Differential mast cell numbers and characteristics in human tuberculosis pulmonary lesions

Tuberculosis (TB) is still a major worldwide health threat and primarily a lung disease. The innate immune response against Mycobacterium tuberculosis (Mtb) is orchestrated by dendritic cells, macrophages, neutrophils, natural killer cells and apparently mast cells (MCs). MCs are located at mucosal sites including the lungs and contribute in host-defence against pathogens, but little is known about their role during Mtb infection. This study investigates the location and characteristics of MCs in TB lesions to assess their contribution to TB pathology. To this purpose, number, location and phenotype of MCs was studied in 11 necropsies of pulmonary TB and 3 necropsies of non-TB infected lungs that were used as controls. MCs were localised at pneumonic areas, in the granuloma periphery and particularly abundant in fibrotic tissue. Furthermore, MCs displayed intracellular Mtb and IL-17A and TGF-β immunostaining. These findings were validated by analysing, post-mortem lung tissue microarrays from 44 individuals with pulmonary TB and 25 control subjects. In affected lungs, increased numbers of MCs expressing intracellularly both tryptase and chymase were found at fibrotic sites. Altogether, our data suggest that MCs are recruited at the inflammatory site and that actively produce immune mediators such as proteases and TGF-β that may be contributing to late fibrosis in TB lesions.

www.nature.com/scientificreports/ phase culminates in irreversible lung tissue damage manifested by necrosis and fibrosis 8,9 . Although many cells are involved in this process, little is known about the contribution of MCs in this pathology. MCs are distributed in lungs and mucosal tissues and contribute to host-defence against bacterial infections 12 . In humans, MCs are classified as tryptase + MCs (MC T ), chymase + MCs (MC C ) or both tryptase + and chymase + MCs (MC TC ) 13 . Upon bacterial exposure, they release a wide variety of cytokines and chemokines, including IL-17, TNF-α, IL-8 and TGF-β either by degranulation or canonical secretory pathways 14 . Additionally, during lung infections MCs are altered in numbers, phenotype and localization 15,16 . For instance, MC numbers are decreased in lungs of Streptococcus pneumoniae infected patients 17 . Furthermore, MCs are capable to phagocytose bacteria and present antigens 18 . Although hypothetically MCs have an important role in TB 3 , little has been explored. For instance, it is unknown if MCs are localized at pulmonary tuberculous human lesions and what is the predominant MC phenotype. It is also unclear whether MC cytokines, e.g. IL-17A, and TGF-β contribute to the inflammatory process, and formation and maintenance of granulomas and fibrogenesis.
In this study we examined number, distribution and phenotype of MCs in human TB-infected lungs and MC cytokine expression at granulomas and fibrotic sites in TB-infected lung tissue. Our descriptive findings demonstrated that MCs are likely to participate to the early inflammatory phase, as well as to the late fibrosis formation during TB pathology.

