Organotypic endothelial adhesion molecules are key for Trypanosoma brucei tropism and virulence

Article Organotypic endothelial adhesion molecules are key for Trypanosoma brucei tropism and virulence Graphical abstract Highlights dOur study investigates the blood vasculature for T. brucei reservoir establishment dWe show the pancreas is a large extravascular reservoir dWe establish that T. brucei tropism is linked to organotypic adhesion molecules dInterfering with adhesion molecules impacts parasite virulence and host survival Authors Mariana De Niz, Daniela Bra ´s, Marie Ouarne ´, ..., Lenka Henao Misikova, Claudio A. Franco, Luisa M. Figueiredo Correspondence [email protected] In brief Trypanosoma brucei are parasites that cause severe disease in mammals. De Niz et al. investigate how T. brucei interacts with blood vessels to preferentially traverse into certain organs, where they establish vast reservoirs. Selective removal of organ-specific vascular receptors ultimately alters parasite virulence and host survival. De Niz et al., 2021, Cell Reports 36, 109741 September 21, 2021 ª2021 The Author(s). https://doi.org/10.1016/j.celrep.2021.109741 ll Article Organotypic endothelial adhesion molecules are key for Trypanosoma brucei tropism and virulence Mariana De Niz, 1,4 Daniela Bra ´s, 1,5 Marie Ouarne ´, 1 Mafalda Pedro, 1,2 Ana M. Nascimento, 1,3 Lenka Henao Misikova, 1 Claudio A. Franco, 1,6 and Luisa M. Figueiredo 1,6,7, * 1 Instituto de Medicina Molecular Joao Lobo Antunes, Faculdade de Medicina, Universidade de Lisboa, Lisboa 1649-028, Portugal 2 Departamento de Ci^ encias da Vida, Faculdade de Ci^ encias e Tecnologia, Universidade Nova de Lisboa, Caparica 2825-149, Portugal 3 Bioimaging Unit, Instituto de Medicina Molecular Joao Lobo Antunes, Faculdade de Medicina, Universidade de Lisboa, Lisboa 1649-028, Portugal 4 Present address: Institut Pasteur, Paris 75015, France 5 Present address: Instituto Gulbenkian de Ci^ encia, Oeiras 2780-156, Portugal 6 Senior author 7 Lead contact *Correspondence: [email protected] https://doi.org/10.1016/j.celrep.2021.109741 SUMMARY Trypanosoma brucei is responsible for lethal diseases in humans and cattle in Sub-Saharan Africa. These extracellular parasites extravasate from the blood circulation into several tissues. The importance of the vasculature in tissue tropism is poorly understood. Using intravital imaging and bioluminescence, we observe that gonadal white adipose tissue and pancreas are the two main parasite reservoirs. We show that reservoir establishment happens before vascular permeability is compromised, suggesting that extravasation is an active mechanism. Blocking endothelial surface adhesion molecules (E-selectin, P-selectins, or ICAM2) significantly reduces extravascular parasite density in all organs and delays host lethality. Remarkably, blocking CD36 has a specific effect on adipose tissue tropism that is sufficient to delay lethality, suggesting that establishment of the adipose tissue reservoir is necessary for parasite virulence. This work demonstrates the importance of the vasculature in a T. brucei infection and identifies organ-specific adhesion molecules as key players for tissue tropism. INTRODUCTION Tissue-specific tropism within vertebrate hosts has been the focus of great interest in the field of parasitology in recent years. However, the cellular and molecular adaptations that allow parasite tropism are still poorly understood. For many parasites, tropism to specific organs is an essential step of their life cycle, because the organs provide a niche for persistence, latency or dormancy, massive replication and/or growth, protection from the host immune responses, or differentiation into alternative stages essential for completion of the life cycle, among others (Boyett and Hsieh, 2014;Fernandes and Andrews, 2012;Gue ´rin and Striepen, 2020;Lima and Lodoen, 2019;Onyilagha and Uzonna, 2019;Prud^ encio et al., 2006;Re ´nia and Goh, 2016;Silva Pereira et al., 2019;Venugopal et al., 2020). Trypanosoma brucei is a parasitic organism transmitted by tsetse flies (Glossina spp.), responsible for human African trypanosomiasis (HAT) in humans and nagana in other mammals. It requires two hosts to live and reproduce, namely, the insect vector and the mammalian host (Centers for Disease Control and Prevention, 2019). T. brucei invades the bloodstream and lymph and disseminates across the host body (Kr€ uger and Engstler, 2018). The adipose tissue, skin, and brain have been identified as important extravascular reservoirs of T. brucei because of the numbers of parasites harbored, the role they play in parasite transmission, or the associated pathology, respectively (Casas-Sa ´nchez and Acosta-Serrano, 2016;De Niz et al., 2019a–c;Silva Pereira et al., 2019). Yet, the relative contribution of parasite tissue reservoirs to overall parasitemia and disease outcome is unclear. The relevance ofthe vascular endothelium for tropism and parasite dissemination has been studied in detail in the context of other parasites. For instance, Plasmodium-infected red blood cells (RBCs) undergo sequestration in the vasculature of several organs. Multiple endothelial cell (EC) receptors, including intercellular adhesion molecule 1 (ICAM1), endothelial protein C receptor (EPCR), platelet/endothelial cell adhesion molecule 1 (PECAM1), and CD36, have been associated with sequestration in specific tissues in both human and rodent malaria infections (Fonager et al., 2012;Hviid and Jensen, 2015;De Niz et al., 2016;Smith et al., 2001). While Plasmodium asexual blood stages are unable to cross the vascular endothelium, T. brucei extravasation has been studied in the brain. Two key findings include that a trypanosome-derived cathepsin L-like cysteine protease (brucipain) is required for traversal of the brain-blood barrier, and that vascular permeability does not correlate with parasite dissemination in the brain (reviewed in Kristensson et al., 2010). Cell Reports 36, 109741, September 21, 2021 ª2021 The Author(s). 1 This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/). ll OPEN ACCESS Here, we used intravital imaging to study the role of blood vessels in the establishment of T. brucei tissue reservoirs in vivo.We have characterized the dynamics of host-parasite interactions with the vascular endothelium and its relationship to parasite tropism. We identified the pancreas as another reservoir for T. brucei, which, together with the gonadal white adipose tissue (g-WAT), constitutes the organs with highest parasite density. We showed that g-WAT and pancreatic tropism depend on several adhesion molecules expressed at the surface of ECs. We found that extravasation into the adipose tissue is specifically affected by the adhesion molecule CD36, and we demonstrated that the establishment of this reservoir is a key virulence mechanism with great impact in disease outcome. Our analysis shows that the vasculature plays an essential role in T. brucei infection, and that the establishment of tissue reservoirs increases the fitness of parasites and disease severity. RESULTS g-WAT tissue and pancreas are largest reservoirs of extravascular parasites Little is known about the dynamics of T. brucei reservoir