Repetitive schistosoma exposure causes perivascular lung fibrosis and persistent pulmonary hypertension

Pulmonary hypertension (PH) is characterized by vascular remodeling, right ventricular (RV) hypertrophy, and perivascular fibrosis. Many pulmonary vascular diseases develop after prolonged or repeated exposures, such as after many years of autoimmune disease, or with repeated hypoxia exposures in the setting of obstructive sleep apnea. Irreversibility of pulmonary pathology is a major challenge in PH, as the therapies currently available improve symptoms and slow progression but do not reverse established disease.

Dysregulated host inflammation probably contributes to the pathogenesis of many forms of PH, including that due to schistosomiasis, which due to its high prevalence is a common cause of PH worldwide. The Schistosoma mansoni parasite lays its eggs in the portal venous system of the host. The chronic and repeated infection causes pre-portal liver fibrosis, resulting in portal hypertension and the opening of portocaval shunts. These shunts enable the passage of eggs around the liver, from the portal system to the systemic venous circulation, where the eggs embolize into the lungs. Schistosoma eggs have a diameter of ∼100 µm, lodging in pre-capillary vessels and resulting in significant localized Type 2 inflammation. This intrapulmonary inflammation is likely critical to the pathogenesis of vascular disease in Schistosoma-PH.

In mice experimentally exposed to Schistosoma, sensitization prior to intravenous egg challenge is required for the PH phenotype [1], indicating a requirement for adaptive immunity. Further, mice lacking B and T cells are protected from Schistosoma-PH [2]. Activating CD4 T cells in the lung to a Th2 phenotype results in significant localized expression of IL-4 and IL-13, and mice absent these cytokines globally or in CD4 T cells specifically are protected from Schistosoma-PH [2,3]. IL-4/IL-13 expression results in the recruitment of CCR2⁺ monocytes to the adventitial space, where these cells express thrombospondin-1 (TSP-1), which functionally activates TGF-β causing vascular remodeling [4].

While inflammation is a clear driver of schistosomiasis-PH, it is unclear how host inflammation evolves and contributes to the persistence of the PH disease after the initial triggering event. Clinical data indicate that despite treatment with antihelminthics such as praziquantel, schistosomiasis PH in humans persists and progresses [5–7]. Furthermore, modern autopsy case series of individuals who died of Schistosoma-PH reveal substantial vascular remodeling but no significant egg antigens, consistent with parasite eradication but the progressive vascular disease [8].

Here, we explore the concept that the duration of disease and parasite exposure contribute to the development of persistent pulmonary pathology. We tested the specific hypothesis that repeated exposure to Schistosoma antigens contributes to the development of persistent PH, using a mouse model of repeated intravenous egg administration. Overall, we observed that repeated Schistosoma exposure leads to persistent PH accompanied by perivascular fibrosis, a pathologic feature we also found present in human lung tissue from individuals who died of Schistosoma-PH.

C57BL6/J mice were purchased from Jackson Laboratories (#00664, Bar Harbor, ME). Female mice were used between 6 and 8 weeks of age at the commencement of the experiment, as PH is more prevalent in females [9] and thus may be more representative of the human condition; we have not previously observed a difference in Schistosoma-PH phenotype between males and females; and we have found older male mice may develop spontaneous mild PH which would be confounding in the longer time course experiments. All animals were housed under specific pathogen-free conditions in an American Association for the Accreditation of Laboratory Animal Care-approved facility at the University of California, San Francisco (UCSF). All animal studies were approved by the UCSF Institutional Animal Care and Use Committee. The animal procedures followed were approved by institutional guidelines and conducted at UCSF.

