Mast cells (MCs) are primarily known for their prominent role in immunoglobulin E (IgE)-mediated allergies [1
]. However, they are evolutionary old and humans with a complete lack of MCs have not been identified, indicating an important physiological role of these cells. They are found in all studied vertebrates, and MC-like cells have also been identified in tunicates, a chordate, which represents an early evolutionary branch leading to all vertebrates, including mammals [3
]. Mast cells most likely act as early sentinel cells, inducing inflammation and attracting other inflammatory cells to the area of inflammation. They are primarily resident hematopoietic tissue cells that are distributed along both external and internal surfaces of the body where they act in the first line of defense [1
]. Mast cells are frequently found in the connective tissue of the skin and around blood vessels and nerves, as well as in the mucosa of the airways and intestines. Mature MCs are generally not found in the circulation. Two main types of MCs have been identified both in rodents and humans: Connective tissue MCs (CTMCs) and mucosal MCs (MMCs). Connective tissue mast cells are, as the name implies, primarily found in the connective tissue and in the peritoneal cavity whereas MMCs are mainly found in the intestinal mucosa of mice and rats and to a lesser extent in the lungs [4
Mature MCs store large numbers of cytoplasmic granules that are rapidly exocytosed following activation. The majority of proteins found in these granules are serine proteases, which can generally be subdivided into chymases and tryptases [10
]. Chymases are chymotrypsin-like and cleave substrates after aromatic amino acids whereas tryptases are trypsin-like in their specificity, with a preference for Arg and Lys in the P1 position [10
]. Very high amounts of these proteases are found in MCs, where the levels can reach 35% of the total cellular protein [14
]. Phylogenetic analyses of the chymases have led to the identification of two distinct subfamilies, the α-chymases and the β-chymases [6
]. Connective tissue mast cells also express an MC-specific carboxypeptidase- carboxypeptidase A3 (Cpa3). These proteases have been shown to inactivate snake, bee, and scorpion toxins; regulate blood pressure by angiotensin II generation; and control inflammation by cleaving a selective panel of cytokines [1
]. Mast cell proteases likely have several other important physiological functions, including connective tissue turnover and regulating coagulation [21
]. A potential role of the mouse MC tryptase (Mcpt6) to link innate and adaptive immunity in the chronic phase of Trichinella spiral
infection has also been observed [23
Mast cell granules also contain high levels of heavily sulfated, and thereby negatively charged, glycosaminoglycans, such as heparin or chondroitin sulfate, and also large amounts of vasoactive low molecular weight compounds, including histamine and serotonin [14
]. Mast cells also express a number of cell surface receptors and other molecules, of which many are MC specific or restricted to a few cell types. Such important receptors are the high affinity receptor for IgE (FcεRI) and the receptors for stem cell factor (SCF) and interleukin -3 (IL-3) [28
]. Connective tissue mast cells and MMCs show major differences in their expression of granule proteases and also in their levels of receptors and in numerous other aspects, indicating that they have, at least partly, different physiological functions.
To obtain a multi-faceted view of the phenotype of mouse MCs, here we performed a transcriptome analysis of purified mouse peritoneal MCs and analyzed MCs from other tissues for the expression MC-restricted compounds. By using a larger number of cells, as opposed to single cell analysis, we also increased the fidelity of the study to obtain a highly variable estimate of 20% to 30% of the top transcripts and a good quantitative estimate of the transcription levels of approximately all 21,000 mouse genes. Single cell analysis, with 1 to 1.5 million reads per cell, and analysis of the combined information from more than 50 cells may give similar data. However, no such information is available for a more detailed comparison.
Our results showed that MCs of different tissue locations are highly specialized, differing substantially in their expression of major granule compounds and enzymes involved in the processing of such compounds. This analysis has also resulted in the identification of a number of interesting genes expressed at much higher levels in MCs compared to the other 12 different tissues included in this study. Analyses of the roles of these newly identified proteins in MC biology may also give new insights into the complex physiological roles of this medically important cell. This detailed analysis of the MC transcriptome, and the abundance and heterogeneity of MCs in various mouse organs, can also serve as a solid base for future studies concerning the roles of MCs in vertebrate immunity.
