Introduction

Hematopoietic stem cells (HSC) cannot be distinguished from Hematopoietic Progenitors (HP) by phenotypic and/or molecular analysis. Indeed, even though human CD34⁺CD38^-/lowCD90⁺CD45RA^- hematopoietic cell population [1] is very enriched in stem cells, it still contains a non-negligible proportion of HP. In vivo xenotransplantation assays are thus mandatory to reveal and quantify human HSC. This test is based on the unique functional capacity of HSC to durably engraft recipients in which they self-renew, expand and give rise to a progeny of more mature hematopoietic cells. To address human HSC, CB17-SCID and NIH Beige-Nude-Xid mice strains were used up to 1995. Afterwards, NOD/SCID model, exhibiting a more severe immune deficit due to the combined alteration of NK cells, complement and macrophages, was used thereafter. The recently developed NOG (NOD/Shi-scid/IL-2Rγ^null) and NSG (NOD.Cg-Prkdc^scid Il2rg tm1Wjl/SzJ) mouse strains are now considered as the most efficient model for studying human HSC [2-6], commonly named “Scid Repopulating Cells” (SRC).

The immuno-deficient models are able to evidence the functional heterogeneity of HSC population that contains short term (ST-) and long term (LT-) SRC. These functional SRC subpopulations are distinguished by the period between cell injection and analysis of their progeny (6-8 weeks and 12-14 weeks for ST- and LT-SRC, respectively) or by using serial transplantations (limited primary or serial (secondary and more) engraftment capacity for ST- and LT-SRC, respectively). This xenotransplantation procedure, based on the functional characteristic and definition of HSC, is presently the gold standard to study these cells. Successful transplantation of HSC requires a partial or total suppression of the recipient hematopoiesis. One of the critical point is represented by the pre-conditioning of the animals before they receive infusion of cell suspension. Indeed, it facilitates the homing of HSC in specific BM niches and their later self-renewal, proliferation and differentiation inside the whole bone marrow space by a combined suppression of endogenous bone marrow hematopoiesis and stimulation of a major feedback hematopoietic regeneration process. Despite the fact that it is tedious to organize, sub-lethal or lethal X or Gamma rays irradiation is the most common conditioning regimen. Alternatively, some authors proposed recently to use chemotherapeutic agents such as busulfan and 5-fluorouracil that induce similar hematopoietic effects [7,8] while being easier to set-up and less expensive than irradiation.

Despite the use of the recently developed NOG (NOD/Shi-scid/IL-2Rγ^null) and NSG (NOD.Cg-Prkdc^scid Il2rg tm1Wjl/SzJ) immunodeficient mouse strains that lowered the engraftment threshold [2,6], the use of non standardized procedures leads to frequent inter-laboratories variations. Based on the recent conditioning regimen that proposes two i.p. injections of busulfan 48 and 24 hours before IV injection of HSC [8], we tried to reduce the time schedule constraints of this procedure and to delineate in the meantime the best procedure to address human cord blood CD34+ HSC. For this purpose we compared the NSG engraftment by using the classical procedure or modifications targeting busulfan dose, gender of animals, injection route, and the period between NSG mouse conditioning and cell injection. Among the results we obtained, we showed most interestingly that prolonging the period up to 7 days between conditioning busulfan regimen (25mg/Kg (x2)) and cells injection did not decrease the HSC engraftment.

