Morphological variability in the mucosal attachment site of Trichuris muris revealed by X-ray microcomputed tomography

Parasitic infections can be challenging to study, because light and electron microscopy are often limited in visualising complex and inaccessible attachment sites. Exemplifying this, Trichuris spp. inhabits a tunnel of epithelial cells within the host caecum and colon. A significant global burden of this infection persists partly because available anthelminthics lack efficacy, although the mechanisms underlying this remain unknown. Consequently, there is a need to pioneer new approaches to better characterize the parasite niche within the host and investigate how variation in its morphology and integrity may contribute to resistance to therapeutic intervention. To address these aims, we exploited 3D X-ray micro-computed tomography (microCT) to image the mouse whipworm T. muris in caeca of wild-type C57BL/6 and SCID mice ex vivo. Using osmium tetroxide staining to effectively enhance contrast of worms, we found that a subset exhibited preferential positioning towards the bases of the intestinal crypts. Moreover, in one rare event, we demonstrate whipworm traversal of the lamina propria. This morphological variability contradicts widely accepted conclusions from conventional microscopy of the parasite niche, showing Trichuris in close contact with the host proliferative and immune compartments that may facilitate immunomodulation. Furthermore, by using a skeletonization-based approach we demonstrate considerable variation in tunnel length and integrity which may represent an indicator of tunnel “health”. The qualitative and quantitative observations provide a new morphological point of reference for future in vitro study of Trichuris-host interactions and highlight the potential of microCT to more accurately characterise enigmatic host-parasite interactions. Author Summary Parasites are often difficult to observe once established within host tissues, presenting a barrier to biological understanding and therapeutic innovation. Whipworms (Trichuris spp.) affect 500 million people worldwide, causing significant disability, and appear partially resistant to widely used “deworming” drugs. However, the inaccessibility of worms within the cells of the host intestine makes them highly challenging to image and study. By investigating Trichuris attachment sites in 3D, using X-ray micro-computed tomography, we found that the niche is highly variable in size and, contrary to reports in all previous studies, can also penetrate different layers of intestinal tissue. By showing that worms are positioned much closer to host immune cells that previously appreciated, we provide a morphological reference point for future studies on how Trichuris effectively avoids clearance by the host. The non-invasive imaging approach used represents an excellent opportunity to clarify the lifecycles of other difficult-to-study parasites.


4
Existing anthelminthic drugs lack efficacy in controlling Trichuris infection [7][8][9]. Therefore, 72 continuing development of novel anthelminthics remains an important contemporary objective in 73 Trichuris research [10]. Trichuris muris, which infects mice, is commonly used as a model to 74 investigate trichurid host-parasite interactions and therapeutic strategies due to its well-75 characterized development and the genetic tractability of its host [11,12]. Preclinical drug and 76 vaccine trials have consistently shown partial reduction of T. muris worm burdens [13][14][15], 77 suggesting that worms within the same host are differentially susceptible to their delivery, toxicity 78 or immunogenic actions. Why some worms are more susceptible to therapeutic intervention than 79 others remains unclear, but the phenomenon may be related to the success of niche 80 maintenance, including the integrity of the host epithelium overlying the worm. Improving our 81 understanding of the Trichuris survival within the host, and examining how the niche varies 82 between worms during chronic infection is therefore an important objective for therapeutically 83 addressing worm resistance [11,12].

84
Most of the knowledge of the trichurid tunnel structure and composition so far has been obtained 85 by electron and light microscopy. Adult worms appear to occupy a tunnel of 'dead' epithelial cells, 86 with bacteria and damaged, fragmented organelles present in the more posterior tunnel regions 87 [2,5]. The survival of Trichuris within this niche in thought to be promoted by modulation of the 88 host immune response. Specifically, secreted proteins [16] promote a pro-inflammatory milieu 89 whilst suppressing elevated epithelial turnover, a key immune effector mechanism which ejects 90 Trichuris from its niche [17,18]. Despite an increasingly complete picture of the immunologic 91 bases for host susceptibility and resistance [12], the nature of tunnel formation and development 92 remains mysterious. During the earliest hours of invasion, larvae actively interact with the live 93 cells at the base of the crypts of Lieberkühn to form the multi-intracellular niche [19,20]. After the 94 sequential moults leading to adulthood, the tunnel becomes much longer and more tortuous,

95
presenting an obstacle to effective imaging [3]. However, microscopic studies of niche 96 morphology to date have all concurred that adult Trichuris live exclusively within the epithelial 97 cells at the surface of the mucosa, directly adjacent to the lumen [2,3,21,22]. X-ray micro-98 5 computed tomography (microCT), which is inherently 3D and is capable of providing images of a 99 comparable resolution to histological approaches [23,24], provides unique opportunities to 100 interrogate the structure and morphology of the tunnel in detail [25]. Capturing a volume, rather 101 than sampling a random 2D plane, offers the opportunity to accurately characterize the 102 morphology of multiple attachment sites in their entirety, as well as the degree of variation 103 between them, potentially revealing new information about the Trichuris life cycle.

