Immunoprofiling Reveals Novel Mast Cell Receptors and the Continuous Nature of Human Lung Mast Cell Heterogeneity

Background Immunohistochemical analysis of granule-associated proteases has revealed that human lung mast cells constitute a heterogeneous population of cells, with distinct subpopulations identified. However, a systematic and comprehensive analysis of cell-surface markers to study human lung mast cell heterogeneity has yet to be performed. Methods Human lung mast cells were obtained from lung lobectomies, and the expression of 332 cell-surface markers was analyzed using flow cytometry and the LEGENDScreen™ kit. Markers that exhibited high variance were selected for additional analyses to reveal whether they were correlated and whether discrete mast cell subpopulations were discernable. Results We identified the expression of 102 surface markers on human lung mast cells, 23 previously not described on mast cells, of which several showed high continuous variation in their expression. Six of these markers were correlated: SUSD2, CD49a, CD326, CD34, CD66 and HLA-DR. The expression of these markers was also correlated with the size and granularity of mast cells. However, no marker produced an expression profile consistent with a bi- or multimodal distribution. Conclusions LEGENDScreen analysis identified more than 100 cell-surface markers on mast cells, including 23 that, to the best of our knowledge, have not been previously described on human mast cells. The comprehensive expression profiling of the 332 surface markers did not identify distinct mast cell subpopulations. Instead, we demonstrate the continuous nature of human lung mast cell heterogeneity.


INTRODUCTION
Heterogeneity among mast cells has been known for a long time and was first attributed to differential expression of proteoglycans in rodent mast cells, which gave them distinct staining patterns (1). This led to the division of rodent mast cells into connective tissue mast cells and mucosal mast cells. In humans, mast cell heterogeneity has been based on the expression of mast cell proteases, i.e., cells expressing tryptase only (MC T ) and those expressing both tryptase and chymase (MC TC ) as well as carboxypeptidase A (2,3). These subtypes have been defined using immunohistochemistry, a method that produced binary results, that is, the absence or presence of expression. The MC TC subtype is more predominant in connective tissues such as the skin, while the MC T subset is more prevalent in mucosal surfaces such as the airways and gastrointestinal tract (4).
Mast cells are found in the human lungs in all different compartments, i.e., under the epithelium, in smooth muscle bundles, around pulmonary vessels, in the parenchyma and in close proximity to sensory nerves (5). Human lung mast cells (HLMCs) have several important functions in health and diseases, such as host defense, induction of acute inflammatory responses, vascular regulation, bronchoconstriction and tissue remodeling (6)(7)(8)(9). The heterogeneity of HLMCs was first suggested to be related to differences in their size and function (10), where a heterogeneity in the response to secretagogues also was reported (11). It was described that the protease expression within HLMCs differ, with the MC T being the predominant subtype except around pulmonary vessels, where the MC T and MC TC subtypes are found in equal numbers (2). However, the heterogeneity among HLMCs goes beyond size and protease expression, as demonstrated by the differential expression of certain mast cell-related markers (FceRI, IL-9R, 5-LO, LTC 4 S, etc.) among the MC T and MC TC populations in different lung compartments (12).
Mast cell heterogeneity has primarily been studied in a binary manner using immunohistochemistry, describing the absence or presence of a given marker. Here, we used a quantitative flow cytometry-based approach to study HLMC heterogeneity, profiling the expression of 332 surface markers and intracellular staining of the proteases tryptase, chymase and CPA3. None of these markers distinctly divided the HLMCs studied into subpopulations. However, several markers showed a high degree of variation within the mast cell population with a nonclustered gradient expression pattern. Six of these markers correlated with each other, revealing the continuous nature of HLMC heterogeneity rather than separation into distinct subpopulations.

Ethical Approval
The local ethics committee approved the collection of lung tissue from patients undergoing lobectomy, and all patients provided informed consent (Regionala Etikprövningsnämnden Stockholm, 2010/181-31/2).

Cell Preparation
Single-cell suspensions were obtained from macroscopically healthy human lung tissue as previously described (13). Briefly, human lung tissue was cut into small pieces and enzymatically digested for 45 min with DNase I and collagenase. Thereafter, the tissue was mechanically disrupted by plunging through a syringe, the cells were washed, and debris was removed by 30% Percoll centrifugation. After preparation, the cells were stained and analyzed by flow cytometry.