Schistosoma mansoni fresh and live eggs were obtained by homogenizing and purifying the livers of Swiss-Webster mice pre-infected with S. mansoni cercariae from the Biomedical Research Institute (Rockville, MD). Similar to prior publications [3,4,10], the experimental mice were intraperitoneally (IP) sensitized with 240 S. mansoni eggs/gram body weight, and then 2 weeks later intravenously (IV) challenged with 175 S. mansoni eggs/gram body weight, and right heart catheterization (RHC) was performed at +7 days (Group B; see Figure 1A), +14 days (Group C), +21 days (Group D), or +28 days (Group E) after the IV challenge. Control mice were unexposed to S. mansoni eggs (Group A1), only IP +14 days (Group A2) or only IP +28 days sensitized (Group A3). In the multiple IV challenge model, IP sensitized mice were IV challenged with three doses of Schistosoma at 2 weeks intervals for 3 total IV doses. RVSP was measured at +7 days (Group F; see Figure 2A) or +21 days (Group G) following the third IV challenge. The dose of IV eggs was the same 175 S. mansoni eggs/gram body weight each time.

We administered neutralizing antibodies targeting PD-1 to selected mice, or isotype control antibodies to control animals. For PD-1 intervention, mice were injected with 200 µg of anti-PD-1 antibody (clone 29F.1A12, BioXcell) or rat IgG2a isotype control (clone 2A3, BioXcell), administered intraperitoneally with the regimens outlined in Supplementary Figure S10A, with the dose based on prior publications [11].

To measure hemodynamics, the mice underwent terminal RHC by an open chest method as described previously [3,4,10]. Briefly, the mice were sedated with intraperitoneal ketamine-xylazine injection, and a tracheostomy tube was placed, and mechanical ventilation was initiated at 6cc/kg. The abdomen and diaphragm were opened by careful dissection to avoid any blood loss, and a 1 Fr pressure–volume catheter (PVR-1035, Millar ADInstruments, Houston, TX) was placed directly into the right ventricle (RV), and then the left ventricle (LV) chambers through the free walls.

Frozen lung tissues were macerated and sonicated in RIPA buffer (Thermo Scientific, Rockfield, IL, U.S.A., cat. # 89900) containing protease and phosphatase inhibitor cocktails (EMD Millipore, MA, U.S.A., cat. #s 539131 and 524625, respectively). Total protein concentration was determined by Bradford assay. IL-4, IL-13, and TGF-β1 levels were quantified in the whole lung lysates by ELISA using the commercial kits (Mouse IL-4 Quantikine ELISA Kit, M4000B; Mouse IL-13 Quantikine ELISA Kit, M1300CB and TGF-β1 Quantikine ELISA Kit, MB100B from R&D Systems, Inc. Minneapolis, MN; Mouse C3a ELISA Kit, MBS268280, MyBioSource, San Diego, CA). Active TGF-β1 was quantified using the lung lysates directly; total TGF-β1 was quantified by acid-treating the samples prior to quantification, per the manufacturer’s directions, using the same kit.

A mouse cytokine array was performed on lung tissue lysates per the manufacturer’s guidelines (Proteome ProfilerTM Array, R&D System, MN, U.S.A.). In brief, 200 μg protein from whole lung lysates was mixed with binding buffer and incubated overnight with a membrane containing 40 antibodies for inflammatory markers. Chemiluminescent imaging of protein blots using a BioSpectrum™ 510 Imaging System integrated with a VisionWorksLS image capture and analysis software package (Ultra-Violet Products, LLC, Upland, CA) were used to capture the images and perform quantification.

Real-time polymerase chain reaction (RT-PCR) was used to quantify the expression of mRNA transcripts in lung lysates using the primers in Supplementary Table S1. iScript cDNA Synthesis kit (Bio-Rad Laboratories, Hercules, CA, U.S.A., cat.#: 1708891) was used for first-strand cDNA synthesis using 1 µg RNA. RT-PCR was performed in duplicate for each gene and each sample. TaqMan™ Gene Expression Master Mix (Applied Biosystems, cat.#: 4369016) was used to see the expression of target genes using StepOnePlus™ Real-Time PCR System (Applied Biosystems) using the following conditions: Holding stage: 55°C for 10 min, 95°C for 3 min; Cycling stage: 95°C, 5 s and 60°C, 10 s; no. of cycles = 40. A reaction volume of 20 μl was used. To calculate relative transcript quantities, the 2^−ΔCt method was used with β-actin as the endogenous reference genes. In addition, Col1a1 mRNA had been previously quantified using a HiSeq 2000 RNA sequencing (RNA-seq) system (Illumina, San Diego, CA); data are deposited in National Center for Biotechnology Information’s GEO dataset GSE49116 [12].

