Mouse models of human diseases are crucial for our understanding of the mechanisms governing disease induction and progression, in particular for studies of cancer. The existing mouse models have some advantages but many important disadvantages. In this study, we compared oral and pancreatic tumor studies in NSG and hu-BLT mice. In addition, we compared the immune function of humans to hu-BLT mice to learn about the differences and similarities of the systems.

The NSG mouse model is the most frequently used mouse model for tumor studies mainly due to good engraftment of the tumors, ease of use and price point; however, as shown in this study, it is far from being the most appropriate model due to a lack of functioning immune cells, in particular, NK cells, monocytes, dendritic cells, and T and B cells. Immunodeficient mice such as the NSG do not recapitulate the complexity of the tumor microenvironment or the exact dynamics of tumor growth, since immune–tumor interaction shapes not only the growth dynamics of the tumor, but also its differentiation and survival, none of which takes place in NSG mice. To demonstrate the differences, we implanted oral and pancreatic tumors in NSG and hu-BLT mice and then compared the dynamics of tumor growth, rate of expansion, and differentiation among these two mouse models. Tumors grew much faster, and larger tumors formed in NSG mice when compared to hu-BLT mice. When tumors were dissociated, and single-cell tumor cells were cultured, the rate of expansion was much greater for tumors obtained from NSG mice than those from hu-BLT mice. Indeed, the rate of fold expansion for oral tumors ranged from 4.5–9.5 for NSG mice and 2.1–6.3 for hu-BLT mice when half a million tumor cells were cultured initially. The rate of tumor cell growth is greatly correlated with the increased MHC-class I expression on the tumor cells obtained from the hu-BLT (27–50% increase (Figure S5) and decreased NK cell-mediated cytotoxicity, a profile that we have previously shown to correlate with NK cell-mediated differentiation of the tumor cells in hu-BLT mice injected with allogeneic super-charged NK (sNK) cells [11,12]. Therefore, it is likely that immune cells in hu-BLT mice are not only suppressing the tumor growth by direct lysis of the stem-like tumors, but also by differentiating the tumor cells in vivo via secretion of IFN-γ and TNF-α, which limits the growth and expansion of tumor cells [19]. Indeed, when supernatants from NK cells are added to both oral and pancreatic stem-like tumors, the expression of differentiation antigens such as MHC-class I, CD54, and PDL-1 increases, the tumors grow much more slowly, and they become resistant to NK cell-mediated cytotoxicity [11,27,28]. This is directly relevant to induction, expansion, and tumor growth in vivo, since having competent immune cells is necessary for immune surveillance, lysis, and differentiation of the tumors. In the absence of any competent immune effectors, the tumors in NSG mice grow and expand, whereas in hu-BLT mice, there are enough competent cells to decrease the expansion and progression of the cancer. Indeed, in vivo NSG mice grow larger tumors, and the mice lose weight rapidly. Hu-BLT mice grow smaller tumors and lose weight at later times than NSG, but the tumors are still able to be established in them. This could be due to a number of reasons. Even though hu-BLT mouse peripheral blood has competent B and T cells, and the numbers and percentages are similar to humans [12], the percentages of NK cells are at approximately 50% those seen in human peripheral blood. Further, since both oral and pancreatic tumors are stem-like/poorly differentiated tumors, the percentages of NK cells are not enough to contain all the growing tumors. Indeed, when a million sNK cells were injected after tumor implantation, the tumor growth and expansion were substantially decreased [11,12]. In addition, the killing function of human NK cells in hu-BLT mice is approximately 50% less than those seen in the peripheral blood of humans, even though the levels of IFN-γ secretion are higher than those seen in the peripheral blood of humans (Table 1). Therefore, the decrease in both the percentages of NK cells and their killing function may facilitate the growth and expansion of stem-like/poorly differentiated aggressive tumors. There is considerable debate regarding the mechanisms of decrease in NK cells’ function in hu-BLT mice. It was hypothesized that deficiency in IL-15 may be one underlying mechanism for the decreased function and numbers of NK cells in hu-BLT mice [29,30]. Supplementation of IL-15 increased both the numbers and function of NK cells in hu-BLT mice [31]. Similar to hu-BLT, supplementation of IL-15 can also increases the cytotoxic function of NK cells and their numbers in human cancer patients [32,33]. In addition, we have not observed significant defects in the autologous NK cells’ function in hu-BLT mice, other than a reverse in the levels of cytotoxicity and IFN-γ secretion, which may point to the split anergy in NK cells in this mouse model system, since human NK cells may also have to deal with the mouse stromal cells. Split anergy denotes the concept of decreased cytotoxicity in the presence of increased IFN-γ secretion, which occurs after NK cells are cultured with tumor cells or after crosslinking of important receptors such as CD16 or NKp46 on NK cells [34]. Indeed, the CD56bright population in the tissues, which most likely represents an activated phenotype of NK cells in humans, has also been shown to have decreased cytotoxicity in the presence of increased IFN-γ secretion [35,36]. Therefore, the conditioning of human NK cells in the context of the mouse tissue microenvironment may create the differences we observed between human-derived NK cells and hu-BLT-derived NK cells. To demonstrate the expansion and activation potential of NK cells from hu-BLT, we enriched splenocyte-derived NK cells and cultured them with the autologous osteoclasts differentiated from the monocytes. After different days of culture, we determined the expansion potential, as well as functions of NK cells, and found them to have excellent responsiveness, indicating that their activation and expansion potential is largely intact in the hu-BLT mice. The slight differences that were noted between the sNK cells from humans or hu-BLT mice is likely due to the compartments from which the NK cells were obtained: peripheral blood from the humans and spleen for hu-BLT. Due to a lack of an adequate volume of peripheral blood from hu-BLT, side-by-side comparisons of hu-BLT peripheral blood-derived NK cells with human peripheral blood-derived NK cells are challenging. However, these experiments are ongoing in our lab at the moment.