Background
Although many human pathogens currently under research have been adapted to the mouse model, some are strictly human pathogens [
1]. Further, mice are evolutionarily distinct from humans and are short-lived [
1]. Within the last 30 years, however, researchers have been developing and improving upon a hybrid mouse model wherein human tissues are housed in the background of a mouse. The generation of the humanized mouse model now allows researchers to directly study the dynamics of human immune cells in a biological system.
The first humanized mouse models engrafted a human immune system into SCID mice [
1,
2]. Since then, many protocols have used immunocompromised mice with varying knockout combinations of NOD, SCID, IL-2R, Rag1, and Rag2 genes, wherein the complete absence of all lymphocytes has thus far yielded the best human immune cell engraftment efficiency [
1,
2]. Furthermore, there are a variety of hematopoietic stem cells sources (HSCs; i.e. peripheral blood, fetal liver, and cord blood) along with a number of different methods to engraft those progenitor cells, such as intrahepatically or intraperitoneally into newborn mice or intravenously into adult mice. Each method tends to yield different reconstitution rates and immune cell distributions [
1,
2]. In regards to HSCs, progenitor CD34+ cells from cord either blood or fetal liver are a superior HSC source compared to PBMCs or bone marrow as there is a higher level of engraftment and limited development of graft vs. host disease [
3].
A large variety of protocols for generating humanized mice now exist and encompass an array of murine recipients with different sources of HSCs and the methods of engraftment. This results in varying reports of human immune cell engraftment efficiencies and human immune cell subset distribution [
1,
2]. In particular, there is very little data comparing the engraftment of HSCs into mice as newborns or as adults. Interestingly, Brehm et al. [
4] has shown that mice engrafted as newborn mice have an overall greater reconstitution of human CD45+ cells in the blood compared to those engrafted as adults. Furthermore, newborn engrafted mice had a greater proportion of T-cells and a lower proportion of B-cells in the blood compared to mice engrafted as adults [
4]. Unfortunately, the comparison between mice engrafted as newborns or adults was limited to the blood and it is unknown whether this trend prevails in all other immunological organs, including the liver.
Much of the data analyzing human immune cell reconstitution in mice has been limited to the blood, spleen, bone marrow, and thymus. Little is know about human immune cells in lymph nodes or liver. Ishikawa et al. [
5] as well as Traggai et al. [
6] both observed mesenteric lymph node development in humanized mice engrafted as newborns, though using different strains of recipient mice. Traggai et al. [
6] further showed that the mesenteric lymph nodes contained both B and T-cells. Recently, researchers have developed a new humanized mouse model in which human liver cells are engrafted into immunocompromised mice, allowing for the study of human liver tropic viruses such as hepatitis C virus and malaria, with the ultimate goal of creating a humanized mouse model with both human liver and a human immune system [
7‐
9]. However, the human immune cell composition in the livers of humanized mice is not well characterized.
In this article, we have conducted a comprehensive comparative analysis of human immune cell engraftment and distribution between mice given HSCs as newborn pubes (1–2 days old) intrahepatically or intravenously as adult mice in a variety of organs, including the lymph node and liver.
Discussion
In this study, we compared human immune cell reconstitution as well as the distribution of various immune subsets between two different methods of HSC engraftment: intravenously in adult NRG mice or intrahepatically in newborn NRG mice. Similar to the findings of Brehm et al. [
4], mice engrafted as newborns had higher levels of human CD3+ T-cells and lower levels of CD19+ B-cells in the blood compared to mice engrafted as adults. We observed very little difference in the level of human CD45+ cells in the blood, spleen, bone marrow, thymus, lymph node, or liver between mice engrafted as adults or pups, suggesting that there is very little difference in human immune cell reconstitution between either methods. Furthermore, we observed a long-term reconstitution (up to 28 weeks) of human immune cells in mice engrafted as either adults or newborns. Human T- and B-cell populations were found in all tissues examined, with lower levels of myeloid, NK, and NKT cells, similar to the results of others. Lastly, similar to the human thymus, the humanized mouse thymus contained a high population of human T-cells.