Immunostaining for α-smooth muscle actin was performed on FFPE mouse lung tissue using the reagents listed in Supplementary Table S2. To quantify fractional media thickness, images of 10–12 vessels of each specimen from the α-smooth muscle actin-stained slides were acquired using a Nikon Eclipse 80i microscope (Nikon, Melville, NY) with an Olympus DP74 color camera (Olympus, Waltham, MA). The external and internal perimeters of the media were identified using image processing software (ImagePro v10, Media Cybernetics, Bethesda, MD), the radii of each layer calculated (r = sqrt[A/π]), and the fractional thickness calculated from the difference between the two radii divided by the external radius. Peri-egg granuloma volumes were measured using the optical rotator stereological method, with hematoxylin and eosin stained FFPE tissue, using 8–10 images of granulomas with a single visible egg acquired for each sample, and the rotator method for object volume estimation applied using the egg as the central point (ImagePro) [3,13].

Immunostaining for pSmad2/3, collagen-producing cells such as fibroblasts, and macrophages was performed on FFPE mouse lung tissue using the reagents listed in Supplementary Table S2. Collagen-producing cell density quantification as performed as the ratio of HSP47⁺ cells in the adventitial space, or the intima plus media space, divided by the area analyzed, using Stepanizer v1.8 (stepanizer.com), and macrophage density using CD68⁺ cells in the same manner. Picrosirius red staining was used to quantify perivascular fibrosis in unexposed and IP/IV egg-exposed mice, using standard protocols [14]. Ten images of each mouse were acquired at 10× magnification and superimposed on a 1700-point grid using ImagePro software. Adventitial regions of interest (ROI) were identified, and points intersecting with collagen (positive picrosirius red polarization) were counted, as compared with all points falling within the ROI, and used to determine the volume fraction of vascular fibrosis (ImagePro).

Mice were killed in CO₂ and then dissected to expose the thoracic cavity. The right ventricle was perfused with PBS. Lungs were inflated using 2% low melting point agarose at 37degC. After inflation, lungs were fixed in 10% neutral buffered formalin for 24 h, then transferred to 70% ethanol in water. After 48 h, the lungs were paraffin embedded and cut into 5 μm sections. The RNA Scope [15] Multiplex Fluorescent V2 Assay (ACD Bio, Newark CA, #323136) protocol for formalin-fixed paraffin-embedded tissue was followed as per the manufacturer's directions. Hybridization probes for Col1a1 (Mm-Cola1a1-C2, #319371-C2) and Tgfb1 (Mm-Tgfb1-C1, #407751) were used. Probes were imaged using the following Opal fluorophore dyes (Akoya Biosciences, Marlborough, MA): Opal 570 (FP1488001KT) for Tgfb1 and Opal 620 (1495001KT) for Col1a1, respectively. After completing the RNA scope protocol, slides were washed in phosphate buffered saline (PBS) and then immediately stained overnight with pre-conjugated Acta2-Alexa488 antibody (Clone 1A4, Invitrogen 53-9760-82) at a 1:200 dilution along with Alexa 750 conjugated hydrazide (Abnova U0261) at a 1:400 dilution in a buffer containing 5% goat serum and 1% bovine serum albumin in PBS. Slides were washed in PBS containing 0.5% Tween-20, stained with DAPI at a 1:500 dilution in PBS, and then mounted in Prolong Gold (Invitrogen P36930). Sections were imaged on a Leica Sp8 inverted confocal microscope with a white light laser system using a 20× glycerol immersion objective. Z-stacks (1 μm sections) at 1024 × 1024 resolution were taken of all observable blood vessels. Images were processed using ImageJ [16], with published images representing a maximal intensity projection of the two best z-stack slices.

Picrosirius red staining [14] and immunostaining (using the reagents listed in Supplementary Table S3) of human lung tissue was performed on FFPE samples of two types of human lung tissue: (1) Schistosoma-PH, obtained at autopsy from patients who died of schistosomiasis-associated PH in Recife, Brazil; and (2) control explanted lung tissue, from unsuccessful lung donors collected by the Pulmonary Hypertension Breakthrough Initiative in the US. The UCSF IRB approved all human studies.