establishment in vivo. To evaluate parasite distribution in mice with high temporal resolution, we used a transgenic parasite line expressing the red-shifted firefly luciferase PpyREH9 and TdTomato fused with a TY1 tag (Calvo-Alvarez et al., 2018). We injected 2 310 3 parasites intraperitoneally in each mouse, and we followed the parasitemia in vivo in C57BL/6 and C57BL/6 Albino mice, the latter being a more suitable model to measure bioluminescence (Curtis et al., 2011;Doyle et al., 2004)(Figure S1A, left panel). In both mouse strains, we observed high blood parasitemia at days 5–8, and then later in infection, as previously reported (Calvo-Alvarez et al., 2018). Both strains of mice showed similar survival times, with 50% survival being reached by days 22–23 post-infection (Figure S1A, middle panel). Therefore, we limited the infection to day 20 for all subsequent experiments. The relative whole-body bioluminescence measured in Albino mice (Figure 1A) correlated very well with blood parasitemia (R 2 = 0.82, p < 0.001) (Figure S1A, right panel). To determine relative parasite density in individual organs, we dissected each organ following injection of luciferin and performed ex vivo bioluminescence imaging (Figures 1B and 1C; Figure S1B). Three white and two brown adipose tissues (WATs and BATs, respectively) were collected (schematic shown in Figure S1C). Animals were not perfused prior to organ excision; therefore, the signal measured includes both intravascular and extravascular parasites (Figure 1C; Figure S1B). g-WAT, pancreas, and lungs (marked with arrows, Figure 1C) were the organs that presented the highest bioluminescence values throughout infection (Figure 1C; Figure S1B). For quantitative comparison, we show pooled bioluminescence values of all organs during the first peak of parasitemia (days 5–8 post-infection) (Figure 1D). The lungs, pancreas, and g-WAT showed around 30-fold higher bioluminescence than the average of other organs. Without perfusion, among adipose tissues, bioluminescence signal in the g-WAT was around 60to 250-fold higher than other WAT depots (p < 0.001). Global comparisons between WAT and BAT depots showed that bioluminescence in BAT depots was significantly lower than WAT depots (p = 0.04) (Figure S1C). The pancreas was the organ with the highest accumulated daily intensity, being on average 2.8-fold higher than g-WAT and 22-fold higher than lungs considering the 20 days of infection. In the skin, we identified high heterogeneity in parasite distribution (Figure S1D), whereby some regions are highly enriched in parasites and others are not. This heterogeneity is consistent with previous observations (Capewell et al., 2016). In the bloodstream, parasites are at very high concentrations (Figure S1A). To discriminate the bioluminescent signal between intravascular and extravascular compartments, we perfused mice prior to measuring bioluminescence to remove signal from intravascular parasites on day 6 post-infection (Figure 1E). In perfused mice, only g-WAT and pancreas showed significantly high bioluminescence. Upon perfusion, g-WAT was 200to 4,800-fold higher than the four other WAT or BAT depots (p < 0.001 for all comparisons with g-WAT). At day 6 post-infection, the pancreas was equally enriched as the g-WAT (p = 0.25) and 1,000-fold higher than lungs (p < 0.001). Notably, the signal intensity in lungs decreased 200-fold after perfusion compared with non-perfused mice (p < 0.001) (Figures 1D and 1E), suggesting that most parasites in the lungs remained intravascular. These results were independent of the route of infection used (Figure S1E). To confirm the bioluminescence results, we determined the intravascular and extravascular location of parasites in each organ at a single-cell level, using intravital microscopy (IVM) and ex vivo confocal microscopy (Figures 1F–1J; Video S1). Intravascular parasite density (parasites/10 mm 2 ) measured by intravital microscopy showed a similar pattern of parasite density in all organs, with two parasitemia ‘‘waves’’ in peripheral blood (Figure 1G), consistent with measurements performed by hemocytometer (Figure 1C; Figure S1A). Quantification of intravascular and extravascular parasites during the first parasitemia wave (days 5–8) (Figures 1I and IJ) confirmed a significant enrichment in the g-WAT and pancreas. Throughout infection, in total, the number of parasites per 10 mm 2 found in the extravascular space of g-WAT and pancreas corresponds to around 40% of the sum of extravascular parasites counted per 10 mm 2 in the 13 organs, while the three WAT depots together contribute to 48%. The intravital imaging analysis identified four different patterns of intravascular and extravascular T. brucei distribution across organs (Figures 1G–1J; Figure S2). In group 1 (Figure S2A), there is proportionally more parasites in the extravascular space (blue) than intravascularly (red) throughout most of the infection. Group 2(Figure S2B) has a pattern similar to group 1, in which the intravascular and extravascular compartments present two waves of parasite populations with a minimum around day 11, with the difference being that group 2 has consistently more intravascular than extravascular parasites. Given the consistency of patterns between the BAT and WAT groups, henceforth we use only one representative of each tissue category (gonadal for WAT and interscapular for BAT [i.e., isc-BAT]). In a third group (Figure S2C), the vast majority of parasites reside inside vessels, with a small population colonizing the extravascular space mostly after the first peak of blood parasitemia. Finally, the last group of organs (Figure S2D) showed an early 2Cell Reports 36, 109741, September 21, 2021 Article ll OPEN ACCESS ABC ED FG H IJ Figure 1. Identification of WAT and pancreas as the two main T. brucei extravascular reservoirs (A) Representative images of whole-body bioluminescence in ventral (right panel) and dorsal (left panel) positions of a mouse infected with luciferase-expressing T. brucei at day 6 post-infection. (B) Representative images of non-perfused organs collected from infected mice (day 6 post-infection). (C) Mean bioluminescence (extravascular and intravascular parasites) values of individual mouse organs expressed as log 10 light units per tissue area. (D and E) Mean parasite density of individual organs during the first wave of parasitemia (days [d] 5–8) in non-perfused mice (D) and at day 6 post-infection in perfused mice (E) measured by bioluminescence (log 10 light units per tissue area). All bioluminescence quantifications (A–D) correspond to values obtained from a minimum of nine mice. (F) Immunofluorescence representative images obtained by intravital live imaging of parasites (TdTomato reporter, red) in g-WAT, pancreas, brain, and spleen at d3, d6 and d10 post-infection. Vessels are labeled with anti-CD31 antibody (yellow), interstitial space with FITC-Dextran, and nuclei with Hoechst (blue). Dyes and antibodies were administered by retroorbital intravenous injection 10–30 min prior to imaging. Scale bars, 50 mm. (G and H) Heatmap of parasite density (number of parasites per tissue area) inside (G) or outside vessels (H). Dotted white lines denote the beginning