Skip to main page content
Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2019 Apr 23;10(1):1823.
doi: 10.1038/s41467-019-09816-4.

Spatial and Temporal Localization of Immune Transcripts Defines Hallmarks and Diversity in the Tuberculosis Granuloma

Affiliations
Free PMC article

Spatial and Temporal Localization of Immune Transcripts Defines Hallmarks and Diversity in the Tuberculosis Granuloma

Berit Carow et al. Nat Commun. .
Free PMC article

Abstract

Granulomas are the pathological hallmark of tuberculosis (TB) and the niche where bacilli can grow and disseminate or the immunological microenvironment in which host cells interact to prevent bacterial dissemination. Here we show 34 immune transcripts align to the morphology of lung sections from Mycobacterium tuberculosis-infected mice at cellular resolution. Colocalizing transcript networks at <10 μm in C57BL/6 mouse granulomas increase complexity with time after infection. B-cell clusters develop late after infection. Transcripts from activated macrophages are enriched at subcellular distances from M. tuberculosis. Encapsulated C3HeB/FeJ granulomas show necrotic centers with transcripts associated with immunosuppression (Foxp3, Il10), whereas those in the granuloma rims associate with activated T cells and macrophages. We see highly diverse networks with common interactors in similar lesions. Different immune landscapes of M. tuberculosis granulomas depending on the time after infection, the histopathological features of the lesion, and the proximity to bacteria are here defined.

Conflict of interest statement

The authors declare no competing interests.