2. Materials and Methods
Female BALB/c mice were purchased from Taconic Biosciences ((Europe) Ejby, Denmark) and maintained at the animal facility in the Biomedical Center (Uppsala University) or the Swedish Veterinary Institute (Uppsala, Sweden). The animal experiments were approved by the local ethical committee (Uppsala djurförsöksetiska nämnd; Dnr 5.8.18-05357/2018).
2.2. Generation of Bone Marrow-Derived MCs (BMMCs) and the effect of LPS Stimulation
Bone marrow cells were isolated from the femur and tibia of mice and grown in Dulbecco’s modified Eagle’s medium (Sigma-Aldrich, Saint Louis, MO, USA) containing 30% WEHI-3B-conditoned medium, 10% heat-inactivated fetal bovine serum (BSA) (Gibco, Carlsbad, CA), 100 U/mL penicillin, 100 μg/mL streptomycin, 2 mM L-glutamine (all from Sigma-Aldrich, Saint Louis, MO, USA), and 10 ng/mL IL-3 (PeproTech, Rocky Hill, NJ, USA). The medium was changed twice every week and cells were cultured at a concentration of 0.5 × 106 cells/mL in a humidified 37 °C incubator with 5% CO2 for at least 4 weeks to obtain mature and pure BMMCs. The cells were divided into two separate fractions, one was directly frozen in liquid nitrogen for the preparation of total RNA and the second was incubated in the above medium with the addition of 1 μg/mL of LPS per ml for 4 h, after which the cells were pelleted and frozen in liquid nitrogen for subsequent RNA preparation.
2.3. Peritoneal Cell Extraction and Sorting of Peritoneal Mast Cells
For the extraction of peritoneal cells, mice were euthanized by isoflurane overdose and neck dislocation, the abdominal skin was removed, and 9 mL of ice-cold phosphate-buffered saline (PBS) was injected into the peritoneal cavity. After shaking the abdomen, peritoneal lavage fluid was collected, and the cells were pelleted by centrifugation at 400× g for 10 min. The cells were resuspended in magnetic-activated cell sorting (MACS) buffer, containing 0.5% BSA in PBS pH 7.2, and 2 mM EDTA, followed by incubation with 20 μL of anti c-kit MicroBeads (Miltenyi Biotec, Bergish Gladbach, Germany). After 30 min, cells were washed, resuspended in MACS buffer, and passed through an LS column (Miltenyi Biotec, Bergish Gladbach, Germany). Magnetically labeled (c-kit+) cells were collected and used for RNA isolation and assessment of mast cell purity. Alternatively, to enhance the purity of peritoneal MCs, collected peritoneal cells were incubated in the first step with 1 μL of primary PE-Cy5-conjugated anti-lineage antibodies CD3 (17A2), CD4 (GK1.5), CD8b (eBioH35-17.2), CD11b (M1/70), CD19 (ebio1D3), B220 (RA3-6B2), Gr-1 (RB6-8C5), and TER-119 (TER-119). After 20 min, the cells were washed, resuspended in MACS buffer, and incubated with 20 μL Anti-PE-Cy5 MicroBeads (Miltenyi Biotec, Bergish Gladbach, Germany). After 30 min, cells were washed, resuspended in MACS buffer, and passed through an LD column (Miltenyi Biotec, Bergish Gladbach, Germany) according to the manufacturer’s instructions. Subsequently, the unlabeled (Lin−) cells were collected, washed, and in the second step incubated with 20 μL of anti c-kit MicroBeads for 30 min. After washing, cells were resuspended in MACS buffer and passed through an LS column (Miltenyi Biotec, Bergish Gladbach, Germany). The unlabeled cells were discarded while magnetically labeled (Lin− c-kit+) cells were collected and used for RNA isolation and assessment of mast cell purity. The anti-lineage antibodies were purchased from BD Biosciences (Franklin Lakes, NJ, USA) or eBioscience (Hatfield, United Kingdom).
2.4. Image Analysis
The magnetically sorted single run CD117+ cells and the Lin− CD117+ cells were cytospun onto glass slides using a Shandon Cytospin 2 (Thermo Fisher Scientific, Inc., Waltham, MA, USA) and were allowed to dry before staining with toluidine blue using a standard protocol to assess the purity of sorted Lin− CD117+ MCs. The cells were imaged using a Nikon Eclipse Ni-U microscope (100× or 200× magnifications).
