Activation Induced Cytidine Deaminase (Aid) Acts As a Gate Keeper in Pro-B Cells and Prevents PB-ALL

Introduction: Activation induced cytidine deaminase (AID) mediates somatic hypermutation and class switch recombination in splenic germinal center B cells and is implicated in retaining central B cell tolerance in the bone marrow (BM) (Cantaert et al., Immunity, 2015). Moreover, there is recent in vitro evidence that AID is upregulated in precursor B cells after exposure to LPS, contributing to the clonal evolution of pB-ALL (Swaminathan et al., Nat Immunol, 2015) (Greaves M. and Müschen M., Cancer Discovery, 2015). These studies were carried out in pre-BII / early immature B cells, which are the first B cell compartments with detectable intrinsic AID expression. However a functional role of AID in pro-B cells is still controversially discussed and a functional role of AID in leukemogenesis remains speculative. We designed an in vivo model which allowed us the investigation of intrinsic Aid expression in tumor prone pro-B cells. Our data indicate that Aid is a gate keeper at the early stage of B cell development and its loss of function facilitates the development of pB-ALL.

Methods: We crossed a Rag1 deficient tumor prone mouse model (p19Arf^-/-/Rag1^-/-) (Hauer et al., Blood, 2011) on an Aid deficient background to obtain Aid knockout (p19Arf^-/-/Rag1^-/-/Aid^-/-) and heterozygous (p19Arf^-/-/Rag1^-/-/Aid^+/-) mice. Healthy and diseased mice were characterized by immunohistochemistry, Flow cytometric analysis, genome and transcriptome profiling. Cell cycle analysis was performed with pro-B cells of healthy mice.

Results: P19Arf^-/-Rag1^-/- mice display a B cell developmental arrest at the pro-B stage and develop pB-ALL at an incidence of 26 %. Surprisingly, an additional loss of Aid in these cells accelerated the pB-ALL incidence to 98 % (44/45, median onset 25 weeks). Moreover our model reproduces the dose dependent effect of AID on regulating B cell tolerance in humans, since Aid^+/- mice on the same background displayed significant disease reduction (83 %, 15/18, median onset 33 weeks, Mantel-Cox Test p=0.0175). The leukemia displayed a pro-B cell phenotype (CD19⁺B220⁺ckit⁺IgM^-) and manifested with splenomegaly, dissemination of blast cells to the BM, peripheral blood (PB) and spleen. Pro-B tumors from p19Arf^-/-Rag1^-/- mice expressed Aid on transcript (qRT-PCR) and protein (western blot) level, indicating that Aid expression is not restricted to CD19⁺ BM cells with co-expression of a functional IgM heavy chain product but rather occurs at earlier stages of B cell development. Again this effect was dose dependent, since in pB-ALLs of p19Arf^-/-Rag1^-/-Aid^-/+ mice Aid expression was significantly reduced. To identify the second hit we performed whole exome sequencing of murine tumors, which revealed accumulation of recurrent somatic Jak3 (R653H, V670A) and Dnm2 (G397R) mutations. To extend these findings further, Sanger sequencing of these regions displayed a mutational pattern of somatic Jak3 mutations in 60 % of Aid^+/- and 80 % of Aid^-/-pB-ALLs, while Dnm2 was somatically mutated in 96 % of all pB-ALLs analyzed. The detected Jak3 variants are known to induce a constitutive active downstream signaling. Loss of function mutations in DNM2 can increase the IL-7R cell surface expression, which highlight the relevance of the IL7R signaling in the context of tumor progression. However we did not observe detectable Aid expression in healthy pro-B cells of p19Arf^-/-Rag1^-/- animals in line with findings from Cantaert et al. On the other hand loss of Aid expression accelerates the repopulation capacity starting at the pro-B cell compartment (Kuraoka et al., Proc Natl Acad Sci, 2011). In our model Aid loss produces a dose dependent increase in proliferation and BrdU assays of B220⁺ sorted pro-B cells of healthy mice from the different cohorts (30 % cells in S-Phase in p19Arf^-/-Rag1^-/-compared to 50 % S-Phase with additional Aid loss), although Aid expression is below the detection limit.

Conclusion: We present in vivo evidence that Aid has a gate keeper function in pro-B cells, which allows aberrant IL-7 dependent pro-B cells without a functional receptor to be eliminated through Aid induction. This further extends the observation that Aid mediates the clearance of autoreactive early immature B-cell clones and is required to prevent pB-ALL. In this regard Aid overexpression but also loss of Aid expression can facilitate pB-ALL development.