T. gondii infection induces IL-1R dependent chronic cachexia and perivascular fibrosis in the liver and skeletal muscle

Cachexia is a progressive muscle wasting disease that contributes to death in a wide range of chronic diseases. Currently, the cachexia field lacks animal models that recapitulate the long-term kinetics of clinical disease, which would provide insight into the pathophysiology of chronic cachexia and a tool to test therapeutics for disease reversal. Toxoplasma gondii (T. gondii) is a protozoan parasite that uses conserved mechanisms to infect rodents and human hosts. Infection is lifelong and has been associated with chronic weight loss and muscle atrophy in mice. We have recently shown that T. gondii-induced muscle atrophy meets the clinical definition of cachexia. Here, the longevity of the T. gondii-induced chronic cachexia model revealed that cachectic mice develop perivascular fibrosis in major metabolic organs, including the adipose tissue, skeletal muscle, and liver by 9 weeks post-infection. Development of cachexia, as well as liver and skeletal muscle fibrosis, is dependent on intact signaling through the type I IL-1R receptor. IL-1α is sufficient to activate cultured fibroblasts and primary hepatic stellate cells (myofibroblast precursors in the liver) in vitro, and IL-1α is elevated in the sera and liver of cachectic, suggesting a mechanism by which chronic IL-1R signaling could be leading to cachexia-associated fibrosis.


T. gondii infection leads to sustained cachexia in mice.
Oral infection with T. gondii causes severe intestinal inflammation, leaky gut, and commensal-induced inflammation during the first 2 weeks of infection 16,19,20 . We previously demonstrated that mice orally infected with T. gondii develop chronic cachexia that was sustained even after intestinal inflammation resolved 13 . We have also shown that T. gondii-induced cachexia occurs when the gastrointestinal tract is bypassed entirely using intraperitoneal infection 14 . To understand the molecular mechanisms controlling chronic cachexia, 10-14 week old male C57BL/6J mice were intraperitoneally injected with 10 Type II T. gondii (Me49 strain) bradyzoite tissue cysts. T. gondii-infected mice lost 15-20% of their initial body mass during the first 3 weeks of infection (Fig. 1a,b), and remained significantly wasted compared to uninfected controls for up to 21 weeks (Fig. 1b). Chronic infection was confirmed by counting cysts in brain homogenate at 9 weeks post-infection (Fig. 1c). Infected female mice exhibited similar kinetics of weight loss during infection with T. gondii (Supp. Figure 1a). Although infected mice went through a period of acute anorexia 8-10 days post-infection, they regained eating relative to uninfected controls, indicating that sustained weight loss was not due to prolonged anorexia (Fig. 1d). Growth stunting or developmental malnutrition is a distinct disease from cachexia, so to avoid inducing this biology, 10-14 week old, mature adult mice were used in all experiments. Infected mice trended towards eating more chow per 24-h period than uninfected mice, although this was not significant (Fig. 1d,e, left). Bomb calorimetry on fecal pellets confirmed that infected mice were absorbing a similar number of calories as uninfected mice (Fig. 1e, right).