2.5. RNA Isolation
RNA isolation from purified cell fractions: Total RNA was prepared from MACS-sorted cells, cultured BMMCs plus 4-h LPS-treated cells using the Nucleospin RNA kit (from Macherey-Nagel, Germany), according to the manufacturer’s recommendations. The RNA was eluted with 30 μL of DEPC-treated water, and the RNA concentration was determined by using a Nanodrop ND-1000 (Nano Drop Technologies, Wilmington, Delaware, USA). The integrity of the RNA was confirmed by visualization on 1.2% agarose gels using ethidium bromide staining.
RNA isolation from tissue: Ear, lungs, brain, tongue, heart, liver, kidney, pancreas, duodenum, proximal part of the colon, spleen, and uterus tissues were dissected from the mouse. Immediately after removal, the tissues were frozen in liquid nitrogen, and made into a fine powder by grinding with a pestle and mortar. The tissue powder was then used for total RNA isolation using the protocol described above.
2.6. Analysis of the Transcriptome by RNA-Seq and by the Thermo Fisher Ampliseq Chip and PCR-Based Method
Total RNA from the different cell fractions and whole tissues were used for transcriptome analyses (GATC-Biotech, Ebersberg, Germany). mRNA was purified by poly-A selection following fragmentation of the RNA and sequencing of 30 million fragments was performed. Individual read lengths of around 50 to 100 nucleotides were then matched against a reference transcriptome library. The FPKM values were normalized based on the total number of reads and sent to us as a large Excel file with all the different cell and tissue samples in one file for easier comparison. The number of reads per gene was, for the RNA-seq data from GATC, adjusted to the transcript length as longer transcripts generate more fragments per mRNA and therefore a higher number of reads. Following the sorting of the 33,915 different transcripts from the most highly expressed to the lowest, we manually went through the entire list to identify the MC-specific highly expressed genes and the genes that were markedly more highly expressed in MCs compared to other tissues. The results from the RNA seq analysis matched well with previous cDNA library screenings and also with the Thermo Fisher chip-based Ampliseq transcriptomic platform (Ion-Torrent next-generation sequencing system- Theromofisher.com). The Thermo Fisher mouse Ampliseq transcriptome analysis platform is based on the purification on a chip of the individual mRNAs (as cDNAs), which are then PCR amplified and sequenced individually. The RNA is not fragmented but copied into cDNAs before binding to the chip. In the Ampliseq analysis, every mRNA is read only once and the number of reads then corresponds to the expression level more directly. Here, as for the RNA seq analysis, all the data from 16 different samples were sent to us as a large Excel file containing all the transcripts for all the 16 samples in one file, which enabled an easy comparison of the transcript levels between cell fractions and tissues. Following the sorting of the 23,931 different transcripts from the most highly expressed to the lowest, we manually went through the entire list to identify the MC-specific highly expressed genes and the genes that were markedly more highly expressed in mast cells compared to other tissues. In total, 265+ such genes were identified for the pure MC preparation and all of these genes are listed in the Supplementary Material
, with the tissue and the expression level in that tissue that showed the second highest expression of this gene being named in the table for an easier comparison.
Analyses of total transcriptomes by different techniques offers powerful tools in the analyses of cells, and also allows studies of the development of cells and tissues. For such studies, several alternative techniques have recently been developed. However, remarkably few quantitative studies using such tools have been published. In general, only the fold increase or decrease of transcripts during stimulation is presented and absolute values of expression are rarely given. Data are often also not presented by real numbers but rather as heatmaps, which lack information concerning the abundance of the various transcripts. The presentation of only a fold change can also be misleading concerning their relevance. Frequently, transcript levels can increase 10- or even 100-fold, but when their baseline expression is starting at an extremely low level, it is questionable if such increases are biologically relevant. For example, here we showed that the expression of granzyme C increases by almost 1000-fold after LPS stimulation of BMMCs. However, the original levels were very low (0.1 reads), and we can therefore not be certain that the increase in granzyme C has any biological impact. For this reason, we propose that the absolute values of gene expression (rather than fold changes) are more informative concerning the biological impact. In a hallmark study by Dwyer et al., many MC- and basophil-specific transcripts were identified [30
]. Many of the transcripts we describe in this study were also identified in that study. However, it is important to note that the former study by Dwyer et al. did not provide absolute values of expression. Almost all the MC-related transcripts are marked in the same bright red in their heatmaps even if they differ by two to three orders of magnitude in expression levels in our analysis. One of the aims of this study was therefore to generate a deeper understanding of the MC transcriptome by providing data on the absolute expression levels for the various transcripts. This study also involved many more transcripts than the previous study and identified many new proteins that may have major impacts on MC biology, which could be the target for future studies. In contrast to previous studies, this study also involved the analysis of the abundance and heterogeneity of MCs in 12 different mouse organs.
