.6%, respectively; P < .05 for G₀, G₁, and S/G₂/M) (day 1 unir vs day 0). In contrast to unirradiated cells, in vitro–irradiated cells incubated with 4F for 24 hours showed the maintenance of their initial cell cycle distribution (P > .05 for G₀ and S/G₂/M, day 0 vs day 1, 2.5 Gy) (P < .05 for G₀, G₁, and S/G₂/M, day 1 unir vs day1, 2.5 Gy). From day 3 of culture with 4F, irradiated CD34⁺ cells were stimulated to cycle as did unirradiated cells. At 3 days, the relative cell cycle distributions were similar even though, in terms of absolute cell number, unirradiated cells exhibited a greater expansion rate (K_4F = 2.6 ± 0.7) than irradiated cells (K_4Firrad = 0.55 ± 0.34). At 7 days, the respective cell amplification rates were K_4F = 12.5 ± 0.7 and K_4Firrad = 2.9 ± 0.6.

Monkey BM CD34⁺ cells harvested 6 hours after 5 Gy TBI exhibited a notable increase in mitochondrial transmembrane potential compared with their own initial value (Figure3). In contrast, this hyperpolarization did not occur in animals given 4F 2 hours after TBI.

Data regarding 300-day survival are presented in Figure 1B and Table 1. Delayed death occurred in all treated groups and in the few surviving controls, mainly from day 150 after TBI, well beyond the classical period of RI hematologic death. In contrast, age-matched unirradiated control (UC) animals experienced no deaths in 360 days of follow-up. Despite delayed death in irradiated animals, all treatments except 4F10 and SDF-1 induced significantly improved long-term survival rates. Multiple cytokine combinations remained the most effective treatments, resulting in survival rates of 43.8% (5F), 50% (4F), and 41.7% (4Fsingle) versus 29.2% (TPO), 12.5% (SDF-1), and 3.3% (control). Moreover, 4F injected only 2 hours after TBI at 50 μg/kg (4Fsingle) resulted in greater long-term survival (41.7%) than when injected 2 hours and 24 hours after TBI at 10 μg/kg (4F10, 8.3%), in accordance with short-term results. The benefit conferred by 5F and 4F versus TPO was not significant.

Autopsies carried out on animals that died between 30 days and 300 days after TBI showed a pattern of cachexia and multiorgan failure. Autopsies performed at 360 days in long-term survivors revealed signs of macroscopic anemia associated with liver disease in most animals and 4 cases of cancer (1 myosarcoma, 2 thymomas, 1 leukemia) in different groups. Spleens and femurs from mice of groups 5F, 4F, TPO, SDF-1, and UC were collected at that time. Spleen weights are reported in Table4. In contrast to the spleens of groups 4F and TPO, those of group 5F were significantly smaller than UC spleens. BM TNC counts (Tables 4 and 5) decreased in all treated groups to values equal to 64.5% (5F), 77.4% (4F), 76.3% (TPO), and 61.5% (SDF-1) of the UC level. However, this reduction in BM TNC was not significant for 5F (P = .06 vs UC). The most striking unexpected observation was that all treated animals proved to be significantly depleted of BM colony-forming units (CFUs) in comparison with UC. BM CFU of treated groups were 27.8% (5F), 56.1% (4F), 59.1% (TPO), and 44.1% (SDF-1) of UC CFU. In fact, though group 5F appeared significantly more depleted of CFU than other treated groups (Table 5), BM megakaryocytic progenitors were twice as numerous in group 5F as in other groups, UC included (4.1% of TNC vs 2% on average; P < .01). At 360 days, there was only one surviving irradiated control mouse whose BM displayed a marked reduction in TNC and clonogenicity (Table 4).

The results of this study support the concept that cytokine cooperation is important for hematopoietic recovery from radiation injury. They also suggest that short-term treatment is efficient. This work not only accounts for the short-term hematopoietic activity of cytokines on the irradiated organism, it also accounts for the long-term response to cytokine treatment. To date the prolonged administration of GM-CSF or G-CSF remains the criterion standard of cytokine therapy for irradiated victims in spite of their granulocyte-restricted activity.7,26 In most preclinical trials, HGF administration starts from the day after myeloablative injury, and the duration of treatment covers 15 to 21 days until neutrophil or platelet recovery.2-6,27 Few combinations have been used despite promising studies in myeloablated nonhuman primates that entailed 2 cytokines4-6 or synthetic dual cytokine receptor agonists.3,27 Thus far, nuclear accident victims have been treated with GM-CSF and IL-328 or G-CSF + GM-CSF.7 More recently, the victims of the criticality accident in Tokai-mura were given G-CSF, MGDF, and erythropoietin in combination.8 However, in these dramatic situations, cytokine therapy was initiated after a variable delay following radiation exposure, which might have reduced its efficacy. Furthermore, the prolonged administration of HGF may induce notable proinflammatory29 or immunogenic activity.3 9 