Prism (v9, GraphPad, San Diego, CA) was used for statistical analyses and graphs. Differences between two groups were assessed by t-test; for ≥3 groups, differences were assessed by ANOVA followed by post-hoc Tukey test. Non-normally distributed data or groups with samples <5 were analyzed by non-parametric analysis. P values less than 0.05 were considered statistically significant.

To understand the long-term impact of Schistosoma exposure on PH phenotypes, we performed a time course analysis of a well-established experimental mouse model of Schistosoma PH, which uses intraperitoneal (IP) egg sensitization followed by intravenous (IV) egg challenge [4,10], by analyzing groups of mice at different times following the IV egg challenge (Figure 1A). Compared with either totally unexposed mice (Group A1), sensitized but unchallenged mice at either 14 days (Group A2) or 28 days (Group A3) did not develop a PH phenotype. Following IV challenge, we observed significantly higher right ventricle systolic pressure (RVSP), medial remodeling, and right ventricle (RV) hypertrophy in sensitized mice that were challenged with a single intravenous (IV) dose of Schistosoma eggs at +7 days after the IV eggs were given (Group B; Figure 1B,C and Supplementary Figure S1A), particularly in comparison to the unexposed mice (Group A1), but also in contrast with the IP challenged mice (Groups A2 and A3). These phenotypes were maintained through +14 days after the intravenous Schistosoma challenge (Group C). Thereafter, the pathologic phenotype substantially, but not entirely, resolved at days +21 and +28 after intravenous Schistosoma exposure (Groups D and E, respectively). Even at +28 days, there remained a modest increase in medial thickness and RV hypertrophy compared with the unchallenged mice.

We then investigated the timecourse of Th2 inflammation. We observed maximal IL-4 and IL-13 cytokine levels, and peri-egg granuloma volumes, at the +7 day timepoint after the Schistosoma IV challenge, which progressively declined over the next 21 days (Figure 1D–F). By day +28, the cytokine levels had returned to approximately baseline values. These data are consistent with the concept that host inflammation leads to vascular remodeling, and that resolution of inflammation begins prior to the resolution of vascular remodeling.

We did not observe any difference in left ventricular pressures, heart rate or animal body weight in the Schistosoma exposed mice analyzed at the various timepoints (Supplementary Figure S1B–E). Because the mice exposed to IP eggs alone (Group A2) had essentially the same phenotype as unexposed mice (Group A1), we used unexposed mice for the control mice for the remainder of our experiments.

To investigate the impact of repeated exposure to Schistosoma, and to better model the repeated and chronic infection in humans, we challenged previously sensitized mice with three consecutive doses of IV Schistosoma eggs every 2 weeks, as compared with mice that were sensitized and then challenged with a single dose of Schistosoma. We observed at +7 days after the last IV egg challenge, the group that received 3 IV doses (i.e., Group F, Figure 2A) had a higher RVSP and greater vascular remodeling as compared with the group that had received just 1 IV dose (as in Group B above; Figure 2B,C). There was no difference in RV hypertrophy (Supplementary Figure S2A). Overall, these data indicate the degree of peak PH severity was mildly greater after multiple Schistosoma exposures.

We then investigated whether the PH pathology resolved in the chronically challenged mice, studying the +21 day timepoint after the final IV egg dose in mice that received 3 IV egg challenges (termed Group G, Figure 2D). A comparison was made to the same timepoint in mice challenged just once, as the PH phenotype had largely resolved by this time in the mice that received a single IV egg dose (i.e., Group D in Figure 1B). We found the increased RVSP and media remodeling now persisted in mice challenged with three doses (Figure 2E,F, and Supplementary Figure S2B and C).

We did not observe any difference in left ventricular pressures, heart rate, or animal body weights between the single versus thrice challenged mice at either the +7 or +21 day timepoint (Supplementary Figure S3).