of extravascular colonization (S [start], day 2), first peak of infection (P1, day 6), and the end of the first wave of parasitemia (E1 [end], day 10). For each day of infection, at least 100 fields of view were quantified, and the mean value is shown in the heatmap. (I and J) Mean parasite density (number of parasites per tissue area) of individual organs inside (I) or outside (J) vessels during the first wave of parasitemia (d5–8), measured by intravital imaging. For all relevant figures, dotted red lines show the interquartile range considering all organs. Organs above the upper threshold are considered significantly enriched, using an ANOVA statistical test. ***p < 0.001, **p < 0.01, *p < 0.05. All relevant data used to generate this figure are included in Data S1 (tabs 1–4). p/s, photons per second. See also Figures S1 and S2 and Videos S1 and S2. Cell Reports 36, 109741, September 21, 2021 3 Article ll OPEN ACCESS extravascularization of parasites prior to the peak in the peripheral blood but little enrichment later in infection. Altogether, these results indicate that parasite distribution in intravascular and extravascular spaces is very heterogenous across organs and varies with time. Importantly, in terms of establishment of an early tissue reservoir, we conclude that g-WAT and pancreas parasite reservoirs are established since day 3 post-infection, and that both remain the most parasitized organs throughout infection. T. brucei enrichment in the pancreas and WATs is not correlated to vascular density Next, we investigated the reasons for g-WAT and pancreatic tropism. First, we explored if there were differences in vascular density between organs. Vascular density indicates the percentage of the imaged area filled with blood vessels (marked by CD31/PECAM1, an EC receptor, or fluorescein isothiocyanate [FITC]-Dextran, a large polysaccharide that stays inside vessels unless there is increased vascular permeability) (Figure 2A). In non-infected animals, we found that organs such as brain, lungs, liver, spleen, and lymph nodes are highly vascularized (30%–60% of their area consists of vessels), while adipose tissue depots, pancreas, heart, and kidneys are less vascularized (12%–30%) (Figures 2B and 2C; Figure S3), consistent with the literature (Cook, 1965). Figure 2C depicts vascular density during the first parasitemia wave (days 5–8, black boxes) and the last parasitemia wave (days 15–18, red boxes). During infection, vascular density remained unchanged in most organs, including A BC Figure 2. Changes in organ vascular density during T. brucei infection (A) Vasculature is labeled with an antibody antiCD31 (yellow), while nuclei are labeled by Hoechst (blue) upon retroorbital injection. Vascular density corresponds to the percentage of the imaged area filled with blood vessels based on the marker CD31 (middle and right panels). Scale bar, 50 mm. (B) Heatmap of mean vascular density throughout 20 days of infection. At least 100 fields of view per day. (C) Boxplot of mean vascular density for d5–8 (black) and d15–18 (red) depicted in (B). For all graphs, significance is shown as ***p < 0.001, **p < 0.01, or *p < 0.05. All relevant data used to generate this figure are included in Data S1 (tabs 6 and 7). See also Figure S3. pancreas. In contrast, vascular density in g-WAT significantly increased from 13% at day 1 to 49% at day 20 post-infection (Figure 2B; Figure S3), and it significantly decreased in spleen and lymph nodes from 51% and 36%, respectively at day 1 to 16% and 8%, respectively, at day 20 post-infection (Figure 2B; Figure S3). The increase in vascular density observed in the g-WAT coincides with and might be caused by tissue shrinkage by up to 82% probably because of extensive lipolysis during infection (also observed in other WATs) (Trindade et al., 2016). Conversely, the decrease in vascular density observed in the spleen and lymph nodes could be because ofthe dramatic tissue enlargement (increase of 60% and 210%) that these organs undergo as infection progresses. When we compared the variation over time of the organ vascular density with the organ extravascular parasite density during 20 days of infection (Figure 1H), we found a significant correlation for some organs (g-WAT, pancreas, kidneys, iscBAT, and spleen), but not others (heart, brain, lungs, liver, and lymph nodes). However, when we considered whether vascular density correlated with reservoir establishment (i.e., considering extravascular parasite density only the first 6 days of infection), we found only a significant correlation for the spleen (R 2 = 0.92, p < 0.01), whereas the g-WAT (R 2 = 0.13, p = 0.54) and pancreas (R 2 = 0.006, p = 0.9) showed no strong correlation. Altogether, we conclude that vascular density is not the reason why WAT and pancreas are the main reservoirs early in infection. Parasite distribution across blood vessels is heterogeneous Capillaries and post-capillary venules of several organs have been shown as important locations for Plasmodium sequestration (Franke-Fayard et al., 2010). Thus, we asked whether the type of vessel allowing parasite traversal is important for T. brucei tropism. In mammals, the vasculature is comprised of vessels of different diameters, of either arterial or venous origin (Figure 3A) (Betts et al., 2013;Tucker et al., 2020). We 4Cell Reports 36, 109741, September 21, 2021 Article ll OPEN ACCESS A C B D EF GH Figure 3. Distribution of T. brucei across arterial and venous vasculature is organ specific (A) Schematic of arterial and venous vasculature, their deriving branches, and respective diameters. (B) Representative microscopy images of carboxyfluorescein diacetate succinimidyl ester (CFDA-SE)- and 7-hydroxy-9H-(1,3-dichloro-9,9-dimethylacridin-2one) succinimidyl ester (DDAO-SE)-labeled RBCs. Parasite concentration was calculated as the number of parasites relative to labeled RBCs. Scale bars, 20 mm. (C) Representative immunofluorescence images show arteries as Ephrin B2-stained vessels, while veins are Ephrin B4-stained vessels. CFDA-SE-labeled RBCs are shown in green (marked by arrows). Scale bars, 50 mm. (D) Violin plots of global parasite concentration relative to RBC in arterial or venous vasculature for the 10 organs analyzed in this study (green dotted line indicates that there are 100 parasites per 100 RBCs). Significance relative to RBC normalizers is shown as green asterisks. (E–H) Heatmaps of parasite distribution in vasculature, relative to labeled RBCs. Panels are divided in arterial and venous vasculature. Violin plots of parasite concentration relative to RBC according to vessel diameter and vessel type. Data shown in the heatmaps and violon plots refer to the representative organ of each group, which is underlined. For each day of infection, parasites were quantified in at least 100 fields of view. Significance is defined as follows: ***p < 0.001, **p < 0.01, or *p < 0.05. All relevant data used to generate this figure are included in Data S1 (tabs 8–9). See also Figures S4 and S5. Cell Reports 36, 109741, September 21, 2021 5 Article ll OPEN ACCESS hypothesized that in the WAT and pancreas, parasites may accumulate in specific vessels, which could favor crossing from the blood