The potential therapeutic benefit of early and short-term administration of antiapoptotic cytokines has been suggested by some authors whose studies have been primarily focused on TPO.12,15 For some years the latter has emerged as a key factor, not only for its thrombopoietic activity but also as a survival factor on the HSPC.30 TPO/MGDF counteracts lethal myelosuppression when given early after radiobiologic or radiocombined damage.12,15,31-33 Moreover, this beneficial effect can be obtained with a single administration,12,15,31-34 which is as effective as 7 repeated injections for hematopoietic recovery and prevention of death.32 The single administration of TPO to humans has been reported to maintain hematopoiesis without inducing adverse effects.35 As demonstrated by Pestina et al,15 the survival effect involves the mitigation of the p53 apoptotic pathway, but TPO/Mpl-ligand appears to act downstream of p53.

IL-1 and tumor necrosis factor (TNF) also exhibit radioprotective effects, but these cytokines have been discarded because of high toxicity.36 In fact, given that hematopoiesis is regulated by a subtle balance between numerous positive and negative regulators whose cooperation is required to optimize cell survival, proliferation, and differentiation, we chose to evaluate in vivo the therapeutic efficacy of combinations of 5 different cytokines—SCF, FLT-3 ligand, TPO, IL-3, and SDF-1—whose antiapoptotic activity has been screened in vitro or in vivo.13,18,20 The SFT3 combination exhibited the highest antiapoptotic efficacy toward in vitro–irradiated CD34⁺ cells (dose range, 2.5-4 Gy) in comparison with the same cytokines used as single factors or other combinations.13 For instance, IL-3 has been demonstrated to reduce p53-dependent apoptosis in response to DNA damage by inducing the expression of Bcl-2 and Bcl-X_l.15 The SFT combination (3F) has recently been described to increase antiapoptotic survivin protein levels in quiescent CD34⁺cells.37 Moreover, the same authors have reported an inverse correlation between survivin and active caspase 3.38 In our in vitro study,13 the SFT3 combination reduced RI Fas/Fas ligand up-regulation and maintained the Bcl-2 expression level. Cell cycle analysis of irradiated cells treated with SFT3 showed almost no activation after 24 hours of incubation, unlike treated unirradiated cells. Altogether 3 days of incubation were necessary for irradiated cells to display a similar activation pattern of cell cycle as unirradiated cells, which resulted in subsequent CD34⁺ cell expansion within 7 days of liquid culture. Regarding SDF-1α, this is a promising stromal factor locally induced at the BM level, namely after irradiation, which keeps in vitro CD34⁺ cells from apoptosis after serum deprivation.20 Recently, SDF-1 has been described as a survival factor for CD34⁺ cells because it counteracts spontaneous apoptosis through an autocrine/paracrine mechanism.18,19 Preliminary studies suggested a multilineage activity in vivo because SDF-1 administration to nonobese diabetic/severe combined immunodeficiency (NOD/SCID) mice grafted with human cord blood mononuclear cells resulted in the mobilization to peripheral blood and spleen of CD41⁺ cells and myeloid/monocytic cells.39 

In the present study, we first focused on RI short-term hematopoietic cell death evaluated by the classical 30-day survival end-point. We have based our 2 injection treatments on the in vitro capacity of SFT3 to exert a strong antiapoptotic effect when added early to HSPC and to initiate the production of radioprotective progenitors. The cytokine dose (50 μg/kg for each cytokine of the combination) was chosen in accordance with our in vitro study, which demonstrated the necessity for using high concentrations to counteract RI apoptosis (ie, 50 ng/mL is better than 10 ng/mL for each cytokine of the combination),1 and the Pestina et al15 study, in which a dose range of 50 to 65 μg/kg was selected to reduce in vivo RI apoptosis. Here we show that administering the SFT3 combination at 50 μg/kg (each cytokine), supplemented or not with SDF-1α, 2 and 24 hours after acute, lethal TBI results in significantly improved survival.

To confirm that 4F efficacy was related at least partially to apoptosis reduction at the HSPC level, we used a monkey model of 5 Gy TBI. BM CD34⁺ cells were harvested 6 hours after irradiation, and we looked for mitochondrial membrane hyperpolarization as an early marker of active cell death.25 We brought evidence that 4F preserved CD34⁺ cells in vivo from RI apoptosis because no hyperpolarization occurred. On the contrary, BM CD34⁺ cells from untreated irradiated animals exhibited significant hyperpolarization in comparison with the preirradiation value, accounting for the initial phase of apoptosis. These data and those on the cell cycle sustain the hypothesis that the predominant effect of 4F treatment during the first 24 hours after irradiation consists of apoptosis reduction rather than cell proliferation.