Figures

Fig. 1
Fig. 1
Preferential localization of immune sequences in granulomas during infection with M. tuberculosis. a At 3, 8, and 12 weeks post M. tuberculosis infection (wpi), the fixed lung tissue sections were stained with hematoxylin–eosin (H&E) and signals for Cc10, Cd68, Inos, and Cd3e were plotted on DAPI labeling as background. Pseudo-color density XY positional log2 plots of transcript representations are shown below for Cd68, Inos, and Cd3e transcripts. One representative of three consecutive sections is displayed. Scale: 1000 μm. b The ratio of amplified transcripts in granulomas vs. unaffected regions was calculated for each transcript. The mean log2 ratio of transcript density in the granuloma in relation to the density in unaffected region + SEM in three consecutive sections is depicted. Sections from lungs at 3 and 12 wpi are compared. *Differences in relative transcript density at 3 and 12 wpi are significant (p < 0.05 unpaired Student’s t-test). Source data are provided as a Source Data file
Fig. 2
Fig. 2
Maturation of the granuloma. a The spatial co-expression relationships between the in situ sequencing data were converted into network-based visualization, where each unique transcript is a node in the network and edges represent the interactions using InsituNet. To identify spatially co-expressed transcripts, InsituNet analyzed the co-occurrence of transcript detections within 30 pixels (10 μm). Euclidean distance for each pairwise combination of transcripts (see Supplementary Table 1 as example). Representative examples of one lesion per time point as defined in Supplementary Fig. 1 were selected. The significantly co-expressed sequences that are common in the same lesion from at least two consecutive sections are here depicted. Actb mRNA was excluded from network analysis due to broad expression in different cellular populations. b The tissue section plane was uniformly tiled into 200 pxls (70 μm) radius hexagons and the density of the multiple sequences in each hexagon was aggregated by binning and displayed into a 2D-hexbin map. The densities of the sequences were organized by clustering the hexagons into 3 (3 and 8 wpi) or 4 (12 wpi) different expression patterns. The H&E staining of one representation lung section and their hexbin maps at 3, 8, and 12 wpi are shown. Scale: 1000 μm. c The mean centroid normalized transcript counts in each hexagon was compared for the different clusters. The color-code used for the bars corresponds to that in the 2D-hexbin map. Note that the red clusters corresponding to unaffected areas contained less counts for of all specific sequences. Note also at 12 wpi that although some sequences were dominant in single clusters (i.e., Cd19 mRNA in the yellow cluster), other sequences were more evenly distributed, or predominated in the blue clusters. Source data are provided as a Source Data file
Fig. 3
Fig. 3
Distinct localization of Cd19 mRNA within the lymphoid-rich areas in the granuloma. a In situ detection of Cd19 mRNA transcripts in lungs from M. tuberculosis-infected animals. For one representative of three consecutive sections per time point, the DAPI staining, Cd19 mRNA raw signals, and pseudo-color log2 density plots are shown. Scale: 1000 μm. b Immunohistochemical labeling of CD45R/B220. Note the similar pattern of the labeling as compared with the in situ staining for Cd19 mRNA in a consecutive section at 12 wpi in a. Scale bar: 1000 μm. c The expression of Cd68, Cd3e, Cd19, and Ccr6 mRNA in one representative granuloma from C57BL/6-infected mice at 12 wpi is shown. Cd68 and Cd19 sequences locate in distinct areas of the granuloma, whereas Cd3e mRNA locates in both Cd68 and Cd19 mRNA-rich areas. Scale bar: 1000 μm. d The density of sequences in the epithelioid or lymphoid areas as defined in Supplementary Fig. 1 were quantified in three consecutive sections at 12 wpi. The ratios of sequence densities in lymphoid/epithelioid areas were calculated per section and the mean is depicted. The ratios were color coded accordingly if their relative frequency was higher (red) to Actb mRNA or if lower than the frequencies in epithelioid cells (green). Transcripts with ratios of lymphoid/epithelioid >1 and less than the ratio of Actb mRNA are depicted in blue. Source data are provided as a Source Data file
Fig. 4
Fig. 4
Identification of transcripts co-localizing with M. tuberculosis in tissue sections. a The networks of co-expressed transcripts in one representative lymphoid area of granulomas at 8 and 12 wpi are shown. b The networks of co-expressed transcripts in one representative epithelioid area of granulomas at 8 and 12 wpi are shown. c The auramine–rhodamine-stained M. tuberculosis bacteria were aligned with in situ transcript signals in tissue sections at 8 wpi. M. tuberculosis bacteria were identified by an automated cellprofiler pipeline and a 30 μm radius around those identified is depicted. d The Inos and Cd68 transcript signals at <30 μm from identified M. tuberculosis bacteria are shown together with M. tuberculosis and DAPI staining for the same selected region as in c. Scale bar: 200 μm. e The sequences located within a 3, 10, 30, 100, and 200 μm radius from M. tuberculosis bacteria were identified. The frequency of each sequence within a given distance was determined in relation to the total transcript count for that distance. The fold increase of this frequency with respect to that observed for the total lung section is depicted. Thus, whether a certain transcript is over- or under-represented within the defined distances from M. tuberculosis was determined. f The Cd68 and Inos transcripts located at <30 μm are shown aligned with the H&E staining of one representative section (of three sections). Note that most of the extracted sequences are present in the epitheloid region in relative proximity to the lymphoid areas. Scale bar: 1000 μm. Source data are provided as a Source Data file
