Differential susceptibility of C57BL/6NCr and B6.Cg-Ptprc^a mice to commensal bacteria after whole body irradiation in translational bone marrow transplant studies

Bone marrow transplantation (BMT) is a common therapy for many neoplastic and non-neoplastic conditions [1]. Current research is focused on the development of more efficient transplant regimens and therapies [1]. Before receiving a donor bone marrow (BM) graft, BMT patients undergo preparatory regimens (which often include irradiation) to deplete the host BM. Complications of BMT include viral, bacterial or fungal infections secondary to bone marrow ablation, and relapse of neoplastic disease [2, 3]. Few human patients die directly from the preparatory regimens (chemotherapy, radiation) because of the availability of supportive care. However, debilitated high risk transplant patients require careful titration of the myeloablative therapy to prevent multiorgan disease [4, 5]. Some of these complications are related to failure of engraftment. One of the most deleterious side effects of allogeneic BMT (when donor and recipient are genetically different) is graft-versus-host disease (GVHD) [6]. Grafted donor cells recognize the host (recipient of BMT) tissues as foreign leading to GVHD. Acute GVHD is defined as a disease appearing within 100 days post-BMT which targets the skin, liver, lung, thymus and secondary lymphoid tissues [7].

The mouse is one of the most utilized animal models for translational BMT studies. With the development of inbred strains of mice, many discoveries in the field of immunology, immunogenetics, and radiation oncology have taken place [8]. Inbred strains of mice permit the study of BMT in genetically identical and genetically variable systems. Transplantation between animals within the same strain allow for modeling of autotransplants or transplants between genetically identical twins (syngeneic). Genetically different donor and recipient strains allow modeling of transplantation between genetically different individuals (allogeneic). The mice in this report came from a number of BMT studies conducted by three laboratories where they utilize inbred strains for both syngeneic and allogeneic BMTs. There are three groups of BMT recipient per experiment: 1) syngeneic bone marrow recipients (syngenic BMTs) in which all animals are expected to survive. 2) allogeneic bone marrow recipients (allogenic BMTs) in which animals are expected to eventually succumb to GVHD, and 3) experimental allogenic transplant recipients (experimental transplants) which receive allogeneic bone marrow with genetically induced modifications or treatments thought to prevent or decrease GVHD. Experimental animals are expected to have improved survival compared to the allogeneic control groups.

In general, mice irradiated for BMT studies experience lethargy and malaise of one week duration due to radiation induced multisystem organ damage. Commonly, all animals, whether syngeneic or allogeneic survive the post-irradiation period. Syngeneic animals usually recover within two weeks, while allogeneic groups experience a biphasic illness. After the first week of irradiation sickness, allogeneic recipients undergo a relatively short recovery period prior to becoming sick for the second time due to GVHD. Animals developing GVHD, die as early as 12–14 days post-BMT and as late as 30–40 days post-BMT depending on the degree of genetic mismatch. Mice developing GVHD may exhibit diarrhea, skin lesions (ulcerations) and overall lethargy and anorexia as a consequence of the damage to the classical GVHD target organs (skin, gastrointestinal tract, liver, thymus) [9] similar to man.

In our studies, we observed an unexpected increase in mortality very early after transplantation. One of the most striking findings was facial swelling in the affected animals. Premature deaths in experimental murine models of BMT cause significant complications in interpretation of the experimental results and delay it's translation to the clinic. Furthermore, it greatly increases expense due to the necessity of using large numbers of mice, as well as the effort of the investigators. These economic factors, the impact on experiment outcomes, and the humane considerations all indicate the need to minimize premature deaths and better understand the limitations of the animal models utilized for human research. In our studies because of the difference in the non-GVHD related early mouse deaths in both syngeneic and allogeneic BMTs, we carefully evaluated the environmental, genetic, treatment, and pathologic factors associated with early death in a large group of BM transplanted mice.