Materials and methods
Ethical statement and tissue procurement. All methods were performed in accordance to relevant guidelines and regulations. Lung tissue sections from 11 necropsies of deceased TB patients and 3 controls from non-TB infected necropsies were obtained from the Pathology Department files at the National Institute of Medical Sciences and Nutrition "Salvador Zubiran" in Mexico City. A specific informed consent was not required for this study, but every medical autopsy was performed with a signed authorization of the representative legal family member and tissue samples were obtained during legally authorized autopsies with signed permission by a relative who agreed tissue sample donation for the present and previous studies 19 . The microarray tissue study from 44 individuals with pulmonary TB and 25 control subjects were taken from the Pathology Department of the General Hospital of Mexico 'Eduardo Liceaga' at Mexico City, and ethical statement was approved by the local ethic committee of the Hospital Infantil de Mexico, Federico Gomez (No. HIM/2008/015). This study including all experimental protocols were approved by the in-house ethical committee at the National Institute of Medical Sciences and Nutrition "Salvador Zubiran". TB tissue processing. Macroscopically, lung tissue from 11 necropsies of deceased TB patients showed extensive cavitary bilateral lung disease, surrounded by numerous white nodules with irregular shapes and size that alternated with pneumonia patches. Extensive sampling was performed, obtaining several tissue fragments from different lesions and were embedded in paraffin blocks sectioned at 3 µm and mounted in glass slides. Additionally, we used post-mortem lung tissue, essentially granulomas, from 44 individuals with pulmonary TB and 25 control subjects (subjects who died as a result of any other cause without significant pulmonary disease). The lung tissues (TB and control tissue samples) were organized in a tissue microarray (MTA) as previously described 20 . One sample of each section (lung sections and MTA) were stained with haematoxylin and eosin (HE) to select lung lesions for study. Spare sections were left at room temperature before being used for immunoperoxidase and immunofluorescence staining. The lung tissues in this study correspond to active TB cases, not latent TB cases were included. Immunoperoxidase staining. Lung sections were deparaffinized and treated with antigen retriever (1X; Bio SB, Santa Barbara, California) for 5 min under microwave heating. Endogenous peroxidase was blocked incubating tissue with methanol-H2O2 (9:1) for 10 min. After three washes, unspecific sites were blocked using a background sniper (BIOCARE MEDICAL; Pacheco, California) for 30 min. Slides were washed and incubated with either a rabbit anti-human chymase antibody (Ab186417, Abcam; Cambridge, United Kingdom) or a mouse anti-tryptase antibody (Ab2378, Abcam; Cambridge, United Kingdom) for 2 h. After three washes, tissue was processed using a mouse/rabbit PolyDetector DAB (3-3′-diaminobenzidine)/HRP (horseradish peroxidase) brown detection system (BSB0219, Bio SB; Santa Barbara, California) following manufacturer´s instructions. Micrographs were acquired using a LEICA DMLS microscope with a 2.5X and 40X dry objectives equipped with a LEICA DFC295 camera and analysed using an automated image analyser (QWin Leica; Wetzlar, Germany). Immunofluorescence staining. To visualize MC phenotypes, lung tissue sections were deparaffinized, treated with DNA retriever (1X; Bio SB; Santa Barbara, California) for 5 min under heating and incubated with blocking buffer (goat serum 1:10 in PBS + tween 0.1%) for 30 min. After three washes, tissue were incubated with a rabbit anti-human chymase antibody (Ab186417, Abcam; Cambridge, United Kingdom), and a mouse antitryptase antibody (Ab2378, Abcam; Cambridge, United Kingdom) for 1 h followed by 1 h incubation with a goat anti-mouse antibody conjugated to Alexa 488 fluorophore (Ab150117, Abcam; Cambridge, United Kingdom) and a goat anti-rabbit antibody conjugated to Alexa 647 fluorophore (Ab150083, Abcam; Cambridge, United Kingdom). After three washes, tissue was mounted using a fluoroshield mounting media containing 4′,6-diamidino-2-phenylindole (DAPI, Abcam; Cambridge, United Kingdom). Slides were analysed using a fluorescent microscope OlympusBX41 with either 40 × and 10 × dry objectives. Images were acquired using a Zen 2.6 blue and analysed using Fiji.

MC quantification.
Forty-four MTAs from TB-infected lung sections from autopsy cases and 22 non-TB infected controls contained in 4 different slides were stained with HE and analysed. Five MTAs were selected as control lung tissue and 10 MTAs presenting fibrosis were selected as representative fibrotic tissue. Selected MTAs were immunostained with tryptase and chymase (as described above) and studied at 10 × magnification using a fluorescent microscope OlympusBX41 and acquired using Zen 2.6 blue software system. One high power field was taken for each MTA and all single positive (MC C or MC T ) and double positive (MC TC ) cells were counted per field using Fiji. MC numbers were graphed using GraphPrism.

