Schistosoma mansoni infection reprograms the metabolic potential of the myeloid lineage in a mouse model of metabolic syndrome

Despite evidence that helminths protect from metabolic disease, a major gap exists in understanding the underlying mechanism(s). Here we demonstrate that bone marrow derived macrophages (BMDM) from S. mansoni infected male ApoE-/- mice have dramatically increased mitochondrial respiration compared to those from uninfected mice. This change associates with increased glucose and palmitate shuttling into TCA cycle intermediates and decreased accumulation of cellular cholesterol esters. Moreover, systemic metabolic modulation by schistosomes is a function of biological sex, where infection protects ApoE-/- male, but not female, mice from obesity and glucose intolerance. Sex-dependence extends to myeloid cells, where reprogramming leads to opposite cholesterol phenotypes in BMDM from females and males. Finally, the metabolic reprogramming of male myeloid cells is transferrable via bone marrow transplantation to an uninfected host, indicating maintenance of reprogramming in the absence of sustained antigen exposure. This work reveals that S. mansoni systemic reprograming of myeloid metabolism is sex-dependent.

, 4B, 4C). Moreover, side by side OCR analysis in females and males with palmitate 276 as a substrate (glucose limiting conditions) showed that BMDM from infected male, but not from 277 infected females had an increased ability to oxidize exogenous palmitate ( Figure 4D). In 278 addition, BMDM from infected males but not females had significantly increased palmitate basal 279 OCR and palmitate spare respiratory capacity, suggesting exogenous free fatty usage as a carbon 280 source for OXPHOS ( Figure 4E, 4F). Similar to the male only data, we observed no differences 281 in macrophage lipid content in either group, suggesting that global lipolysis may not underlie 282 OCR and spare respiratory capacity in our model ( Figure 4G). Additionally, analysis of 283 mitochondrial mass in females and males showed that similar to blood monocytes, BMDM from 284 infected male mice, but not females have a significantly higher Mitotracker MFI. These data 285 suggest that S. mansoni infection improved mitochondrial biogenesis in males, but not females.  Figure 5A). Further, analysis of total CE (identified as a VIP species in 296 males in Figure 2) showed that infection led to decreased levels of CE in cells derived from 297 infected male mice. In contrast, infection led to significantly increased total CE in BMDM 298 derived from infected female mice ( Figure 5B). Similar to males, infection in females induces a 299 unique lipid signature ( Figure 5C), with two species of plasmanyl-PE, three species of 300 plasmanyl-PC, BMP and CE as the main drivers of the altered lipid profile in females during 301 infection ( Figure 5D). Since we observed BMP's as a VIP compound driven by S. mansoni 302 infection in females, but not males, we went back and performed targeted lipidomics to quantify 303 BMPs at the class level. Indeed, total BMPs were significantly increased by infection in BMDM 304 from females but not males ( Figure 5E). Our Seahorse data indicates that S. mansoni infection 305 induced differential oxidation of exogenous palmitate in BMDM derived from males and 306 females. In order to understand the metabolic flux of exogenous palmitate, we performed 307 metabolic tracing analysis of palmitate, where macrophages were differentiated in the presence 308 of normal glucose and then switched to C 13 -labeled palmitate for 24 hours. Similar to what we 309 documented with glucose, we observed increased shuttling of heavy labeled palmitate into 310 succinate, malate, and fumarate in BMDM derived from infected males, with decreased shuttling 311 into myristate (a marker of fatty acid elongation). BMDM derived from infected females 312 displayed the opposite phenotype, with decreased shuttling of heavy labeled palmitate into 313 succinate, malate, and fumarate, and increased shuttling into myristate compared to BMDM from 314 uninfected controls. These data suggest that BMDM derived from infected male ApoE -/mice 315 have increased basal utilization of exogenous palmitate for beta-oxidation, while the female data 316 suggests increased lipid synthesis, strengthening the data obtained from our palmitate 317 extracellular flux analysis. 318 To further understand the role of biological sex in the S. mansoni induced reprograming of the 319 BMDM lipidome, we isolated cellular lipids from male and female BMDM either unstimulated (media), or following LPS stimulation, and then performed untargeted lipidomics as described in 321 Figure 5. We analyzed the lipidomic data with machine learning using a two-step selection 322 process (see Methods). This approach identified ~20 features (lipid compounds) contributing to 323 an elastic net regression model that can predict the infection status of samples-originating 324 animals (Zou, 2005). These informative features are listed in descending order of importance 325 (Supplemental Figure 2A). Interestingly, some key features (for instance, the top scoring one) are 326 not represented as significant in the univariate tests (univariate adjusted p values, Supplemental 327 Figure 2B). However, these features are important contributors to the multivariate machine 328 learning model. This is not unexpected -some features that would be considered uninformative 329 individually may be very useful in improving predictions if combined with other features. When 330 we examine these features individually, we can see a differential response to infection for males 331 and females (for instance, PG 18:1_18:2, a glycerophospholipid) represented in Supplemental 332 Oiknine and Aviram, 1992), so we quantified the acute inflammatory effectors nitrite and iNOS, 341 along with pro-inflammatory cytokines/chemokines IL-12p70, CXCL1, and IL-6 342 (chemokines/cytokines with known pathogenic roles in obesity, insulin resistance and atherosclerosis) following stimulation with LPS. S. mansoni infection increases LPS induced 344 nitrite and iNOS production in BMDM from male, but not female ApoE -/mice on HFD ( Since we documented significant sex dependent shifts in functional metabolism in BMDM from 354 male and female S. mansoni infected mice, we sought to determine if differential transcriptional 355 modulation underlies these shifts. In order to investigate the genes and respective pathways that 356 were associated with specific conditions and the ones that were differentially regulated by sex we 357 performed mRNAseq on unstimulated BMDM derived from male and female ApoE -/mice at 10-358 weeks post S. mansoni or mock infection. We identified different subsets of genes that were 359 preferentially upregulated in males (p<0.05, Figure 6A). Among 1448 genes upregulated in 360 males regardless of infection status with a p value <0.05, the majority of these (238 genes) were 361 associated to metabolic functions by pathway analysis. Of these 238 genes involved in 362 metabolism we identified hexokinase 1 (hk1), citrate synthase (cs), apolipoprotein A2 (apoa2), 363 aldehyde dehydrogenase 3 family member A2 (aldh3a2), lipoyltransferase (lipt1), solute carrier 364 family 19 member 1 (slc19a1), LDL receptor related protein 1 (lrp1), many of these are involved 365 in lipoprotein and cholesterol metabolism. These data suggest that myeloid metabolism is differentially regulated by sex, at least in the context of HFD. In addition, we found 216 genes 367 involved in immunity. Among these genes with immune function we identified interleukin 10 (il-368 10), Toll like receptor 5 (tlr5), NLR family pyrin domain containing 3 (nlrp3), inducible T cell 369 costimulatory (icos), which have diverse pro and anti-inflammatory function in the immune 370 system. Next, we surveyed the genes that are differentially regulated in males and females 371 following S. mansoni infection. We found 66 genes involved in metabolism, hemostasis, the 372 adaptive immune system, collagen degradation and not annotated to specific pathways. 373 Following the genes with known function, the majority (10) of differentially regulated genes 374 have documented roles in metabolism. Among these, we identified type II iodothyronine 375 deiodinase (dio2), which has been implicated in the regulation of diet induced obesity 376 (Kurylowicz et al., 2015;Vernia et al., 2013). Moreover, we found that hexokinase 3 (hk3), fatty 377 acid binding protein 4 (fabp4), sphingomyelin synthase 2 (sgms2), solute carrier family 6 378 member 8 (slc6a8) were all upregulated in male and downregulated in female BMDM following 379 S. mansoni infection ( Figure 6B). In addition, we performed gene ontology analysis from the 66 380 genes that were differentially regulated in females and males. The most significant pathways 381 were response to glucocorticoids, glycoprotein metabolism, and fatty acid metabolism, again 382 suggesting that there is a sex-specific metabolic response to S. mansoni infection in myeloid 383 cells. These pathways may help explain the differential metabolic modulation induced by 384 schistosomiasis in males and females ( Figure 6C). 385 386