104
By exploiting microCT imaging at "histological" resolutions, we visualized tunnel morphology in 105 the developing and adult worms. The aims of the study were to 1) describe the 3D morphology of 106 the attachment site throughout worm development; and 2) to quantify the degree of morphological 107 heterogeneity across multiple Trichuris attachment sites during chronic infection. By exploiting the 108 3D nature of the data, we morphometrically explored multiple tunnels using a skeletonization-109 based method order to estimate tunnel length, which we suggest may constitute a basis for future 110 in situ screening tools of anthelminthic drugs. We also describe the presence and frequencies 111 distinct sets of tunnel morphologies. To achieve these goals, we developed a sample preparation 112 protocol adopted from scanning electron microscopy which facilitates quantitative analyses, is

118
In order to investigate qualitatively the morphology of the epithelial tunnel throughout the 119 development of chronic infection in the caecum, we examined the caeca of mice at different 120 stages of infection, roughly corresponding with the T. muris moulting pattern. Male C57BL/6 mice 121 were infected orally with a low dose of T. muris eggs in order to ensure a chronic infection [26]. At 122 7, 14, 21 and 35 days post-infection (PI) with T. muris, 3-5 mm tubular fragments of caecum were 6 dissected and processed for microCT imaging ( Fig. 1 A). In caecum fragments taken 35 days PI,

124
we identified an optimal staining protocol utilizing Osmium Tetroxide (OsO4), which allowed 125 superior contrast between the worm and surrounding tissue when compared to worms immersed 126 in aqueous potassium triiodide (Lugol's Iodine) and PTA (Fig. 2 A-C In caeca from C57BL/6 mice taken 35 days PI, microCT scanning of whole fragments (pixel size 135 = 3.62 µm, Fig. 1 B) revealed the attachment sites of multiple worms. When the magnification 136 was increased in order to visualize regions of interest in more detail (pixel sizes 1.56 µm and 0.59 137 µm), multiple aspects of worm and host morphology were apparent. The crypts of the host 138 mucosa were visible as layers of reduced X-ray attenuation ( Fig. 1 C). The bacillary band, 139 stichocytes and oesophagus of the worm and their orientations were all distinct, exhibiting higher 140 X-ray attenuation than the surrounding host tissue (Fig. 1 C, D). Cuticular inflations, another

219
The worm here appears to be crossing the lamina propria (black arrowhead). B) Haematoxylin In addition to C57BL/6 mice, we were interested in observing the positioning of worms in Severe

225
Combined Immunodeficient (SCID) mice. These mice lack the adaptive components of the 226 immune response, and are susceptible to T. muris at both high and low doses. In this study, SCID 227 mice were infected with a low dose (<40 eggs) of T. muris in order to replicate the infection 228 burden of the C57BL/6 mice examined in Figs 1 + 2. When microCT of SCID caeca was carried 229 out using Osmium tetroxide staining, further head-down morphology was observed, similarly to 230 C57BL/6 mice. However, in one SCID mouse we observed one worm which appeared to travel 231 laterally between crypts, half-way between the lumenal surface and the muscularis mucosae

250
In order to examine the occurrence of the head-down behavior across mouse strains and sexes,

251
we undertook a series of scans using the staining methodology described above. Frequencies of 252 head down behavior in C57BL/6 (male, female) and SCID (female only) mice are recorded in 253 Table 2. Statistical analysis was not attempted due to the low frequencies of worms sampled in 254 each group. However, in the 13 mouse donors examined, we note that qualitatively there was no 255 trend evident in the occurrence of the head down behavior between sexes or strains.   In order to understand whether heterogeneity in the niche could also be extended to include 280 variations in the integrity and lengths of the tunnels themselves, we utilized skeletonization to 281 quantitatively assess the number and size of previously described "breaks" in the epithelium 282 covering the tunnel [3], as well as the length of the embedded portion of the stichosome. In the 283 caecum images from the C57BL/6 mouse at day 35 P.I. examined in Fig 1 and Fig 3, worms   284 were assigned a unique ID and subsequently the proportions of the stichosome which were 285 covered and uncovered by the epithelial layer measured. Worms were segmented, and 286 skeletonization facilitated length measurements of the tortuous stichosome ( Fig. 5 A-C). The

288
Therefore, the length of worm covered by the epithelial tunnel could be quantified, and was 289 expressed as a percentage as "epithelial coverage". Epithelial coverage was measured in seven 290 worms in the caecum of the male C57BL6 mouse (pixel size = 3.619 µm), and was highly 291 variable; of the seven worms analyzed only one had its stichosome totally covered by epithelium,

294
The cumulative tunneled length of the stichosome calculated for seven worms (Table 3), was 295 also highly variable, ranging from 2.96 mm to 8.95 mm. Furthermore, the number of breaks varied 296 between worms (range: 0 to 5). Importantly, the tunneled length of the stichosome was not

359
The "head-down" and "head curled" behaviors have not been previously described in Trichuris

381
In this work, we demonstrated the utility of microCT in identifying rare events by screening

395
In addition to the morphology described here, we also developed a method by which the tunneled

405
These are also prospective measures of susceptibility to anthelminthic, since they represent the 406 extent to which the worm's metabolically active stichosome is directly exposed to anthelminthics

485
For measurements of epithelial coverage, the total tunneled length of the stichosome, including 486 epithelial breaks was estimated by use of AVIZO's "centerline tree" algorithm on embedded 487 portions of stichosome. Centerline tree extracted a non-branching spatial graph from the 488 segmented worm which ran directly through the centre of the worm along the anterior-posterior 489 axis. The proportion of the stichosome which was exposed by breaks in the epithelium overlaying 490 the tunnel was calculated by manually separating the spatial graph of the embedded stichosome 491 into either "exposed" or "unexposed" lengths which were subsequently measured. The cumulative 492 length of the exposed stichosome was calculated and subtracted from the total tunneled length to 493 obtain the proportion of the tunneled region of the stichosome which was exposed by epithelial 494 breaks.

502
BrDU staining was carried out by the ABC method using the mouse on mouse kit 503