Tissue Distribution and Phenotype of Human Mast Cells in BRGSF and NSG‐SGM3‐IL15 Humanized Mice

To the Editor,

Mast cells (MCs) are tissue resident cells that play key roles in allergies and other inflammatory diseases, and thus represent important targets for drug development. MCs are found in many tissues and are particularly abundant at sites exposed to the external environment, such as the skin, the gastrointestinal tract, and the lung [1]. Recent scRNAseq data indicate that human MCs (hMCs) are highly heterogeneous within and across organs and might extend far beyond the classical mucosal and connective tissue dichotomy [1]. Thus, testing the efficacy of novel drugs targeting hMCs in vivo will require models that can recapitulate, at least in part, the complex phenotypes and tissue distribution of these cells.

Several groups have used “humanized mice,” which are highly immunodeficient mice engrafted with human stem cells to allow the development of hMCs and other hematopoietic cells. In particular, immunodeficient NOD‐scid Il2rγ ^−/− mice, which express transgenes for hIL3, hGM‐CSF, and hSCF (called NSG‐SGM3 mice) develop high numbers of hMCs upon engraftment with stem cells [2, 3, 4]. However, relatively little is known about the phenotype and tissue distribution of MCs in such humanized mice.

Here we sought to characterize the MC compartment in two commercially available humanized mouse strains: NSG‐SGM3‐IL15 and BALB/c Rag2 ^−/‐ Il2rγ ^−/‐ Sirpα ^NOD Flk2 ^+/− (BRGSF) mice engrafted with cord blood‐derived hCD34⁺ progenitors. NSG‐SGM3‐IL15 mice represent a “next generation” NSG‐SGM3 strain with the addition of a hIL15 transgene to further improve lymphoid cell engraftment. BRGSF mice also support strong human immune reconstitution, particularly in the intestine [5], but their MC compartment has not yet been characterized.

BRGSF mice had an overall increased percentage of circulating hCD45⁺ immune cells 10 to 15 weeks after engraftment as compared to NSG‐SGM3‐IL15 mice (Figure S1A). There were also significantly more B and T cells among hCD45⁺ blood cells in BRGSF mice (Figure S1B,C), which confirms that this strain permits good development of abundant human lymphoid cells [5]. Interestingly, the frequency of circulating human basophils was significantly higher in NSG‐SGM3‐IL15 mice (Figure S1D), likely reflecting the presence of the hIL3 transgene, since IL3 is an important growth and survival factor for basophils. NSG‐SGM3‐IL15 mice had slightly increased hCD45⁺ cell frequency in the skin, lung, spleen, and bone marrow (Figure S1E), which highlights distinct tissue‐homing characteristics for hCD45⁺ cells between both strains.

Both strains of humanized mice had detectable baseline levels of serum tryptase, a protease synthesized and stored in hMCs (and to a much lesser extent basophils), which was not detectable in non‐engrafted NSG‐SGM3‐IL15 controls (Figure 1A). Tryptase levels were significantly higher in NSG‐SGM3‐IL15 mice, but these levels remained within the reference range found in healthy humans (1–15 ng/mL) (Figure 1A). We identified hMCs by flow cytometry in the skin, lung, gut, and peritoneal cavity of both strains, with very low levels in the bone marrow and spleen (Figure 1B). The frequency of hMCs was higher in NSG‐SGM3‐IL15 mice in all organs, with the exception of the gut, where similar levels were observed between both strains (Figure 1B). Importantly, although these mice are highly immunodeficient, we still observe a high frequency of mouse MCs (mMCs) in the skin and peritoneal cavity of both strains (Figure 1C). Skin mMC frequency was even higher in engrafted mice compared to non‐engrafted controls, suggesting human–mouse immune cell interactions. These parameters need to be taken into consideration for the analysis of any MC‐mediated in vivo disease models in those mice, since both mMCs and hMCs might concomitantly play a role.

FIGURE 1.

Phenotype and tissue distribution of mast cells in NSG‐SGM3‐IL15 and BRGSF humanized mice. (A) Serum tryptase level measured by ImmunoCAP in humanized mice (n = 8–9 mice per group) and non‐engrafted NSG‐SGM3‐IL15 controls (n = 5). (B–C) Percentages of human MCs (B) and mouse MCs (C) among human CD45⁺ cells in different organs. (D–F) Quantification of KIT (CD117) (D), FcεRI (E) and MRGPRX2 (F) expression (delta mean fluorescence intensity [ΔMFI]) in human MCs from different organs. (G) Representative confocal microscopy images of Avidin (red), tryptase (green), and DAPI (gray) fluorescent signals of skin, lungs, and gut. Scale bars: 50 μm or 20 μm for enlargements of the dashed areas. Some avidin⁺tryptase⁻ mouse MCs, avidin⁺tryptase⁻ and avidin⁺tryptase⁺ human MCs are indicated with arrows. (H‐J) Quantification of avidin^‐tryptase⁺ and avidin⁺tryptase⁺ human MCs in skin (H), lung (I), and gut (J) tissue. Data in A‐F and H‐J show individual values, with bars indicating means ± SEM. P‐values were calculated using a Kruskal–Wallis test, with Dunn's correction in A‐C or Mann–Whitney U test in D‐F. PLF: Peritoneal lavage fluid. BM: Bone marrow. (K) Mice were sensitized intradermally in the ear skin with human anti‐NP IgE (or with PBS as control in the other ear) followed by intravenous challenge with NP‐BSA 24 h later. Data show changes in ear thickness over 1 h after challenge with NP‐BSA. For statistical analysis, area under the curve was calculated for each individual, and P‐values were calculated using a Kruskal–Wallis test, with Dunn's correction.