Design and Methods

Results

Discussion

The Immuno-deficient NSG mouse strain is now widely used for the in vivo detection and quantification of human hematopoietic stem cells since it allows an enhanced engraftment compared to the “classical” NOD/SCID mice xenotransplantation model [2-6]. The first step of the procedure is to “condition” the receiver before the injection of human cells in order to ensure the best conditions for supporting the BM homing and development of HSC. Total body non-lethal irradiation of mice with x- or γ-rays, the most common conditioning regimen triggers several mechanisms, including the secretion of stem cell factor (SCF) which is critical for HSC engraftment, proliferation and survival [10]. However, it requires complex and expensive aseptic handling procedures (since irradiators are commonly located out of the animal facilities) and specific materials (manipulation, transfer and transportation of animals in sterile conditions) to protect these immunodeficient animals from microbiological agents. The use of such pre-conditioning procedure is thus very stressing for animals. From that point of view, the conditioning of NSG mice by busulfan injections, represents an interesting alternative to irradiation [8]. Moreover, it allows adapting the dose of drug delivered according to each animal weight whereas irradiation (effected in compartmentalized cassettes) could result in slightly different absorbed doses if the individual mice are of very different weight. Based on this approach, we chose to test the effects of different doses of busulfan on NSG mice sensitiveness and engraftment level upon injection of cord blood CD34⁺ cells. We confirmed the data from Robert-Richard et al. [8] obtained with NOD/Scid mice since we showed that 2x25 mg/kg Busulfan provided the optimal huCD45 BM chimerism in mice. However, the fact that busulfan dose did not influence the number of huCFU in recipients’ bone marrow (Figure 1E) suggests that the conditioning could be realized in a relatively large dose-range without affecting the engraftment of more immature SRC. This point is supported by two studies showing that mouse pre-conditioning regimen might not be critical for long-term SRC engraftment even in particular contexts in which the combination of the robustness NOD/SCID IL-2Rgamma-/- (NOG) mouse strain and the sensitivity of intra bone marrow transplantation technique is mandatory [11] or when classical NOD/SCID mice received a multiple-day injection of cells [12]. We also found that males were more sensitive to lower doses of busulfan treatment than females (weight loss (16% vs 10%) and mortality (11% vs 8%)) (Figure 1A to C). In spite of fact that these data are not original they confirmed those from Notta et al. [13] showing that females were more “permissive” than males to engraftment of human cells (Figure 4) whatever the cell number injected (from 100 to 1000 cells). Here, our data suggest that the 100 CD34⁺ cord blood cells dose corresponds to the limiting dilution for NSG males since around 50% of mice were negative for huCD45 whereas more than 95% of females were positive. The lower level of the limiting dilution threshold in females argues for a better sensitivity of female NSG mice to human HSC engraftment. This result is completely in accordance with previous published study in which the very SRC-enriched population “Lin^-CD34⁺CD38^-CD90⁺CD45RA^-“ was injected using the strongly efficient intra femoral route in sub-lethally irradiated NSG mice. In this context, the authors also showed that females were better engrafted than males [13], evidencing thus that this difference between males and females mice in terms of engraftment sensitivity persists whatever the injection route, pre-conditioning technique or SRC enrichment. In addition to the better immunodeficient status of females compared to males, and some sex-associated factors, such as steroid hormones that can positively or negatively regulate human HSCs considered by these authors [13], we propose that the weight of animal could also be involved in this phenomenon. Indeed, even we usually used males and females having the same age, we observed a 16% higher weight of males compare to females (mean weight was 24.4g (n=75) and 20.3g (n=40) for male and female cohorts respectively). In this situation, even if the same number of cells per mice was injected, the cell dose/kg of body weight is in fact, lower in males. This could partly explain the difference in engraftment level. Taken together, our data presented in Figure 1 led us to use female mice conditioned with 2 injections of 25 mg/kg of busulfan in most of our further experiments. Very interestingly, using these conditions, we were able to increase the delay between pre-conditioning and cell injections from 24 hours up to 7 days without significantly altering the huCD45 chimerism, the balance between myeloid and B lymphoid cells and the huCFU number. This observation is of major practical interest because it largely increases the experimental time-schedule flexibility that is frequently requested in human SRC experiments. Intra venous, intra bone marrow and intra peritoneal injections of human cells in NOD/SCID, NOG or NSG mice are described in the literature for SRC studies [14]. Here, we confirmed that the huCD45 chimerism in mice bone marrow is not significantly affected by the type of intra venous injection i.e intra caudal and retro orbital routes (Figure 3) [15,16].

However we noticed that 2 mice out of 16 injected in the caudal vein did not show any human chimerism whereas all the 21 mice injected in the retro orbital sinus exhibited human engraftment. Moreover, the significantly higher number of huCFU detected in the bone marrow of recipients injected in the retro orbital sinus compared to the one in the caudal vein suggests that this route allows a better HSC homing or “nichage”. Despite frequent subjective reluctances of researchers to its use, the retro-orbital injection route is in fact less painful and stressful for animals [16,17] than alternative intravascular injection techniques. Indeed, the technically challenging injection in the caudal vein leads to some failures even for skilled experimentators (2 non engrafted mice in our study). Moreover, this procedure requires the heating of mice (under a lamp or a slightly heating block) to promote peripheral vasodilation that facilitates the vein catheterism and cell injection [17] and it causes distress in the nonanesthetized animals, especially if the initial venipuncture is unsuccessful and repeated attempts are made. Taken together our data lead us to propose that SRC studies have to be performed with female NSG mice, conditioned with 2 IP injections of 25mg/kg of busulfan before CD34+ cells injection in the retro orbital sinus. Very interestingly, we provide the first comparative study which demonstrates that increasing the usual 24h period between conditioning and cell transplantation of NSG mice up to 7 days resulted in similar percentages and types of BM chimerism at 8 weeks post transplantation. Our data could facilitate the development of in vivo studies with human CD34⁺ cells by improving the flexibility of experimental protocols without altering the reliability of the results.

Acknowledgments

The authors would like to thank Dr B. Rousseau and the staff of A2 animal facility of Bordeaux Segalen University. We are also very thankful for the staff of Cellular Therapy Department of EFS-AqLi who provide us cord blood samples. We finally thank Mrs Elisabeth Doutreloux-Volkmann for the reading of the manuscript.