We then explored potential mechanisms underlying the persistent PH phenotype observed in the thrice IV challenged as compared with the single IV challenged mice, interrogating for differences in the pulmonary vasculature that might have been induced by the Th2 inflammation and TGF-β signaling that occurs in Schistosoma-PH [10,17,18]. Clinical and experimental pulmonary fibrosis can be complicated by PH [19,20], and peri-vascular fibrosis is a pathological feature of PH [21]. Thus, fibrosis of the pulmonary vasculature is a plausible contributor to the persistent PH observed after multiple Schistosoma exposures.

We analyzed the density of collagen-producing cells, such as fibroblasts by immunostaining around and within the remodeled vessels, using an anti-HSP47 antibody; HSP-47 is a chaperone protein required for collagen protein synthesis [22]. We found a significant increase in the number of collagen-producing cells in the adventitial space, around the vessels, in the thrice challenged mice compared with the singly challenged mice analyzed 21 days after the final egg dose (e.g., Group D vs. Group G; Figure 3A,B). We did not find a difference in the density of collagen-producing cells within the combined media and intima compartments of the pulmonary vasculature (Figure 3C).

To determine if there was a potential functional difference in extracellular matrix deposition resulting from a greater number of collagen-producing cells, we studied the perivascular collagen protein density using picrosirius staining, analyzed using cross-polarization, which provides spatial quantification of Type 1 collagen specifically [14]. We observed that the collagen content in the perivascular space of the thrice challenged mice was significantly greater than in those challenged just once (Figure 3D,E). Significant fibrosis was not observed in the combined media and intima compartments. Following a single Schistosoma egg challenge, we observed no difference in the perivascular collagen deposition analyzed at different timepoints (Supplementary Figure S4).

Pulmonary fibrosis can be triggered by Th2 inflammation and TGF-β [23,24]. To assess the possible Th2-regulation of collagen in the Schistosoma-PH model, we analyzed a previously published whole lung mRNA dataset from our group (GSE49116), which revealed an increase in Col1a1 expression in sensitized and singly IV challenged mice, at a timepoint 7 days after the IV eggs were administered, as compared to unexposed mice (equivalent of Groups A1 and B; Supplementary Figure S5). The increased Col1a1 observed in Schistosoma-exposed wild-type mice in this dataset was suppressed in Schistosoma-exposed Il4^−/−Il13^−/− mice, which have blocked Th2 inflammation, implicating Th2 inflammation as a potential driver of the vascular fibrosis in Schistosoma-PH.

Having identified a difference in vascular fibrosis in the repetitively challenged mice after 21 d, we next sought to characterize the degree of inflammation at the same timepoint. We first looked at the +7 day timepoint after the final dose of IV eggs (Group B versus Group F), at a time of peak pathology. We found that the estimated peri-egg granuloma volumes, as a readout of overall inflammation, were not different between the two groups (Figure 4A and Supplementary Figure S2D), nor were there differences in IL-4 and IL-13 in whole lung lysates as measured by ELISA (Supplementary Figure S2E and F, and Supplementary Figure S6A and B). We then analyzed the thrice and singly challenged mice 21 days after the final egg dose (Groups D and G). Again, we found no difference in peri-egg granuloma volume (Figure 4B and Supplementary Figure S2D), nor a difference in IL-4 or IL-13 concentrations between the two groups (Supplementary Figure S2E and F, and Supplementary Figure S6C and D).

Suspecting that repeated exposure to the same antigen could induce fatigue of the immune system, we interrogated the expression of the T cell inhibitory receptor PD1 and its ligands PDL1 and 2 by RT-PCR of lung lysates. We found increased expression of PD1 (expressed by Pdcd1) and PDL1 (Cd274), but lower expression of PDL2 (Pdcd1lg2) (Figure 4C–E). Overall, this is consistent with a phenotype of T-cell exhaustion in the repetitively challenged mice.

We also looked for changes in T-cell phenotype or complement as potential mechanisms underlying the persistent PH by analyzing mRNA expression. We found no differences in Foxp3, Il17a, or Il17f between the two groups suggesting unchanged Th17 and Treg CD4 T cells (Supplementary Figure S7A–C). We also found no change in the mRNA expression of complement members C3, C5, or C1q or protein level of C3a between the two groups (Supplementary Figure S7D–G).