Loss of lean muscle mass is the primary diagnostic marker of cachexia. By 2 weeks post-infection, infected mice had a significant reduction in both fat (Fig. 1f, left) and lean (Fig. 1f, right) body mass by EchoMRI Whole Body Composition Analysis. Importantly, while fat mass partially recovered by 6 weeks post-infection, lean body mass wasting progressed from 2 to 6 weeks post-infection (Fig. 1f). When individual tissues were dissected and weighed, we found that inguinal subcutaneous white adipose tissue (scWAT) and epigonadal visceral white adipose tissue (vWAT) were significantly reduced at 2 weeks post-infection. Acute hepatomegaly was also observed, consistent with published reports of acute T. gondii infection in the liver (Fig. 1g) 21 . By 9 weeks post-infection, scWAT had recovered in mass but vWAT, quadriceps muscle (quad), and liver were significantly reduced in size relative to uninfected littermate controls (Fig. 1g). Gastrocnemius (GA), tibialis anterior (TA), and extensor digitorus longum (EDL) muscles were also weighed. The TA and EDL were significantly smaller in infected mice relative to uninfected; the GA from infected mice trended smaller, although this was not significant due to variability in GA mass (Supp. Figure 1b). T. gondii burden was measured in the tibialis anterior at 10 weeks post-infection by qPCR of parasite DNA (Fig. 1h). Consistent with previous reports that parasite burden is low in skeletal muscle, Toxoplasma genomic DNA was detected in 3 out of 5 infected samples, suggesting that while there may be a chronic parasite presence in muscle tissue, this is not sufficient to explain such widespread muscle wasting 22 . Elevated circulating inflammatory cytokines are a hallmark of cachexia that contribute to disease pathology 23 . At 1 week post-infection, when T. gondii infection is systemic, mice had elevated circulating IFN-γ, IL-1β, IL-6, and TNF-α compared to uninfected controls (Fig. 1i, grey background). Although TNF-α, IL-6, and IFN-γ were reduced by 5 weeks post-infection, they were still significantly elevated relative to uninfected controls (Fig. 1i, 25 . Additionally, we did not observe the morphological changes in vWAT histology associated with fat browning (Supp. Figure 2d) 28 . These data indicate that non-shivering thermogenesis or fat browning are likely not the central drivers of sustained cachexia in T. gondii infection. Consistent with other cachexia models, T. gondii-infected mice had small but significant reductions in fasting and non-fasting blood glucose levels (Fig. 2c,d) 29,30 . However, no significant differences in serum insulin levels were observed (Fig. 2e), and glucose clearance in response to a bolus of insulin was similar in cachectic mice and uninfected mice (Fig. 2f), suggesting that insulin resistance is not the primary driver of metabolic dysfunction during T. gondii-induced chronic cachexia.
To assess systemic metabolic function during T. gondii-induced cachexia, mice were individually housed in Comprehensive Laboratory Animal Monitoring System (CLAMS) metabolic cages. Cachectic mice had significantly reduced nighttime activity compared to uninfected mice (Fig. 3a), as well as decreased calculated heat production (Fig. 3b), consistent with cachexia-associated fatigue. However, the reduced activity confounds our ability conclude that the reduced respiratory exchange ratio (RER) observed in cachectic mica was due to a shift towards beta-oxidative rather than glycolytic metabolism (Fig. 3c-e). To better address this question, we next measured levels of key lipolytic and metabolic signaling enzymes in vWAT, liver, and muscle by western blot. Although there was some animal-to-animal variability, no consistent differences were observed between uninfected and cachectic mice in the levels of hormone sensitive lipase (HSL), phospho-HSL (Ser660), AKT, phospho-Akt (Ser473), phospho-ACC, adipose triglyceride lipase (ATGL), phospho-ATGL, and perilipin in the WAT (Supp. Fig. 3a,b), muscle, or liver (Supp. Fig. 3c). Atglistatin, a pharmacological inhibitor of ATGL, has been shown to block muscle wasting in acute murine cancer cachexia 31 ; however, it inhibits T. gondii growth 32 and could not be used to confirm the conclusion that increased lipolysis was not the major driver of chronic T. gondii-induced cachexia. Together, these data are consistent with the observation that scWAT weight rebounds and vWAT weight stabilizes during chronic cachexia (Fig. 1f,g), clinical observations that cachexia can co-occur with obesity, and the observation that not all cachectic patients present with adipose tissue loss 1 .
T. gondii-induced cachexia is associated perivascular fibrosis in metabolic organs. In the course of probing protein expression by western blot, we noticed that the β-actin loading control was significantly upregulated in infected mice relative to uninfected controls in all the tissues assessed (Supp. Fig. 3a-c). β-Actin has over 90% sequence homology with alpha smooth muscle actin (α-SMA), and most commercially  Error bars are standard error of the mean *P < 0.05; **P < 0.01; ***P < 0.001, by unpaired Student's t test (mean data) or *P < 0.05; # P < 0.01; † P < 0.001, § P < 0.0001 (traces).