into the extravascular space. To address this question, we injected uninfected and infected mice with exogenously labeled RBCs (as the basis for normalization for quantifications) and monitored their distribution across all vessel types, which were marked either with the pan-EC marker CD31, the arterial marker Ephrin B2, or the venous marker EphB4, conjugated to the Alexa 647 fluorophore (Figure 3B). It has been shown that the distribution of RBCs is proportional to the caliber and hemodynamics in each vessel (Carlson et al., 2008;Chien et al., 1966;Coppola and Caro, 2009;Secomb, 2016). For each day of the infection and for each organ, we surgically stopped blood flow and we quantified the number of parasites relative to the number of labeled RBCs. Vessels were classified according to their caliber either considering the total vasculature (Figure S4) or by venous and arterial identity (Figures 3E–3H). The longitudinal results of parasite concentration in each type of vessel are displayed as heatmaps, while the overall distribution for each condition is represented by violins plots (Figure 3;Figures S4 and S5). Vessel caliber was the most important variable that contributed to heterogenous parasite distribution in the vasculature, when looking at total vasculature. We identified four phenotypes of parasite enrichment across organs (shown in Figure S4): group 1, enrichment in medium and small vessels (Figure S4B) in the brain and g-WAT; group 2, enrichment in all vessels (Figure S4C) in the pancreas and lungs; group 3, enrichment in medium and large vessels (Figure S4D) in the heart, kidneys, and isc-BAT; and group 4, no relative enrichment (Figure S4E) in the liver, spleen, and lymph nodes. When taking into account arterial and venous identity, across the 20 days of infection, we saw that in g-WAT, parasite concentration is similar in arteries and veins (p = 0.12), as is the case also for the brain (p = 0.16), lungs (p = 0.14), isc-BAT (p = 0.93), liver (p = 0.68), spleen (p = 0.74), and lymph nodes (p = 0.74). In contrast, in pancreas (p = 0.02), heart (p < 0.001), and kidneys (p = 0002), parasites were globally more concentrated in arterial vessels (Figure 3D). In g-WAT (Figure 3E), parasite concentration was highest in vessels of small and medium caliber of both arterial and venous identity, while in the brain the enrichment was mostly in arterial vasculature (Figure S5A). In the pancreas and lungs (Figure 3F; Figure S5B), parasites were highly enriched across arterial and venous vessels of all calibers. Although the heart, kidney, and isc-BAT (Figure 3G; Figure S5C) showed a higher parasite concentration in vessels of medium and large caliber, the heart mostly displayed enrichment in the arterial vasculature, while the kidneys and isc-BAT showed enrichment in both arterial and venous vasculature. Finally, in the spleen and lymph nodes (Figure S5D), parasite concentration was equivalent to the labeled RBC markers, showing no significant enrichment in any specific vessel type, while in the liver (Figure 3H), parasites seemed to be enriched in the arterial vasculature. Although parasite accumulation in small-diameter vessels of some organs can be explained by flow conditions characteristic of these organs, parasite enrichment in large vessels of the heart, kidneys, and isc-BAT was unexpected given the high flow characteristic of these vessels. This implies that parasites should exert significant forces to adhere to ECs. In conclusion, our in vivo imaging demonstrates that the distribution of parasites across vessels is extremely heterogeneous, organ dependent, and time dependent. Our results are consistent with the notion that ECs have unique features in each organ, a concept termed organotypic vasculature (Augustin and Koh, 2017). However, organs with similar parasite distribution (i.e., g-WAT and brain, or the pancreas and lungs) do not form equivalent extravascular reservoirs. Thus, additional features should contribute to reservoir establishment. Extravascular reservoirs are established before vascular permeability is compromised Vascular permeability is typically increased during an infection and has previously been related to facilitating immune cell extravasation (Schnoor et al., 2015;Vestweber, 2015). Yet, changes in vascular permeability have not been previously studied in a T. brucei infection. We hypothesized that increases in vascular permeability could regulate parasite extravasation preferentially in the g-WAT and pancreas, allowing tissue tropism. To investigate vascular permeability, we intravenously injected mice with CD31 and 70 kDa FITC-Dextran and measured both the intravascular and extravascular mean fluorescence intensity (MFI) of FITC-Dextran using established methodology for intravital microscopy (Egawa et al., 2013). At basal conditions, albeit depending on each organ’s inherent vascular permeability, most of the FITC-Dextran remains inside blood vessels. If the organ’s vasculature becomes more permeable, FITC-Dextran will leak into the organ’s parenchyma (Figure 4A). The baseline extravascular MFI values in uninfected mice (controls) for FITC-Dextran depend on each organ’s endogenous vascular permeability (Figure S6A). g-WAT and isc-BAT were the least permeable organs, followed by the brain, heart, and lungs. Spleen, pancreas, and lymph nodes showed intermediate levels of basal permeability, while the most permeable organs were the liver and kidneys (Figures S6B and S6C). Organ-specific permeability values in uninfected mice were used to calculate fold changes in permeability throughout infection (Figure 4). Vascular permeability increased in all organs throughout infection (Figure 4B). However, we observed significant differences in the extent of such increase and in the day at which leakage was first detected (Figures 4B–4D; Figure S7). In g-WAT and pancreas, vasculature becomes more permeable from day 5 post-infection, followed by a progressive increase in permeability, and at the end of the infection (day 20), g-WAT is the tissue that shows the highest increase in permeability (14-fold). Figure 4B shows organs in increasing order of the onset of increased vascular permeability. Among these organs, the onset of increased vascular permeability happens on day 3 in spleen and lymph nodes, on day 5 in liver, kidney, and isc-BAT, on day 6 in brain and heart, and on day 7 in lungs. For most organs, maximum permeability is reached close to the end of the infection (day 20) (Figure 4B; Figure S7). Interestingly, in the brain, permeability remains low until day 7 and increases dramatically (2.6to 5-fold, p < 0.001) on days 8–9, suggesting a significant disruption of the blood-brain barrier around this time. This is consistent with our IVM observations of petechiae and pools of 6Cell Reports 36, 109741, September 21, 2021 Article ll OPEN ACCESS FITC-Dextran across the brain from day 8 of infection onward (Figure S7). When we compare the variation over time of the increase in vascular permeability of each organ with the extravascular parasite density, we found a significant correlation for most organs (R 2 values ranging between 0.2 and 0.83; p values ranging from 0.04 and <0.001), with the exception of the isc-BAT (R 2 = 0.07, p = 0.24) and lymph nodes (R 2 = 0.07, p = 0.26). g-WAT and pancreas showed significant correlations throughout the 20 days of infection (R 2 = 0.42, p = 0.002 for g-WAT; R 2 = 0.64, p < 0.001 for pancreas), and the highest fold change in the first parasitemia wave (Figure 4C). These results suggest that increased vascular permeability favors parasite extravasation in most organs. HowAB CD EF Vascular Figure 4. Changes in vascular permeability during T. brucei infection (A) Vascular permeability was measured as an increase in the mean fluorescence intensity (MFI) emitted by the 70-kDa FITC-Dextran in the extravascular space per organ. Scale bar, 50 mm. (B) Heatmap of vascular permeability every day of infection. Vascular permeability is calculated as the FITC-Dextran MFI in the extravascular space at each day post-infection relative to FITC-Dextran MFI in uninfected control (baseline values are shown in Figure S6). The white dotted lines show the infection events defined in Figure 1G. (C) Boxplot showing quantitative values of vascular permeability corresponding to d5–8. (D) Graphical representation comparing the time of increased T. brucei extravasation in each organ (red dot) and the time of increased vascular permeability (blue dot). (E) Heatmaps of vascular permeability upon induction by histamine treatment 2 days prior to infection up to day 6 post-infection (left), an extravascular parasite density (right). (F) Boxplots of extravascular parasite density at d3–6 post-infection in mice with (values from E, right panel) or without histamine treatment (values from Figure 1G). For all graphs, significance is shown as ***p < 0.001, **p < 0.01, or *p < 0.05. For (B) and (H), FITC-Dextran was injected intravenously immediately prior to imaging, and at least 100 fields of view were used for measurements. Heatmap shows the mean of these values. All relevant data used to generate this figure are included in Data S1 (tabs 10–14). See also Figures S6 and S7. ever, we noted that the sharp increase in parasite density in g-WAT and pancreas (at day 3 post-infection) precedes the increase in vascular permeability in these tissues (day 5 post-infection) (Figure 4D). Thus, parasites preferentially enter and accumulate in the extravascular spaces of g-WAT and pancreas before vascular integrity is compromised. Considering only days 1–5 of infection, we found no correlation between extravascular parasite density and vascular permeability in g-WAT and pancreas, suggesting that tissue tropism is independent of vascular permeability. To further confirm that vascular leakage is not involved in the early colonization of g-WAT and pancreas, we induced vascular leakage with histamine, which induces vasodilation and increases leakiness (Egawa et al., 2013)(Figure 4E, left panel). Upon treatment of mice with histamine 2 days prior to infection with T. brucei, and during the first 3 days of infection, the 70kDa FITC-Dextran presence in the organ’s parenchyma revealed a generalized increased vascular permeability, leading to a 1.2to 5-fold increase in permeability relative to untreated mice during the same days post-infection. We capped the analysis to day 6 post-infection, because this is the first time point by which the Cell Reports 36, 109741, September 21, 2021 7 Article ll OPEN ACCESS tissue reservoirs are already established. In general, the treatment with histamine results in only a slight increase in the number of parasites in the parenchyma of most organs (with only the spleen showing a significant difference relative to untreated mice, p < 0.01) (Figure 4F). Importantly, the establishment of the tissue reservoirs in the g-WAT and pancreas does not start earlier than in non-treated conditions (Figure 4F). Altogether we conclude that although vascular permeability is likely a key player for the presence of parasites in the extravascular space of various organs, it does not explain the initial establishment of the reservoirs in g-WAT and pancreas. These results suggest that the mechanism for tissue tropism is active and independent of vascular permeability. Endothelial adhesion molecules are upregulated during infection Our intravital imaging approach showed that parasites frequently contact ECs (Video S2). ECs are covered by many surface molecules that play important communication roles with circulating cells, including leukocytes or cancer cells (Muller, 2002,2013; Vestweber, 2015;Wettschureck et al., 2019). Moreover, many of these adhesion molecules play a role in host-parasite interactions (Smith et al., 2001). To investigate whether EC surface proteins are involved in parasite extravasation, we began by measuring the protein expression levels of seven EC adhesion molecules in all the organs of non-infected or infected mice. We measured P-selectin, E-selectin, ICAM1, ICAM2, VCAM1, PECAM1, and CD36 by injecting fluorescently labeled antibodies against those adhesion molecules in non-infected mice and animals at day 6 post-infection and by measuring the fluorescence intensities by intravital microscopy (Figure 5;Figure S8). The route of injection (retroorbital or caudal intravenous) did not result in differently labeled vasculature (Figure S9). The heatmap shows the intensity of antibody staining in animals infected for 6 days relative to non-infected animals (Figure 5B). We observed that several adhesion molecules become more abundant during infection, but the extent of this effect was very variable among organs and among adhesion molecules. Of note, E-selectin is the adhesion molecule that is more upregulated during infection (11.3-fold increase on average [p < 0.001], ranging from 2.5 to 26.7 across organs). In the g-WAT of infected animals, CD36 and E-selectin were significantly upregulated during infection (1.9and 10.6-fold; p < 0.001). In the pancreas, E-selectin showed the highest upregulation following infection (13.6-fold, p < 0.001), while P-selectin (1.8-fold, p < 0.001) and CD36 (1.7-fold, p < 0.001) upregulation were also significant. We conclude that during a T. brucei infection there are significant changes in the pattern of expression of the selected EC adhesion molecules, which indicates organ-specific molecular alterations in the vasculature during infection. Adhesive molecules modulate T. brucei tissue colonization Next, we investigated the effects of blocking each of these adhesion molecules on parasitemia and mouse survival. In addition, we targeted P-selectin in combination with E-selectin, because compensation mechanisms for both molecules have been reported (Hickey et al., 1999). We treated mice with the respective blocking antibodies for 8 days (from 2 days prior to infection until 6 days post-infection). We monitored peripheral parasitemia (Figures 6A and 6B). In general, when the effect of blocking a receptor led to a significant reduction in parasitemia in peripheral blood, it also led to an increase in mouse survival (Figure 6C). We observed that blocking E-selectin, P-selectin, ICAM2, CD36, and PECAM1 allowed 50% of animals to survive past day 40. Interestingly, simultaneous blocking of Pand E-selectin appears to have cured animals because no animals died (Figure 6C). In contrast, blocking ICAM1 and VCAM1 did not change parasitemia and resulted in a mixed survival phenotype (with some mice dying at similar times as control mice, while others had extended survival at around 10 days compared with controls). A B Figure 5. Changes in expression of vascular EC receptors during a T. brucei infection (A) Expression of seven EC receptors (P-selectin, E-selectin, ICAM1, ICAM2, VCAM1, PECAM1, and CD36) measured by intravital imaging after injection of antibodies coupled to A647 fluorophore. Representative intravital images relative to g-WAT and pancreas. Scale bars: 50 mm. (B) Heatmap shows MFI values for each receptor measured in each organ on uninfected and d6-infected mice. Measurements are the mean value of at least 100 vessels measured per condition. All relevant data used to generate this figure are included in Data S1 (tabs 15 and 16). See also Figures S8 and S9 and Video S2. 