In our studies we document the clear differences in mortality between C57BL/6NCr (Ly5.1) and B6.Cg-Ptrpc^a/NCr (Ly5.2)C57BL/6Cr mice in BMT studies that utilized irradiation as the main immunosuppressive regimen for translational transplant studies. When syngenic animals (positive controls), which would otherwise survive, started dying, this made the transplants uninterpretable. The death of immunocompromised mice such as SCIDs and Nude (nu/nu) mice are common when exposed to bacterial, viral, parasitic or protozoal pathogens similar to immunocompromised individuals [13‐17]. However following BMT, despite transient immunosuppression, immunocompetent transplanted animals usually survive for greater than seven days after the immunoablative treatment [10, 18‐23]. If there is lack of engraftment the animals will eventually succumb to infection secondary to BM aplasia [24, 25] and anemia or thrombocytopenia. Premature death may occur if the BM inoculum is not successfully delivered, or the delivered cells are unable to engraft and repopulate the hemolymphatic system of the new host. In the cases reported here, all animals were given a BMT dose and were expected to survive the initial immunosuppressive period. Failure to deliver an appropriate dose of viable cells is unlikely because the intravenous tail vein injections were given by experienced individuals and the Ly5.2 mice that were not affected were given the same dose, by the same route, and by the same personnel as those mice that died prematurely.

Animals might die early if aged, debilitated or if burdened by a subclinical infection that was not identified prior to myeloablation. This effect has also been observed in the clinic with elder or very critical patients [4]. In our studies, the mice undergoing transplantation were 3–8 months old and healthy (based on NCI's health reports). As previously mentioned, mice that receive myeloablative conditioning prior to BMT survive the time period immediately after the preparatory regimen (5–7 days post irradiation). By day +14 post-BMT, there is immunological recovery which is demonstrated by normal peripheral immune reconstitution [26]. As is expected in in vivo studies, some intra-animal variation may occur, however if a mortality over 20% in the syngeneic controls is observed, one must re-assess the experimental protocol. Mice receiving allogeneic BM, and T cells may die of severe GVHD after 2 weeks post-transplantation. Depending on the T cell and BM dose and the degree of mismatch, GVHD may develop before 14 days post-BMT. However, 1) syngeneic controls should survive and 2) facial swelling should not be observed as it is not a feature of GVHD in mice.

The premature animal deaths in this study were independent of the transplant details (donor T cell source, personnel, housing, cell preparation, irradiation source, age and sex of recipient or housing facilities). Some mouse viruses have been documented to be immunosuppressive and impede stem cell differentiation [27‐32] possibly having an influence in immune reconstitution and engraftment. Serology of 177 naïve animals to be used at a later time as either donor or recipients, were negative for all bacteria and viruses including parvoviruses. The strain of the recipient and the dose of irradiation were the two predominant factors influencing facial swelling and mortality in these BMT models (Figure 3A,B).

We identified E. coli, Pseudomonas spp., Enterococcus and Streptococci spp in the blood and tissues. These pathogens have commonly been reported to affect immunocompromised animals [33]. In order to prevent infections in BMT mice some groups recommend treating the animals with antibiotics in the water in order to minimize bacterial contamination post irradiation. In our institution we utilize acidified (hyper-chlorinated) water with a pH = 2.9–3.0 [19, 20] and have rarely experienced problems in mice in these study protocols.

Based on previous studies [22, 23], gut damage in animals that are SPF (not germ free), have some degree of translocation of bacterial toxins (e.g. LPS) and possibly bacteria post-irradiation, which in turn causes a severe inflammatory response and severe hypotension. This is known to be mediated by TNFα and IL-1 inflammatory mediators [9]. However, and as previously mentioned, during uncomplicated bone marrow transplants, syngeneic animals recover successfully after this period. Based on the blood and tissue cultures and histopathology, we clearly document that the affected mice were severely septic. Even though death from sepsis may be expected in severely immunocompromised mice if challenged with a pathogen, early death from sepsis on transplanted animals, though possible, it is not common. In our studies, the systemic peri-BMT preparatory regimen caused a degree of damage where even syngeneic recipients were unable to successfully recover from the early post-transplant period. As documented, their deaths were likely caused by the translocation of gastrointestinal bacteria and their toxins into the circulation, and subsequently into other organ systems. Such effects continue to be an important co-morbidity factor in human BMTs [1].