We previously performed a study on rat peritoneal MCs where we measured the absolute values of expression. At that time, the only reliable available method was the use of unamplified cDNA libraries. We therefore hybridized individual filters with a panel of probes able to obtain accurate quantitative information concerning transcript levels for approximately 26 different transcripts (Figure 2
However, this technique is extremely time consuming and labor intensive, which hampered a more extensive analysis. However, by using the new transcriptome platforms it is now possible to obtain reliable information of almost all the 21,000 genes of the mouse or the human transcriptome, and also information on the abundance of different splice variants. Despite this, these technologies are still relatively poorly validated, mainly because appropriate reference libraries have not yet been established. Our earlier cDNA library data was therefore valuable as a reference material for the optimization and validation process in this study. The possibility of using two independent technologies with the same sample also increased our confidence that the data obtained was reliable. By using the information obtained by both the RNA-seq as well as the chip and PCR-based transcriptome analyses, we can now confirm that the general pattern observed from our previous studies using the unamplified cDNA libraries holds remarkably well.
The transcript levels for the major MC granule proteins are remarkably high, in the range of several percent of the total mRNA pool of the cell [7
]. The levels of the most characteristic MC cell surface receptors, including FcεRI, c-kit, the IL-33 receptor, and the IL-3 receptor, were, in general, close to two orders of magnitude lower than for the major granule proteins. The expression of processing enzymes, including cathepsin C (involved in activation of the granule serine proteases), enzymes involved in histamine, serotonin, and heparin synthesis, histidine decarboxylase, tryptophane hydroxylase, and NDST-2, was also relatively low. An estimate of the house-keeping genes indicated that they represent a large fraction of the total transcriptome. A calculation of the transcriptomes of the MCs and the different tissues (skin and lungs), based on the 300 top transcripts, indicated that the house-keeping genes accounted for 60% to 80% of the total mRNA pool of a cell or tissue, with the remaining transcripts being tissue specific. Among the latter, a few major transcripts accounted for the absolute majority, where in the case of the MCs, these transcripts coded for the major granule proteases and the heparin core protein (serglycin).
During MC development, the first proteases to appear are Mcpt5 and Cpa3, whereas Mcpt4 and Mcpt6 appear later and are therefore markers for more mature MCs. An interesting finding in this study was the very low levels of Mcpt4 expression in BMMCs grown in the presence of IL-3. The levels of Mcpt4 were ~1000 times lower than in peritoneal and skin MCs. Mast cells grown in the presence of SCF, a major MC growth factor, have also been shown to express low levels of Mcpt4, indicating that yet unknown factors are needed to obtain full differentiation into mature tissue-resident MCs [45
] The level of Mcpt6 was also considerably lower in BMMCs compared to the peritoneal or skin MCs, whereas the IgE receptor α-chain was expressed very highly in BMMCs compared to peritoneal or skin MCs, indicating an immature phenotype of BMMCs. Interestingly, the addition of SCF to the MC cultures seems, at least in C57BL/6 mice, to upregulate Mcpt6 but not to increase the levels of Mcpt4 [45
When examining the transcriptome of the ear skin and the tongue, we noted that the skin MCs were almost exclusively of the CTMC subtype. The skin and tongue MCs expressed remarkably high levels of four different proteases, of which three were serine proteases and one a carboxypeptidase: Mcpt4, Mcpt5, Mcpt6, and Cpa3 (Table 3
). We detected almost no expression of the mucosal MC-specific proteases Mcpt1 and Mcpt2, and no expression of the basophil-specific protease Mcpt8 (Table 3
When assessing the lungs, we found very low levels of MC- and basophil-related transcripts, confirming the previous findings of such levels in mouse lungs [34
]. The situation is somewhat different with human lungs, which have relatively high numbers of MCs [48
]. This is very relevant for the discussion when using mice as a model for studying the impact of MCs on lung function. However, recently it has been found that MC numbers increase quite substantially in the mouse lung upon different viral infections or allergen challenge [49
]. Viral infections also lead to increased MC numbers in human lungs [51
]. During these infections, it seems that primarily MCs of the mucosal type increase and that this increase is transient, which most likely originates from MC precursors originating from the bone marrow [49
]. Quite dramatic increases in MC numbers can also appear in other tissues upon infection. Massive increases in mucosal MC numbers are generally observed upon intestinal worm parasite infections [9
]. An interesting observation comes from Schistosoma mansoni
-infected BALB/c mice, where a strong upregulation of both MMC and CTMC proteases occurs in the intestinal region, and a very strong upregulation of the basophil-specific transcript for Mcpt8 is seen the lungs, indicating a marked infiltration of basophils [8
]. These findings indicate that both MCs and basophils can fluctuate quite dramatically in numbers upon various viral, parasite, and bacterial infections.