8Cell Reports 36, 109741, September 21, 2021 Article ll OPEN ACCESS Smith, J.D., Gamain, B., Baruch, D.I., and Kyes, S. (2001). Decoding the language of var genes and Plasmodium falciparum sequestration. Trends Parasitol. 17, 538–545. Theron, M., Hesketh, R.L., Subramanian, S., and Rayner, J.C. (2010). An adaptable two-color flow cytometric assay to quantitate the invasion of erythrocytes by Plasmodium falciparum parasites. Cytometry A 77, 1067–1074. Trindade, S., Rijo-Ferreira, F., Carvalho, T., Pinto-Neves, D., Guegan, F., Aresta-Branco, F., Bento, F., Young, S.A., Pinto, A., Van Den Abbeele, J., et al. (2016). Trypanosoma brucei Parasites Occupy and Functionally Adapt to the Adipose Tissue in Mice. Cell Host Microbe 19, 837–848. Tucker, W.D., Arora, Y.M.K., and Mahajan, K. (2020). Anatomy, Blood Vessels (StatPearls). Urban, B.C., Hien, T.T., Day, N.P., Phu, N.H., Roberts, R., Pongponratn, E., Jones, M., Mai, N.T.H., Bethell, D., Turner, G.D.H., et al. (2005). Fatal Plasmodium falciparum malaria causes specific patterns of splenic architectural disorganization. Infect. Immun. 73, 1986–1994. Venugopal, K., Hentzschel, F., Valki unas, G., and Marti, M. (2020). Plasmodium asexual growth and sexual development in the haematopoietic niche of the host. Nat. Rev. Microbiol. 18, 177–189. Vestweber, D. (2015). How leukocytes cross the vascular endothelium. Nat. Rev. Immunol. 15, 692–704. Wettschureck, N., Strilic, B., and Offermanns, S. (2019). Passing the Vascular Barrier: Endothelial Signaling Processes Controlling Extravasation. Physiol. Rev. 99, 1467–1525. Zhang, F., Hao, G., Shao, M., Nham, K., An, Y., Wang, Q., Zhu, Y., Kusminski, C.M., Hassan, G., Gupta, R.K., et al. (2018). An adipose tissue atlas: an imageguided identification of human-like BAT and beige depots in rodents. Cell Metab. 27, 252–262.e3. Cell Reports 36, 109741, September 21, 2021 15 Article ll OPEN ACCESS STAR+METHODS KEY RESOURCES TABLE REAGENT or RESOURCE SOURCE IDENTIFIER Antibodies Alexa Fluor 647 anti-mouse CD36 clone HM36. Isotype Armenian Hamster IgG 0.5mg/ml BioLegend Cat# 102610; RRID: AB_528794 Alexa Fluor 647 rat anti-mouse CD62P (P-selectin) clone RB40.34 0.2mg/ml BD PharMingen Cat# 563674; RRID: AB_2738366 Alexa Fluor 647 anti-mouse CD102 (ICAM2) clone 3C4 (MIC2/4). Isotype Rat IgG2a,k BioLegend Cat# 105612; RRID: AB_2122182 Alexa Fluor 647 anti-mouse CD54 (ICAM1) clone YN1/1.7.4. Isotype Rat IgG2b,k BioLegend Cat# 116114; RRID: AB_493495 Alexa Fluor 647 anti-mouse CD31 (PECAM1) clone 390. Isotype Rat IgG2a,k BioLegend Cat# 102416; RRID: AB_493410 Mouse CD106/VCAM-1 FITC conjugate Invitrogen Cat# RMCD10601; RRID: AB_2556575 Purified Rat Anti-mouse CD102 BD PharMingen Cat#553326; RRID: AB_394784 Blocking antibody RB40.34 against CD62-P BD Biosciences Cat# 553742; RRID: AB_2254315 Blocking antibody P2H3 against CD62-E eBioscience Cat# 14-0627-82; RRID: AB_1210768 Blocking antibody CBR 1C2/2 against CD102 eBioscience Cat# BMS109; RRID: AB_10598677 Blocking antibody YN1/1.7.4 against CD45 eBioscience Cat#16-0541-85; RRID: AB_468980 Blocking antibody against CD31 abcam Cat#ab32457; RRID: AB_726369 Blocking antibody against CD36 (185-1G2) ThermoFisher Cat# sc-21772; RRID: AB_627042 Blocking antibody Mouse IgG1 K isotype control (P3.6.2.8.1) eBioscience, ThermoFisher Cat# 16-4714-82; RRID: AB_470161 Blocking antibody Mouse IgG2 K isotype control (eBM2a) eBioscience, ThermoFisher Cat# 14-4724-82; RRID: AB_470114 Bacterial and virus strains Viral envelope vector VSVG Addgene Plasmid #45494 Chemicals, peptides, and recombinant proteins mCD36/Fc chimera (recombinant mouse, CHO cell-derived) R&D systems Cat# 2519-CD Critical commercial assays Chemotaxis 24-well cell migration assay kit Merck ECM505 Experimental models: Cell lines HEK293T ATCC CRL3216; RRID: CVCL_0063 HUVECs Lonza C2519A Experimental models: Organisms/strains Mouse: C57BL/6 Charles River Strain code: 027 Mouse: C57BL/6 Albino Charles River Strain code: 493 Mouse: CD36/Jackson laboratories Stock code: 019006 Mouse: CD36+/+ In house breeding Mouse: CD36+/In house breeding Oligonucleotides pLenti-C-mGFT-P2A-Puro OriGENE PS100093 Cd36 (NM_007643) Mouse Tagged ORF Clone Lentiviral Particle OriGENE MR227663L4 (Continued on next page) e1 Cell Reports 36, 109741, September 21, 2021 Article ll OPEN ACCESS RESOURCE AVAILABILITY Lead contact Further information and requests for resources and reagents should be directed to and will be fulfilled by the lead contact, Luisa Miranda Figueiredo ([email protected]). Materials availability This study did not generate new unique reagents. Data and code availability dAll data used for the generation of figures in this paper has been included in Data S1. dThis paper does not report original code. dAny additional information required to reanalyze the data reported in this paper is available from the lead contact upon request. EXPERIMENTAL MODEL AND SUBJECT DETAILS Animals used Animal experiments were performed according to EU regulations and approved by the Animal Ethics Committee of Instituto de Medicina Molecular (IMM) (AEC_2011_006_LF_TBrucei_IMM). Mice used across this study included wild-type male C57BL/6J mice, and C57BL/6 Albino obtained from Charles River, France. Transgenic B6.129S1-Cd36 tm1Mfe /J (CD36 / ) mice were obtained from The Jackson Laboratory. Breedings were generated in-house at IMM. A homozygous KO line was kept over multiple generations. A heterozygous (CD36 +/ ) line was generated by crossing CD36 / mice with C57BL/6J mice, and a homozygous CD36 +/+ line was generated through backcrossing. Mice were backcrossed over multiple generations, and littermates of all 3 genotypes were used throughout this work. All mice were 6–9 weeks old males, with an average weight ranging between 25 and 30 g. Trypanosoma brucei parasites and infections Mice were infected by intraperitoneal injection of 2,000 T. brucei AnTat 1.1 E chimeric triple reporter parasites (Calvo-Alvarez et al., 2018) expressing the red-shifted firefly luciferase protein PpyREH9, TdTomato and ty1. For parasite counts, blood samples were taken daily by tail vein puncture, using 1ml of blood diluted in 200 ml of HMI11 medium, and 10 mm loaded in a hemocytometer for parasite quantification. For mouse survival, mice were followed up until clinical signs indicating no recovery. After determining the time of highest disease recrudescence