In the present work we established a new approach to generate fully human mAbs. We immortalized B cells from vaccinated HIS mice by transduction with BCL6 and BCL-XL followed by expansion in presence of IL-21 and CD40L. Antigen-specific B cell clones were obtained that expressed the BCR on their cell surface and secreted antigen-specific antibodies. Similarly to methods based on the immortalization of human memory B cells from individuals that were either vaccinated or exposed to pathogens, our strategy exploits the antibody repertoire of human B cells which is likely to be different from that of B cells of mice expressing human Ig gene segments.
Naïve HIS mice display an extensive human IgM-expressing B cell repertoire. Based on the analysis of the length of the CDR3 regions, this IgM B cell repertoire is similar to the repertoire of healthy individuals. Thus, HIS mice have no obvious limitations for the generation of human IgM mAbs against any possible antigen. Upon intramuscular vaccinations with either TT or HBsAg, approximately 40% of the HIS mice were able to mount an antigen-specific antibody response. Human IgM-producing B cell lines against both antigens were obtained after isolation of memory B cells followed by ex vivo differentiation into plasmablast-like cells. It is important to highlight that the selection of the antigen-specific human B cell clones relied on relevant bioassays (e.g., ELISA or neutralization test). In contrast to EBV-based approaches, human B cell immortalization using transduction with BCL-6 and BCL-XL preserves the expression of the BCR at the surface and antigen-specific B cell clones can also be selected by binding of labeled antigen to the BCR of immortalized memory B cells (e.g. by using a labeled antigen).
Even when IgG was used as a selection criterion, we were unable to establish antigen-specific IgG+
human B cell clones. The reason for this might be that T cell help in this system is suboptimal as indicated by the absence of antigen-specific T cell responses after vaccination (not shown). We also observed that the BCR of the B cell clones had a close to germ-line sequence, suggesting that also the induction of somatic hyper-mutation is sub-optimal in HIS mice. In our hands the great majority of the vaccinated HIS mice showed a defective formation of germinal centers 
, which further explains the absence of antigen-specific Ig-class switched B cells. So far, humanized mouse models based on the transplantation of human HSC only – i.e. without additional human tissues – share these limitations, and immunization strategies result in the limited generation of class-switched antigen-specific B cell responses 
. Similar patterns are observed in human HSC-transplanted immunodeficient mice infected with lymphotropic pathogens, such as HIV 
or EBV 
, although Dengue virus infection in HIS mice was reported to induce an IgG response in a majority of the responder animals 
. It is not clear why IgG antigen-specific responses are limited while serum IgG can accumulate efficiently, considering the low frequency of IgG+
B cells in HIS mice. It remains to be determined whether this apparent discrepancy might be explained by the conjunction of particularly effective IgG production on a cell basis by IgG+
B cells (which might occur in a T cell independent manner, such as in the case of the IgG3 subclass), long-term stability of human IgG in the HIS mouse serum as compared to human IgM, and/or defective survival of IgG+
B cells under specific conditions (e.g. after antigen-specific triggering of the BCR).
Although IgM mAbs might already be useful for some specific applications or could be modified by Ig class swapping to obtain IgG mAbs 
, optimized humanized mouse models with improved B cell function are highly desirable. One reason for the suboptimal interaction of T and B cells may be the poor survival resulting in a high turnover of human T cells (discussed in 
), making it very likely that procedures leading to improved accumulation of human T cells may promote B cell responses and isotype switching. It was already shown that human B cells undergoing isotype switching can be obtained in humanized mice, provided that a human environment supporting this process is present, e.g. in SCID mice transplanted with human fetal bones, thymus and lymph nodes 
. Consistent with this notion, enhanced human peripheral T cell accumulation was observed in NOD/SCID mice transplanted with human bone marrow HSC, fetal liver and fetal thymus tissues (referred to as BLT mice), as compared to conventional humanized mouse systems 
. Interestingly, BLT mice consistently generated an antigen-specific IgG response after HIV-infection 
. Although it is yet unknown whether the isotype switch observed in BLT mice is truly T cell dependent, those data might support the idea that improved T cell homeostasis has a positive impact on B cell responses. To obtain humanized mouse models with improved B and T cell homeostasis, alternative strategies not relying on the transplantation of human fetal tissues – which are not necessarily easy to access to, for ethical, legal or practical reasons – will likely be favored in the future. The replacement of mouse genes involved in the hematopoietic system by their human equivalent is a valuable strategy to improve development, maintenance and/or function of several hematopoietic cell subsets in humanized mouse models, as shown with cytokines, such as IL-7 and Il-15 
, or MHC molecules (N.D.H and J.P.D., manuscript submitted) 
. The fact that the human CD47 was shown to be unable to properly interact with the mouse SIRPα indicates that re-introducing a functional phagocyte inhibition mechanism via the CD47/SIRPα signaling axis is another strategy of potential interest 
In conclusion, our results show using two standard vaccine antigens the general applicability of an innovative B cell immortalization method in combination with the HIS mouse model to generate human mAbs. Similarly to methods based on the immortalization of human memory B cells from vaccinated or convalescent individuals 
, our approach exploits the broad antibody repertoire of human B cells, overcoming the potential limitations of conventional humanized murine mAbs such as laboriousness or impaired biological properties, synthetic antibody libraries that require a known target antigen, and transgenic mice bearing the human Ig locus that have limited B cell repertoires. In addition, our method enables to exploit experimental infection models and immunization regimes that would be unethical or untenable in humans. Considering the upcoming advances in HIS mice models 
, this new approach will provide a powerful tool to generate human mAbs for either diagnostic or therapeutic purposes.