In vivo analysis of Trypanosoma cruzi persistence foci at single cell resolution

Infections with Trypanosoma cruzi are usually life-long despite generating a strong adaptive immune response. Identifying the sites of parasite persistence is therefore crucial to understand how T. cruzi avoids immune-mediated destruction. However, this is a major technical challenge because the parasite burden during chronic infections is extremely low. Here, we describe an integrated approach involving comprehensive tissue processing, ex vivo imaging, and confocal microscopy, which has allowed us to visualise infected host cells in murine tissue, with exquisite sensitivity. Using bioluminescence-guided tissue sampling, with a detection level of <20 parasites, we show that in the colon, smooth muscle myocytes in the circular muscle layer are the most common infected host cell type. Typically, during chronic infections, the entire colon of a mouse contains only a few hundred parasites, often concentrated in a small number of cells containing >200 parasites, that we term mega-nests. In contrast, during the acute stage, when the total parasite burden is considerably higher and many cells are infected, nests containing >50 parasites are rarely found. In C3H/HeN mice, but not BALB/c, we identified skeletal muscle as a major site of persistence during the chronic stage, with most parasites found in large mega-nests within the muscle fibres. Finally, we report that parasites are also frequently found in the skin during chronic murine infections, often in multiple infection foci. In addition to being a site of parasite persistence, this anatomical reservoir could play an important role in insect-mediated transmission, and have implications for drug development. IMPORTANCE Trypanosoma cruzi causes Chagas disease, the most important parasitic infection in Latin America. Major pathologies include severe damage to the heart and digestive tract, although symptoms do not usually appear until decades after infection. Research has been hampered by the complex nature of the disease and technical difficulties in locating the extremely low number of parasites. Here, using highly sensitive imaging technology, we reveal the sites of parasite persistence in experimental mice at single-cell resolution. We show that parasites are frequently located in smooth muscle cells in the circular muscle layer of the colon, and that skeletal muscle cells and the skin can also be important reservoirs. This information provides a framework for investigating how the parasite is able to survive as a life-long infection, despite a vigorous immune response. It also informs drug-development strategies by identifying tissue sites that must be accessed to achieve a curative outcome.

INTRODUCTION The intracellular protozoan parasite Trypanosoma cruzi is the 75 etiological agent of Chagas disease, and can infect a wide variety of mammalian hosts. 76 Transmission to humans is mainly via the hematophagous triatomine insect vector, 77 which deposits infected faeces on the skin after a blood-meal, with the parasite then  In humans, infection normally results in mild symptoms, which can include fever and 85 muscle pain, although in children the outcome can be more serious. Within 6 weeks, 86 this acute phase is usually resolved by a vigorous CD8+ T cell response (4, 5), and in 87 most cases, the infection progresses to a life-long asymptomatic chronic stage, where 88 the parasite burden is extremely low and no apparent pathology is observed. However, 89 in ~30% of individuals, the infection manifests as a symptomatic chronic condition, 90 although this can take many years to develop. The associated cardiac dysfunction, 91 including dilated cardiomyopathy and heart failure, is a major cause of morbidity and 92 mortality (6, 7). In addition, ~10% of those infected display digestive pathologies, such tract, particularly the colon and/or stomach, was found to be a major site of parasite 125 persistence during chronic stage infections, but it has not so far been possible to 126 identify the infected host cell types in these complex tissues. The immune-mediated 127 restriction to the GI tract was not absolute, with both host and parasite genetics  Here, we describe how these enhanced imaging procedures, coupled with 152 modifications to tissue processing, have allowed us to identify the sites of parasite 153 persistence during chronic murine infections. We reveal that the circular muscle layer 154 is the major reservoir of infection in the colon, that skeletal muscle can be an important 155 site of persistence, although this phenomenon appears to be strain-specific, and that 156 the skin can harbour multiple infection foci. parasites could be determined with precision using full-thickness serial Z-stacking ( Fig.   199 1e, Fig. S1). This allowed us to establish that the total number of parasites persisting 200 in the external colonic wall (tunica muscularis) of a chronically infected mouse is 201 typically in the range of a few hundred (697 + 217, n=16), although this number can 202 be higher if the tissue contains one or more "mega-nests" (Fig. 1c, highlighted in 203 yellow, as example).

205
When we compared parasite distribution in the external gut wall during acute and 206 chronic murine infections, the most striking difference was the presence in the chronic 207 stage of some host cells that were infected with >200 parasites (Fig. 2). The existence 208 of these mega-nests resulted in a significant alteration in parasite number distribution 209 at the level of single infected host cells (Fig. 1, Fig. 2b-d). In acute infections, parasites 210 were spread between many more host cells, with the average parasite content per cell  The number per infected cell was determined by Z-stacking, which could be done with   with >200 parasites, were refractive to staining with any of the three markers (Fig. 4b). fur from the carcass allowed the whole of the skeletal muscle system to be exposed 260 and imaged ( Fig. 5b and d). The skin could also be placed fur side down and imaged 261 in its entirety after the removal of adipose tissue. All adipose tissue harvested during 262 the dissection process was combined to be imaged separately. consistently close to background levels, whereas with the C3H/HeN mice, more than 274 half displayed a detectable signal (>2SDs above background radiance) (Fig. 5c).

275
Following bioluminescence-guided excision (Fig. 5d) (Fig. 6b). For all four parasite:mouse strain combinations, we observed 294 a wide range of skin parasitism patterns, as judged by both the number and the 295 intensity of the bioluminescent foci ( Fig. 6a and b). There was some evidence that 296 C3H/HeN mice had more CL Brener skin parasites than BALB/c mice ( Fig. 6b and c).

297
Infections with the CL Brener strain produced more discrete foci and a higher inferred    Here, we exploited parasites that express fusion proteins containing bioluminescent 342 and fluorescent domains. Together with improved tissue preparation techniques, this 343 has enabled us to achieve a limit of detection by ex vivo imaging that is less than 20 344 parasites ( Fig. 2d and e). By facilitating the routine detection of parasites in their tissue  (Fig. 3) and that smooth muscle myocytes are the main infected host cell 353 type. Enteric neurons can also be parasitized, but these infections are much less 354 common (Fig. 4). The extent to which this apparent tropism is determined by a  C3H/HeN mice than in the BALB/c strain (Fig. 5c), a phenomenon which is associated 370 with increased cardiac pathology (22). In the human population, this highlights that 371 genetic diversity affecting the functioning of the immune system and its ability to 372 restrict the tissue range of T. cruzi to reservoir sites could be a major determinant of 373 Chagas disease pathology. Within C3H/HeN mice, skeletal muscle was also found to 374 be an important site of persistence during the chronic stage, whereas in the BALB/c 375 strain, parasites were far less evident in this location (Fig. 5c). Some T. cruzi strains 376 have been reported to be myotropic in mice, with pathological outcomes that include       were drawn to quantify bioluminescence expressed as radiance (photons/s/cm 2 /sr).

497
Because different tissue types from uninfected control mice have different background 498 radiances, we normalized the data from infected mice using matching tissues from 499 uninfected controls (n=4) and used the fold-change in radiance, compared with these 500 tissue-specific controls, as the final measure of ex vivo bioluminescence. Detection 501 thresholds for ex vivo imaging were determined using the fold-change in radiance for 502 ROIs in images obtained from infected mice compared with matching empty ROIs in 503 images from uninfected control mice of comparable age. parasites, samples were imaged in 3-dimensions, with the appropriate scan zoom 524 setting, and the files exported for analysis using image J software (see Fig. S1).