Statistical analysis.
A Shapiro-Wilk tests was performed to determine normality during phenotype quantification. Statistical analysis was achieved using Kruskal-Wallis test and a Dunn's multiple comparison post-test (adjusted p ≤ 0.01) using GraphPad Prism 8th edition.
Compliance with Ethical standards. Authors declare that no conflict of interest happened during the present study. The present research involved the use of lung sections obtained from deceased individuals. Autopsies were selected from the Pathology Department files at the National Institute of Medical Sciences and Nutrition "Salvador Zubiran" in Mexico City. A specific informed consent was not required for this study, but every medical autopsy was performed with a signed authorization of the representative legal family member and tissue samples were obtained during legally authorized autopsies with signed permission by a relative who agreed tissue sample donation for this and previous studies 19  Experimental procedures. All methods were performed in accordance to relevant guidelines and regulations. This study including all experimental protocols were approved by the in-house ethical committee at the National Institute of Medical Sciences and Nutrition "Salvador Zubiran".

Results
Tryptase positive mast cells are the most abundant phenotype in non-TB infected human lungs. In physiological conditions MCs expressing either tryptase or chymase or both proteases reside in alveolar parenchyma. However, in pulmonary infections MC numbers and phenotype are altered 21 . To investigate lung MC distribution and their characteristics we studied their number, location and phenotype in autopsies from control lungs (non-TB infected). Control cases had heart attacks as death cause with lungs showing overall a normal structure with some focal patches of centrilobular emphysema (Fig. 1A). These tissues showed MC T (Fig. 1B) and lesser MC C (Fig. 1C) at alveolar walls. Both MC T and MC C were preferentially located in blood vessels adventitia. Moreover, MC T were the most abundant phenotype (median = 5 cells per field) ( Fig. 1D) followed by MC TC (median = 2 cells per field), whereas MC C were not detected. In fact, all MC C observed were also tryptase + therefore MC TC (Fig. 1E). Thus, MCs expressing only chymase were rare whereas tryptase positive MCs were predominant in human lung parenchyma.

Mast cells are located at active inflammatory sites of TB-infected lungs and show intracellular mycobacterial antigens.
MC numbers, location and phenotype found in non-infected lungs were used as controls to investigate MC characteristics in TB lung lesions. A table summarizing clinical data from the studied TB active necropsies is shown in Supplemental Table 1. Both MC phenotypes (MC T and MC C ) were abundant at inflammatory (granulomas, pneumonia, vascular and airways walls) and fibrotic areas (Fig. 2), but absent in the proximity of necrotic sites (Supplemental Fig. 1). At pneumonic areas ( Fig. 2A), numerous MCs were seen in alveolar walls and alveolar lumen (Fig. 2B,C), while in blood vessels (Fig. 2D) MCs were found in the adventitia and in bronchial airways (Fig. 2E,F). Both phenotypes were positioned below the epithelium, in the submucosa and in the muscular wall between smooth muscle cells. Furthermore, as shown in Fig. 2G, MC T stationed in inflammatory regions of TB-infected lung sections showed vacuoles containing Mtb antigens. Thus, both MC T and MC C reside in TB lung lesions and store Mtb.
Mast cells are located at the periphery of granulomas and express IL-17. Granulomas are characterized by a necrotic core containing Mtb surrounded by macrophages and lymphocytes and a fibrotic external layer 20 . As shown in Fig. 3, granulomas at different stages were analysed. In early or incipient granulomas (Fig. 3A), characterized by small nodular congregates of inflammatory cells, occasional MCs were seen intermixed with lymphocytes and macrophages (Fig. 3B,C). However, MCs were not observed to infiltrate typical or www.nature.com/scientificreports/ mature well-organized granulomas (Fig. 3D) but were located at their periphery (Fig. 3E,F). Indeed, MCs were abundant at the fibrotic outer layer of necrotic granulomas (Fig. 3F). Representative TB-infected tissue containing typical granulomas or incipient granulomas was incubated with anti-IL-17A and anti-tryptase antibodies followed by a fluorescent staining. As shown in Fig. 3G, IL-17 positive MCs were observed at neighbour inflammatory tissue near to typical granulomas or mixed with inflammatory cells in incipient granulomas. Thus, MCs are virtually absent inside mature granulomas but located at their periphery and expressing IL-17.