Acutely, Th2 inflammation triggers TGF-β activation to cause Schistosoma-PH [4], but with quelling of the Th2 inflammation it could be that TGF-β signaling could also be now suppressed. We found no differences in active or total TGF-β protein or the mRNA expression of prototypical TGF-β targets Pai1 or Smad7 in whole lung lysates, or phospho-Smad2/3 protein by immunostaining (Group D vs. Group G, Supplementary Figure S8).

We sought to identify additional inflammatory proteins which could be different between the singly and thrice challenged mice at this late timepoint by using a protein array for mouse cytokines on the lung lysates. Seven additional proteins were consistently seen in all samples (Supplementary Figure S9A): CXCL12, CCL5, IL-1Ra, IL-16, sICAM-1, TIMP-1, and C5/C5a. However, none of these proteins differed between the two groups (Supplementary Figure S9B).

We also sought to identify if there might be a difference in macrophage density between the singly and thrice challenge mice at the 21 day timepoint by immunostaining for CD68. We did not observe any difference in macrophage density in either the adventitial space or the intima plus media space (Supplementary Figure S10).

Together, these data suggest that while Th2 inflammation and TGF-β signaling are critical to early disease pathogenesis in Schistosoma-PH, these drivers may no longer be necessary for the persistent PH accompanied by vascular fibrosis observed at later timepoints.

Hypothesizing that PD-1/PD-L1 signaling causing T-cell exhaustion may impact the persistent PH phenotype, we attempted to target PD-1 using a neutralizing antibody approach. However, this treatment did not modify the persistent PH phenotype (Supplementary Figure S11).

To additionally validate the observation of increased peri-vascular fibrosis, we used in situ RNA hybridization (RNAScope). We began with probing tissue from single IV challenged mice at the +7 day timepoint, as compared with unexposed mice (Group A1 and Group B). We observed that the Schistosoma-challenged mice had evidence of increased TGF-β1 mRNA expression, consistent with the known mechanism of Th2-driven inflammation causing TGF-β signaling [4], as well as increased expression of Col1a1 mRNA (Figure 5A).

We then probed the thrice IV challenged versus single IV challenged mice, both at 21 d after the final IV eggs were given (Group D and Group G). We observed that the expression of Col1a1 mRNA was prominent in the pulmonary vascular compartment in both groups, but was particularly more prominent in the perivascular space in the thrice challenged mice (Figure 5B,C). The expression of TGF-β1 appeared similar in the two groups, corroborating the absence of a difference in TGF-β signaling observed above.

To determine if the perivascular fibrosis we observed in multiply challenged mice may also be relevant to the chronic disease in human Schistosoma-PH, we analyzed perivascular fibrosis in four lung specimens collected at autopsy from patients that died of this condition in Brazil, also using picrosirius red staining. We observed that 2 of the 4 specimens had substantial perivascular fibrosis, particularly around pulmonary vascular lesions and in the vicinity of schistosomal pigment (which is frequently observed near pulmonary vascular lesions [25]), and to a much greater extent than was observed in three control specimens (Figure 6). Demographic data were unavailable for the four Schistosoma-PH specimens; brief demographic data for the three control specimens are listed in Supplementary Table S4.

We also interrogated for a potential change in fibroblast density in human Schistosoma-PH by immunostaining for vimentin, a marker of mesenchymal cells including fibroblasts. As compared with control human tissue (Supplementary Figures S12A–C), we observed an apparent increase in mesenchymal cell density, particularly in the media vascular compartment in Schistosoma-PH specimens (Supplementary Figures S12D–E). We also observed significant vimentin staining in plexiform lesions in the Schistosoma-PH tissue (Supplementary Figure S12F), as previously reported in idiopathic PH [26].