A unique feature of MCs is the co‐expression of the SCF receptor KIT and the high‐affinity IgE receptor FcεRI, which can vary across organs (Figure S1F). We found that the expression of KIT was much lower, and that of FcεRI much higher in hMCs from all organs in NSG‐SGM3‐IL15 mice as compared to BRGSF mice (Figure 1D,E), which might reflect a more mature hMC phenotype in NSG‐SGM3‐IL15 mice. In humans, skin MCs also express the human G‐protein‐coupled receptor MRGPRX2, which is responsible for MC activation by a range of cationic substances called basic secretagogues [6]. We found that only skin hMCs from NSG‐SGM3‐IL15 mice expressed high levels of MRGPRX2 (Figure 1F), which makes this strain particularly interesting for the study of pseudoallergic skin reactions and therapeutic proof of concept of targeting MRGPRX2. However, it should be noted that mouse connective‐tissue type MCs (CTMCs) found mostly in the skin and peritoneal cavity express Mrgprb2, the orthologue of MRGPRX2 [1]. Therefore, interpretation of results from pseudoallergic reactions should be taken carefully due to the fact that most MRGPRX2 agonists cross‐react with this mouse ortholog.

In mice, two mutually exclusive MC subpopulations have been described: CTMCs and mucosal MCs (MMCs) found in the lung and gut. CTMCs but not MMCs can be stained with avidin [1]. While hMCs are much more heterogeneous [1], staining with avidin can also be used to distinguish between hMCs at connective versus mucosal sites. We observed similar numbers of human tryptase⁺avidin⁺ hMCs in skin tissue in both strains of mice (Figure 1G,H). In agreement with our flow cytometry data (Figure 1C), we also observed human tryptase⁻avidin⁺ mMCs in the skin of both strains and the non‐engrafted controls (Figure 1G). In lung and gut tissues, we found that avidin⁻ hMCs were much more frequent than avidin⁺ hMCs (Figure 1G,I,J). Both strains exhibited a relatively high proportion of hMCs among all human CD45⁺ immune cells in the gut, providing an interesting opportunity to study the effect of therapeutic molecules in the context of GI tract‐associated disorders. Finally, the number of lung hMCs was significantly higher in NSG‐SGM3‐IL15 mice as compared to BRGSF (Figure 1G,I,J), which makes this strain potentially interesting for the study of hMCs in the context of lung inflammatory responses.

We then performed hIgE‐mediated passive cutaneous anaphylaxis (PCA) in both strains. As expected, the hIgE PCA in the non‐engrafted controls did not induce swelling (Figure 1K). Although hMCs from NSG‐SGM3‐IL15 mice expressed higher levels of FcεRI, we observed marked PCA responses in both strains, with a more sustained ear swelling in BRGSF mice (Figure 1K). In line with this, important hMC degranulation was observed in the ear tissue sensitized with IgE in both strains, and little to no degranulation was observed in the non‐engrafted mice (Figure S1G,H). Interestingly, although hIgE does not bind mFcεRI, mMCs were also largely degranulated in IgE‐sensitized ears. It is thus likely that mMC degranulation occurred indirectly in response to mediators released by hMCs and participated in the overall strong PCA response (Figure S1I,J).

hMCs in NSG‐SGM3‐IL15 are more abundant in most organs and appear more mature than in BRGSF mice, expressing high levels of FcεRI and MRGPRX2 (in the skin), making this strain particularly suited for the study of hMCs function via these receptors and test of novel MC‐targeted therapies in vivo. Thus, to further delineate the physiological relevance of the NSG‐SGM3‐IL15 model, we carried out single‐cell RNA sequencing on skin and gut hMCs using 10X GEM‐X sequencing technology (Figure 2 and Figure S2), in order to determine if the hMCs recapitulate at least in part the transcriptomic heterogeneity that has previously been reported in humans [1]. We isolated human CD45+ immune cells from the skin and gut and generated a unique UMAP composed of 26,182 single cells (Figure S2A). We next identified a unique cluster composed of 4855 MCs based on the signature of the cardinal MC genes such as KIT, CPA3, GATA2, and CMA1 (Figure S2A–D). We then projected all of the identified MCs on the same UMAP (Figure 2A). We could not observe an obvious CTMC/MMC transcriptomic dichotomy (Figure 2B,C and Figure S2D).