Mast cells are in high numbers in fibrotic tissue that surrounds granulomas and cavitary lesions.
MCs were constantly detected at fibrotic areas around granulomas or wall cavities supporting a number of studies that have shown MCs participate in the fibrotic process 22,23 . To characterize the nature of the MCs residing at fibrotic areas of TB infected tissues we determined their proteases expression. Although MC T , MC C and MC TC phenotypes were all seen in fibrotic areas (Fig. 4A), MC TC were the most abundant (median = 8.5 cells per field), followed by MC T (median = 2 cells per field) (Fig. 4B). Since non-TB infected lung controls are colonised by MC T (Fig. 1D), our data suggests a switch in proteases expression with an increase of chymase at TB-induced healing and fibrotic sites. Furthermore, by a double immunofluorescence labelling for TGF-β and tryptase, we could demonstrate that some MCs in fibrotic tissue not only express TGF-β ( Fig. 4D) but exhibit a conserved but disorganized granular content suggesting their activation or partial degranulation (Fig. 4C). Thus, MCs at TB-induced lung lesions show an increased expression of chymase and pro-fibrogenic TGF-β and both may contribute to the fibrotic process.

Discussion
Although MCs may play a role in TB infection 3 , little has been explored and the available studies were performed in rodent models. Furthermore, considering the differences in MC numbers and phenotypes between mouse and human lungs it is unclear whether the described findings are relevant for human pathology 21 . Our study demonstrates that in TB-infected human lungs MCs are located at pneumonic areas, in proximity to granulomas and particularly abundant at fibrotic sites. Besides, we observed a MC colocalization with Mtb antigens indicating an intracellular localisation of the pathogen. In addition, MCs located in proximity to granulomatous lesions were found to express IL-17, while TGF-β positive MCs were found embedded in the fibrotic tissue.  www.nature.com/scientificreports/ Our observations in lung controls showing MCs surrounding blood vessels and throughout the alveolarcapillary interstitial tissue with MC T as the most abundant population are in line with current evidence that shows MC T as predominant phenotype in large airways including bronchial and alveolar, followed by MC TC which are more common at blood vessels, whereas MC C are rare or sporadically observed in human lungs 24,25 . Although it is not clearly understood yet, this suggests that MC T reside in the alveolar parenchyma prone to activation or differentiation upon environmental stimulus 16 .
Since MCs release a wide variety of cytokines, chemokines and antimicrobial molecules, they are likely to be involved in TB pathogenesis at different stages 3 . For instance, since our findings show that MCs locate at inflammatory areas of TB-infected human lungs, here MCs may contribute to the initial TB inflammatory stage by releasing TNF-α, IL-6, MCP-1, IL-1β, GM-CSF and IL-8. Muñoz et al. support this concept by showing the release of MC-dependent TNF and IL-6 upon Mtb infection 26,27 . Furthermore, the ability of MCs to contribute to immune cell recruitment was observed in Chlamydia pneumoniae lung infection characterised by a MCdependent cellular infiltration in the lung airways 12,28-30 . Instead, the appearance of MCs at incipient granulomas and surrounding mature granulomas suggests their contribution in orchestrating granuloma formation and maintenance potentially by the secretion of IL-17, IL-6, IL-8, MCP-1, IL-10, IFN-γ and IL-1β 31,32 . However, MC mediators including TGF-β, IL-33, chymase and IL-1β are known to induce excessive inflammation and fibrosis and have detrimental effects 31 . Thus, suppressing MC-mediators release, e.g.IL-1β and TGF-β may diminish MC activation 33 as proposed in SARS-COV2 infection 34,35 and serve to prevent TB fibrosis.