Fig. 5
Fig. 5
Heterogeneity of granulomas in the same lung. ad Heat map analysis depicting the transcript density in different granulomas as defined in Supplementary Fig. 1 at 3 (a) and 12 (b) wpi normalized per area. To simplify the description of the data, results were processed by a multivariate principal component analysis (c, d), showing that although some granulomas cluster together with respect to their transcripts titers, others segregate from these as well as from the unaffected area. One representative result of three consecutive sections per time point is shown. e The density of sequences in lymphoid and epithelioid areas of individual granulomas of one tissue section 12 wpi was extracted. The ratio of sequence density of epithelioid and lymphoid areas per granuloma in the three different granulomas in the same lung was quantified and depicted. Source data are provided as a Source Data file
Fig. 6
Fig. 6
Heterogeneity of transcript networks in areas of the granuloma during infection with M. tuberculosis. a, b Diversity of interacting networks of transcripts in the lymphoid (a) and epithelioid (b) area of three granulomas from one representative lung 12 wpi is depicted. Nodes in the network represent unique transcripts and node size is proportional to the number of transcript detections. Edges represent significant spatial co-expression between transcripts. The more statistically significant the co-expression is, the greater the weight (thickness) of the edge in the network. Connections that occur in at least two of the three networks per region appear in gray, unique connections in red. c, d Core networks, defined as co-expressed transcripts at <10 μm in consecutive sections found in all annotated regions (Supplementary Fig. 1) for each time point, calculated as described in the Methods section, are depicted. Diverse core networks were defined for whole granulomas, for epithelioid and lymphoid regions of the granulomas at different time points after M. tuberculosis infection in C57BL/6 mice (c). The core networks from non-organized and organized granulomas C3HeB/FeJ mice are also shown (d). No core networks for unaffected regions could be found
Fig. 7
Fig. 7
In situ sequencing in encapsulated granulomas. a, b. H&E (a) and Auramine–Rhodamine stain of M. tuberculosis bacteria together with DAPI staining (b) of a C3HeB/FeJ lung section 10 wpi. Scale bar: 1000 μm. c In situ signals of Inos, Cd68, Tcrb, Foxp3, Il10, and Cd19 mRNA transcripts in a lung section from M. tuberculosis-infected C3HeB/FeJ, aligned with the DAPI staining, are depicted. d, e Heat map analysis (d) depicting the sequence density in the annotated areas of region 1 of the C3HeB/FeJ lung (Supplementary Fig. 1D). A multivariate principal component analysis of signals shows proximity between those in annotated areas from all regions sharing histopathological features, but the distances between those of different kinds (e). The predictive ellipses displayed have a 95% probability that a new observation from the same group will fall inside the ellipse. The areas correspond to the center (GC) and edge (GO) of encapsulated granulomas, non-encapsulated granulomas with relatively low mycobacterial density (NOG), small-encapsulated granulomas with high bacterial numbers (SG), and an unaffected area (Unaff). f The density of sequences in the different areas of the C3HeB/FeJ lung were extracted. The log2 ratio of sequence density ± SEM in the center (GC) and edge (GO) of the encapsulated granuloma and in the center vs. the non-organized granuloma (NOG) are depicted. The correlation between the GC/GO and GC/NOG is significant (Pearson’s p < 0.001; r: 0.67). g The log2 ratio of sequence density ± SEM in the non-organized (NOG) vs. the surrounding area from the encapsulated granuloma (GO) or the small necrotic granuloma (SG) are depicted. The correlation between the NOG/GO and NOG/SG is significant (Pearson’s p < 0.002; r: 0.65). h The sequences in lungs from C3HeB/FeJ in 70 μm radius hexabins were clustered into three different expression patterns. These overlap with regions with encapsulated and non-encapsulated granulomas or with unaffected lung areas (Supplementary Fig. 1D). i The mean centroid normalized sequence counts in each hexagon was compared for the clusters. The ratio of sequence densities in red and green clusters is depicted. Source data are provided as a Source Data file

Similar articles

See all similar articles

Cited by 4 articles

References

    1. WHO. Global Tuberculosis Report, http://www.who.int/mediacentre/factsheets/fs104/en/ (2016).
    1. Cohen SB, et al. Alveolar macrophages provide an early Mycobacterium tuberculosis niche and initiate dissemination. Cell Host Microbe. 2018;24:439–446 e434. doi: 10.1016/j.chom.2018.08.001. - DOI - PMC - PubMed
    1. Ehlers S, Schaible UE. The granuloma in tuberculosis: dynamics of a host-pathogen collusion. Front. Immunol. 2012;3:411. - PMC - PubMed
    1. Cadena AM, Fortune SM, Flynn JL. Heterogeneity in tuberculosis. Nat. Rev. Immunol. 2017;17:691–702. doi: 10.1038/nri.2017.69. - DOI - PMC - PubMed
    1. Canetti G. Dynamic aspects of the pathology and bacteriology of tuberculous lesions. Am. Rev. Tube. 1956;74:13–21. - PubMed

Publication types

MeSH terms

Feedback