Facial swelling has been documented in immunodefficient SCID mice after severe infections in their oropharyngeal regions [17, 33]. Therefore, the swelling in our animals can, in part, be explained by the autoinoculation of bacteria during their vulnerable immunocompromised period as described in SCID mice.

In our studies the dose of radiation was one major factor in causing the premature deaths. It has been documented that splitting a single radiation dose into two equal half doses causes less GI damage [34]. Based on our retrospective study, we observed that only the animals that died at the early time points post-BMT (4–8 days) and exhibited facial swelling, were the mice that received the highest radiation doses performed in the laboratories (1100–1300 cGy). High radiation doses such as 1000 cGy have been documented to be relatively toxic when given to larger species such as dogs and man [35, 36]. However, many strains and stocks of mice have historically been more resistant to irradiation (e.g. C3H strains) [37] while others are significantly more sensitive (e.g. BALB/c) [38] once again emphasizing the importance of genetics not only in BMT match/mismatches, but in the susceptibility to the preparatory regimens. These factors are also likely to impact human BMT outcomes.

In our retrospective analysis we have been able to identify that most of the animals affected have been C57BL/6NCr (Ly5.1) (Table 1 and Figure 3). B6.Cg-Ptrpc^a/NCr (Ly5.2) syngeneic recipients receiving 1000–1100 cGy irradiation doses were less affected (Table 2 and Figure 3) by irradiation and the differences were statistically significant (p < 0.001). This suggests that B6-Ly5.2 congenic mice are therefore more resistant to irradiation. One group of B6-Ly5.2 mice in Table 2 exhibited a 29% mortality. It is possible that 1100 cGy given in a single dose may be the radiation limit that Ly5.2 mice can withstand in our hands. We must mention that certain variability is always observed when performing in vivo experiments (even in age matched and using inbred strains) and there can be unknown factors which may impact (hopefully to a lesser degree) the experimental outcomes. It is possible in the case of Ly5.2 animals mentioned above that human error calculating the irradiation dose may have influenced the outcome of this experiment, especially if we assumed that we are at the cusp of the maximal radiation dose tolerated by Ly5.2 mice. It has been our experience that Ly5.2 animals that received 1300 cGy as a single dose had a 90% mortality. In any case, and based on the number of animals included in this study, we show that Ly5.2 mice are more resistant to radiation and that no mortalities are observed when either splitting the 1100 radiation dose or decreasing the overall amount of radiation. There are a plethora of studies that have documented the link between the susceptibility of ionizing radiation and the genetic makeup [37‐46] between different strains of mice and also man. For example DNA-protein kinases (DNA-Pks), which are important for DNA repair, were found to be diminished in BALB/c and 129/SvJ compared to C57BL/6 mice after being exposed to ionizing radiation [47]. In our specific study, we document for the first time a difference in sensitivity between two nearly genetically identical strains commonly used in biomedical research. Despite their nearly identical genetic makeup supported by immunological tolerance by the absence of GVHD [48] in allogeneic BMTs or rejection of skin grafts, there may still remain enough genetic disparity between the two strains to produce this observed difference in response to radiation. The mechanism of this differential sensitivity is unkown. It is possible that these two strains have differences in their DNA repair capabilities or radioprotective processes. It has been shown that the release of toxic nitric oxide radicals (NOr) can contribute to irradiation damage and that low NOr release post-irradiation is radioprotective [49]. It is also possible that Ly5.2 recipients may have decreased NOr release post irradiation compared to Ly5.1.

In conclusion, we have demonstrated for the first time a measurable difference in radiation sensitivity and its effects between two nearly genetically identical strains of mice commonly used for translational studies where C57BL/6NCr die of bacteremia more readily than B6.Cg-Ptprc^a/NCr congenics at radiation doses over 1000 cGy.

NIH/NCRR T32-RR07008-21A1 (R.D.S) and Unit for Laboratory Animal Medicine internal funds.