Here, we observed that a single exposure to intravenous Schistosoma eggs in sensitized mice induces a PH phenotype that is substantially self-resolving. In contrast, repeated exposures cause a persistent PH phenotype (Figure 7). This result suggests a progression to persistent disease as is observed in humans with chronic and repeated Schistosoma exposure as occurs in endemic settings. We observed the persistent disease correlated with the emergence of increased density of perivascular fibroblasts and fibrosis, even at a timepoint when the inflammation was largely resolved. Perivascular fibrosis is a pathologic feature that we also found in lung specimens from humans who died of this condition, although not universally so.

These data help interpret a discrepancy previously observed between the murine Schistosoma-PH model and human Schistosoma-PH, namely the treatment effect of antihelminthics such as praziquantel. Mice experimentally exposed to Schistosoma can model aspects of the human disease, but there are key limitations, including the absence of any schistosomal liver disease, and a relatively short time course. In mice infected with Schistosoma for a relatively short duration, praziquantel causes reversal of the PH phenotype [27]. In mice, it has been observed that the severity of the lung remodeling depends upon the duration and the number of eggs deposited into the lungs [28,29]. In contrast, in humans chronically infected with Schistosoma, the PH phenotype is not significantly altered with praziquantel treatment [5–7]. We suspect the timing is important: in humans, the PH condition develops after years or even decades of chronic infection [30], whereas mice develop Schistosoma-induced PH in weeks to months [1,28]. It should be emphasized that diagnosing and treating schistosomiasis is still important in patients with Schistosoma-PH, as eradicating the parasite may slow the rate of PH progression, and likely helps reverse Schistosoma-induced fibrosis in the liver [31].

Schistosoma egg deposition can lead to tissue fibrosis [32]. Adventitial fibrosis has been observed in other clinical forms and experimental models of PH [33–35], and thus appears to be shared vascular pathology that fundamentally contributes to PH pathobiology. In particular, the adventitial fibroblast has been found to contribute to pulmonary vascular pathology through the secretion of extracellular matrix and cytokines that cause activation of other vascular cells [34]. We observed an increase in the density of collagen-producing cells around the vessels, likely but not definitively fibroblasts, as marked by HSP47 protein and Col1a1 mRNA expression. Increased stiffness of the pulmonary vasculature is clinically observed, along with increased pulmonary vascular resistance as a sign of vascular remodeling in PH [36]. PH can also be a complication of parenchymal lung fibrotic diseases, such as idiopathic pulmonary fibrosis [20], and experimental lung fibrosis can cause PH in animal models [37–39].

As a strong driver of Th2 inflammation, schistosomiasis infection has been observed to result in fibrosis around deposited eggs in the liver, which is dependent on IL-13 but independent of TGF-β [40]. In contrast, we have previously found the pulmonary vascular phenotype to be dependent on both IL-4/IL-13 and TGF-β [3,4]. We observed that collagen transcript expression is regulated in an IL-4/IL-13-dependent manner, as was described in the liver.

We observed an increase in whole lung PD1/PDL1 expression, suggestive of T cell exhaustion, a phenotype reasonably expected to occur following repeated antigenic challenges. As we have previously found CD4 Th2 T cells to be critical to the pathogenesis of Schistosoma-PH, we wondered if T cells may be involved in the resolution of Schistosoma-PH, a role that could be impaired by an exhausted phenotype. However, blocking PDL1 by administering a neutralizing antibody did not modify the persistent PH phenotype, suggesting that T-cell exhaustion while present may not directly contribute to PH persistence or resolution after repeated challenges.

We have previously observed evidence of persistently increased IL-4 and IL-13 in Schistosoma-PH human autopsy tissue, including increased IL-13, IL-4Rα, and phosphorylated STAT6. However, other Type 2 inflammation receptors and cytokines were unchanged, including IL-13Rα1, IL-13Rα2, IL-4, and RELM-α [3,4,10]. Additionally, the plasma concentration of IL-13 was not elevated in subjects with Schistosoma-PH [41]. We also observed increased TSP-1 and TGF-β signaling in Schistosoma-PH autopsy tissue [3,4,10], and patients with Schistosoma-PH have increased TGF-β in their peripheral blood [41]. Overall, this suggests Type 2 inflammation and TGF-beta signaling may persist in human subjects with Schistosoma-PH, which is not fully recapitulated in the mouse model we are using. Furthermore, there may be other immune pathways now activated that are autonomously contributing to a persistent vascular remodeling phenotype.