FIGURE 2.

Transcriptomic profiling of human mast cells in the skin and gut of NSG‐SGM3‐IL15 mice. (A) UMAP of 4855 human MCs isolated from skin and gut tissue of NSG‐SGM3‐IL15 humanized mice visualized by tissue of origin. (B) UMAP of the different MC clusters. (C) Heatmap of the top 25 most differentially expressed genes in each MC cluster. (D) Calculated metacluster signatures, visualized over the MCs of this study. “MC1_Signature” translates as MC1 Signature from Tauber et al. [1].

We thus decided to adopt an unbiased approach to better understand MC heterogeneity and performed an unsupervised nearest‐neighbor analysis (Figure 2B) together with a correlation heatmap (Figure S2E) to directly visualize the strength of relationships between the different clusters. We could identify the presence of 4 potential MC states (Figure 2B) with a total of 8610 statistically significant DEGs that defined the transcriptomic heterogeneity between each state of MCs (Figure 2C). As expected, we only observed MRGPRX2 gene expression in the skin‐related MC cluster #4 (Figure 2C). Importantly, none of these 4 clusters matched with the transcriptomic signature of the 6 MC clusters identified in human [1] (Figure 2D). These data demonstrate that at least 4 MC populations/states exist in the skin and gut in humanized mice, but they only partially recapitulate the transcriptomic heterogeneity found in humans. Humanized mice are thus suitable models for the preclinical development of MC‐targeted therapies and basic functional responses; however, they are limited in capturing the full extent of MC heterogeneity, developmental dynamics, and fine biological responses observed in human tissues.

Author Contributions

W.P.M.W., N.S., J.J.K., C.E.‐S., J.B.J.K., C.H., E.L., A.L., and P.A.A. carried out experiments. W.P.M.W., N.S., L.L.R., and N.G. designed experiments. W.P.M.W., N.S., C.E.‐S., R.E., and N.G. carried out analysis. W.P.M.W., N.S., L.L.R., and N.G. drafted the manuscript. All authors read, edited, and approved the manuscript.

Conflicts of Interest

Disclosure of potential Conflicts of Interest: Laurent L. Reber is or recently was a speaker and/or advisor for and/or has received research funding from Argenx, Novartis, Ceva, and Neovacs. Nicolas Gaudenzio is (or was) collaborating or consulting or a member of the scientific advisory board for Genoskin (CSO, shareholder), Escient Pharmaceuticals, Aikium, CEVA, MaxiVax, Boehringer Ingelheim, Novartis, Sanofi, and ArgenX. The rest of the authors declare that they have no relevant conflicts of interest.

Supporting information

Figure S1: all70058‐sup‐0001‐FiguresS1.docx. Frequency of human CD45⁺ cells, T cells, B cells, and basophils in NSG‐SGM3‐IL15 and BRGSF humanized mice. (A–D) Percentages of human CD45⁺ cells (A), CD19⁺ B cells among human CD45⁺ cells (B), CD3⁺ T cells among human CD45⁺ cells (B), and CD45⁺ CD123⁺HLA‐DR⁻ basophils in the blood. (E) Percentages of human CD45⁺ cells in different organs. (F) Representative FACS plots of hMC gating strategy. (G) Representative toluidine blue staining and (H) quantification of MC degranulation in sections of ear skin following hIgE‐mediated PCA. (I) Representative avidin staining. (G) Representative toluidine blue staining and (H) quantification of MC degranulation in sections of ear skin following hIgE‐mediated PCA. (I) Representative confocal microscopy images of avidin (red), tryptase (green), and DAPI (gray) fluorescent signals of ear skin from anti‐NP hIgE‐sensitized mice 1 h after challenge with NP‐BSA. Scale bars: 50 μm. Some individual avidin⁺tryptase⁻ mouse MCs and avidin⁺tryptase⁺ human MCs are indicated with arrows as representative examples. (J) Quantification of degranulation of mMCs (avidin⁺tryptase⁻) and hMCs (avidin⁺tryptase⁺) 1 h after challenge with NP‐BSA. Data show individual values, with bars indicating means ± SEM. P‐values were calculated using a one‐way ANOVA with Kruskal–Wallis multiple comparisons test. PLF: peritoneal lavage fluid. BM: bone marrow.

Figure S2: Gating strategy used for the identification of human mast cells in various organs. A. UMAP of 26,182 QC‐passed single cells visualized by their tissue of origin. B. Density plot of PTPRC (CD45) and mast cell marker genes. C. Visualization of the mast cell signature module score. D. Density plot of mast cell marker genes among the mast cell clusters. E. Correlation heatmap of the different mast cell clusters with relational distances between them indicated by dendrograms on the sides.

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.