The benefit provided to mice by SFT3 was also demonstrated in comparison with TPO-treated mice, which argues in favor of an additive or synergistic effect of the combination. In fact, TPO-treated mice in our model exhibited a survival rate consistent with that shown by Mouthon et al,12,32 who administered TPO 2 hours after TBI. By contrast, Pestina et al15 and Abushullaih et al33 obtained a higher survival rate by giving TPO immediately after lethal myelosuppression, which is unfortunately incompatible with the medical management of accidentally irradiated victims.15,33 Yet, their result is in agreement with the benefit that should be expected from the early counteraction of apoptosis. Preliminary data with the 4F combination in irradiated monkeys suggest that SFT3 can sustain multilineage recovery, including lymphopoiesis (F.H., M.D., manuscript in preparation). In the study by Neelis et al,31 single administration 2 hours after TBI proved to be significantly more efficient than treatment at 24 hours; therefore, we evaluated the efficacy of SFT3 given only at 2 hours after TBI. The benefit conferred by 2 injections of 4F versus a single injection was not actually significant. Interestingly, the dose administered appeared crucial, as did the multiple cytokine composition of the combination. In our model, dose reduction of SFT3 to 10 μg/kg or IL-3 withdrawal (2-hour and 24-hour schedules) only allowed for reduced survival rates. Finally, SDF-1 alone was less efficient than TPO, but recent studies suggest that SDF-1 bioavailability has to be improved before development for clinical use.39 

In this study we then focused on long-term follow-up to assess whether the cytokines selected and their combinations could behave as long-term protectors. In fact, if numerous studies conducted in mice have accounted for long-term follow-up,40-43 data on the effects in aging irradiated animals of early cytokine administration are lacking. In our study, delayed death mainly observed from day 150 after TBI was likely to occur from hematologic failure (hematopoietic and thrombotic disorders) and nonhematopoietic defects (fibrosis-associated multiorgan failure). In the Barnes et al42 study, most late deaths of untreated irradiated mice occurred between 9 and 32 weeks after mid-lethal X-irradiation (earlier than in our treated irradiated mice), and the most frequent syndromes observed were liver necrosis, diarrhea, pleural effusion, and aplastic anemia. In our study, no predominant etiology for death was observed. In particular, oncogenesis, which could be expected to be highly increased after antiapoptotic treatment, affected only a minority of long-term survivors, at least from a macroscopic analysis. From a clinical point of view, these survivors were comparable to healthy animals regarding body weight, food intake, and locomotion. However, when compared with UC, hematopoiesis was impaired in most of the treated animals, as shown by the reduction in BM TNC and clonogenic cell counts. This impairment was observed for all treated groups but appeared greater for 5F than for TPO, 4F, and SDF-1. These results are consistent with lasting impairment of marrow stem and progenitor cell compartments previously reported in various species following irradiation.40,41,43 44 

BM CFU depletion could be explained by 3 mechanisms: (1) RI apoptosis, because our in vitro results showing that SFT3 protects approximately 12.3% of CD34⁺ cells irradiated at 4 Gy from apoptosis13 support the hypothesis that such cytokine combinations prevent apoptosis only partially in vivo; (2) partial stem cell exhaustion following cytokine-induced overstimulation.31 In fact, whole-body survival provided by multiple cytokines involves the production of short-term radioprotective progenitors that occurs, in the range of lethal radiation doses, at the expense of long-term stem and progenitor cells. As proof of the contribution of radioprotective progenitors, we clearly observed 2 striking waves of platelet production 8 and 23 days later, issued respectively from megakaryocytes and megakaryoblasts in 5 Gy γ-irradiated cynomolgus monkeys receiving a single 4F injection 2 hours after TBI. In contrast, irradiated control animals displayed no response, but a transient profound phase of thrombocytopenia (F.H., M.D., manuscript in preparation) (3) reduced effectiveness of residual stem cells.41,45 One can wonder whether the level of HSPC depletion could be related to the multiple cytokine composition of the combination (4F and 5F) or short-term treatment and whether in this context high survival rates paradoxically are obtained at the expense of the HSPC compartment. We found that 5F-treated animals exhibited increased BM megakaryocytes. This result is consistent with the megakaryopoietic activity of SDF-1 reported by some authors.39 Finally, we showed that despite delayed death occurring from day 150 on, half the protective effect of selected cytokines was preserved at 300 days.

A crucial issue is the selection of the ideal treatment for patients who have sustained severe radiobiologic injury. In our study, the short-term injection of multiple cytokine combinations was well tolerated and proved to be more efficient than single factors in promoting short-term survival in mice. It is possible that the administration of factors such as protease inhibitors, p53 chemical inhibitors,46 and negative regulators of hematopoiesis, with or without sequential cytokines such as antiapoptotic factors followed by colony-stimulating factors, may result in improved long-term hematopoiesis.

We thank S. Doulay, C. Dupuy, N. Grenier, V. Leroux, and L. Crépin for their technical assistance.