Interleukin-15 (IL-15) is critical for the development and homeostasis of natural killer and memory CD8+ T lymphocytes.However, when expressed as transgene, IL-15 can cause fatal leukemia.IL-15 also facilitates the activation of autoreactive CD8+ T cells and promotes inflammatory and autoimmune diseases.We have recently shown using the non-obese diabetes (NOD) mouse model that IL-15 plays a pathogenic role in autoimmune type 1 diabetes by activating diabetogenic CD8+ T cells.Adoptive transfer of splenocytes from diabetic 8.3-NOD mice, which express a transgenic major histocompatibility complex class-I-restricted autoreactive T-cell antigen receptor (TCR) called 8.3-TCR, induced type 1 diabetes in NOD mice with severe combined immunodeficiency (NOD/SCID) but not in IL-15-deficient NOD/SCID (NOD/SCID.Il15-/-) mice.Intriguingly, the NOD/SCID.Il15-/- recipients showed massive enlargement of lymph nodes and spleen with high cellularity and became sick within 12 wk, characterized by kyphosis and piloerection.These NOD/SCID.Il15-/- recipients showed massive mononuclear cell infiltration of lymph modes, liver and kidneys, and abundant lymphoblasts with large nuclei in the peripheral blood.These findings indicated that IL-15 plays an important role in preventing leukemia development.Adoptively transferred splenocytes from 8.3-NOD donors populate the lymphoid organs of NOD/SCID recipients with CD8+ and CD4+ T cells, which express the transgenic 8.3-TCR.In contrast, leukemic cells accumulating in NOD/SCID.Il15-/- mice showed CD4hiCD8lo phenotype and did not express the 8.3-TCR (Figure 1d, lower panels), suggesting that the IL-15-deficient lymphopenic environment in NOD/SCID.Il15-/- mice gave rise to abnormal CD4+CD8loTCR- cells with leukemic properties.To test whether these leukemic cells outnumbered autoreactive CD8+ T cells, the leukemic cells were adoptively transferred to NOD/SCID mice along with splenocytes from diabetic NOD donors.The leukemic cells not only prevented type 1 diabetes in NOD/SCID recipients but also expanded robustly, causing lymphadenopathy, splenomegaly, kidney enlargement and death within 4-6 wk.The expanding cells retained the CD4+CD8int phenotype, expressed CD5 but not major histocompatibility complex-I or B220.These observations indicated that the leukemic cells outnumbered the diabetogenic CD8+ T cells and retained aggressive growth potential even in an IL-15-sufficient environment.As the leukemic cells showed CD4+CD8int phenotype, we examined whether they originated from CD4+ or CD8+ donor T cells.For this, we compared total splenocytes and purified CD4+ or CD8+ T cells for their ability to induce leukemia in NOD/SCID.Il15-/- recipients.Total splenocytes from pre-diabetic (3 wk-old) and diabetic 8.3-NOD donors induced leukemia within 12 wk after adoptive transfer.Surprisingly, purified CD4+ T cells from 8.3-NOD donors caused leukemia within 6 wk, whereas CD8+ T cells failed to cause disease.CD4+ T cells, but not total splenocytes, from non-TCR transgenic diabetic NOD mice also induced leukemia, albeit at a lower frequency and delayed kinetics.Similar to leukemic cells induced by total splenocytes, those arising after the transfer of 8.3-NOD CD4+ T cells showed the CD4+CD8int TCR- phenotype.We maintained the leukemic cells that developed in NOD/SCID.Il15-/- mice in NOD/SCID recipients by serial passage, which increased their pathogenic potential, causing rapid mortality even in wild-type NOD mice.These leukemic cells rapidly colonized lymph nodes, liver and kidneys and occasionally the thymus in recipient mice.Cells colonizing the thymus also displayed the CD4+CD8loCD3/TCR- phenotype and expressed CD25.These cells readily grew in vitro, from which we established the SID-T-cell line that retained the surface phenotype.We established several other leukemic cell lines directly from lymph nodes and spleen of NOD/SCID.Il15-/- recipients without passaging in NOD/SCID mice.The leukemic cell lines grew exponentially without requiring any cytokines or growth factors.Therefore, we examined whether these leukemic cells displayed constitutive activation of signaling mols.Murine T-leukemic cell lines EL4 and P1798, IL-2-dependent CTLL-2 and naive CD3+ T cells served as controls.All the leukemic cell lines displayed increased protein tyrosine phosphorylation and elevated expression of LCK and FYN kinases.As CTLL-2 cells showed STAT5 phosphorylation (induced by IL-2), STAT3 and ERK, the leukemic cell lines only showed moderate activation of STAT3.However, these cell lines (except SID-T) showed AKT phosphorylation similar to EL4 and P1798.These findings suggested that constitutive activation of diverse signaling mols. might contribute to leukemogenesis in NOD/SCID.Il15-/- recipients.Human