To determine if tissues with elevated α-SMA were fibrotic, tissue sections were stained with Picrosirius Red, an anionic dye that labels elongated collagen fibers. Significantly more perivascular collagen was observed in the liver (Fig. 4d), muscle (Fig. 4e), and vWAT of cachectic mice relative to uninfected controls (Fig. 4f). Increased levels of collagen I and collagen III in the liver in proximity to α-SMA-expressing cells was confirmed by immunofluorescence staining (Fig. 4g). These data indicate that perivascular fibrosis in major metabolic tissues occurs during T. gondii infection-induced chronic cachexia.
IL-1α and IL-1R are expressed in the fibrotic liver microenvironment. Transforming growth factor beta (TGF-β) is a canonical inducer of tissue remodeling and fibrosis, so we hypothesized that TGF-β would be elevated in fibrotic tissues. As expected, TGF-β was significantly increased in liver lysates ( Fig. 5a) and was trending higher in muscle and vWAT lysates (Supp. Fig. 5a,b). We also evaluated expression of IL-6, TNF-α, IFN-γ, IL-1α, and IL-1β because these inflammatory cytokine are frequently observed in cachexia as well as fibrotic diseases 37 . We found that IL-1α was significantly higher in liver lysates (Fig. 5a) and the serum (Fig. 5b) of cachectic mice at 9 weeks post-infection relative to uninfected controls. Of note, IL-6, TNF-α, and IFN-γ, which are also elevated in sera during chronic T. gondii infection (Fig. 1h), play a well-established and essential role in controlling chronic T. gondii burden. The role of IL-1α/IL-1R axis in T. gondii infection is comparatively understudied. IL-1 has been implicated in the development of liver fibrosis [38][39][40] ; and altered liver biology, which is central to systemic metabolism, has been observed in a number of experimental and clinical cachexia models [41][42][43][44] . To determine if IL-1α was localized to areas of fibrosis, liver sections were stained with an IL-1αspecific antibody for immunofluorescence assays. Cells expressing IL-1α were observed within collagen I-rich perivascular fibrotic lesions, but the majority of IL-1α positive cells did not co-stain with immune cell marker CD45, suggesting that these were likely liver-resident cells (Fig. 5c). IL-1α signals through the type I IL-1 receptor (IL-1R). IL-1R staining was also observed within perivascular fibrotic lesions on a subset of α-SMA positive cells with fibroblast morphology (Fig. 5d), suggesting that α-SMA positive cells could directly respond to locallyreleased IL-1α in the liver.

IL-1 is sufficient to stimulate myofibroblast contractility in vitro.
To determine if IL-1α was sufficient to promote myofibroblast activation in vitro, murine embryonic fibroblasts (MEFs) were treated with media, IL-1α, IL-1β, or TGF-β (positive control) for 48 h then stained for immunofluorescent imaging (Fig. 6a). Compared to media alone, IL-1α treatment induced α-SMA expression and increased cell surface area (a measure of increased cell contractility) to a similar level as TGF-β (Fig. 6b,c) although these were not significant when accounting for multiple comparisons analysis 45 . IL-1β treatment also increased cell spreading. By contrast, neither IL-1α or IL-1β promoted MEF proliferation or survival under serum starvation conditions (Supp. Fig. 6a).