By extended analyses of a more than 95% pure MC population from the peritoneum and a simultaneous analysis of 10 additional tissues, we could look deeper into the MC transcriptome and also obtain a more general picture of the presence of MCs and their subtypes in a large panel of different tissues. As for the initial analysis, we noted that the four CTMC proteases are the major tissue-specific transcripts of the CTMC. Next was the core protein for heparin and chondroitin sulfate synthesis, serglycin. When examining the surface receptors, the cell adhesion molecules, and the enzymes involved in histamine, leukotriene, prostaglandin, and PAF synthesis, we noted expression of at least two orders of magnitude lower than for the highest expressed genes. Interestingly, we also saw that FcγRIII was more abundant than the IgE receptor, and that the inhibitory receptor FcγRIIb showed relatively low expression compared to the other two dominating Ig receptors. Also of note, the Ig receptor-related receptor allergin-1 or Milr1 was expressed at a level similar to the alpha chain of the IgE receptor (345 and 283 reads, respectively). This receptor may, together with the inhibitory receptor FcγRIIb, control signaling from the IgE receptor [37
]. This analysis has also made it possible to examine in more detail the presence of known and previously unidentified transcription factors, signaling molecules, and molecules involved in vesicular transport. These analyses identified a number of genes that had previously not been linked to MCs, for example, Fxyd5, mucin 13, Myb, Zeb2, Tox2, Runx3, and Fli1. We also found several that have previously been identified as being MC expressed, including Gfra2, histamine receptor 4, and CREB3l1 [30
]. Further work will be required to establish the importance of these genes in MC development/function.
Most of the immune-related cytokines, including interleukins and colony-stimulating factor, were expressed at very low levels. They were expressed higher in MCs than in other tissues but were nevertheless very low, indicating that MCs primarily produce cytokines upon activation. The low level of cytokine mRNA thereby confirmed our previous findings from rat peritoneal MCs [7
] (Figure 2
) and has recently also been confirmed in another study of the effect on chromatin opening by various MC stimuli, where the levels of cytokine mRNA were found to be very low in the absence of stimulatory signals [53
]. Interestingly, and in contrast to the interleukins, some chemokines were expressed at relatively high levels in MCs, including Ccl2 and Ccl6. Ccl2, also named monocyte chemoattractant protein 1 (MCP1), may have a role in maintaining sufficient numbers of monocytes/macrophages in the tissue, also during non-inflammatory conditions. Three cytokines of importance for fibroblast growth and differentiation were also found to be relatively highly expressed: TGF-β, fibrosin 1, and TGF-β activator (Nrros), indicating a role of MCs in connective tissue homeostasis. Mast cells may also participate in the normal connective tissue turnover via their proteases. These can be released by piecemeal degranulation during non-inflammatory conditions and can, by degrading fibronectin and activating pro-collagenases, remove damaged connective tissue components and provide space for new such components produced by fibroblasts. The production of fibroblast-activating cytokines by MCs may thus be important for tissue maintenance. In line with this, Mcpt4 knockout mice accumulate excessive connective tissue in an age-dependent manner [31
]. TGF-β also has a potent anti-inflammatory effect and MCs may thereby have a potent anti-inflammatory role in the tissue, in addition to their pro-inflammatory impact. It is also interesting to note that MCs express high levels of galectin 1, which also has anti-inflammatory properties, and suggests that this expression also adds to the potential anti-inflammatory effects of MCs [55
]. The detection of very high levels of the initial enzyme in the steroid hormone synthesis pathways (P450scc) was unexpected. The expression of P450scc in MCs was at least 300 times higher than in any of the other tissues studied. However, low levels of the other enzymes in steroid hormone production were detected, indicating that the first product in this cascade, pregnenolone, may have a physiological role in the connective tissue.