T-acute lymphoblastic leukemia (T-ALL) cells often overexpress the MYC proto-oncogene downstream of NOTCH1, which is activated by gain-of-function mutations.NOTCH1 activation also occurs in mouse T-ALL with concomitant upregulation of MYC.MYC gene activation can also occur by t(12:15) chromosomal translocation in T-ALL.Among the leukemic cell lines that we have established, only SID-T showed elevated MYC expression comparable to EL4 thymoma, suggesting that Myc gene activation is unlikely to be a key mediator of leukemogenesis in NOD/SCID.Il15-/- recipients.Next, we carried out karyotype analyses by multiplex fluorescence in situ hybridization.Even though many of the leukemic cell lines showed t(12:15) translocation, they also harbored other chromosomal abnormalities such as 1p+ and t(10:13) translocation.Clearly, the t(12:15) translocation did not always result in increased MYC expression.These data suggest that multiple genetic and signaling abnormalities may contribute to leukemia induction in NOD/SCID.Il15-/- mice.Defective thymic maturation in NOD mice has been reported to facilitate thymoma development under lymphopenic conditions caused by SCID mutation (67% at 40 wk) or recombination activating gene deficiency (44% at 25 wk), raising the possibility that the leukemic cells in NOD/SCID.Il15-/- recipients might have originated from host cells.To test this idea, we examined the leukemic cell lines for the presence or absence the Il15 knockout allele, due to the lack of donors and recipients bearing different allelic markers.Surprisingly, all the leukemic cell lines harbored the Il15 knockout allele, indicating that they arose from the recipient's cells.Further supporting the host origin of leukemic cells, long-term observation of NOD/SCID mice lacking either IL-15 revealed that they all developed spontaneous leukemia over a period of 32 wk, whereas control NOD/SCID mice did not.These spontaneously developing leukemia also gave rise to leukemic cell lines.It is noteworthy that NOD/SCID.Il15ra-/- mice lacking the ligand binding α-subunit of the IL-15R complex also developed spontaneous leukemia that was accelerated by adoptively transferred of splenocytes from diabetic 8.3-NOD mice.The SID-15RaL cell line derived from NOD/SCID.Il15ra-/- recipients, harbored the wild-type Il15 allele as expected.Finally, we examined the expression of terminal deoxynucleotidyl transferase (TdT), a widely used marker for T-ALL.TdT, involved in DNA repair during TCR gene rearrangement, is expressed in CD4+CD8+ double pos. thymocytes, but is repressed in CD4+ or CD8+ single pos. T cells.Similarly to EL4 and P1798 cells, all leukemic cells expressed TdT.Collectively, our findings indicate that lack of IL-15 signaling in NOD/SCID mice due to the absence of IL-15 or IL-15Rα facilitated leukemogenesis of T-cell precursors within the recipients.Animal models play an important role in deciphering mol. mechanisms of leukemogenesis, and in developing new drugs and treatment methods.In this study, we have shown that NOD/SCID mice lacking IL-15 or IL-15Rα develop leukemia with 100% penetrance within 32 wk, and this process is accelerated by adoptively transferred CD4+ T cells (100% within 12 wk).Moreover, the leukemic cells from NOD/SCID.Il15-/- mice readily grow in vitro without requiring in vivo passage.Overall, the leukemia model described in our study provides a unique opportunity to characterize the genetic determinants and non-genetic modifiers that promote leukemia development and to better understand the protective role of IL-15 against precursor T-cells that harbor or develop a potential to cause leukemia.Indeed, constitutive expression of IL-15 by thymic stromal cells, raise the possibility that IL-15 signaling within the thymic microenvironment may prevent neoplastic transformation of thymocytes, either directly or via attenuating the induction or upregulation of oncogenic growth factors.Leukemia relapse following bone marrow/stem cell transplantation remains a challenging issue, and IL-15 has been suggested as a potential treatment option.Indeed, IL-15 has been shown to improve immune reconstitution of NK, NK-T and CD8+ T cells following bone marrow transplantation, and can be safely administered to patients.We posit that the conditioning treatment used to suppress the immune system before bone marrow transplantation may promote aggressive growth of leukemic cells in patients with defects in the IL-15 signaling pathways.Even though our study was carried out in an unconventional exptl. setting, this leukemia model could be used to further understand the protective functions of IL-15 and to refine the IL-15-based therapeutic strategies to prevent leukemia relapse following bone marrow transplantation.