To determine if IL-1α was sufficient to induce myofibroblast differentiation in primary cells, we took advantage of a recent protocol to isolate primary hepatic stellate cells (HSCs), the major liver myofibroblast precursor cell type in the liver, based on the endogenous autofluorescence of vitamin A-containing granules (Supp. Fig. 6b) 46 . Mechanosensing of rigid tissue culture plastic can spontaneously trigger HSC activation 47 , potentially masking an activating effect of IL-1α, so HSCs were plated on a 4kPA hydrogel to approximate normal liver rigidity 48 . IL-1α or TGF-β stimulation for 48 h led to a trend in increased cell area (Fig. 6d,e left panel) and intracellular α-SMA staining (Fig. 6f) relative to HSCs treated with media alone when single cell measurements were averaged within a biological (mouse) replicate. However, liver cell functions are specialized by lobe, zone, and proximity to arteriolar or venous blood, indicating that heterogeneity in HSC responsiveness may be expected. When single cell data was pooled across replicates, we found that response to IL-1α and TGF-β was bi-modal and that there was an increase in cell area and α-SMA staining, suggesting that at least some of the primary HSCs could be activated in response to IL-1α (Fig. 6e,f, right panels). HSCs stimulated with IL-1β for 24 h also had elevated cell area, and trended towards an increase in α-SMA which was not significant over this shorter time frame (Supp. Fig. 6c). Taken together, these data indicate that IL-1α is sufficient to promote myofibroblast activation in vitro.
In comparison to WT mice, which had fibrotic tissues at 9 weeks post-infection, infected IL-1R −/− mice had levels of liver α-SMA that were substantially lower than infected wildtype mice, and muscle α-SMA that was comparable to uninfected controls (Fig. 8a,b). Infected IL-1R −/− vWAT, which was wasted in chronic infection (Fig. 7e), had more α-SMA than uninfected controls, but α-SMA levels were slightly lower than infected wildtype mice (Fig. 8c) (full-length blots available in Supp. Fig. 7a-c). When collagen was measured directly, infected IL-1R −/− mice had significantly less picrosirius red collagen staining in the liver (Fig. 8d) and skeletal muscle (Fig. 8e) compared to infected WT mice. Although IL-1α and IL-1β were not significantly increased in skeletal muscle lysate by ELISA (Supp. Fig. 5a), the low levels of circulating IL-1α (Fig. 5b) may be sufficient to promote perivascular fibrosis in the skeletal muscle. A not mutually exclusive possibility is that pockets of IL-1α and/or IL-1β may exist in muscle fibrotic microenvironments which are below the level of detection in whole tissue lysate. We have previously demonstrated that IL-1R antagonist protein levels, which are elevated in response to IL-1R signaling, are significantly elevated in the skeletal muscle of cachectic mice, suggesting that sustained IL-1R signaling may be occurring in the muscle 14 . In the vWAT, picrosirius red staining was similar between www.nature.com/scientificreports/ WT and IL-1R −/− mice at 9 weeks post-infection (Fig. 8f). This was consistent with vWAT wasting and α-SMA upregulation and suggest that IL-1R-independent pathways regulate collagen deposition and/or turnover in this tissue. Together, these data indicate that IL-1R drives chronic cachexia and the associated liver and skeletal muscle fibrosis during chronic T. gondii infection.