Only very low levels of TLRs were observed. We only detected the expression of two TLRs, TLR-4 and TLR-13, which indicates a low response by normal tissue MCs to TLR ligands during non-inflammatory conditions. Very low levels of CD40 and the CD40 ligand were also observed, indicating that MCs have a minor direct role in B-cell responses, as CD40-CD40L interaction is a key component for such responses. However, cytokines produced by MCs during inflammatory conditions may, if present in sufficient concentrations, have a role in isotype switch regulation. The expression of complement factor 2 (C2) with 159 reads in MCs was also interesting, which was three times higher than in the liver sample with 50 reads. Liver is thought to be the primary site for the production of most complement and coagulation components, and this finding indicates that this is not the situation for all of these factors and that MC actually can be an important production site for C2.
Concerning the role of transcription factors in MC and basophil development, several, including GATA-2 and MITF, are well established as key regulators, and others, like IRF8, STAT5, and C/EBPa, are also central to this process [44
]. Here, we noted that GATA-2 and MITF are relatively MC specific, whereas IRF8, STAT5, and C/EBPa were more broadly expressed, indicating that some factors may be guiding whereas others are necessary as supporting factors. It will be interesting to determine the role of several of the factors with high or medium MC specificity identified in this transcriptome analysis, for example, Myb, Zeb2, Tox2, Runx3, CREB3l1, and Fli1, for their roles in MC development. Further, a number of factors with a potential role in cell signaling, as well as the control of intracellular calcium signaling and granule release, were identified. Some of these are already well characterized whereas the roles of others in regulating MC function remain to be investigated.
It is apparent from this study that MCs express an extremely complex network of receptors, signaling molecules, and transcription factors. High levels of expression were seen for the activating receptors FcεRI, FcγRIII, and CD200R3, the Mas-related receptors (e.g., Mrgprb2), receptors for ATP, and endothelin receptors. These activating receptors are in turn controlled by a number of negatively acting receptors, including FcγRIIb, CD300a, and Allergin-1, as well as possibly by the soluble factor galectin-1. TGF-β secreted by MCs also has an anti-inflammatory role. Interestingly, the neurotrophin receptor Gfra2 was highly expressed in MCs, and one of its ligands, neuroturin, has been shown to have potent anti-inflammatory effects [56
]. Relatively high expression levels were also seen for PD-L1, indicating that MCs also can have a dampening function on cytotoxic T cells (Supplementary Table S8
). The proteases released by MCs also regulate the activity of endothelin-1 and of a set of Th2-promoting cytokines, including IL-18, IL-33, TSLP, IL-15, and IL-21, by cleavage. Endothelin-1 is inactivated primarily by CPA3, whereas chymase cleaves IL-18, IL-33, and IL-15, and tryptase cleaves TSLP and IL-21 [19
Analysis of the different tissues for MC-specific protease transcripts also gave an insight into the presence of MCs in different organs. Very low levels of MC-specific transcripts were seen in many tissues, including the brain, liver, kidney, heart, pancreas, and colon. In contrast, relatively high levels of transcripts corresponding to CTMCs were seen in the ear, tongue, and uterus, and high levels of MMC transcripts were found in the duodenum. We observed relatively large amounts of a panel of keratin transcripts in the tongue, indicating some similarities between the tongue and skin, which may have a role in the preferential presence of CTMCs in these tissues. Interestingly, the keratin genes expressed in the two tissues were quite different, with keratins 10, 2, and 14 dominating in the ear and 13, 4, and 16 dominating in the tongue.