There are several limitations of the present work. We do not have primary demographic data of the Schistosoma-PH human specimens, making it unclear how comparable they are to the control specimens. It remains unclear whether perivascular fibrosis is truly causal of the persistent PH or an epiphenomenon. To test this mechanistically, the role of fibrosis could be tested using antifibrotic small molecule inhibitors, or interrogating candidate drivers of fibrosis such as IL-4/IL-13 signaling, such as by deletion of STAT6 in fibroblasts specifically. It is also possible that the murine PH phenotype could resolve at later timepoints, although if so, it is likely that adding additional intravenous challenges to further approximate the human condition of chronic and recurrent exposure would result in a progressively more persistent phenotype.

In conclusion, we observed that repeated Schistosoma exposure leads to a persistent PH phenotype associated with perivascular fibrosis. The development of perivascular fibrosis may contribute to the persistent nature of Schistosoma-PH in some humans with this disease.

• Schistosomiasis is a common cause of pulmonary hypertension worldwide, but why the disease persists despite antihelminthic therapy is unknown.

• We found that exposing mice to schistosomiasis causes pulmonary hypertension, which became persistent after repeated exposures, and was associated with the development of fibrosis around the lung blood vessels.

• We also observed perivascular fibrosis in human lung tissue from individuals that died of this condition.

• These results indicate that repeated schistosomiasis exposure can contribute to the development of persistent pulmonary hypertension, which may benefit from targeting the fibrosis that develops around the lung vasculature.

The data that support the findings of this study are available from the corresponding author upon reasonable request.

The authors declare that there are no competing interests associated with the manuscript.

Grant funding was provided by American Heart Association [grant number 19CDA34730030]; ATS Foundation/Pulmonary Hypertension Association Research Fellowship; the Cardiovascular Medical Research Fund and United Therapeutics Jenesis Innovative Research Award (to R.K.); the UCSF Nina Ireland Program in Lung Health [grant number W81XWH2210457] from the US Department of Defense, and Fulbright Scholarship [grant number 12398-BR (to M.H.L.)]; and NIH [grant numbers R01HL135872 (to B.B.G.); P01HL152961 (to B.B.G. and R.M.T.); K08HL168310 (to M.H.L.); R01HL128734 and R01HL158868 (to E.S.); and R01HL135872-06S1 (to D.F.B.)].

Open access for this article was enabled through a transformative open access agreement between Portland Press and the University of California.

Rahul Kumar: Conceptualization, Data curation, Formal analysis, Funding acquisition, Investigation, Methodology, Writing—original draft. Michael H. Lee: Data curation, Formal analysis, Investigation. Biruk Kassa: Data curation, Investigation. Dara C. Fonseca Balladares: Data curation, Investigation. Claudia Mickael: Data curation, Investigation. Linda Sanders: Data curation. Adam Andruska: Data curation, Formal analysis. Maya Kumar: Data curation, Formal analysis. Edda Spiekerkoetter: Data curation, Formal analysis. Angela Bandeira: Data curation. Kurt R. Stenmark: Formal analysis. Rubin M. Tuder: Data curation. Brian B Graham: Conceptualization, Resources, Data curation, Formal analysis, Supervision, Funding acquisition, Investigation, Methodology, Writing—original draft.

Schistosome-infected mice were provided by the NIAID Schistosomiasis Resource Center at the Biomedical Research Institute (Rockville, MD) through NIH-NIAID Contract HHSN272201700014I for distribution through BEI Resources.

CCR2

C-C chemokine receptor type 2

CD4

cluster of differentiation 4

IL-13

interleukin 13

IL4

interleukin 4

PD-1

programmed cell death protein 1

PDL2

programmed cell death ligand 2

PDL1

programmed death ligand 1

PH

pulmonary hypertension

RV

right ventricle

RVSP

right ventricular systolic pressure

TGF-β

transforming growth factor-β

Th2

T helper 2

TSP-1

thrombospondin-1