absence of antigen 388
Our metabolic and transcriptomic data from BMDM differentiated in vitro from male S. mansoni 389 infected mice suggested that metabolic modulation may be long-lived in the absence of ongoing 390 antigen exposure. In order to determine durable nature of modulation we transferred bone 391 marrow from either 10-week S. mansoni infected male ApoE -/mice on HFD, or uninfected male 392 controls into busulfan treated recipient ApoE -/mice on HFD. At 10 weeks post-bone marrow 393 transfer we assayed glucose tolerance via an i.p. GTT. Mice that received bone marrow from 394 infected males have a significantly lower glucose area under the curve (AUC) than those that 395 received control bone marrow. We harvested bone marrow from all recipients and differentiated 396 BMDM in M-CSF for 6 days and then performed real-time extracellular flux analysis. BMDM 397 from recipients of bone marrow from S. mansoni infected mice had significantly higher basal 398 oxygen consumption and a trend towards increased spare respiratory capacity as compared to 399 BMDM generated from recipients of uninfected control bone marrow. Additionally, BMDM 400 from recipients of bone marrow from S. mansoni infected mice had a significantly higher 401 Mitotracker MFI than BMDM from recipients of control bone marrow. These data suggest that S. 402 mansoni induced metabolic modulation of the myeloid lineage in males is long-lived even in the 403 absence of ongoing exposure to egg antigens, and that hematopoietic cells are at least partially 404 responsible for the regulation of whole-body glucose metabolism in the HFD ApoE -/model. 405