As mentioned already in the beginning of the discussion, a large number of the above described MC-related transcripts were also identified in the hallmark study by Dwyer et al. and in the very important papers from the Immgen consortium [30
). The ones identified in these studies include Plau, Adamts9, C2, Hpgds, the Mrgpr receptor members, CD200r3, CD34, and histamine receptor 4. In these studies, different MC proteases and IgE receptor components as well as the enzymes involved in histamine and heparin synthesis and transport, the chemokine Ccl2, and also a number of the abovementioned transcription factors and signaling molecules. including GATA-2 and MITF, were identified as being MC expressed. The coherence between these different studies makes us confident that we now have a quite detailed view of the transcriptome of the mouse MC. An important study of the human MC transcriptome has also been published by Motakis et al., which makes it possible to look at similarities and differences between human and mouse MCs [59
]. In this study, the transcriptome of human skin MC was analyzed by deep-CAGE sequencing. In MCs isolated from human skin without further culturing, high levels of transcripts for the human MC proteases, including the β-tryptases, Cma1, and CPA3, and interestingly, almost four times higher levels of cathepsin G transcripts than for the second most abundant MC protease, the β-tryptase, were found. This in marked contrast to the mouse, where cathepsin G is expressed at a level of only 1% of the level of the mouse tryptase Mcpt6 (Table 1
). They also find very high levels of several of the important cytokine receptors for MC development, including kit and IL1RL1(ST2), and also the α and β chains of the IgE receptor (FcεRI). The levels of kit transcripts in these human MC were actually even higher than the level of transcripts for the β-tryptase [59
]. They also found high levels of the human counterpart to the mouse substance P receptor Mrgprb2, the Mrgprx2 [59
]. The levels of the different toll-like receptors were also found to be very low in the human skin MCs, as we observed in the mouse [59
]. So, although there are clear similarities, there are also major differences between mouse and human MCs as, for example, the transcript levels of the receptors, which appear to be in the same range as the proteases in human MCs whereas they are almost two orders of magnitude lower than the protease transcripts in the mouse.
In conclusion, new methods for transcriptome analysis can now give us high quality, quantitative information concerning the entire transcriptome in various tissues and cell preparations. These methods can serve as powerful tools to study the biology and phenotype of a cell type or an organ. Additionally, the use of tissue-specific expressed genes makes it possible to study individual cell populations within a tissue without the need to purify them from the tissue, which reduces the error in the analysis resulting from difficulties in obtaining quantitative recovery of minor cell populations. Here, these techniques were used to obtain quantitative information concerning MCs in various tissues, high quality data on types and numbers of these cells in different organs, and indications for still unknown tissue factors that are of major importance for MC development. The immature phenotype of BMMCs, which are often used as tissue equivalents in studies of MC biology, also raises questions of how relevant they are for such studies. The detailed analysis of a relatively pure population of peritoneal MCs and its comparison with transcript levels in 10 different mouse tissues also made it possible to identify a large number of new MC-specific proteins, of which many may have a major role in MC biology. However, in-depth analyses of the roles of these newly identified proteins in MC differentiation, inflammatory, anti-inflammatory responses, and in normal tissue homeostasis will be needed before we have the complete picture of their precise role in MC biology. This detailed quantitative analysis of the MC transcriptome in the mouse and the abundance and subtypes of MC in different organs can also serve as a solid base for future studies of the role of MCs in vertebrate immunity. As previously mentioned, MC-like cells have been identified in early chordates, as represented by the sea squirt Ciona intestialis
(a tunicate) [3
]. These cells have been shown to stain positive with toluidine blue, as mammalian MCs, and to contain heparin and histamine, as well as expressing a serine protease with tryptic activity and being able to produce prostaglandin D2 [60
]. However, the relationship of this tryptic protease to the human MC tryptase and mouse Mcpt6 or Mcpt7 is not yet established. Zebrafish MC-like cells have also been shown to stain positive with toluidine blue, to express a carboxypeptidase that shows 38% identity with human CPA3 and 64% identity to human pancreatic CPA1, and to stain positive for c-kit [61
]. Zebrafish MC-like cells are also dependent on GATA2 and Pu-1 for their development, similarly to mammalian MCs, but do not need GATA-1 [61
]. Interestingly, zebrafish MC-like cells also express lysozyme C and MPO, two markers that are not normally associated with mammalian MCs but instead are associated with neutrophils in humans, indicating that there are also clear differences between fish and mammalian MCs [61
]. A clearly identifiable homolog to the IgE receptor alpha chain has also been identified in the opossum, a marsupial, with possible similar receptor components in the platypus also [28
]. However, it is not yet known if they are expressed in MCs. These studies show that information is also starting to accumulate concerning MCs not only in humans, mice, and rats but also other vertebrates. The recently developed transcriptome platforms can now be used to obtain a more detailed picture concerning the evolution of MC and MC-like cells in different branches of the vertebrate evolutionary tree. This detailed analysis of the mouse MC transcriptome can thereby be used as a solid reference for such studies.