in the survival experiments, all other experiments were capped to 20 days of infection. HUVECs cell culture Human umbilical vein ECs (HUVECs – Lonza, C2519A) were routinely cultured following the manufacturer’s guidelines at 37C and 5% of CO 2 with complete medium EGM-2 Bulletkit (Lonza, CC-3162) without antibiotics. All the experiments were performed between passages 2 and 4. When passaging HUVECs for experiments, cells were washed twice in sterile PBS (137mM NaCl, 2.7mM KCl, 4.3mM Na2HPO4, 1.47mM KH2PO4, pH7.4) and then incubated for 5min in TrypLE Express Enzyme (1X) (Alfagene, 12605028) at 37C, 5% CO2. When 95% of the cells detached, complete medium was added to each flask to inhibit the activity of the TrypLE Express Enzyme and the cell suspension was transferred into a falcon tube. Cells were then centrifuged at 700rpm for 5min at 10C and the pellet re-suspended in fresh complete medium. HUVECs were then seeded at the desired concentration, depending on the experiments. HUVECs viral transduction Replication-incompetent lentiviruses were produced by transient transfection of HEK293T co-transfected with the viral packaging vector D8.9 and the viral envelope vector VSVG. Medium was replaced with fresh culture medium 4-6h post transfection. 48h after post transfection, lentiviral particles were concentrated from supernatant by ultracentrifugation at 90000 g for 1h30 and re-suspended in 0.1% BSA PBS. HUVECs were transduced 24h after seeding with 1% lentiviral plasmids containing GFP (control) (pLenti-C-mGFT-P2A-Puro, Cat # PS100093, Origene) or CD36-GFP (NM_007643, Cat # MR227663L4, Origene) sequences. 24h Continued REAGENT or RESOURCE SOURCE IDENTIFIER Software and algorithms ImageJ (Schneider et al., 2012)https://imagej.nih.gov/ij/ Cell Reports 36, 109741, September 21, 2021 e2 Article ll OPEN ACCESS after viral transduction, media was replaced by fresh complete medium. Cells were then kept in culture until 48h post-transduction and then processed for T. brucei co-culture and imaging. METHOD DETAILS Bioluminescence imaging (in vivo and ex vivo) For whole body imaging, infected mice were injected with 200 ml of RediJect D-luciferin (Xenolight, Perkin Elmer) prior to imaging. All measurements were performed in an IVIS Lumina imaging system. A kinetic curve was established to determine the peak of bioluminescence, which was found to be at 10 minutes post-injection, with a plateau lasting a further 10 minutes. For in vivo imaging, mice were anaesthetized with Isofluorane (Isotroy) and imaged using an exposure time of 1 minute and a FOV D (12.5 3 12.5 cm). For ex vivo imaging of non-perfused organs, mice were injected with RediJect as described above, and sacrificed using CO 2 within 3 minutes following this injection. Organs were then extracted, washed in 1x PBS, and placed in a plastic Petri dish for imaging at the IVIS Lumina instrument, using an exposure time of 1 minute and a FOV C (10 310 cm). For ex vivo imaging of perfused organs, mice were injected with RediJect as described above, and sacrificed using CO 2 within 3 minutes following this injection. The chest was then exposed, the inferior vena cava was cut, and the heart was injected with 40 mL of warm 1x PBS. Organs were then imaged as described above. Image acquisition was obtained, and image analysis performed using the Living Image software version 3.0.4.6. For all bioluminescence measures, the results of 9 animals (3 biological replicates in triplicate) are expressed. Intravital and ex vivo imaging For intravital imaging, surgeries were separately performed in the brain; the lungs and heart; the liver; the pancreas, spleen and kidney; the adipose tissues and lymph nodes, as described in De Niz et al. (2019a,2019b,2019c,2020). Briefly, mice were anaesthetized with a mixture of ketamine (120 mg/kg) and xylazine (16 mg/kg) injected intraperitonially. After checking for reflex responses and ensuring none occurred, mice were intraocularly injected with Hoechst 33342 (stock diluted in dH 2 O at 100 mg/ml; injection of 40 mg/kg mouse), 70 kDa FITC-Dextran (stock diluted in 1x PBS at stock concentration of 100 mg/ml; injection of 500 mg/kg), and vascular markers of interest conjugated to A647 (CD31 (Biolegend, used at 20 mg), Ephrin B2 (R&D systems, used at 20 mg), Eph-B4 (R&D systems, used at 20 mg), or VEGFR3 (R&D systems, used at 20 mg). A temporary glass window (Merk rectangular coverglass, 100 mm x 60 mm or circular coverglass (12 mm)) was implanted in each organ, and secured either surgically, with surgical glue, or via a vacuum, in order to enable visualization of the organ surface. For intravital microscopy, all imaging relative to parasite quantifications, vascular density and vascular leakage was done in a Zeiss Cell Observer SD (spinning disc) confocal microscope (Carl Zeiss Microimaging, equipped with a Yokogawa CSU-X1 confocal scanner, an Evolve 512 EMCCD camera and a Hamamatsu ORCA-flash 4.0 VS camera) or in a 3i Marianas SDC (spinning disc confocal) microscopy (Intelligent Imaging Innovations, equipped with a Yokogawa CSU-X1 confocal scanner and a Photometrics Evolve 512 EMCCD camera). Laser units 405, 488, 561 and 640 were used to image Hoechst in nuclei, extravascular and intravascular FITC-Dextran, TdTomato in T. brucei, and CD31, respectively. The objective used to image vascular density, vascular leakage, and proportion of intravascular and extravascular parasites was a 40x LD C-Apochoromat corrected, water immersion objective with 1.1 NA and 0.62 WD. The objective used to classify T. brucei movement phenotypes was a 100x plan-apochromat, oil immersion objective with 1.4 NA and 0.17 WD. Between 20 and 100 images were obtained in any one time lapse, with an acquisition rate of 20 frames per second. For vascular density measurements and parasite quantification, in order to gain access to the full organ, we performed ex vivo imaging from different organ regions. For this, we performed z stacks consisting of 16 stacks covering up to 200 mm of tissue depth. For all acquisitions, the software used was ZEN blue edition v.2.6 (for the Zeiss Cell Observed SD) allowing export of images in .czi format, and 3i Slidebook reader v.6.0.22 (for the 3i Marianas SD), allowing export of images in TIFF format. Intravascular and extravascular parasite quantification In order to quantify intravascular and extravascular parasites, we took as reference, the vascular marker CD31-A647 and 70 kDa FITC-Dextran, and we quantified numbers of T. brucei parasites within the confines of the regions marked by the vascular marker, and numbers of parasites outside these regions. We then normalized the total quantity to numbers per mm 2 so as to be able to compare both measurements. Measurements were repeated throughout 20 days of infection, and at least a total of 100 fields of view were quantified. Vascular density and diameter quantification In order to quantify vascular density, we took as reference the vascular marker CD31-A647 and 70 kDa FITC-Dextran. We obtained 100 fields of view, and for each field of view the total area was defined as 100%. Using the CD31 signal we were able to segment out the vascular regions using Fiji software. We calculated the percentage of vascular area