Discussion
Here we show that intraperitoneal T. gondii infection is a robust model for chronic cachexia that recapitulates critical aspects of clinical disease. This is supported by several previous reports describing chronic weight loss and muscle dysfunction during T. gondii infection 13,52-55 . A major advantage of the T. gondii cachexia model is its longevity, which opens the door to studying mechanisms of disease reversal and testing therapeutic tools 11,56 . www.nature.com/scientificreports/ One outstanding question for the field is whether chronic infection is necessary to sustain cachexia. Recently, a master regulator of bradyzoite formation, Bradyzoite Factor for Differentiation 1 (BFD1), has been identified. Parasites lacking this gene have similar kinetics of acute infection but do not form cysts in the brain, suggesting that this may be a definitive tool to study the requirement for chronic infection in cachexia 57 . Our data indicate that pathways critical to acute anorexia-cachexia progression (lipolysis and non-shivering thermogenesis) may not be central drivers of chronic cachexia. Moreover, the longevity of our model allowed us to observe fibrosis in the liver, adipose tissue, and skeletal muscle, a process that typically takes many weeks to develop. This finding complements recent studies from the Seelaender group, who has recently reported increased tumor and adipose tissue fibrosis in cachectic gastrointestinal cancer patients relative to age and disease-matched, weight stable cancer patients 58,59 . Skeletal muscle fibrosis has also been reported in the quadriceps of cachectic patients with chronic heart failure 60 , as well as in the rectus abdominus of cachectic patients with pancreatic ductal adenocarcinoma (PDAC) compared to weight stable controls 61 . In the latter study, skeletal muscle collagen levels directly correlated with weight loss and mortality. While only a handful of studies have assessed fibrosis in cachectic patients, there is a strong association between cachexia and fibrotic diseases 2,62 . Emerging clinical data show a clear correlation between muscle wasting, disease progression, and mortality in liver cirrhosis, non-alcoholic fatty liver disease, and hepatocellular carcinoma, the 4th leading world-wide cause of cancer related mortality [63][64][65][66][67][68][69][70][71][72][73] . The liver is a central regulator of nutrient absorption, storage, and metabolism, and as such, liver fibrosis may influence aberrant metabolism in distal tissues. Additionally, fibrosis in the skeletal muscle physically restricts muscle regeneration 74 and fibrosis in other muscles like the heart and diaphragm could lead to dysfunction that is ultimately fatal. Historically, peripheral fibrosis during cachexia may have been overlooked because of the difficulty in obtaining biopsies from these tissues. However, muscle biopsy and assessment of adipose tissue removed during surgery are becoming increasingly more common, which may enable a more thorough understanding of cachexia-associated fibrosis in the future.
Our experiments showed that IL-1R deficient mice are protected from developing cachexia. A historical evaluation of the literature reveals that IL-1R signaling can directly induce muscle wasting; however, many of these studies were conducted before the modern, clinical definition of cachexia was established. Peripheral and central nervous system administration of IL-1 is sufficient to acutely recapitulate aspects of muscle wasting in vivo 75,76 , and in vitro, IL-1 treatment of cultured myotubes induces MuRF1 and atrogin1, the E3 ubiquitin ligases primarily responsible for muscle catabolism 77,78 . Other studies have shown that NF-κB, which is downstream of IL-1R and Toll-like receptor (TLR) signaling, is both necessary 79,80 and sufficient to cause muscle catabolism; and constitutive NF-κB activation in skeletal muscle recapitulates a muscle wasting phenotype 81 . TLR4, which upregulates IL-1, has been shown to control muscle wasting during Lewis lung carcinoma-induced cachexia 82 ; and mice deficient in MyD88, the signaling adaptor downstream of IL-1R and most TLRs, are protected from www.nature.com/scientificreports/ fat and muscle loss, fatigue, and mortality in a pancreatic cancer model of cachexia 83 . Our observation that IL-1R −/− mice are protected from perivascular fibrosis in the liver and muscle during chronic Toxoplasma cachexia are consistent with these data and previous studies showing that pharmacological or genetic blockade of IL-1R signaling in mice attenuates fibrosis in the liver 38,84,85 , heart 86,87 , and lungs 87,88 . Additionally, clinical studies report that SNPs in the IL-1R antagonist gene resulting in elevated IL-1 signaling are associated with an increased risk of developing