In our study, we observe that human-lung resident MCs display intracellular Mtb fragments. This is in line with reported data 26,27 showing MCs able to uptake Mtb via membrane binding through lipid rafts and CD48 receptor engagement 27 . Although MCs are not professional phagocytic cells, they are known to phagocytose 36 and kill bacteria via acidified vacuoles 37 . These findings would support the role of MCs in uptaking and eliminating Mtb through phagocytosis or serve as a reservoir of intracellular Mtb during the active phase of TB infection. Although is not within the scope of this study, further research is needed to investigate whether MCs contribute to T cell activation as described in other conditions 36,38,39 via Mtb-antigen presentation.
Mtb persistence promotes an ongoing cellular recruitment 40 that results in the formation of early granulomas 6,40 . The association between MCs and TB-induced granulomas is still controversial. A positive correlation between MCs and granuloma formation was observed in tuberculous lymphadenitis tissue 41 but not in tuberculous liver tissue 42 . Our data demonstrate that MCs infiltrate incipient granulomas and locate at the www.nature.com/scientificreports/ periphery or close proximity in mature or necrotic granulomas. Since different immune cells including T lymphocytes are part of the granuloma structure, a direct interaction between MCs and T cells is likely to occur at this site. MC cytokines including IL-17, and TNF-α are necessary for granuloma maturation and maintenance in mouse TB infection 43 . Although MC-TNF-α association was not observed (data not shown), we showed IL-17A positive MCs in the periphery of granulomas. Using an IL-17A gene-knockout mouse model, Okamoto-Yoshida et al. reported that IL-17A is necessary for granuloma maturation with γδ T cells as the major IL-17 source 44 . Also, granuloma formation was impaired in an IL-17A-deficient mouse model of sarcoidosis 45 . Therefore, we would like to propose the concept that MCs contribute to the initial step of cellular recruitment at the infection site, remain outside the inflammatory core during the adaptive immune stage, and orchestrate granuloma maturation via IL-17 expression.
In severe and chronic TB infection, fibrosis is the result of excessive inflammation 11 . MCs are abundant in fibrotic sites in non-infectious lung diseases 46 , including idiopathic pulmonary fibrosis 47 and cystic fibrosis 48 . In addition, MCs products such as the fibroblast growth factor 2 (FGF-2) 49 , prostaglandin E2 (PGE2) 50 , IL-1β 51 , TGF-β 10,52,53 , tryptase 2223,54,55 and chymase 47,56,57 are known to contribute to fibrogenesis 50,53,58 . Our study reproduces findings described in idiopathic pulmonary fibrosis where MCs are increased in numbers and are partially degranulated at fibrotic sites 46,59 . Furthermore, we found a switch in MC phenotype from MC T to MC TC in fibrotic areas. In line with this, Andersson et al., correlated high MC TC numbers with lung function, tissue  www.nature.com/scientificreports/ remodelling and TGF-β1 expression 48 , suggesting MC TC as important fibrosis mediators. In addition, mucosal MCs (MC T ) in coculture with fibroblast lead to differentiation of MC T into connective tissue MCs (MC TC ). This process is coupled with an increase in fibroblast proliferation and an enhanced collagen synthesis necessary for fibrosis generation 60 . Moreover, some antimicrobial drugs are known to activate MCs 61 or induce differentiation 62 however, none of the treatments used for the patients from the reported autopsies in this study seems to have a direct effect on MC activation or differentiation. Thus, phenotypic change from MC T to MC TC with an increase in chymase expression induced by Mtb active infection would initiate or promote fibrosis together with the release of additional fibrogenesis-specific molecules.

Conclusions
In conclusion, our results demonstrate that MCs expressing IL-17 are localizing in TB-induced human lung injuries at inflammatory sites while TGF-β positive and chymase rich MCs are stationed in the proximity of mature granulomas and embedded in fibrotic tissue. Although this is a descriptive study, our data suggest that MCs probably contribute to both, early immune cellular recruitment via IL-17A release and late fibrosis formation with differentiated MC T into MT TC .

Data availability
The datasets generated during and/or analysed during the current study are available from the corresponding author on reasonable request.