covered using the following formula: AV%= Av3AT% AT, where A T is the total area of the field of view, A T% is 100, and A Tv is the total area marked by CD31. Vascular density measurements were performed throughout 20 days of infection. Vessel diameters were measured using Fiji. e3 Cell Reports 36, 109741, September 21, 2021 Article ll OPEN ACCESS Vascular permeability quantification In orderto quantify vascularpermeability changes, we took as reference themarker 70 kDa FITC-Dextran as previously published methodology (Egawa et al., 2013). We measured FITC-Dextran in intravascular and extravascular regions in uninfected mice, and then at each time post-infection with T. brucei. The permeability ratio was calculated using the following equation: Permeability Ratio Dn = Mean MFIDn Mean MFI D0, where mean MFI Dn is the extravascular MFI at a specific Day n, and the mean MFI D0 is the extravascular MFI in uninfected mice (Day 0). To induce vascular hyper-permeability, 5 mg/ml of histamine (Sigma-Aldrich) were prepared in 1 x PBS, and 200 ml were injected intravenously every 2 days, starting 2 days prior to infection. Erythrocyte labeling and parasite quantification normalization by vascular type To quantify and normalize parasite numbers per vascular type, red blood cells were extracted from uninfected mice, and pre-labeled ex-vivo with intracellular dyes CFDA-SE or DDAO-SE as previously described (Theron et al., 2010). The required volume of erythrocytes at 5% hematocrit were resuspended in 500 ml HMI11. RBCs were centrifuged and the pellet resuspended either in 20 mM carboxfluorescein diacetate succinimidyl ester (CFDA-SE) (Invitrogen), or 10 mM 7-hydroxy-9H-(1,3-dichloro-9,9-dimethylacridin2-one) succinimidyl ester (DDAO-SE) (Invitrogen) in HMI11 and incubated for 20 min at 37C. The suspension was washed 3x in HMI11, and resuspended at concentrations equaling those observed for parasites at each day post-infection. The labeled RBCs were then injected into infected mice. The number of RBCs per vessel type were quantified, and the number of parasites expressed as a percentage of labeled RBCs in any one vessel as follows: Relativeparasite %Dn= ParasitenumbersDn3labeledRBCT% labeledRBCnumbersDn The value of Relative parasite % at Day n(D n ) was plotted and color coded in heatmaps based on total vasculature, arterial vasculature, or venous vasculature. Labeling endothelial receptors To investigate the relative expression of EC receptors in different organs, we intravenously injected 20 mg of antibodies against Eselectin, P-selectin (BD PharMingen), ICAM1, ICAM2, PECAM1, CD36 (BioLegend) and VCAM1 (Invitrogen) conjugated to FITC (VCAM1) or A647 (all other antibodies), into uninfected or day 6 infected mice, as previously described in the context of parasitology for CD31 (Hopp et al., 2015). We measured MFIs of at least 100 different vessels per organ in 3 separate mice using an LSM 710 Zeiss microscope, a 40x objective (1.3 NA). Blocking endothelial receptors To investigate the effects of blocking various EC receptors, antibodies used included RB40.34 against P-selectin (BD Biosciences, 30 mg per mouse); P2H3 against E-selectin (R&D systems, 20 mg per mouse), a 1:1 combination of both; CBR IC2/2 against ICAM2 (Invitrogen, 20 mg per mouse); YN1/1.7.4 against ICAM1 (eBioscience, 20 mg per mouse); CD31 antibody (abcam, 20 mg per mouse); and 185-1G2 against CD36 (Abcam, 20 mg per mouse) blocking antibodies were used. Mouse IgG1 and IgG2 isotype controls were used as controls (eBioscience). Antibodies were injected intravenously daily by tail vein injection, starting two days prior to infection, and continuing until day 6 post-infection. Recombinant CD36 binding assays For in vitro measurement of T. brucei attachment to CD36, we coated dishes with recombinant mouse CD36/Fc chimera (R&D). Triplicate plastic dishes were prepared for each concentration. They were first washed with 1x PBS, and then blocked with 1% (wt/v) bovine serum albumin (BSA) at 4C overnight. After washing once with 1x PBS, dishes were coated with recombinant mouse CD36 (2nM, 5nM, 10nM, 50nM, 100nM and 200nM), or with 1x PBS (control 1) or 1x PBS containing 1% BSA (cell culture grade; control 2) for 3h at 37C in a cell culture incubator. Following this time, plates were washed 3 times and incubated again with 1% BSA in PBS for 30 minutes. One million T. brucei in HMI-11 were overlayed on the dishes for 1h. After 1h, total parasites were quantified by microscopy prior to performing washouts with fresh HM1-11 media. After washouts, bound parasites were quantified by microscopy. Each condition was done in triplicate and each experiment repeated 3 times. CD36 binding assays in HUVECs T. brucei parasites were diluted to a concentration of 0.5 million parasites per ml, and 1ml was overlayed on wells in 24-well plates containing control and CD36-expressing HUVECs for 0.5, 1, 2, 4, 6 and 8 hours. Separate wells were used for each time point. At each time point, total parasites were quantified by microscopy prior to performing washouts. After washouts to remove any unbound parasites, attached parasites were quantified by microscopy. CD36 transmigration assays in HUVECs For transmigration assays, a QCM chemotaxis cell migration assay (24 well-plates, 3 mm) (MERCK, ECM 504) was used. WT, GFPand CD36-expressing HUVECs were seeded in the base of the upper chamber of the transwell set, in complete medium. The lower chamber was also filled with complete medium. Half a million T. brucei parasites suspended in 250 ml of HMI-11 were overCell Reports 36, 109741, September 21, 2021 e4 Article ll OPEN ACCESS layed on the ECs in the upper chamber, and the co-culture was maintained for 1 hour. At time point zero, parasites were quantified in the upper and lower chambers by microscopy. At 1h post-co-culture, parasites in the upper and lower chambers were again quantified. QUANTIFICATION AND STATISTICAL ANALYSIS Data were displayed in graphs and heatmaps generated using Prism 9 software (GraphPad). Means, medians, survival, correlation tests, comparison tests, and error measures were calculated from triplicate experiments with 3 biological replicates each, and/or at least 100 images per condition. For comparisons of bioluminescence measures between organs of different groups we performed multiple t tests in addition to a one-way ANOVA (differences were considered significant when p < 0.05). Pearson correlations measures (R), and R 2 values were calculated to determine the strength of linear association between parasite density and either vascular permeability or vascular density. For comparison of proportions between parasites in vessels of different diameters and proportions in vessels of equal diameter across vascular groups, a linear model was performed, and p values < 0.05 were considered significant. For comparisons of survival, a log-rank (Mantel-Cox) test was performed, and p values < 0.05 were considered significant. Statistical details of experiments are included in the figure legends, the results section, and a Supplementary Table excel file. All data used for the generation of the figures is included as a supporting file. e5 Cell Reports 36, 109741, September 21, 2021 Article ll OPEN ACCESS