both respiratory 89,90 and liver fibrosis 91 , suggesting that IL-1R driven fibrosis may be conserved between mice and humans. In our model, low levels of IL-1α are sustained in circulation and IL-1α-expressing cell types were observed within perivascular fibrotic lesions in the liver. This observation, in conjunction with the observation that IL-1R is also expressed within fibrotic regions, suggests that local IL-1α is sufficient to control tissue-wide development of fibrosis in the liver. Although we did not see IL-1α or IL-1β elevated in skeletal muscle lysate, we have previously shown IL-1R antagonist (which is regulated downstream of IL-1R signaling) levels to be elevated in the skeletal muscle, suggesting that IL-1R in the muscle may be responding to circulating IL-1α 14 . As an alarmin, bioactive IL-1α can be released via a broad range of cell damage or death inducers, and future work will be necessary to determine the source and mechanism of release of IL-1α during cachexia. Studies are also underway to identify precise myofibroblast cell or precursor populations that express IL-1R and test whether selectively eliminating or driving IL-1R signaling on these cell types is sufficient to influence fibrosis and cachexia during T. gondii infection. There are several possible explanations for why IL-1R −/− mice are not protected from adipose tissue fibrosis. First, we have previously observed that during acute T. gondii infection, IL-1R −/− mice have more visceral white adipose tissue pathology compared to infected wildtype mice 14 . It is possible that the tissue damage incurred during acute infection is sufficient to induce IL-1R-independent fibrosis. Second, "myofibroblast" is a general term that describes heterogenous populations of cells based on ⍺-SMA expression and collagen production. Lineage tracing experiments have identified diverse myofibroblast precursor cell types, and myofibroblast gene expression and cell signaling is highly dependent on tissue residency 92 . It is possible that other inflammatory signals chronically elevated in Toxoplasma infection induced cachexia drive IL-1R independent myofibroblast activity in the fat. Finally, IL-1⍺ and IL-1β remain elevated in the brain during chronic T. gondii infection. We www.nature.com/scientificreports/ have not ruled out the involvement of central nervous system IL-1 signaling in the regulation of peripheral tissue homeostasis 14 .
Like cachexia, fibrosis has been notoriously difficult to target and reverse in the clinic. Our data suggest that these conditions may be linked, which poses a potential explanation for why cachexia has been so difficult to reverse with nutritional supplementation and anabolic steroids. Additional experiments will be necessary to determine if fibrosis is causal for cachexia and/or limits recovery from acute cachectic weight loss, although a recent study found that experimental models of cirrhosis lead to muscle wasting 93 . The relationship between IL-1R signaling, fibrosis and cachexia is an exciting new area for exploration in both animal models and clinical cachexia.

Materials and methods
infections. To generate cysts, 8-10 week female CBA/J mice were infected with 3-10 Me49 or Me49 stably expressing green fluorescent protein and luciferase (Me49-GFP-luciferase) bradyzoite cysts by intraperitoneal injection. 4-8 weeks following infection, mice were euthanized with CO 2 and brains were harvested, homogenized through a 70 μm filter, washed 3 times in PBS, stained with dolichos biflorus agglutinin conjugated to either FITC or rhodamine (Vector labs) and the number of cysts were determined by counting FITC-positive cysts at 20 × magnification using an EVOS FL imaging system (Thermo Fisher). 10-14 week male mice (unless otherwise noted) were infected with 10 Me49 or Me49-GFP-luciferase bradyzoite cysts by intraperitoneal infection. Prior to infection, mice were cross-housed on dirty bedding for 2 weeks to normalize commensal microbiota. Brain and tibialis anterior DNA was isolated as described 94 , and used at 100 ng DNA per qPCR reaction. For PCR of the 529 bp Repeat Element (RE), the following Taqman primer/probes were used forward: 5′-CAC AGA AGG GAC AGA AGT CGAA-3′; reverse: 5′-CAG TCC TGA TAT CTC TCC TCC AAG A-3′; probe: 5′-CTA CAG ACG CGA TGCC-3′ (IDT) 95 ; Mm02619580_g (ThermoFisher Scientific) was used as the mouse beta actin reference gene.
Mouse strains/husbandry. C57BL/6 and IL-1R −/− mice were purchased from Jackson Laboratories. All mice were bred in-house. All animal protocols were approved by the University of Virginia Institutional Animal Care and Use Committee (protocol # 4107-12-18). All animals were housed and treated in accordance with AAALAC and IACUC guidelines at the University of Virginia Veterinary Service Center.