Expression of Mast Cell Proteases
The classic division into distinct human mast cell subpopulations is based on the presence or absence of granule-associated antigens, i.e., the proteases tryptase, chymase and CPA3. We therefore first performed intracellular staining of lung cells and used a flow cytometry-based readout to accurately characterize the mast cell population with respect to the expression of tryptase, chymase and CPA3. Flow cytometry analysis identified CD45 + CD14 low CD117 high HLMCs with the characteristic expression of tryptase ( Figures 1A, B). The HLMC population showed a high variation of chymase and CPA3 expression between individual cells and the degree of CPA3 and chymase double positive cells varied considerably from donor to donor ( Figure 1C). However, no distinct subpopulations were discernable, instead there was a continuous spectrum of expression levels.

Immunoprofiling of HLMCs
Whereas intracellular antigens failed to discriminate distinct subpopulations, we set out to perform an extensive mapping of cell surface antigens on HLMCs to potentially identify distinct subpopulations and novel markers. Flow cytometry analysis characterized the expression of 332 surface markers on HLMCs, using a LEGENDScreen ™ human cell screening kit. Since many of the markers are broadly expressed we used CD14 + cells as a reference to selectively enrich for lineagespecific antigens in HLMCs ( Figure 1D). Well-known monocyte markers such as CD11b, CD11c, and HLA-DR were highly expressed on CD14 + cells, verifying the validity in enriching for lineage-characteristic markers. In analogy, HLMCs expressed high levels of CD117 and FceRI. Markers with the most significant differences between the HLMCs and CD14 + cells included CD9, CD59, CD274 and CD226 ( Figure 1D). CD9 is a broadly expressed tetraspanin with a wide variety of functions; in mast cells, it is abundantly expressed and has been implicated in chemotaxis and activation (14). CD59 can prevent complement-induced cytolytic cell death by preventing assembly of the complement membrane attack complex and has also been implicated in T cell activation (15). CD274 is also known as programmed death ligand-1 (PD-L1) and can cause blockade of T cell activation (16). CD226 has received increasing interest in recent years and can play a role in many immunological processes (17), including enhancement of FceRI-mediated activation in mast cells (18).
Of the 332 markers analyzed, HLMCs showed significant expression of 102 markers (Figure 2), of which 23, to the best of our knowledge, have not been described on (non-neoplastic) human MCs before ( Table 1).

Heterogeneous Expression of the High-Affinity IgE Receptor FceRI
One of the principal markers for MCs is the expression of the high-affinity IgE-receptor FceRI. FceRI expression on HLMCs has been shown to differ among compartments in the lungs, as well as between healthy and diseased tissues. HLMCs from healthy individuals present in the parenchyma are negative for FceRI, while patients with concurrent asthma show higher expression of the receptor (12,19). In our study, the expression varied considerably between donors and there was no clear separation between the negative and positive FceRI populations but rather a continuous spectrum of expression and approximately 50% of the donors expressed FceRI on virtually all MCs ( Figures 3A-C).

Heterogeneous Expression of Cell-Surface Markers With a Continuous Distribution
The heterogeneity of HLMCs has primarily been studied using immunohistochemistry in a binary manner and they have been divided into the MC T and MC TC subtypes (2,3). How this heterogeneity is reflected by the heterogeneous expression of cellsurface markers has barely been investigated. None of the surface markers investigated in this study, using quantitative flow cytometry, did distinctly and consistently divide the HLMCs into subpopulations (data not shown). However, several markers were expressed with a considerable continuous variation within the HLMC population ( Table 2). Co-stainings of nine of these markers revealed that expression of six of the markers, SUSD2, CD49a, CD326, CD34, CD66 and HLA-DR correlated (r > 0. 4) ( Figures 4A-H). The correlations seen between markers was not due to autofluorescence since there was no correlations in the FMO controls (Supplementary Figures S1A-E). In addition, the HLMCs showed no significant staining with the 10 isotype controls (Figure 2).
We next investigated if the expression of these markers correlated to the classical mast cell subsets using intracellular staining for CPA3 as a marker for MC TC . Expression of neither SUSD2, CD38 (Figures 4I, J) nor CD344 (data not shown) correlated with CPA3 expression.
Since the expression of SUSD2, a marker for stem/progenitor cells (20), correlated to CD34, that is expressed on circulating mast cell progenitors (21), we investigated if the expression of these markers could identify MCs in different stages of maturation. When mast cells mature they become larger with an increasing number of granules (22). We therefore gated SUSD2 low, intermediate and high cells and compared the FSC and SSC values that reflect the size and granularity of the cells. However, SUSD2high cells had higher FSC and SSC then SUSD2low cells (Figures 4K-M), suggesting that they were  larger with a higher granularity and therefore unlikely to be immature MCs. SUSD2 has been linked to proliferation in cancer cells (23), why we investigated the proliferation status of the cells with the proliferation marker Ki-67. However, in agreement with the fact that mast cells are long-lived cells with low turnover (24), no staining was observed (Supplementary Figure S2).