Discussion 406
Helminth infections in general, and schistosomiasis in specific, have been known to be 407 inversely correlated with obesity and glucose intolerance for over a decade, a phenomenon 408 thought to be associated with Type 2 polarization of macrophages and T cells. In the current 409 study we report that S. mansoni infection induces dramatic metabolic alterations in BMDM from 410 male ApoE -/mice on HFD. Our results indicate that macrophages derived from the bone marrow of infected male mice have increased basal oxygen consumption and spare respiratory capacity 412 compared to those derived from uninfected males. In T cells, an increase in spare respiratory 413 capacity has been linked to mitochondrial biogenesis, and controls the transition to a long-lived 414 memory phenotype (van der Windt et al., 2012). In macrophages, M2 (alternative) activation has 415 previously been shown to lead to increased spare respiratory capacity, a process that also There are significant differences in both the susceptibility and disease presentation 515 between males and females for both diabetes and cardiovascular disease, but few studies have 516 focused on the role of immunometabolism in this dichotomy. Our data suggests apparent 517 biological sex-dependent differences in both the ability of schistosomes to protect from the 518 development of HFD induced metabolic disease parameters, and the ability to modulate 519 macrophage glucose and lipid metabolism. The current epidemiological data strongly indicates 520 an inverse correlation between helminth infections and metabolic diseases such as diabetes and 521 cardiovascular disease, but these studies were not designed to identify sex-specific correlations. 522 Few animal studies have focused on the role of myeloid cells in driving sex-specific metabolic 523 differences. Our current data indicate a clear need for further studies in both humans and animal 524 models to specifically probe the relationship between biological sex and myeloid metabolism, 525 and how chronic helminth infections modulate this. 526

Conflict of Interest 528
The authors declare no competing interests. 529

Acknowledgements 530
The work was supported by The University of Utah, a Scientist Development Grant from the 531 American Heart Association to KCF (14SDG18230012), an American Heart Association Pre-532

LEAD CONTACT AND MATERIALS AVAILABILITY 543
Further information requests should be directed to and will be fulfilled by the Lead Contact, 544 Keke Fairfax (keke.fairfax@path.utah.edu). These studies generated no new reagents. infection when they were transitioned to a high-fat diet (HFD: 21% milk fat, 0.15% cholesterol: 557 TD 88137 Envigo). Bone marrow chimeras were generated by treating ApoE -/mice that had 558 been on high-fat diet for 4 weeks with 20mg/kg of pharmaceutical grade busulfan for 5 days 559 (total dose of 100mg/kg). On day 6 mice were i.v. injected with 2.5-3 x 10 6 bone marrow cells 560 from either 10-week S. mansoni infected or control uninfected ApoE -/mice on high-fat diet.
Reconstitution was validated via flow-cytometry at 3-weeks post-transfer and recipient mice 562 were maintained on high-fat diet for 10-weeks post-reconstitution. 563

S. mansoni infection and glucose tolerance test 565
ApoE -/female and male mice of 6 weeks of age were exposed percutaneously to 75-90 cercariae 566 of S. mansoni or were mocked infected (as controls). At five-and ten-weeks post-infection mice 567 were fasted for 5 hours and baseline blood glucose levels were obtained via lateral tail vein nick. The chromatography gradient for both positive and negative modes started at 15% mobile phase 666 B then increased to 30% B over 2.4 min, then increased to 48% from 2.4-3.0 min, followed by an 667 increase to 82% B from 3-13.2 min, and then to 99% from 13.2-13.8 min where it was held until 668 15.4 min and then returned to the initial conditioned and equilibrated for 4 min. Flow was 0.5 669 mL/min throughout, injection volume was 5µL for positive and 7 µL negative mode. Tandem 670 mass spectrometry is conducted using the same LC gradient at collision energies of 20 V and 40 671 V. 672 673

Sample Preparation
Lipids were extracted from cell pellets (500,000 cells) as described in detail above (Matyash et 676 al., 2008). For targeted lipidomics, lipid extracts were separated on a Waters BEH HILIC column 677 1.7 μm, 100 mm × 3 mm column maintained at 60 °C connected to an Agilent HiP 1290 678 Sampler, Agilent 1290 Infinity pump, and equipped with an Agilent 6490 triple quadrupole 679 (QqQ) mass spectrometer. Lipids were detected using dynamic multiple reaction monitoring 680 (dMRM) in negative ion mode. Source gas temperature is set to 225 °C, with a gas flow of 13 681 L/min and a nebulizer pressure of 30 psi. Sheath gas temperature was 350 °C, sheath gas flow 682 was 11 L/min, capillary voltage of 4000 V, nozzle voltage was 500 V, high pressure RF was 190 683 V and low-pressure RF was 120 V. Injection volume was 2 µL and the samples were The data table exported from MHQuant was evaluated using Excel where initial lipid targets 697 were parsed based on the following criteria. Only lipids with relative standard deviation (RSD) less than 30% in QC samples were used for data analysis. Additionally, targets identified in