Sorting and culturing of primary hepatic stellate cells. Primary murine hepatic stellate cells were isolated as previously described 46 . Following sorting on a BD Influx Cell Sorter in the University of Virginia Flow Cytometry Core, cells were seeded onto fibronectin coated 4 kPa polyacrylamide hydrogels (Matrigen) and stimulated with 10 ng/mL recombinant mouse IL-1α (R&D), 10 ng/mL recombinant mouse TGF-β (R&D), or media alone for 48 h, after which hydrogels were fixed with 4% paraformaldehyde and stained with phalloidin-488 (Invitrogen) and anti α-SMA antibody (Invitrogen, clone IA4). Cells were mounted with ProLong Diamond Anti-Fade mountant (Thermo Fisher Scientific), and imaged at room temperature on a Nikon Eclipse Ti microscope with an UltraView VoX imaging system (PerkinElmer) using a Nikon N Apo LWD 40 × water objective (numerical aperture: 1.15) and cell area and α-SMA intensity were determined using Volocity software.
Mouse embryonic fibroblasts. Transformed mouse embryonic fibroblasts were cultured (DMEM, 10% FBS, 1% l-glutamine, 1% penicillin/streptomycin, 1% HEPES, 1% sodium pyruvate) and used between passage 3-10. 1 × 10 4 cells were seeded overnight onto poly-D-lysine coated glass coverslips in 24-well plates and then stimulated with 10 ng/mL recombinant mouse IL-1α (R&D), 10 ng/mL recombinant mouse TGF-β (R&D), or media alone for 48 h. Coverslips were fixed in 4% paraformaldehyde for 10 min, permeabilized with 0.1% Triton X-100 for 15 min, blocked in 1% BSA for 30 min, and then stained overnight at 4 °C with anti α-SMA antibody (Invitrogen, clone IA4). The next day, coverslips were stained for 1 h at room temperature with phalloidin-eFluor660 (eBioscience) or donkey anti-mouse-AF594 (Jackson ImmunoResearch), and mounted onto slides with Vectashield Mounting Medium containing DAPI (Vector Laboratories). Coverslips were imaged on a Zeiss Imager M2 microscope (Carl Zeiss) with an AxioCam Mrm camera (Carl Zeiss) using a 20 × objective (numerical aperture: 0.80) and ZenBlue software (Carl Zeiss). Cell area and α-SMA was determined by defining each cell as a region of interest by tracing each cell in Fiji software and quantifying cell area and α-SMA intensity. For serum starvation experiments, 2500 MEFs/well were seeded into a 96 well plate overnight and then treated for 48 h with cytokine in media containing either 10% or 1% serum. At the end of the 48-h period, cell proliferation and viability were determined using CellTiter-Glo reagent (Promega).
Body temperature measurements. Mice were anesthetized with isoflurane at 10 days post-infection, and subcutaneously injected with temperature micro-transponders (Bio Medic Data Systems, Seaford, DE). Temperature was monitored daily (at the same time of day) using a telemetric reader.
Bomb calorimetry. Mice were individually housed for 24 h at 10 weeks post-infection. Fecal pellets were collected every 2 h during the daytime and flash-frozen. Pellets were lyophilized and then sent to University of Texas Southwestern Metabolic Phenotyping Core for bomb calorimetric analysis. CLAMS metabolic monitoring. At   picrosirius red. Formalin-fixed, paraffin-embedded tissue was stained with picrosirius red, and slides were imaged on an Olympus BX51 microscope with an Infinity 1 camera (Lumenera) for brightfield or a Zeiss Apotome2 (Carl Zeiss, Germany) under polarized light using the 20 × or 40 × objective. 5-10 blinded fields of view were acquired per mouse. To quantify percent area, images were binarized in Fiji 97 , thresholded, and percentage of positive pixels per area was determined.