DISCUSSION
HLMCs have been shown to be heterogeneous; classically, they have been studied using immunohistochemistry in a binary manner, and they have been divided into the MC T and MC TC subtypes based on whether the mast cell proteases chymase and CPA3 are detectable (2, 3). By using a quantitative flow cytometry based method we indeed found high variation in chymase and CPA3 expression but no distinct subpopulations were discernable ( Figure 1C). The apparent discrepancy between our study and previous immunohistochemical studies is probably due to the fact that we have studied the expression in a quantitative manner using flow cytometry, thus finding that there is a spectrum of different expression levels. In contrast, in previous studies the cells have been classified into MC T /MC TC cells in a binary manner depending on the detection limit of the immunohistochemical technique.
Although attempts have been made to map cell-surface antigens on HLMCs (25)(26)(27)(28)(29)(30), extensive mapping including the Involved in MHC class II antigen processing and a receptor for macrophage migration inhibitory factor CD111 Adhesion molecule CD115 Receptor for M-CSF and IL-34 CD131 Common b subunit of the IL-3, IL-5 and GM-CSF receptors CD143 Metallopeptidase CD148 Tyrosine phosphatase involved in signal transduction CD164 Sialomucin involved in cell adhesion and proliferation CD166 Glycoprotein involved in cell adhesion and migration CD205 Endocytic receptor involved in antigen uptake and processing CD243 Involved in transportation of molecules across cell membranes CD270 Receptor for TNFSF14, BTLA, LTA and CD160 CD277 Regulate T cell responses CD317 Blocks the release of certain viruses from infected cells CD344 (Frizzled-4) Receptor for Wnt proteins and norrin CLEC12A/CD371 C-type lectin-like receptor with an immunoreceptor tyrosine-based inhibitory motif (ITIM) domain Integrin a9b1 Integrin mediating cell adhesion and migration SUSD2 (W3D5, W5C5) Potentially involved in cell adhesion as this transmembrane protein contains functional domains associated with adhesion molecules (W4A5) Antigen has yet to be described Siglec-9 Lectin that binds sialic acid and has ITIM domains SSEA-5 A glycan heterogeneity of cell-surface antigen expression has not been carried out. In this study, we identified significant expression of 102 markers on the HLMC surface (Figure 2), of which, to the best of our knowledge, 23 are novel mast cell markers ( Table 1). Several of these markers, including SSEA-5, SUSD2, W4A5, CD243, CD111, CD131 and CD164, are described as markers expressed on stem cells. The expression of stem cell markers on mast cells is in accordance with results from the FANTOM5 consortium, in which skin mast cells exhibited similarities with stem cells (31). In some cases, our results are in disagreement with previously published data; for example, CD4, CD10, CD36 and CD74 were previously shown to not be expressed by HLMCs (26,27). This discrepancy might be explained by differences in the procedures, as in contrast to published data, we did not purify or culture the studied mast cells prior to analysis (25,27,29,30). Culturing mast cells has been shown to alter their phenotype and expression of cell-surface receptors (31,32). Furthermore, although macroscopically healthy tissue distal from the tumor was used for our analysis, one cannot rule out the possibility that also this part of the tissue is affected by the disease and can have an impact on the results. The expression of FceRI has been shown to differ in different compartments of the lung, with mast cells present in the parenchyma being negative for FceRI (12). We have investigated single cell suspensions from lung tissues without separation of the parenchymal cells. However, we did not observe distinct FceRI positive and negative mast cell populations but rather a continuous spectrum of expression levels ( Figure 3A). Additionally, in about 50% of the donors virtually all mast cells stained positive for FceRI, meaning that we cannot detect any FceRI negative parenchymal mast cell population in these individuals ( Figure 3A). This discrepancy is again likely to be due to the different techniques used, immunohistochemistry and flow cytometry, and the different detection limits of the techniques. We also observed a large variation in expression of FceRI among individuals ( Figures 3A-C), and in line with our results, this has previously been shown also for human skin mast cells (33). The reasons for this variation could be manifold, as the surface expression of FceRI can be regulated in many different ways. FceRI is, for example, upregulated by IL-4 and stabilized on the cell surface by the binding of IgE antibodies (34). Furthermore, it was described recently that IL-33 downregulates the expression of FceRI (35,36), indicating that the state of inflammation in the tissue can influence FceRI expression.