As reported by others, we also observed the development of mesenteric lymph nodes in the humanized mice, engrafted either as adults or pups [
5,
6]. Interestingly, we found that the mesenteric lymph nodes were the only lymph nodes to develop in the humanized mice. The presence of microbiota in the gastrointestinal tract likely facilitates the preferential development of mesenteric lymph nodes. It has been previously shown, however, that humanized mice are capable of developing lymph nodes during infection. Kwant-Mitchell et al. [
10] found that humanized mice in a Rag2
-/-γc
-/- background that were immunized or infected with herpes simplex virus type 2 developed iliac lymph nodes, whereas naïve mice were only able to develop mesenteric lymph nodes.
Recently, a number of research groups have developed a humanized mouse model in which human hepatocytes have been engrafted into mice, thus providing the opportunity to study human liver tropic pathogens, such as hepatitis B and C in a biological system [
7]. Furthermore, double-chimeric humanized mice have been developed, in which both the human immune system and human liver hepatocytes are engrafted onto a mouse background, thereby enabling the study of the human immune response against these pathogens [
8,
9]. Human livers contain a distinct T-cell population in the liver, with a higher proportion of CD8+ T-cells over CD4+ T-cells. In our models, we observed a higher ratio of CD4/CD8+ T-cell population. Similar to all other tissues, there were very low levels of human NK and myeloid cells, which has been observed in other studies [
7,
11]. Using newborn engraftment, Washburn et al. [
7] observed a predominant human T-cell population in the livers of humanized mice as well as small populations of NK cells and plasmacytoid dendritic cells. Likewise, Billerbeck et al. [
11] observed high levels of human B- and T-cells in the liver of humanized transgenic A2 mice, in which mice have been genetically modified to express HLA-A2. Indeed, the humanized mice we generated have decreased levels of NK cells, not only in the liver, but in all other tissues we examined.
The immune population within the human liver is largely composed of innate immune cells. Approximately 30–50 % of the immune cells in the human liver are NK cells, which have been observed to express a decreased level of CD16 compared to peripheral blood NK cells [
12‐
14]. It is evident that the current humanized mouse models, in which HSCs are transplanted into immunocompromised murine recipients, are unable to support a identifiable level of human NK cells or other human innate immune cells, likely due to the lack of compatibility between mouse and human cytokines that are required to stimulate innate immune cell development [
15]. While researchers have attempted to circumvent the problem by administering specific human cytokines allowing for innate immune cell development, this resulted in abnormally high levels of human cytokines that could cause artificial responses [
16‐
19]. Recently, Rongvaux et al. [
20] developed a transgenic mouse model recipient in which they knocked-in five human cytokines into the mouse genome, and found that the human cytokines allowed for the greater development of human monocytes and NK cells in their humanized mouse model. In particular, the generation of human monocytes were able to trans-present human IL-15 to foster the development of NK cells [
20]. In double-chimeric mice, the presence of human hepatocytes encouraged the establishment of human Kupffer cells in the liver, which are known to trans-present IL-15 and could stimulate the development of human NK cells [
9,
21]. Thus, human hepatocytes may promote a greater level of human NK cells within the liver of double-chimeric mice.
Previous studies have found that human NK cells in these mouse models are not only decreased in number but are also dysfunctional [
22]. We found that human NK cells within the bone marrow and liver had a reduced expression of CD16, NKp30 and CD11b, suggesting that these NK cells are less activated and less mature. They also displayed a comparable level of CD16 expression to peripheral blood and splenic NK cells. In contrast, studies examining human liver NK cells found that they express decreased CD16 compared to peripheral blood NK cells [
12‐
14]. Due to the low level of NK cells, however, it would be beneficial to examine NK cell phenotype, maturation, and functionality in a model with a higher frequency of NK cells.
Abbreviations
HSCs, hematopoietic stem cells; NK, natural killer; NRG, NOD-Rag1-/-IL2rγ-/-; NRG: NOD, non-obese diabetic; PBMCs, peripheral blood mononuclear cells; PBS, phosphate buffered saline; Rag, recombination-activating gene; SCID, severe combined immunodeficiency
Acknowledgements
We would like to thank the Nurses at the Labour & Delivery ward for collecting cord blood. This work was supported by grants from the CIHR to Ali Ashkar. Amanda Lee is a recipient of a CIHR Vanier Scholarship (2013–2016). Ali Ashkar is a recipient of a CIHR Tier 1 Canada Research Chair.