Heterogeneous expression of cell-surface markers on mast cells has scarcely been investigated. We investigated the heterogeneity of cell-surface markers in a quantitative manner using flow cytometry and did not find any markers that distinctly and consistently divided the studied mast cells into subpopulations with a bi-or multimodal distribution (data not shown). We did, however, find several markers with considerable continuous variation in expression within the mast cell populations ( Table 2), and co-stainings revealed that six of these markers, SUSD2, CD49a, CD326, CD34, CD66 and HLA-DR, were correlated ( Figure 4). To investigate whether these markers are correlated with the classic mast cell subpopulations MC T and MC TC , we costained for SUSD2 and CPA3, but no correlation was detected, ruling out the possibility that these markers are extracellular markers of the classic mast cell subtypes ( Figure 4I). Recently, CD38 was demonstrated to be differentially expressed in human nasal polyp mast cells where CD38 low were of the MC TC subtype, while CD38 high MCs were a heterogenous pool of both MC T and MC TC cells (37). In our study, CD38 did not distinctly separate the HLMCs into subpopulations and did not correlate with CPA3; i.e., in HLMCs CD38 could not be used to distinguish the MC T and MC TC subtypes ( Figure 4J). CD344 did not correlate with the MC T or MC TC profile either (data not shown). Another marker suggested to be expressed on MC TC is the complement 5a receptor CD88 (38). However, we did not detect significant expression of CD88 on HLMCs ( Figure 2). Thus, we were unable to find an extracellular marker that distinguishes the classical mast cell subsets.
One of the markers that showed high continuous variation in the HLMCs, CD63, is used as a surrogate marker for mast cell activation (39), i.e., degranulation. However, CD63 did not correlate to any of the other eight markers investigated, including the 6 markers that show correlation to each other SUSD2, CD49a, CD326, CD34, CD66 or HLA-DR indicating that these markers do not reflect varying degree of activation ( Figure 4G and data not shown).
Since two of our six heterogeneity markers that correlate, CD34 and SUSD2, is expressed on stem/progenitor cells (20,21,40), we wondered whether these markers could identify cells in different stages of mast cell maturation. However, if that was the case, one would expect cells with high expression of SUSD2/ CD34 to be small and contain few granules similar to mast cell progenitors (21). In contrast, the cells with high expression of SUSD2 had relatively high FSC and SSC values ( Figures 4K-M), suggesting that they were relatively large and granular and therefore unlikely to be immature mast cells. In this context, it is worth noting that mature murine mast cells also express CD34 (41), and in these cells, CD34 inhibits adhesion and is required for optimal migration (42).
SUSD2 is also expressed in certain cancers, in which it has been linked to proliferation (23); thus, one could imagine that  Figure S2).
In summary, we have identified the expression of 102 cellsurface antigens on HLMCs, of which 23 have not been described previously on MCs. Several of these antigens had a high continuous variability in their expression within the HLMC population. The expression of six of these markers correlated to each other and the size and granularity of the cells. Further studies are needed to determine how these cells differ functionally. None of the markers correlated with the intracellular protease expression. Thus, in contrast to the dogma of distinct mast cell subtypes, we demonstrate the continuous nature of HLMC heterogeneity.

DATA AVAILABILITY STATEMENT
The original contributions presented in the study are included in the article/Supplementary Material. Further inquiries can be directed to the corresponding authors. Quantification of FSC and SSC is shown in (M), mean ± SEM, n = 5. Two-way ANOVA with Bonferroni's multiple comparisons test was performed. **p < 0.01; ***p < 0.001; ****p < 0.0001.

ETHICS STATEMENT
The studies involving human participants were reviewed and approved by Regionala Etikprövningsnämden, Stockholm. The patients/participants provided their written informed consent to participate in this study.   the Karolinska Institutet. The funding sources were not involved in the study design, collection and interpretation of data, writing the report or the decision to submit the article for publication.